Thomas Holtgraves - The Oxford Handbook of Language and Social Psychology-Oxford University Press (2014)

Language and Social Psychology: Introduction and Overview
Language and Social Psychology: Introduction and

Overview
Thomas M. Holtgraves
The Oxford Handbook of Language and Social Psychology
Edited by Thomas M. Holtgraves
Print Publication Date: Sep 2014 Subject: Psychology, Social Psychology

Online Publication Date: Jul 2014 DOI: 10.1093/oxfordhb/9780199838639.013.011
Abstract and Keywords
To use language is to engage in a social process. As a result, the social psychological an
tecedents and consequences of language use are extensive. Yet social psychologists have
tended to ignore language, just as cognitive psychologists have tended to ignore the so
cial dimensions of language use. The Oxford Handbook of Language and Social Psycholo
gy is an interdisciplinary volume that focuses specifically on the interface of language and
social psychology. The chapter authors, all well-known researchers in the language and
social psychology domain, address the social underpinnings of language and language
use, the creation of meaning, the linguistic underpinnings of social psychological process
es, and the role of language and social psychology in a variety of applied topics.
Keywords: Social psychology, language use, meaning, coordination, social cognition, culture, pragmatics, attitudes
Language pervades everything we do as social beings. When we interact with others, it is

the medium for exchanging thoughts and feelings, for influencing and being influenced,
for forming impressions of others and manipulating the impressions others are forming of
us. Even when we are alone, language is important because it serves as a platform for so
cial thought, a means for thinking about other people and oneself, for contemplating so
cial situations and relationships, and so on. It is, in fact, difficult to disentangle language
from social life, and hence its importance is often missed. The emergence of new commu
nications technologies makes this even more striking. Consider the frequency with which
relationships develop in electronically mediated settings (e.g., blogs, Facebook, e-mail,
etc.). People come to “know” one another through these interactions without ever having
met face-to-face. Yet rich and detailed and relatively accurate impressions of one another
may be developed. How? Through the words we use and the way we use them. It is what
we write, just as it is what we say, that allows others to form an impression of the kind of
a person we are.
This handbook is relatively unique in that it is devoted to a consideration of the social psy
chological aspects of language. What this means is that language is viewed here as a so
cial activity and that to understand this complex human activity requires a consideration
Page 1 of 11
PRINTED FROM OXFORD HANDBOOKS ONLINE (www.oxfordhandbooks.com). © Oxford University Press, 2018. All Rights
Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a title in
Oxford Handbooks Online for personal use (for details see Privacy Policy and Legal Notice).
Subscriber: UC - Berkeley Library; date: 31 December 2021

of its social psychological underpinnings. Moreover, as a social activity, the use, and in
fact the existence, of language has implications for a host of traditional social psychologi
cal processes. Hence, there is a reciprocal relationship between language and social psy
chology, and it is this reciprocal relationship that defines the essence of this handbook.
All of the chapters in this volume invoke this relationship in some way.
The title of this handbook—The Oxford Handbook of Language and Social Psychology—
would seem to (p. 2) suggest that the chapter authors would most likely be social psychol
ogists. In fact, only about a third of the contributors to this volume would identify as so
cial psychologists, with the remainder identifying primarily as either cognitive psycholo
gists or communication scholars. This distribution of disciplines reflects the relatively
loose boundaries of the language and social psychology domain, a state of affairs deserv
ing of some mention.
Historically, social psychologists have not, for the most part, undertaken extensive re
search on language, even though the importance of language for social psychology has
never been in dispute. In fact, there has been a chapter on language in each of the five
editions of The Handbook of Social Psychology, and there is a society (International Asso
ciation of Language and Social Psychology) and a journal (Journal of Language and Social
Psychology) devoted to the topic. But, in general, language has not been a major focus of
mainstream social psychologists even though it is almost impossible to think of a topic
studied by social psychologists that does not involve language in some manner. Prejudice,
attitudes, group interaction, attributions, and more involve language in some way. Even
the methods used by social psychologists—measures of attitudes and attributions, person
ality scales, judgments of probability, and so on—involve language. The role of language
in social psychological processes is so fundamental and so pervasive that it is often over
looked and taken for granted. And that masks its importance.
Just as social psychologists have tended to ignore language, so, too, have cognitive psy
chologists tended to ignore the social psychological dimensions of language use. Many
psycholinguists have adopted what Clark (1996) refers to as an individualist perspective
and assumed that the production and comprehension of language is the result of isolated
individuals producing and comprehending language in a vacuum. But conversations are
joint accomplishments requiring the coordination of all involved parties. In short, talk,
both its production and interpretation, cannot really be understood apart from the social
context in which it is produced and understood.
The goal of this handbook is to bring together, in one volume, scholarly reviews of re
search that exist at the interface of language and social psychology. To that end, I invited
scholars known for their research in areas with relevance for this domain. Although this
volume does not exhaust the topics that fall under this rubric, it does include most of the
major theoretical and empirical approaches to research in this vein. There are six sec
tions, with chapters in each section resonating in some manner with the reciprocal rela
tionship between language and social psychology. The first two sections focus on the so
cial underpinnings of language; that is, on the various ways in which social psychological
Page 2 of 11

processes are relevant for understanding language use. The next two sections consider
the implications of language for a host of traditional social psychological topics, from at
tribution to stereotyping. The fifth section focuses on the role of language in the creation
of meaning, and the final section illustrates a variety of the important ways in which the
language–social psychology interface contributes to a number of applied areas.
Social Dimensions of Language Variation: So

cial Causes and Consequences of Language
Variability
It is difficult, if not impossible, to disentangle language from the sociocultural milieu in
which it arises, changes, and is used. The language varieties that one speaks are a funda
mental part of one’s identity, including one’s cultural, ethnic, and gender identity. Lan
guage variation both reflects and constitutes these social dimensions, and the chapters in
this section all deal, in various ways, with the manner in which language use reflects and
maintains the sociocultural milieu in which it arises and is used. The first chapter, by
Howard Giles and Tamara Rakić is a sophisticated and up-to-date treatment of language
attitudes (i.e., evaluative reactions to another’s speech style), with a particular emphasis
on standard versus nonstandard varieties. Focusing primarily on accents, these authors
review research demonstrating the powerful effects that accents have, effects that are, in
fact, more powerful than physical cues such as appearance, in terms of how we perceive
others. Accents, of course, do not exist in a vacuum, and one of the main points stressed
in this chapter is the subtle, complex, and dynamic role played by accents in social life.
Expanding the scope, the next two chapters consider the complex relationship between
language and culture. The chapter by Stella Ting-Toomey and Tenzin Dorjee focuses on
the critical role played by language in the creation and maintenance of identities. Using
several different theoretical perspectives, as well as concrete examples from the Tibetan
speech community, these authors persuasively demonstrate how differing intercultural
background assumptions can result in mismanaged and (p. 3) misconstrued identities in
intercultural and intergroup encounters.
The bidirectional relationship between language and culture is addressed in the following
chapter by Yoshihisa Kashima, Emokio Kashima, and Evan Kidd. These authors argue that
language serves as a medium through which culture is transmitted and maintained; lan
guage variability both creates and constitutes cultural variability. Their argument is
based on the existence of cultural variability in contextualizing linguistic practices (some
languages emphasize the context while others do not). They review research demonstrat
ing several linguistic devices that serve contextualizing functions and whose occurrence
co-varies with independent cultural classifications.
The final chapter in this section, by Campbell Leaper, is a comprehensive review of gen
der variability in language, focusing on both abstract structural differences (e.g., gram
matical gender), as well as on behavioral differences (e.g., assertiveness). Like culture,
Page 3 of 11

gender can be viewed as both reflecting and maintaining gender divisions in a society. In
contrast to gender stereotypes, however, gender differences in language behaviors are
relatively small and, echoing the Giles and Rakić chapter, frequently moderated by con
textual variables.
Social Processes Involved in Language-Based

Interaction
Language, like all joint activities, requires coordination, and this is a fundamental prob
lem for language users. How are humans able to coordinate or align their unique per
spectives in order to understand and be understood? The first chapter in this section, by
Herb Clark and Tania Henetz, provides an overview of many of the mechanisms (e.g., con
versational cooperativeness, prospective pairs, embedding, presequences, and so on) that
facilitate coordination. What this chapter illustrates so well is the fundamentally social
essence of language use. Conversations are not comprised of individuals separately pro
ducing and comprehending; rather, conversations are a collaborative joint venture.
One important aspect of coordination is perspective taking, an attempt to consider

another’s goals and thoughts when producing and interpreting utterances. Perspective
taking overlaps with multiple research areas (e.g., theory of mind), and it is critical for
language use. In the next chapter of this section, Dale Barr focuses specifically on issues
involved in the empirical investigation of perspective taking. Rather than focusing on the
external manifestations of perspective taking, Barr reviews research on conversational
perspective taking with an eye on the variety of ways in which it can appear that perspec
tive taking is taking place when in fact it is not.
Much conversational coordination is accomplished nonverbally, by pointing, nodding, and

so on. The critical role played by hand gestures and facial displays in conversational coor
dination is clearly articulated in the chapter by Janet Bavelas, Jennifer Gerwing, and Sara
Healing. Using numerous examples from actual spontaneous dialogues, these authors
document how speech is tightly synchronized with hand gestures and facial displays.
Words, gestures, and facial displays are useful for communicating different kinds of infor
mation, but they work together to form integrated, but not necessarily redundant, mes
sages.
The next chapter, by Martin Pickering and Simon Garrod, is concerned with the cognitive
mechanisms that underlie coordination in dialogue. In their view, interlocutors construct
mental models of a topic, and successful dialogue occurs when these models become
aligned. But how does this happen? In general, alignment occurs via subtle priming and
imitation, with interlocutors mirroring one another’s talk at multiple levels. As a result,
interlocutors speak and think about things in the same way, in other words, their mental
models become aligned. Moreover, this alignment allows interlocutors to predict subse
Page 4 of 11

quent verbalizations, thereby facilitating communicative success and making dialogue ap
pear (deceptively) easy.
The final chapter in this section, by Rachel Giora, addresses the issue of discourse coher
ence, that is, the interrelatedness of sentences (or turns) in a discourse (or conversation).
Speakers and writers may use markers (e.g., “Oh, by the way…”) to help achieve coher
ence, but, in the end, coherence is largely cognitive and constructed by interactants. Gio
ra describes and evaluates models of how this occurs, including the standard pragmatic
(Gricean) model, as well as newer variations and theories, and she notes how social vari
ables (e.g., power) can play a role in this process.
Linguistic Underpinnings of Interpersonal

Processes
The role that language plays in many traditional social psychological topics is often un
derappreciated and frequently overlooked. In the first chapter of the third section, Anne
Maass, Luciano Arcuri, and Caterina Suitner consider some of the ways (p. 4) in which
language can shape interpersonal relations, with a particular focus on stereotypes. The
idea that language can contribute to stereotyping is an old one (Allport, 1954), yet empiri
cal research on this topic is relatively new. These authors argue that language serves a di
rective function, directing our attention to some features of our social world (and not oth
er features) and helping us parse it into social categories. As a result, language can serve
as a vehicle for the transmission of stereotypes, and it does this in various ways, from ex
plicit, in-your-face hate speech to extremely subtle yet tremendously impactful facets (ab
straction, word order, pronoun use, etc.) of which language users are often unaware.
One of the oldest topics in social psychology is persuasion, yet the role of language in per
suasion has been relatively under researched. The chapter by James Dillard provides a
systematic overview of research examining the effects of various dimensions of linguistic
style on the process of persuasion. Dillard takes a bottom-up perspective and considers
seven language variables including both granular (self-referencing, rhetorical questions,
conclusion explicitness) and thematic (vividness, framing, powerful language, dominance)
variables that have been documented to play a role in the persuasion process.
Another classic social psychological topic is relationships, their nature, status, progress,
and so on. The manner in which individuals talk with one another is often seen as a rela
tively accurate barometer of their relationship. Steve Duck and Daniel Usera, the authors
of the next chapter, go further to argue that talk, in essence, is the “lifeblood” of relation
ships. These authors catalogue a variety of language practices and the functions that they
serve for relationships. From self-disclosure to telling stories, these discursive practices
can be regarded as constituting relationships.
The tremendous change in communication technologies over the past 20 years has creat
ed both a challenge and an opportunity for language researchers. The next two chapters
Page 5 of 11

focus on the latter as these researchers advocate for the computerized analyses of natu
rally occurring instances of language use (e.g., tweets, Facebook posts, blogs, and so on)
as a means of studying social and personality processes, an approach with unlimited po
tential. Molly Ireland and Matthias Meehl argue that how a person talks is a linguistic fin
gerprint that displays trait like qualities. Importantly, it is not content but rather style
(function words—pronouns, conjunctions, adverbs, etc.) that is most telling of a person’s
personality. And this is largely because style, relative to content, is less constrained by
topic and situation. These authors provide a comprehensive summary of recent research
demonstrating relationships between various facets of language style and some of the
major dimensions of personality. The chapter by Cindy Chung and James Pennebaker con
tinues with many of the same themes and issues, but rather than focusing on the individ
ual, these authors argue for the systematic analysis of linguistic style as a means of study
ing the course and outcomes of social interactions. They document, for example, how de
gree of linguistic style matching (which is often automatic and unconscious) is related to
group outcomes. Echoing the chapter by Duck and Usera, these authors review research
documenting the manner in which interpersonal variables such as status, cohesiveness,
group stability, and so on, are related to linguistic style.
Linguistic Underpinnings of Intrapersonal

Processes
The chapters in the fourth section also focus on the role of language in social psychologi
cal processes, but with an emphasis on intrapersonal processes. Robert Wyer, Jr.’s chap
ter provides a detailed overview of the role of language in social cognition. Creatively us
ing concepts detailed in other chapters in this volume (e.g., literal meaning, pragmatic
meaning, situational models, etc.), Wyer describes how mental representations of linguis
tic meaning are influenced by situational and individual differences. These variables come
into play as comprehenders attempt to make sense of someone’s communication. Pin
pointing conversation norm violations as the instigator of extra-inferential processing,
Wyer applies this approach to a variety of language situations, including humor, advertis
ing, and psychological experimentation.
The chapter by Klaus Fiedler and Tobias Krüger provides an extensive consideration of
the role played by language in a specific and heavily researched area within social cogni
tion, namely attribution processes. As these authors demonstrate, the process of causal
attribution is very much influenced by the nature and structure of language. This is be
cause the semantic meaning of many words can be described in terms of distinct coordi
nates on the major dimensions of attribution; hence, specific attributions are basically
wired into the lexicon. The authors illustrate how one dimension of language, linguistic
abstractness, is related to the fundamental attribution error, the actor–observer bias, the
egocentric bias, the linguistic-intergroup bias, and the (p. 5) relationship between ab
stractness and psychological distance.
Page 6 of 11

The next chapter, by Monisha Pasupathi, C. D. Mansfield, and T. L. Weeks, deals with sto
rytelling, a ubiquitous and ancient social activity. Telling a story is a social act requiring
considerable coordination; it is an interpersonal process. But it has intrapersonal conse
quences. That is, because telling a story is a social event, various aspects of the social sit
uation (e.g., goals in telling a story) can influence the nature of the story that is told, and
the nature of the story one tells can influence one’s self-view, strengthening some views
of the self and altering or deleting other self-views. These authors argue that the relation
ship between stories and the self is reciprocally influential across developmental time.
The frequency with which people tell stories about the self via online blogs, Facebook
posts, and so on suggests that this may become a topic of increasing relevance.
A relatively new development in language and social psychology is recognition of the im
portant role played by language in emotional experience. Interest in emotions is long-
standing, of course, but there is considerable disagreement regarding fundamental as
pects of how people experience emotions. The chapter by Jennifer Fugate and Lisa Feld
man Barrett reviews an impressive amount of research, both behavioral and imaging,
documenting the important role played by language in the perception of emotional experi
ences. These authors argue that emotions emerge when people make meaning out of sen
sory input and that emotion words provide the “internal context” that helps segment and
constrain this sensory information. In short, “putting one’s feelings into words can clarify
and change them.” As a result, talking (and writing) about one’s emotional experiences
can alter those experiences, an effect with potentially important practical implications
(e.g., Pennebaker & Graybeal, 2001).
The final chapter in this section provides an overview of discursive social psychology, a
broad-based theoretical attempt to specify the role played by language in all social psy
chological phenomena. Authors Alexandra Kent and Jonathan Potter review the basic con
cepts and methods in discursive social psychology and illustrate them with a discursive
analysis of family mealtime interactions. Closely aligned with conversation analysis (e.g.,
Schegloff, 2007), discursive social psychology is a different way of doing social psycholo
gy, an approach that takes talk as its central concern because it is in talk, and in particu
lar the sequential nature of talk, that social psychological processes such as attitudes,
motivations, and so on are revealed.
Language and Meaning

Language can be regarded as a tool for implementing social cognition (Semin, 2000) and
as a symbolic system used for the creation and exchange of meaning (Holtgraves &
Kashima, 2008). And meaning, of course, is central to social psychology; we react not to
others’ actions but rather to our interpretation of those actions. The chapters in this sec
tion all deal in various ways with the creation and exchange of meaning with language.
The first chapter, by Michael P. Kaschak and John L. Jones, considers how language con
veys meaning from the perspective of embodied cognition. Because language is an arbi
trary system of symbols, there must be some mechanism for connecting those symbols
Page 7 of 11

with the outside world, some grounding in our bodily systems of perception and action
planning. In this view, language is understood by constructing internal, sensorimotor sim
ulations of linguistic descriptions, and these simulations are based on past experiences
with the external world. The authors review recent empirical research designed to test
the assumption that the same patterns of neural activity that occur when one experiences
objects and events in the outside world also occur during the comprehension of descrip
tions of those experiences and events.
Linguistic meaning is particularly challenging, of course, because of the possibility that

language users can mean more than they say; that is, they can convey nonliteral mean
ing. The chapter by Rachel Giora deals specifically with this issue. She provides extensive
examples of instances of indirect meaning and reviews much of the existing empirical re
search on how indirect meaning is comprehended. In addition, she argues persuasively
that it is the degree of innovativeness of an expression, rather than its degree of nonliter
alness, that is responsible for how it is processed and interpreted.
The chapter by Raymond W. Gibbs, Jr. deals also with ambiguity in linguistic meaning but
does so by focusing on the role played by communicative intentions. The link between lin
guistic meaning and intentions is far from straightforward, and Gibbs argues persuasively
that linguistic meaning is not the result of an a priori individual intentional state but
rather an emergent property of a dynamically organized system. Hence, the intentionalist
view that meaning is derived from the recovery of a speaker’s intention (Grice, 1957;
Searle, 1969) (p. 6) is misguided. This is because verbal interactions are highly collabora
tive (see chapters by Clark and Henetz; Pickering and Garrod; Bavelas et al., this volume),
with meaning being an emergent property of these interactions.
Although electrophysiological techniques have been used in psycholinguistic research for

several decades (e.g., Kutas & Hillyard, 1980), the use of these techniques to study social
ly situated language comprehension is a relatively new and very welcome development.
The chapter by John Hoeks and Harm Brouwer is a state-of-the-art overview of electro
physiological techniques, in particular event-related potential (ERP) methodology, for
studying language comprehension. Their chapter focuses on three major ERP compo
nents (Nref, N400, and P600) that are linked to specific linguistic functions. These au
thors provide a parsimonious account of these components and link them directly to so
cial and pragmatic factors that can play a role in conversation comprehension.
Jean-François Bonnefon explores the meaning of reasoning terms, specifically quantifiers

(e.g., some) and connectives (e.g., or) when those terms are embedded in a discourse con
text. The manner in which these reasoning terms are interpreted can change when social
processes such as politeness are activated in a situation. This occurs because pragmatic
communication principles such as efficiency are cancelled when politeness concerns
arise, creating a situation of interpretive ambiguity. Importantly, the interpretation of
these terms serves as input to further reasoning and judgment; hence, the consequences
of this effect can be substantial.
Page 8 of 11

Applied Topics
Given the scope and diversity of research in the language and social psychology domain,
it is not surprising that many of these concepts and theories have important real-world
applications. The chapters in the final section involve specific applications in terms of ed
ucation, the law, self-report data collection methods, conflict and conflict resolution, and
automatic tutoring systems.
The chapter by John Edwards considers the role of linguistic diversity in educational set
tings. The immense linguistic diversity in many locales (e.g., 300 languages spoken in
London) makes this a critical issue and one often politically charged. As Edwards points
out, there is often a tension between linguistic pluralism (and the promotion of cultural
variability) and the transmission of general knowledge, a tension complicated by the exis
tence of a pervasive bias against language varieties believed to be substandard (e.g.,
Black English; see chapter by Giles and Rakić, this volume). But, as Edwards points out,
empirically, there are no substandard varieties and the available research supports edu
cational policies of linguistic expansion: there are clear advantages to being multilingual
and/or bidialectical.
Self-report measures are a primary method for data collection in the social sciences. The
chapter by Frederick Conrad, Michael Schober, and Norbert Schwarz provides a compre
hensive overview of research demonstrating the role played by linguistic pragmatics in
self-report data collection methods, with a particular focus on surveys and interviews.
Survey respondents must interpret the intended meaning of survey questions, and they
often do so based on contextual information and conversational rules (e.g., see chapter by
Wyer, this volume). As a result, researchers may not be aware of the extent to which an
interviewee’s responses are influenced by these pragmatic processes. Moreover, with in
terviews, the collaborative nature of the situation can also influence the course of the in
terview and an interviewee’s responses. The authors of this chapter clearly convey the
importance of considering these processes in the construction and interpretation of self-
report instruments.
The chapter by Deborah Davis and Guillermo Villalobos provides an overview of the many
ways in which language is relevant for the law. Drawing on a range of research including
work in pragmatics, linguistic abstraction, and persuasion, these authors describe how
people have difficulty, and are often biased, in remembering their past conversations, an
effect with important implications for legal testimony. They also consider the role of lan
guage in the comprehension of one’s rights (e.g., self-incrimination) and understanding
jury instructions, as well as the role of language in interrogation, classification, and judg
ment. The authors conclude by noting the importance of empirical research in these ar
eas because of the widespread belief in numerous unsupported claims in the legal do
main.
The chapter by Paul Taylor explores the diverse and important role played by language in
conflict and conflict resolution. Taylor considers a variety of conflict situations (e.g., mar
Page 9 of 11

riage, employment disputes, hostage situations, etc.) and reviews relevant research re
garding the effects of major types of language (competitive vs. cooperative vs. avoidant)
on conflict escalation and resolution. Included in this chapter is a consideration of the im
portant role (p. 7) played by interactant coordination and matching, a topic covered ex
tensively in several chapters in this volume (Ting-Toomey and Dorjee, Pickering and Gar
rod, Chung and Pennebaker).
The development of new electronic communications technologies such as texting, Twitter,

Facebook, and so on raises a host of questions regarding communication processes in
these domains. Susan Fussell and Leslie Setlock provide a coherent and informative re
view of communication similarities and differences in these platforms. Using Clark and
Brennan’s (1991) theory of media affordances, these authors review research on the af
fordances of different media, as well as the impact of relevant social psychological vari
ables such as culture. Critical here are issues of collaboration (see chapters in section 2
of this volume) and how different media facilitate or hinder this collaborative process.
The final chapter by Arthur C. Graesser, Fazel Keshtkar, and Haiying Li considers the role
of language in computerized tutoring systems, a relatively recent development but one
with enormous implications for education and training. Automated tutors, of course, in
teract with students via natural language, and the authors of this chapter describe the
role played by social, cognitive, emotional, and motivational mechanisms in successful tu
toring systems. Not surprisingly, these mechanisms have linguistic underpinnings, and
much of the research reviewed in this chapter echoes approaches in other chapters in
this volume, in particular, considerations of common ground and coordination (Clark and
Henetz; Barr; Pickering and Garrod), the possibility of conversation misunderstandings
and the manner in which they are revealed (Kent and Potter) and the role of pragmatics
and politeness in the interpretation of utterances and structure of conversational ex
changes (Bonnefon).
References
Allport, G. W. (1954). The nature of prejudice. New York: Addison-Wesley.
Clark, H. H. (1996). Using language. Cambridge, UK: Cambridge University Press.
Clark, H. H., & Brennan, S. E. (1991). Grounding in communication. In L. B. Resnick, J. M.

Levine, & S. D. Teasley (Eds.), Perspectives on socially shared cognition (pp. 127–149).
Washington, DC: American Psychological Association.
Grice, H. P. (1957). Meaning. Philosophical Review, 67, 377–388.
Holtgraves, T., & Kashima, Y. (2008). Language, meaning, and social cognition. Personali
ty and Social Psychology Review, 12, 73–94.
Kutas, M. & Hillyard, S. A. (1980). Reading senseless sentences: Brain potentials reflect
semantic incongruity. Science, 207, 203–205.
Page 10 of 11

Pennebaker, J. W., & Graybeal, A. (2001). Patterns of natural language use: Disclosure,
personality, and social integration. Current Directions in Psychological Science, 10, 90–
93.
Schegloff, E. A. (2007). Sequence organization in interaction: A primer in conversation

analysis, vol. 1. Cambridge, UK: Cambridge University Press.
Searle, J. R. (1969). Speech acts. Cambridge, UK: Cambridge University Press.
Semin, G. R. (2000). Agenda 2000—Communication: language as an implementational de

vice for cognition. European journal of Social Psychology, 30, 595–612. (p. 8)
Thomas M. Holtgraves
Thomas M. Holtgraves is a professor of Psychological Science at Ball State Universi

ty in Muncie, Indiana. He has conducted interdisciplinary research, supported by the
NSF and NIH, into multiple facets of language and social psychology.
Page 11 of 11

Language Attitudes: Social Determinants and Consequences of Language
Variation
Language Attitudes: Social Determinants and Conse

quences of Language Variation
Howard Giles and Tamara Rakić

Online Publication Date: Mar 2014 DOI: 10.1093/oxfordhb/9780199838639.013.030
This chapter surveys important approaches and findings related to language attitude
studies. It begins by attending to the kinds of evaluations and images people are able to
deduce from voices and attention is focused on judgments conveyed by standard and non
standard features of accents, how they are developed, and their role in subjective com
prehensibility. Thereafter, attitudes toward accents are addressed in a broader context
(e.g., alongside verbal content, facial appearance, and the linguistic landscape), highlight
ing their importance to people’s social identity and emotional expression. Finally, a new
model of language attitudes is introduced that attends to their complexity and role in on
going discourse and information management. In addition and as a means of providing
conceptual coherence to the literature (and particularly in the recent context of so many
emergent models), nine organizing principles of language attitudes inspired by Dragoje
vic, Giles, and Watson’s (2013) complementary innovations are crafted.
Keywords: Language attitudes, accent, dialect, voice evaluations, social identity, emotional expression, compre
hensibility, facial appearance, verbal content
As one of its New Year resolutions for 2012, the (British) Daily Mail (12/29/2011, p. 15)
staunchly advocated: “Classic FM station bosses—fewer Northern accents, if possible,
please, on your airwaves.” This suggests that, even in the twenty-first century and per
haps for good evolutionary effect (Fincher & Thornhill, 2008), accents are still very so
cially significant. In fact, looking at the language attitudes literature, it is clear that of all
the vocal cues investigated (e.g., pitch, lexical diversity), accents have received the most
empirical attention—and it is these issues, as they relate to their determinants and conse
quences, that this chapter addresses. This is timely, to the extent that work in this tradi
tion flourished in the 1970s and 1980s and was, arguably, among the most vibrant do
mains in the social psychology of language. Language attitudes research was rendered
relatively silent in the 1990s through the early 2000s, yet is now experiencing a gratify
ing renaissance (e.g., Giles & Watson, 2013; Gluszek & Dovidio, 2010a; Rakić, Steffens, &
Mummendey, 2011a).
Page 1 of 29

Variation
Accent refers to a manner of pronunciation associated with particular group member
ships, be it social, regional, ethnic, and so on (Giles, 1973). That people have, or rather
are attributed as having, a so-called “accent” has been a concern in everyday parlance,
popular culture, and in many professional contexts for a very long time. Indeed, a stan
dard accent is often referred to as “accent-free” or “accent-neutral” speech, and, con
versely, speakers of many nonstandard and foreign varieties are often saddled with hav
ing “an accent” (Lippi-Green, 1997). This could not, however, be further from the truth:
everyone has their own way of pronouncing words and sentences and, therefore, possess
es (p. 12) an individualized accent (or idiolect). But, more importantly, being attributed as
having an “accent” or “no accent” has immediate consequences for evaluations of a
speaker. The British actor Stephen Fry is reported as saying: “I sometimes wonder if
Americans aren’t fooled by our accent into detecting the brilliance that may not really be
there” (Lane, 2007). Hence, it is not so much objective assessments as it is clusters of be
liefs that govern social judgments of speech styles and their users.
The aim of this chapter is to survey important approaches and findings related to lan
guage attitude studies. We begin by attending to the kinds of evaluations and images peo
ple are able to deduce from voices, and we focus our attention on judgments conveyed by
standard and nonstandard features of accents, how they are developed, and their role in
subjective comprehensibility. Thereafter, we will address attitudes toward accents in a
broader context (e.g., alongside verbal content, facial appearance, and the linguistic land
scape), highlighting their importance to people’s social identity and emotional expres
sion. Finally, we introduce a new model of language attitudes that attends to their com
plexity and role in ongoing discourse and information management. In the conclusion,
and as a means of providing conceptual coherence to the literature, and particularly in
the recent context of so many emergent models (e.g., Dovidio & Gluszek, 2012; Giles &
Marlow, 2012), we end with nine organizing principles of language attitudes inspired by
Dragojevic, Giles, and Watson’s (2013) complementary innovations in this regard.
Socially Diagnostic Language Cues for Social

Evaluation
People not only use language to communicate with others but also to make sense of them,
personally and socially. For instance, almost immediately on hearing someone start to
speak (see Williams, 1976), we are able to tell many things about that person. Just from
their voice, we can gain information about a speaker’s gender, ethnicity, emotional state,
and even physical size (e.g., Honorof & Whalen, 2010; Lass, Hughes, Bowyer, Waters, &
Bourne, 1976). Listeners can also deduce additional attributes from speech, such as per
sonality traits (e.g., Allport & Cantril, 1934; Scherer, 1979) and, indeed, the word person
ality comes from Latin per (through) and sonare (to sound) meaning literally “to sound
through.”
Page 2 of 29

Variation
As McGlone and Giles (2011) commented in their own review of language attitudes re
search, “speakers’ identities are encoded in their voices. And listeners hearing their
speech can, without training, do a remarkable job of decoding this identifying
information” (p. 218). In this vein, there is some evidence that people seem to be fairly
good at judging certain characteristics from voice; for instance, the correlation between
judged and actual socioeconomic status (SES; as measured by the Index of Status Charac
teristics) was found to be +0.80 (Ellis, 1967). However, at other times, as for example
with age, people can be far from accurate in their estimates (Mulac & Giles, 1996).
In the language attitudes literature, many characteristics of speech and language produc
tion have been examined (e.g., powerful/powerless speech, lexical diversity, and speech
rate). Because details concerning speech variation will be covered elsewhere in this vol
ume (see Chapter 5), we will focus our analysis of language attitudes to accented speech.
The Power of Accents

As witnessed, another’s voice affords listeners an array of judgmental options, with
speakers themselves possessing language attitude schemas that can mediate how they
self-present in different situations with various people (see Giles & Street, 1985). Across
studies, different methodologies have been adopted. Whereas some (e.g., Grondelaers,
van Hout, & Steegs, 2010) had genuine speakers of language varieties using free speech
as objects of evaluation, much research in the language attitudes tradition has used Lam
bert, Hodgson, Gardner, and Fillenbaum’s (1960) matched-guise technique (MGT; see al
so, Lambert, 1967; and for other elicitation procedures, see Ryan, Giles, & Hewstone,
1988). Studies employing this procedure present listeners with audio recordings of pur
portedly neutral passages of prose (see Giles & Coupland, 1991) read by bilingual or bidi
alectal individuals who can authentically adopt various guises of the language varieties
under investigation. By using the same speaker to render these guises, other extraneous
vocal variables, such as speech rate, that vary naturally between speakers, are con
trolled. Researchers are thereby able to attribute response differences primarily to the
guises being contrasted. The pros and cons of different techniques continue to be aired
and debated over the decades (e.g., Garrett, 2010) yet, nevertheless, all methods elicit
somewhat the same patterns of social evaluation (Sebastian, Ryan, Keogh, & Schmidt,
1979; see however, Ball, 1983).
Currently, we are witnessing a renewed interest in the methodology for assessing

(p. 13)
language attitudes. Pantos and Perkins (2013), for example, have showed that whereas
people tend to show a pro-foreign accent bias in explicit evaluation of voices, this effect
not only disappears but is reversed (in favor of an American accent) when using an audio
implicit association test of language attitudes. In similar vein, Schoel et al. (2013)
developed a new scale for the assessment of language attitudes, making a clear distinc
tion between dimensions of speaker evaluation (i.e., attitudes toward speakers; see Zahn
& Hopper, 1985) on the one hand and those associated with the language spoken (such as
sound and structure) on the other. Further work in this genre should allow for the un
Page 3 of 29

Variation
packing of even clearer judgmental distinctions between speakers and the speech styles
they adopt. In what follows, we turn our attention to the ways in which accent has been
explored in the language attitudes literature and begin by discussing the role of age.
Development Matters
The fact that we rely so much on voice and language when it comes to making sense of
others should not be very surprising if we consider that hearing is the first fully devel
oped and used sense in newborns (Crystal, 2005; Mehler, Bertoncini, Barrière, & Jassik-
Gerschenfeld, 1978). Indeed, Nazzi, Juscyk, and Johnson (2000) observed that American
children as young as 5 months are already sensitive to dialectical variations (see also Kin
zler, Dupoux, & Spelke, 2007). Such children seem already to have developed an English
reference sound system that helps them immediately notice and react if different accents
(e.g., Italian or Spanish) are presented alongside their native language variety. They are
further able to differentiate Japanese from Italian even without English comparisons, pre
sumably because the former has such a very different rhythmic sound system. It seems
that language-sound connections made during the native language learning process con
tinue to play a central role throughout our communication life. This poses an obstacle to
overcome when it comes to second-language learning at a later point in life, especially
when accents are, purportedly anyway, much more difficult to change (Lippi-Green,
1997).
Hence, by the age of 5, accents can be as potent in children’s social appraisals as gender
or skin color. In this sense, studies show that children prefer friends from the same ac
cent group, regardless of their race (Kinzler, Shutts, Dejesus, & Spelke, 2009), and they
also attribute greater trust to native language speakers (Kinzler, Corriveau, & Harris,
2011). Moreover, these preferences are not driven by impaired comprehension of nonna
tive accents but rather by a simple preference for one’s own accent. Consequently, Kin
zler, Shutts, and Correll (2010) argue that children have a natural tendency to use lan
guage—and even accent—as a meaningful social mechanism to organize the social world
around them (for an evolutionary perspective on this process, see Reid et al., 2012).
Additionally, studies with children are important for understanding how language atti
tudes are learned and developed (for a socialization model of language attitudes, see
Bradac & Giles, 1991). In a French study by Girard, Floccia, and Goslin (2008), 5- and 6-
year-old children were very good at differentiating between-language variation (i.e., na
tive- versus foreign-accented speech), although they seemed not to notice within-lan
guage differences (i.e., northern- and southern-accented French). Several explanations
may account for this phenomenon. First, a foreign accent may vary more from non-for
eign varieties than different regional varieties vary from one another. Second, at that age,
children start to develop national identities; hence, it may be more relevant to distinguish
“us” versus “them” varieties than to distinguish forms within one’s own language; for a
further discussion of this process, see work on the so-called reference framing effect later
in this chapter. This effect diminishes with further development: whereas knowledge of
Page 4 of 29

Variation
regional differentiation increases, the differentiation between native and foreign remains
highly persistent throughout the lifetime.
In contrast, older children (aged 7–10 years) were able to differentiate and evaluate dif
ferent varieties of Dutch (van Bezooijen, 1994), and the results indicated a preference for
standard Dutch over regional varieties. Overall, language attitudes seem to be transmit
ted automatically through socialization, with children demonstrating different patterns of
evaluation of nonstandard varieties depending on their age (e.g., Giles, Harrison, Creber,
Smith, & Freeman, 1983). Welsh children at the age of 7 clearly prefer their own accent
and react negatively toward received pronunciation (RP), although, by the age of 10, they
show an evaluative pattern similar to adults by evaluating the British standard most fa
vorably (Price, Fluck, & Giles, 1983).
Issues of Standardization and Nonstandardiza

tion
During the history of their development, languages usually undergo a process of so-called
(p. 14) standardization (for a framework in which to locate language varieties by this and
other means, see Dragojevic et al., 2013; Ryan, Giles, and Sebastian, 1982), meaning that,
from many different varieties of the same language (e.g., different dialects), one unique
way of defining the grammar, spelling, and pronunciation is established. This variety is
put on a societal pedestal and is typically spoken by culturally and economically powerful
elites. This standard variety is then promoted through the education system as well as
through the media and, in this way, standard language speakers gain high prestige and
status (for the relationship between language ideologies and language attitudes, see
Dragojevic et al., 2013). Intriguingly, the degree of standardization of a given language
has been found to correlate with a country’s economic development (Jones, 1973), al
though operationalizing a standard language is not without its complexities (see Edwards
& Jacobsen, 1987; Kristiansen, 2001). Furthermore, a standard variety’s prestige, aes
thetics, and associations with correctness are actually more context- and norm-dependent
than inherently ascribed as will be evident from what follows.
The notion of standard language seems to be very natural in the minds of people, and yet
it is actually an artificial construct, being a consequence often referred to as the myth of
a standard language (Lippi-Green, 1997). Examples of standard varieties would be RP
(commonly referred to as “the Queen’s or BBC English”), Parisian French, and Castilian
Spanish, whereas nonstandard speakers would be those who speak more regional or non
native varieties of these languages (Gluszek & Dovidio, 2010b; see also Tsurutani, 2012).
Generally, standard language varieties grant people access to political, economic, and ed
ucational forums and opportunities, whereas speakers of nonstandard language varieties
are faced with stigmatization (Gluszek & Dovidio, 2010a); this pattern is even more pro
nounced for nonnative than for other nonstandard accented speakers (i.e., regional ac
cent speakers). Additionally, nonnative accented speakers have reported lower feelings of
belonging in the United States, and this effect can be mediated by perceived problems in
Page 5 of 29

Variation
communication (Gluszek & Dovidio, 2010b). The fact that a standard accent is not in any
sense a given is evident from a recent study in the Netherlands that focused on different
regional varieties of standard Dutch (Grondelaers et al., 2010). Interestingly here, region
al influences on the standard language have started to gain more social acceptance as le
gitimate variations of this speech variety.
The notion of standard language or accent is then somewhat problematic, but even more
so is the notion of “no-accent” or “speech without an accent” referred to at the outset of
this chapter. In many instances of everyday life, nonstandard accented speakers are those
who are accused of having “an accent.” Consequently, ever-growing numbers of courses
and workshops are offered with the promise of helping nonstandard speakers lose their
accent. Whereas at first blush there might not seem to be anything terribly wrong with
this, it sidesteps how relative and highly context-dependent this notion is. For instance,
imagine someone speaking standard American in California. This person would probably
be considered as someone without an accent as compared to someone with a London ac
cent. Once in the UK, however, these inferences would be inverted.
Additionally, the exposure and expertise associated with a given accent determines how
well different within-language variations can be perceived. Quite recently, an American
journalist reporting on the televised 2011 Grammy Awards featured Adele, a singer who
was regularly on stage accepting awards that year. He wrote: “Adele’s clean sweep lent
the night a classy air…hey, anything sounds suave in a British accent” (Wappler, 2012). In
this sense, the singer’s accent was deemed prestigious because it was “British.” In the
UK, however, her Cockney accent would not be evaluated so kindly (see Coupland & Bish
op, 2007).
Generally, it can be concluded that speakers of nonstandard language varieties—and es

pecially those with broader accents (Gluszek, Newheiser, & Dovidio, 2011; Tsurutani,
2012)—are perceived across many diverse international settings (see Giles & Watson,
2013) to be at a social disadvantage vis-à-vis their standard counterparts, particularly in
status-related contexts such as classroom settings or job interviews. Indeed, studies have
shown that different nonstandard varieties can be located hierarchically on aesthetic and
prestige dimensions (e.g., Giles, 1970). Exemplars of such nonstandard forms in the Unit
ed States would be Appalachian (Luhmann, 1990) and Southern (Heaton & Nygaard,
2011; Preston, 1999), as well as African-American vernacular English variants (for a dis
cussion of the complexity of findings relating to the latter, see Ray, 2009). More specifical
ly, nonstandard speakers can be caricatured as downgraded relative to standard accented
speakers on competence (e.g., intelligence, ambition, and confidence) as well as dy
namism traits (e.g., lively, enthusiastic, and talkative); suitability (p. 15) for employment in
high-status jobs (Rakić et al., 2011a); effectiveness as teachers (Williams, 1976); guilt
when under suspicion of committing a crime (e.g., Dixon, Mahoney, & Cocks, 2002); and
the perceived quality of the message content (see Guy, 1988; Riches & Foddy, 1989). As
indicated earlier, it would seem that these attitudes derive from complex schemas about
relationships among language, dialect, ethnicity, and social class (for a French case, see
Page 6 of 29

Variation
Stewart, 2012), and ones, again, that can be acquired very early in life (e.g., Floccia, But
ler, Girard, & Goslin, 2009).
Gluszek and Hansen (2013) also discuss, in ways not usually manifest in the literature,
the relationship between language attitudes and the media. In particular, they look at
how different accents can be strategically adopted in the movies and on TV and radio to
evoke particular social images—as in the case of heroes and villains; the former often be
ing represented by standard and the latter by nonstandard accents (see Lippi-Green,
1997, for representations of native/nonnative accents in Disney movies over 50 years). In
deed, the ways in which language varieties, such as languages, dialects, and accents, can
be associated with distinctive songs, poetry, and other literary forms is a fascinating are
na worthy of systematic analysis.
Language Attitudes and Comprehensibility

Different interpretations have been proposed to explain why nonstandard accented
speakers are evaluated less favorably than standard accented speakers. Some, for us
rather provocatively, have argued that nonnative speakers are downgraded not so much
because of social prejudice or preexisting stereotypes but, rather, due to the increased
cognitive load experienced by listeners trying to understand nonstandard accented mes
sages (e.g., Dovidio & Gluszek, 2012). Whereas this might play a relevant role in a task
where one has to quickly discriminate a short ambiguous statement as true or not (e.g.,
Lev-Ari & Keysar, 2010), it is less likely that the same would apply in real-life encounters
in which other information about a speaker might be present and when speech samples
are longer, thereby offering more time to become accustomed to that language variety. In
terestingly, it is this idea of “incomprehensibility” that is a central argument of standard
language ideology (see Dragojevic et al., 2013); the justification and rationalization for
why nonstandard (and nonnative) speakers are somehow “less.”
Relatedly, Dovidio and Gluszek (e.g., 2012) account for the discrimination of nonstandard
speakers as a result of (nonstandard) accent and nonverbal behavior usage that cause
disfluency and increase intergroup biases in listeners. For these scholars, it seems that
(nonstandard) accent and nonverbal behavior facilitate intergroup bias because they in
crease the strength of intergroup boundaries and in such a way contribute to disfluent in
teractions. Even though this model partially acknowledges the biased perception of oth
ers, at the same time, it relies on the notion that there is something essentially disfluent
in nonstandard (i.e., nonnative) speakers that causes lower comprehension. In our opin
ion, such a view is rather problematic for at least two reasons.
First, research suggests that although unfamiliar accents may initially disrupt processing
fluency, adaptation and normalization quickly follow, and participants show a full recov
ery in processing speed (Floccia, Goslin, Girard, & Konopczynski, 2006), sometimes after
exposure to only 2–4 sentences (Clarke & Garrett, 2004). Second, not only does it assume
that a listener is somewhat of an objective receiver of a message, but it also neglects oth
er aspects that are known to influence our perceptions, such as social identities and emo
Page 7 of 29

Variation
tions aroused. In this regard, there are studies showing how listeners’ prejudice levels
mediate the perception of accent (strength) and overall evaluation of de facto standard
speakers (e.g., Rubin, 1992; Rubin & Smith, 1990). Here, the only cue for ethnicity of
speakers was in their face, which indicated East-Asian origins, whereas all speakers
spoke exclusively with a standard American accent. Remarkably, listeners with high lev
els of prejudice reported that speakers (accompanied with an East-Asian face) were very
hard to understand, reportedly due to their (heavy) accents. Hence for those high in prej
udice, certain foreigners were “heard” to have a “problematic” accent that actually was
not present (see Lindemann, 2002; Munro & Derwing, 1995). Relatedly, Cairns and Duriez
(1976) found that Catholic children performed much poorer than Protestant children on a
recall test after being given a text delivered in RP (the British standard accent variety)
than in a Dublin accent. Tellingly in terms of political identity, the former reported to be
living in Ireland, whereas their Protestant counterparts reported living in Northern Ire
land. Interestingly, the same deficit in recall was observed with Protestant children, but
only after exposure to a Dublin-accented passage of prose. Essentially, these decreases in
comprehension (p. 16) were not due to any lack of understanding because different ac
cented forms were equally well understood by both sets of listeners.
The Value of Nonstandard Accents

On traits of social attractiveness and benevolence, such as friendliness, generosity, and
sincerity, nonstandard speakers have been upgraded, albeit not always consistently (e.g.,
Dailey, Giles, & Jansma, 2005; for a meta-analysis, see Fuertes, Gottdiener, Martin,
Gilbert, & Giles, 2012). Indeed, the values attached to a (foreign) British accent in the
United States, and, in contrast, an American variety in Australia, are illustrative cases in
point (Bayard, Weatherall, Gallois, & Pittam, 2001; Stewart, Ryan, & Giles, 1985). Addi
tionally, male nonstandard speakers have been perceived as more masculine and consid
ered more skilled at manual labor tasks (Giles, Henwood, Coupland, & Harriman, 1992).
In this sense, when ingroup identity and loyalty are salient, nonstandard accents can have
covert prestige for some of its speakers (Marlow & Giles, 2008; Ryan, 1979). This effect is
in line with Cuddy, Fiske, and Glick’s (2008) stereotype content model that shows how
stigmatized groups (e.g., elderly folk) compensate their lack of attributed competence by
elevated attributions of warmth. However, work in this tradition has, to date, focused on
ratings of the conceptual labels of different social groups (e.g., Fiske, Cuddy, Glick, & Xu,
2002) and not invoked vocal characteristic of them. A question then arises (see Ryan,
1979) as to why nonstandard language varieties persist. A plausible answer to that ques
tion is provided by the role that language plays for social identity. Because Chapter 3 is
fully dedicated to the topic of language and identity, we will cover it only briefly here.
People are cognitively complex, and they simultaneously have (or are attributed as hav
ing) different social identities to call on, with (positive) social identities being crucial for
their well-being. Social identity theory (SIT; Tajfel & Turner, 1979) addresses such identi
ty issues and provides a fruitful theoretical frame for understanding language attitudes,
both with respect to the speaker and the listener (see Ryan, Hewstone, & Giles, 1984).
Page 8 of 29

Variation
Ethnolinguistic identity theory (e.g., Giles & Johnson, 1981)—a language-oriented theory
drawing on SIT and communication accommodation theory—underscores the language–
identity bond. According to this model, language performs central psychological functions
of enhancing a speaker’s positive social identity, especially with respect to ethnicity.
Therefore, it is not surprising that the language variety used by others is a very powerful
cue in determining who is a member of one’s in- or outgroup (Yzerbyt, Leyens, & Bellour,
1995; see also Reid et al., 2012). Regarding this last example, participants showed a
strong overexclusion effect (i.e., more promptly excluded those speakers with a not clear
ly identifiable accent) when asked to identify other ingroup members. Consequently, en
counters might be considered more in intergroup terms when the two parties do not ac
knowledge each other as individuals but as members of (different) social groups (see
Cargile & Giles, 1997; for further development of this in terms of so-called interpersonal
communication, see Dragojevic & Giles, 2014).
Thus, whereas it might be objectively difficult to stop speaking with a nonstandard (or
nonnative) accent, from a sociopsychological point of view, it might actually be counter
productive even to try because accent constitutes a fundamental part of an individual’s
social identity and cultural heritage. Indeed, positive social identities are essential for en
suring positive self-esteem and well-being. As previously established, nonstandard (espe
cially immigrant) accented speakers can, over time, feel less of a sense of belonging to
their host society and, as a result, experience alienation and a depleted ingroup image.
However, this could be mediated and alleviated by identification with one’s own ingroup
of nonstandard accent speakers. This was supported for a group of African Americans
who reported being recipients of discrimination in ways that negatively influenced their
well-being but, at the same time, promoted a stronger identification with the ingroup
which, in turn, assisted in restoring a positive well-being (Branscombe, Schmitt, & Har
vey, 1999; see also Wright & Bougie, 2007).
Language Attitudes: Self-Categorization and

Frames of Reference
As witnessed in the foregoing section, language attitudes are not fixed cognitive struc
tures. Language attitude shifts have most often been studied as a function of external fac
tors, such as changes in intergroup relations and shifts in government language policies
(e.g., Woolard & Gahg, 1990). For example, Genesee and Holobow (1989) examined how
Canadians’ language attitudes toward French and English had changed in the years fol
lowing 1977 legislation that made French the official language of Quebec province. They
found that, (p. 17) compared to 30 years earlier (see Lambert et al., 1960), French Canadi
ans were less likely to upgrade English speakers on solidarity traits, although they contin
ued to attribute them with more status relative to French speakers. However, language
attitudes may also change more dynamically as a function of the social categorization
process itself.
Page 9 of 29

Variation
Specifically, language cues (such as accents) can index multiple identities at different lev
els of abstraction. For example, an American Southern English (ASE) accent indexes a re
gional identity (e.g., Southerner) and a national identity (e.g., American). As a result, to
the extent that speakers and listeners share some of these identities (e.g., Californians
and ASE speakers share a national identity, but have different regional identities), the
same language cue can be used to categorize a speaker as an outgroup member or as an
ingroup member, depending on the level of categorization salient to listeners.
These different levels of categorization are likely to become salient to listeners at various
times, as manifest in self-categorization theory (SCT; Hogg & Reid, 2006; Turner et al.,
1987). In other words, social categorization is a subjective process that depends on the
contrasts that are perceptually most obvious and meaningful in a given social context or
frame of reference. In turn, these different categorizations are likely to influence listen
ers’ language attitudes toward the speaker. Relatedly, SIT posits that a part of people’s
self-concept derives from their social group membership (i.e., social identities). As a re
sult, people strive to maintain a positive social identity in an effort to enhance their self-
esteem. One way people can establish positive social identities is through favorable com
parisons of their ingroup against relevant outgroups. This results in ingroup favoritism, or
the tendency of people to show a preference and affinity for members of their own group
over members of other groups (see Hewstone, Rubin, & Willis, 2002). In other words, lis
teners are likely to have more favorable language attitudes toward speakers when they
categorize them as ingroup versus outgroup members, and these categorizations are like
ly to change as a function of the frame of reference in which they occur.
Consistent with this rationale, Abrams and Hogg (1987) found that listener judges in
Dundee, Scotland, evaluated speakers with a Glasgow accent (a local Scottish variety)
less favorably in terms of status and solidarity when they were compared to speakers
with a Dundee accent than when they were compared to speakers with an RP accent (i.e.,
standard British English). The authors argued that the different comparison groups (i.e.,
Dundee or RP accents) provided Dundee listeners with different frames of reference and,
thus, engendered them to categorize the Glasgow-accented speakers differently. Specifi
cally, they argued that the first situation made a local identity (i.e., Dundee vs. Glasgow)
salient in which Glasgow speakers constituted the outgroup. In contrast, the second situa
tion made a regional identity (i.e., Scotland vs. England)—or perhaps national (McIntosh,
Sim, & Robertson, 2004) identity—salient in which Glasgow speakers now constituted the
ingroup (i.e., Scots) and RP speakers the outgroup (i.e., English). In other words, the
more favorable language attitudes toward Glasgow speakers in the latter situation re
flected ingroup favoritism.
Drawing on this paradigm and using the MGT, Dragojevic and Giles (2013) presented Cali
fornian listener-judges with moderate and broad American Southern English accented
guises. In different conditions, these were paired either with a Californian-accented (in
terregional frame of reference) or Punjabi-accented English speaker (international frame
of reference). Listeners reported a stronger sense of connection with the Southern-ac
cented guises, perceived their accents as more similar to their own, and evaluated them
Page 10 of 29

Variation
more favorably in terms of solidarity traits when the frame of reference was international
(i.e., ingroup categorization) rather than interregional (i.e., outgroup categorization).
These scholars labeled this phenomenon the “reference framing effect.”
Although who is in the evaluative frame can influence listeners’ perceptions of linguistic
similarity–dissimilarity with other speakers (foreign and native), the degree of their per
sonally felt vulnerability to disease can mediate this process. Working from a survival
model of disease avoidance and utilizing voices from the International Dialects of English
Archive (http://web.ku.edu/~idea/), Reid et al. (2012) found that Californians who were
high on a scale of pathogen disgust rated Floridian speakers more linguistically similar to
themselves and more dissimilar from foreign-accented Sierra Leonan and Scottish speak
ers than those who were low on pathogen disgust. This effect was most pronounced when
the Californians had just viewed disease-imposed images rather than other kinds of
threatening projections (viz., gun primes). As these authors intriguingly wrote, “our find
ings provide evidence for a novel psychological process—cognition tracks perceived ac
cent dissimilarity because (p. 18) accents are a cue to infection risk….Accents appear to
be used as a cue to track these dynamic infection risks that are posed by outgroups” (p.
477).
Language Attitudes: Beyond the Accent

Voices are rarely present in a social vacuum but rather are embedded in the content of a
message and the name and appearance of the speaker, as evident in matched-guise stud
ies employing virtual visual characters (e.g., Nazarian, Khoosbabeh, & Dehghani, 2012),
as well as the threat they might pose and the emotions they experience as evidenced in
the Reid et al. (2012) study just described.
Generally, language attitudes studies keep the content of what is being conveyed con
stant. Heaton and Nygaard (2011) investigated the evaluation of Southern American and
standard American accents as a function of two different types of messages, either typical
(hunting and cooking) or untypical (medical and investment) of many Southerners’ dis
course. Results indicated that both the accent and the content of the message had an in
fluence on speakers’ evaluations. Both the Southern content of a message and the accent
were rated lower in status (e.g., as less intelligent, competent, and educated) and higher
in sociability attributes (e.g., generous, cheerful, and friendly) than were standard accent
speakers or a non-Southern passage. Interestingly, the sociability attributes varied as a
function of the content of the message only for standard accent speakers; those speaking
about typical Southern topics were perceived as more sociable than those speaking about
non-Southern topics.
In addition to accent, the very name of a speaker can be a potential clue to his or her eth
nic origins. Purkiss, Perrewé, Gillespie, Mayes, and Ferris (2006) showed an interaction
effect of both accent type and a person’s name on the evaluation of speakers, as well as
on their hiring potential. More specifically, less favorable evaluations were attributed to
those speakers who spoke with a Hispanic accent and had a Hispanic name, whereas the
Page 11 of 29

Variation
most positive evaluations were afforded standard accented speakers with a Hispanic
name. Hence, whether the reason for such a positive evaluation of standard accent speak
ers with a Hispanic name was due to different evaluation standards for immigrants and
natives (e.g., Biernat & Manis, 2007) or to a positive reaction to their presumed linguistic
accommodations, these findings show that more secondary cues (e.g., name) can play an
important role in the evaluation of speakers.
In everyday discourse, listeners have access to a speaker’s appearance and facial fea
tures. From research on person perception, it is known that physical attributes, just like
vocal ones, influence the evaluation of a given person (Dion, Berscheid, & Walster, 1972).
For instance, faces considered beautiful are associated with more positive personality
traits, such as intelligence or likeability. However, Zuckerman, Miyake, and Hodgins
(1991) found that a person’s face previously judged as attractive lost out to his or her be
ing viewed as overall socially attractive when presented with what was perceived to be a
less attractive voice (Krahmer & Swerts, 2007; see also, Giessner, Ryan, Schubert, & van
Quaquebeke, 2011).
Relatedly, and referring to their own empirical work regarding ethnic categorization, Rak
ić, Steffens, and Mummendey (2011b) found that “it was rather irrelevant for participants
what targets looked like; it mainly mattered whether they were speaking with an accent
or not. In this case it was almost as if participants became blind to the visual category in
formation in the presence of more meaningful auditory category information” (p. 24). This
does not mean, of course, that visual information about ethnicity is irrelevant. However,
in some cases, ethnicity may be more “visible” and potent when presented through ac
cent than by looks (for an evolutionary explanation of this, see Kinzler et al., 2010).
Whether this would hold with ethnic groups other than those studied in the German con
text of Rakić et al. (2011b) is an empirical question, but there are related findings point
ing to the possibility that it might be easier to ignore visual than vocal information (e.g.,
Kurzban, Tooby, & Cosmides, 2001).
Gratifyingly, systematic investigations of interactions between accent and other cues are
on the upswing (e.g., Freeman & Ambady, 2011a; Ko, Judd, & Blair, 2006). Recently, theo
retical advances have been made in accounting for people’s ability to manage complex so
cial cues, and the so-called dynamic interactive theory of person construal (Freeman &
Ambady, 2011b) suggests that when trying to understand how people are perceived, one
should account for the multiplicity of cues that need to be processed. Therefore, rather
than trying to argue that face or voice is more important, we should strive to use both
these (and other) cues to be able to understand the underlying processes leading to nega
tive evaluation of, for example, nonstandard accented speakers.
In addition, the actual context as manifest in the prevailing linguistic landscape (Bourhis
& Landry, (p. 19) 1997) in which minority members live can influence their evaluations of
standard and nonstandard speakers. Dailey et al. (2005) asked listener-judges to evaluate
a radio speaker who spoke either in standard American or American with a Hispanic ac
cent. Generally, results were a function of the perceived local linguistic climate (i.e., me
Page 12 of 29

Variation
dia, road signs) as being either more Spanish or more English. The perception of a pre
dominately Spanish local climate resulted in less favorable evaluations of the standard
speaker, whereas the opposite was the case for those embedded in a predominately Eng
lish linguistic landscape. These findings reveal radio language as a potentially important
component of the local vocal landscape (see Coluzzi, 2012), as well as open up the rele
vance of dimensions of the environment and social ecology for our wider understanding
of how language attitude processes operate (see Giles, Katz, & Myers, 2006).
Finally, when a social group is linked to a negative stereotype with regard to language
competencies (as in the case of nonstandard accented speakers), it might be suggested
that these individuals will perform a given task below their actual abilities. The literature
on stereotype threat can be fruitfully invoked here. Referring to skateboarders, older
adults, and gang members as exemplars, Steele (1997) wrote that “where bad stereotypes
about these groups apply, members of these groups can fear being reduced to that stereo
type. And for those who identify with the domain to which the stereotype is relevant, this
predicament can be self-threatening” (p. 614). One possible explanation for why stereo
type threat negatively impacts, for example, intellectual performance is given by the fact
that the immediate salience of a negative stereotype triggers a disruptive mental load
that interferes with the actual task performance (e.g., Croizet et al., 2004). Because lan
guage is an essential part of a social identity, negative stereotypes about the given group,
as well as personal relevance of the task at hand (i.e., wish to perform well), might actual
ly hinder performance.
Even an otherwise insignificant reminder of own gender (by providing gender data before
a test) is enough to trigger the negative stereotype about women’s performance in math,
resulting in their lower performance. Not only math performance, but also language-re
lated tasks can be hindered by stereotype threat. For instance, von Hippel, Wiryakusuma,
Bowden, and Shochet (2011) showed that after being exposed to a negative stereotype
with regard to women’s leadership qualities, women pretending to talk to a subordinate
tended to be judged as more direct and instrumental in their communication (i.e., a “mas
culine” communication style). This, however, was not the case for women not being faced
with this negative stereotype or those who, after the exposure, had to think about a value
that was more important to them. These findings indicate that sometimes actively engag
ing in defiance of a given negative stereotype (i.e., lower leadership qualities of women)
can result in counterproductive strategies, resulting in even poorer evaluations and, ironi
cally, stereotype confirmation. One possible explanation could be offered by communica
tion expectations and their violation (Burgoon & LePoire, 1993); women are not expected
to adopt a masculine communication style and are therefore penalized after unexpectedly
doing so.
Relatedly, using a second language can be affected by negative stereotypes (Paladino et

al., 2009). In Italy, there is an official bilingual region where both Italian and German
have the status of official languages. However, only a relatively small proportion of the
population can be seen as truly bilingual, and a minority of native Italian speakers are
negatively stereotyped for their supposedly impoverished German-language (L2) compe
Page 13 of 29

Variation
tencies. After being exposed to a negative stereotype about their own ingroup, those Ital
ian participants who associated high importance to L2 competence performed worse in a
German proficiency test than did those for whom this was not a self-relevant domain or
who were not primed with the negative stereotype. This was also true if the stereotype
threat manipulation was accomplished in a subtler way, as when participants were led to
perceive their group as being generally a less advantaged group in the region. In this
case, the participants’ pronunciations suffered from stereotype threat activation and as
sociated levels of anxiety and arousal. A case in point here might be a traffic stop in
which a Caucasian police officer engages a Hispanic driver. In this instance, the latter’s
driving skills—and subsequently communicative competence in dealing with this multiple
confrontational threat (e.g., perhaps even legal immigrant status)—could stressfully rein
force the person’s nonstandard dialect which, in turn, could yield negative language atti
tudes and consequent nonaccommodation from the officer (see Giles, Linz, Bonilla, &
Gomez, 2012). The interactively cyclical nature of language attitudes featuring in ongoing
discourse is explored more later.
Not only nonnative speakers but also speakers of nonstandard dialects more generally
are faced with (p. 20) similar negative stereotypes about their language competences.
Hence, an unfortunate outcome of an encounter between standard and nonstandard
speaker may be triggered by even a slight or ambiguous comment or even gesture (from
the former), making the negative stereotype salient to the latter. This, in turn, would have
a negative impact on speech production for a nonstandard speaker, which would other
wise not have been the case. Ironically, his comment might actually be a more “positive”
remark about how well a nonstandard accented individual is speaking or how hard it
must be speaking in a second language.
Nonetheless, to announce to a given speaker that he or she “has an accent” (i.e., speaks
with nonstandard accent) means to criticize them. As indicated in the receiver model of
language criticism (Marlow, 2010), there are different ways in which the speaker might
deal with more or less explicit criticism of his or her nonstandard accent (Marlow & Giles,
2010). Whereas some might be more adaptive or accommodative, others might experi
ence repressed anger and avoid similar encounters in future. Moreover, the fact that
speakers of nonstandard varieties are negatively evaluated and often subjected to criti
cism, in the long run, only reinforces the stigma applied to nonstandard speakers
(Gluszek & Dovidio, 2010b). What extreme consequences this might have in some cases
became clear very recently when a Korean student shot his university colleagues who had
allegedly teased and criticized him in regard to his way of speaking English (Elias, 2012).
In sum, both speaker and listener, unwittingly (yet, on occasion, perhaps strategically) in
fluence the reaction and performance of the other, as well as the evaluation of oneself.
Furthermore, this link to the stereotype threat literature underscores the role of emo
tions in processing language attitudes. Indeed, and although only featuring in a small
number of studies to date (e.g., Cargile & Giles, 1997; Giles, Williams, Mackie, & Rosselli.
1995; Reid et al., 2012), it has been regarded as a central part of the attitude concept
Page 14 of 29

Variation
(e.g., Smith, 1993) and can have a very strong influence on our perceptions, as well as on
our interactions with others (see Ufkes, Otten, van der Zee, Giebels, & Dovidio, 2012).
Toward an Interaction-Based Model of Lan

guage Attitudes
In the earliest days of MGT studies, the approach—and hence the language attitude do
main in general—was built on Lambert’s (e.g., 1967) theoretical premise that perceived
language features → social categorization → trait attributions. In like manner, most stud
ies’ endpoints are judgments of speakers on tape via questionnaire rating scales. In other
words, they are typically noninteractive in design. As alluded to earlier, Marlow and Giles
(2010), however, investigated the ways in which listeners can comment across a variety of
social, educational, and organizational settings—sometimes critically and severely—when
encountering the language varieties of others (in this case, English and Hawaiian pidgin)
and how the latter can variably manage such criticisms (see Marlow, 2010). Models have
been developed that explore how language attitudes are integral to processing social in
teractions (e.g., Street & Hopper, 1982), particularly as they relate to the interpersonal
history of speakers and their listeners (Cargile & Bradac, 2001; Cargile, Giles, Ryan, &
Bradac, 1994; Gluszek & Dovidio, 2010a).
Giles and Coupland (1991) presented their own model of the MGT, pointing to the fact
that forming language attitudes can be a constructive, interpretive process. In expanding
and elaborating on this position, as well as in building on the so-called communication
ecology model (see Giles et al., 2006), Giles and Marlow (2011) argued for a dynamic
model of language attitudes that was:
much more than [about] anonymous reports on questionnaires, produced on de

mand about unknown speakers in socially sterile, task-oriented settings. Rather,…
[language attitudes] can also…feature more creatively as cogent elements within
stories told, arguments rendered, and personas performed. In any case, language
attitudes typically fulfill vital personal, rhetorical, and social needs whose origins
are borne out of past, yet evolving, community concerns. (pp. 182–183)
Space precludes a detailed exposition of this complex model of interlocking processes;

however, one central element of it—sense-making—is important to the extent that people
endeavor “to create order and make retrospective sense of what occurs” (Weick, 1993, p.
635). In other words, the model appeals to the ways in which people try to make sense of
their interactions, relationships, and the communicative styles of others based on expec
tations and their preconceived notions. It also appeals to the way in which language atti
tudes can be expressed, not in a verbal vacuum as a simple statement, if you like, but as
an integral component of a measured argument (see Giles, 1992; Painter & Dixon, 2013).
Page 15 of 29

Variation
To develop further sense-making processes, Giles (2011) invoked the theory of mo
(p. 21)
tivated information management (TMIM; e.g., Afifi, 2009) and, as a backdrop, offered a
scenario similar to this:
(German) Local: “Where you from, then?”
(Apparent) OUTSIDER [smiling confidently]: “Jena, Germany”
LOCAL: [indicating somewhat perplexed]: “No, I meant where you really from?”
OUTSIDER [indicating despair]: “Yeh, I’m really from Jena!”
LOCAL: [somewhat agitated] “No, you have an accent, where did you originally
come from?”
OUTSIDER [somewhat reluctantly]: “OK, Croatia, but I have been in Germany for
many years.”
LOCAL: [somewhat surprised and confused]: “Oh?!”
OUTSIDER [somewhat assertive]: “Actually, it’s you who have the accent!”
LOCAL [somewhat indignant]: “What? I don’t have an accent!”
OUTSIDER [somewhat arrogantly]: “I’m a language scholar, believe me, I study

this kind of thing and you, like everyone else, does have an accent!”
As is apparent throughout this chapter, this dialogue conveys a not uncommon conversa
tional event routinely endured by those who have an accent different from local residents
—and this has a number of constituent features that go beyond the traditional MGT para
digm for assessing language attitudes.
First, language attitudes can be a discursive ongoing event in social life—and not merely
a reactive pencil-and-paper task under controlled conditions. The local above positions
herself as linguistically dissimilar from the outsider who actually happens to be a natural
ized citizen of long standing. Second, such a linguistic conundrum can be open to pro
longed negotiation, overt comment, and sometimes interrogation: at the outset, the out
sider does not feel the need to explain herself as anything but “German” and, later, con
tends that she (erroneously) does not have “an accent.”
Third, the local (as indicated by the nonverbal signals) is expressively and emotively in
volved in sense-making: how can this person be a real German? Moreover, and in the lan
guage of TMIM, the local experiences an “uncertainty gap.” In other words, the local per
haps ponders why this outsider has not converged toward sounding like “a proper Ger
man” and/or may wonder why she has held onto her heritage pronunciations—and seem
ingly so tenaciously! Is it because the outsider does not identify with German values and
even brings them implicitly into question?! Fourth, the outsider feels, by being confronted
Page 16 of 29

Variation
and questioned about the integrity of her parlance, that the local is inherently criticizing
her.
Developing these ideas further, and in ways that might theoretically enhance future stud
ies, we call now more directly on the TMIM as a useful heuristic for exploring language
attitudes enacted in situations in which uncertainty gaps are evident and that require in
formation management and sense-making. TMIM describes information management as a
process that unfolds over three stages: interpretation, evaluation, and decision making. In
the interpretation phase, people become aware of a difference between the level of uncer
tainty they have about a topic or, in our case, the relationship between accent usage and
social origins, and the level of uncertainty they would like to have. Our local certainly ex
periences something like this, as indicated by the quoted dialogue (as well as perhaps ex
periencing a quick musing about accent diversity and national identity). This uncertainty
discrepancy is linearly associated with anxiety (or other emotions alluded to in the dia
logue), which is a partial mediator of the relationship between the uncertainty discrepan
cy and the information management process (see Fowler & Afifi, 2011).
Once in the evaluation phase, people assess a range of options open to them by consider
ing the outcomes that a search for information—or, alternatively, a deliberate avoidance
of it—may yield (outcome assessments) and their ability or self-efficacy to gain the infor
mation they seek or not. Clearly, our outsider exuded self-efficacy by unexpectedly trump
ing the local’s initial feelings of being in command of the conversation by alleging their
professional credentials. TMIM predicts a positive relationship between efficacy judg
ments and outcome assessments, with the strength of the relationship determined by the
valence of the outcome expectancies. The ultimate meta-communicative decision as to
whether to exchange views about each other’s accent usages and, if so how, would then
follow. The position being advanced here is that the language attitudes paradigm can in
volve parallel processing by both speaker and listener-judge who can reciprocally assume
each of these roles.
Finally, there are some important features of this TMIM-inspired model of lan
(p. 22)
guage attitudes that also might indicate valuable (and other) avenues of further work:
• Language attitudes can be ecologically richer than that caricatured in the original,
albeit, seminal MGT studies.
• The stance taken reaffirms the importance of emotions in language attitudes by al
luding to the mediating roles of disrespect, shame, disgust, anger, pity, and so forth.
• It affords communicative self-efficacy (as well as the different kinds of it, see Fowler
& Afifi, 2012) a central role in the language attitudes process.
• The model underscores language attitude situations as being often negotiative ones
involving interpretive, sense-making processes. This can lead to decisions about estab
lishing a dialogue about some, and criticizing (or praising) the others’ communicative
practices, or withholding or avoiding such discourse, and so forth.
Page 17 of 29

Variation
Obviously, further theoretical refinements and elaborations are in order for this model, as
well as for its relationships with extant others (e.g., Dovidio & Gluszek, 2012; Giles &
Marlow, 2011) to flourish as a more robust theoretical frame for language attitudes. How
ever, this motivated information management model of language attitudes implicates new
variables and processes that can fruitfully be applied in future empirical research.
Conclusion
As a summary snapshot of work in this arena that is so central to the social psychology of
language, we conclude by proposing a number of heuristic principles that might not only
organize future reviews of language attitudes work but may also excite further refine
ments and elaborations of the propositions listed here. These are reconstituted and em
bellished after an initial foray into this vein by Dragojevic et al. (2013) who framed their
principles in line, in part, with the theme of language ideologies.
• Voices and speech styles can be socially diagnostic, sometimes accurately and other
times stereotypically, of others’ attributes.
• Such language attitude schemas, particularly as they relate to accent and dialect
use, can be learned and judgmentally invoked very early in the lifespan.
• Language attitudes, as they relate to accent, can be more predictive in forging social
judgments and shaping applied decisions than other, quite potent, physical cues (e.g.,
appearance).
• Language varieties can be dichotomized in terms of those imbued with natural, cul
tured sophistication on the one hand and those with innate inferiority on the other.
Consequently, the former speakers are attributed with intellectual competence, where
as the latter can be considered cognitively and communicatively inadequate.
• Stigmatized language varieties can, however, fulfill important social identity-enhanc
ing, community-promoting, and bonding/solidarity functions for their speakers, and
such speakers can be attributed by outsiders with traits of social attractiveness, such
as trustworthiness and kindness.
• Language attitudes are, nonetheless, not immutable social schemas and can be
shaped by other language varieties that are in the comparative frame at the time
wherein different social identities can be triggered for listener-judges. Intergroup
(e.g., fluctuating power relations between ethnic and national group) and other social
dynamics can influence longstanding judgmental profiles.
• Language attitudes can be associated with affective meanings (e.g., irritation or sat
isfaction). Yet they can also be ambivalent (and sometimes seemingly contradictory),
causing stress and anxiety that requires information management depending on the
social groups targeted, methods employed, and the plethora of social contexts in which
they are evoked.
• Threats to a speaker’s beliefs about others’ stereotypical views of their ingroup’s
task and communicative proficiencies can lead the former to accentuate nonstandard
Page 18 of 29

Variation
language forms that can induce their recipients to form negative language attitudes
and unfavorable (and sometimes overtly critical) behavioral inclinations toward them.
• Language attitudes are not only evaluative reactions people have about others’
speech styles—or even triggers to react subtly or directly toward them—but are inte
gral elements in how individuals make sense of and manage information about the situ
ations in which they find themselves, as well as assist in constructing narratives, argu
ments, and explanations about the character and behavior of members of certain so
cial categories.
This chapter has overviewed extant and emergent work in the language attitudes tradi
tion and proposed new directions for future research, and (p. 23) especially as con
stituents of interpretive, discursive, argumentative, and even evolutionary processes. Ele
vated to attention were a number of accent studies that suggest that emotion, message
content, the linguistic landscape, the comparative reference frame, and speakers’ facial
features should all be objects of further empirical scrutiny as well as investigated in con
cert with other speech and nonverbal variables. In tandem, we recommend that more the
oretical energy be devoted to the complex roles of comprehension and stereotype threat
in impacting evaluations of accented speakers and how these, in turn, can shape feelings
of ingroup identification and subjective well-being. Further exploration of the ways in
which nonstandard accented speech is favorably or unfavorably viewed as a function of
whether it is regional, ethnic minority, immigrant, class-related, or foreign (not to men
tion the different classes and social histories of variables within each of these categories)
is deserving of unpacking. Finally, we hope that the previously mentioned renaissance of
language attitudes research and the recent focus on process will be further served by the
research and ideas conveyed in this chapter—and especially with appreciation of the fact
that attitudes toward language varieties and their speakers are changing radically in
some regions of the world (see Davies & Bentahila, 2013).
References
Abrams, D., & Hogg, M. A. (1987). Language attitudes, frames of reference, and social
identity: A Scottish dimension. Journal of Language and Social Psychology, 6, 201–213.
Afifi, W. A. (2009). Uncertainty and information management in interpersonal contexts. In

S. Smith & S. Wilson (Eds.), New directions in interpersonal communication research (pp.
94–114). Thousand Oaks, CA: Sage.
Allport, G. W., & Cantril, H. (1934). Judging personality from voice. The Journal of Social
Psychology, 5, 37–55.
Ball, P. (1983). Stereotypes of Anglo-Saxon and Non-Anglo-Saxon accents: Some explana

tory Australian studies with the matched-guise technique. Language Sciences, 5, 163–
184.
Page 19 of 29

Variation
Bayard, D., Weatherall, A., Gallois, C., & Pittam, J. (2001). Pax Americana? Accent attitudi
nal evaluations in New Zealand, Australia, and America. Journal of Sociolinguistics, 5, 22–
49.
Biernat, M., & Manis, M. (2007). Stereotypes and shifting standards: Assimilation and
contrast in social judgment. In D. A. Stapel & J. Suls (Eds.), Assimilation and contrast in
social psychology. (pp. 75–97). New York: Psychology Press.
Bourhis, R. Y., & Landry, R. (1997). Linguistic landscape and ethnolinguistic vitality: An
empirical study. Journal of Language and Social Psychology, 16, 23–49.
Bradac, J. J., & Giles, H. (1991). Social and educational consequences of language atti
tudes. Moderna Språk, 85, 1–11
Branscombe, N. R., Schmitt, M. T., & Harvey, R. D. (1999). Perceiving pervasive discrimi
nation among African Americans: Implications for group identification and well-being.
Journal of Personality and Social Psychology, 77, 135–149.
Burgoon, J. K., & LePoire, B. A. (1993). Effects of communication expectations, actual

communication, and expectancy disconfirmation on evaluations of communicators and
their communication behavior. Human Communication Research, 20(1), 67–96.
Cairns, E., & Duriez, B. (1976). The influence of speaker’s accent on recall by Catholic
and Protestant school children in Northern Ireland. British Journal of Social and Clinical
Cargile, A. C., & Bradac, J. J. (2001). Attitudes towards language: a review of speaker-
evaluation research and a general process model. In W. B. Gudykunst (Ed.), Communica
tion yearbook 25 (pp. 347–382). Mahwah, NJ: Lawrence Erlbaum.
Cargile, A. C., & Giles, H. (1997). Understanding language attitudes: Exploring listener
affect and identity. Language and Communication, 17, 195–217.
Cargile, A. C., Giles, H., Ryan, E. B., & Bradac, J. J. (1994). Language attitudes as a social
process: A conceptual model and new directions. Language and Communication, 14, 211–
236.
Clarke, C. M., & Garrett, M. F. (2004). Rapid adaptation to foreign-accented English. Jour
nal of the Acoustical Society of America, 116, 3647–3658.
Coluzzi, P. (2012). Modernity and globalization: Is the presence of English and of cultural
products in English a sign of linguistic and cultural imperialism? Results of a study con
ducted in Brunei Darussalam and Malaysia. Journal of Multilingual and Multicultural De
velopment, 33, 117–131.
Coupland, N., & Bishop, H. (2007). Ideologized values for British accents. Journal of Soci
olinguistics, 11, 74–93.
Page 20 of 29

Variation
Croizet, J.-C., Després, G., Gauzins, M.-E., Huguet, P., Leyens, J.-P., & Méot, A. (2004).
Stereotype threat undermines intellectual performance by triggering a disruptive mental
load. Personality and Social Psychology Bulletin, 30, 721–731.
Crystal, D. (2005). The Cambridge encyclopedia of language. In D. Crystal (Ed.), The

Cambridge encyclopedia of language (pp. 229–259). Cambridge, UK: Cambridge Universi
ty Press.
Cuddy, A. J. C., Fiske, S. T., & Glick, P. (2008). Competence and warmth as universal trait
dimensions of interpersonal and intergroup perceptions: The stereotype content model
and the BIAS map. In M. P. Zanna (Ed.), Advances in experimental psychology (vol. 40, pp.
61–149). New York: Academic Press.
Dailey, R. M., Giles, H., & Jansma, L. L. (2005). Language attitudes in an Anglo-Hispanic
context: The role of the linguistic landscape. Language and Communication, 25, 27–38.
Davies, E., & Bentahila, A. (2013). Language attitudes in the Maghreb region of North
West Africa. In H. Giles & B. M. Watson (Eds.), The social meanings of languages, di
alects, and accents: An international perspective (pp. 84–104). New York: Peter Lang.
Dion, K., Berscheid, E., & Walster, E. (1972). What is beautiful is good. Journal of Person
ality and Social Psychology, 24, 285–290.
Dixon, J. A., Mahoney, B., & Cocks, R. (2002). Accents of guilt? Effects of regional accent,
race, and crime type on attributions of guilt. Journal of Language and Social Psychology,
21, 162–168.
Dovidio, J. F., & Gluszek, A. (2012). Accent, nonverbal behavior, and intergroup bias. In H.
Giles (Ed.), The handbook of intergroup communication (pp. 87–99). New York: Rout
ledge.
Dragojevic, M., & Giles, H. (2013). The reference framing effect: An intergroup
(p. 24)
perspective on language attitudes. Human Communication Research, in press.
Dragojevic, M., & Giles, H. (2014). Language and interpersonal communication: Their in
tergroup dynamics. In C. R. Berger (Ed.), Handbook of interpersonal communication (pp.
29–51). Berlin: De Gruyter Mouton.
Dragojevic, M., Giles, H., & Watson, B. M. (2013). Language ideologies and attitudes: A
dynamic foundational framework. In H. Giles & B. M. Watson (Eds.), The social meanings
of languages, dialects, and accents: An international perspective (pp. 1–25). New York:
Peter Lang.
Edwards, J. R., & Jacobsen, M. (1987). Standard and regional standard speech: Distinc
tions and similarities. Language in Society, 16, 369–380.
Page 21 of 29

Variation
Elias, P. (April 3, 2012). Oikos University shooting: Suspect in deadly attack was upset
about being teased over poor English skills, police say. Huffington Post. Retrieved from
http://www.huffingtonpost.com.
Ellis, D. S. (1967). Speech and social status in America. Social Forces, 45, 431–437.
Fincher, C. L., & Thornhill, R. (2008). A parasite-driven wedge: Infectious diseases may
explain language and other biodiversity. Oikos, 117, 1289–1297.
Fiske, S. T., Cuddy, A. J. C., Glick, P., & Xu, J. (2002). A model of (often mixed) stereotype
content: Competence and warmth respectively follow from perceived status and competi
tion. Journal of Personality and Social Psychology, 82, 878–902.
Floccia, C., Butler, J., Girard, F., & Goslin, J. (2009). Categorization of regional and foreign
accent in 5- to 7-year-old British children. International Journal of Behavioral Develop
ment, 33, 366–375.
Floccia, C., Goslin, J., Girard, F. d. r., & Konopczynski, G. (2006). Does a regional accent
perturb speech processing? Journal of Experimental Psychology: Human Perception and
Performance, 32, 1276–1293.
Fowler, C., & Afifi, W. A. (2011). Applying the theory of motivated information manage
ment to adult children’s discussions of caregiving with aging parents. Journal of Social
and Personal Relationships, 28, 507–535.
Freeman, J. B., & Ambady, N. (2011a). A dynamic interactive theory of person construal.
Psychological Review, 118, 247–279.
Freeman, J. B., & Ambady, N. (2011b). When two become one: Temporally dynamic inte
gration of the face and voice. Journal of Experimental Social Psychology, 47, 259–263.
Fuertes, J. N., Gottdiener, W., Martin, H., Gilbert, T. C., & Giles, H. (2012). A meta-analysis
of speakers’ accents on interpersonal evaluations. European Journal of Social Psychology,
42, 120–133.
Garrett, P. (2010). Attitudes to language. Cambridge, UK: Cambridge University Press.
Giessner, S. R., Ryan, M. K., Schubert, T. W., & van Quaquebeke, N. (2011). The power of
pictures: Vertical picture angles in power pictures. Media Psychology, 14, 442–464.
Genesee, F., & Holobow, N. E. (1989). Change and stability in intergroup perceptions.
Journal of Language and Social Psychology, 8, 17–38.
Giles, H. (1970). Evaluative reactions to accents. Educational Review, 22, 211–227.
Giles, H. (1973). Communicative effectiveness as a function of accented speech. Speech

Monographs, 40, 330–331.
Page 22 of 29

Variation
Giles, H. (1992). Current and future directions in sociolinguistics: A social psychological
contribution. In K. Bolton & H. Kwok (Eds.), Sociolinguistics today: International perspec
tives (pp. 361–367). London: Routledge.
Giles, H. (2011). Language attitudes: The role of motivated information management.

Canadian Issues, Fall, 28–34.
Giles, H., & Coupland, N. (1991). Language attitudes: Discursive, contextual, and geron
tological considerations. In A. G. Reynolds (Ed.), Bilingualism, multiculturalism, and sec
ond language learning: The McGill conference in honor of Wallace E. Lambert (pp. 21–
42). Hillsdale, NJ: Lawrence Erlbaum.
Giles, H., Harrison, C., Creber, C., Smith, P. M., & Freeman, N. H. (1983). Developmental
and contextual aspects of children’s language attitudes. Language and Communication, 3,
141–146.
Giles, H., Henwood, K., Coupland, N., & Harriman, J. (1992). Language attitudes and cog
nitive mediation. Human Communication Research, 18, 500–527.
Giles, H., & Johnson, P. (1981). The role of language in ethnic group relations. In J. C.
Turner & H. Giles (Eds.), Intergroup behavior (pp. 199–243). Oxford, UK: Blackwell.
Giles, H., Katz, V., & Myers, P. (2006). Language attitudes and the role of communication
infrastructures: A communication ecology model. Moderna Språk, 100, 38–54.
Giles, H., Linz, D., Bonilla, D., & Gomez, M. (2012). Police stops of and interactions with
Latino and White (Non-Latino) drivers: Extensive policing and communication accommo
dation. Communication Monographs, 79, 407–427.
Giles, H., & Marlow, M. L. (2011). Theorizing language attitudes: Past frameworks, an in
tegrative model, and new directions. In C. Salmon (Ed.), Communication yearbook 35 (pp.
161–197). Thousand Oaks, CA: Sage.
Giles, H., & Street, R. L., Jr. (1985). Communicator characteristics and behavior: A re
view, generalizations, and a model. In M. L. Knapp & G. R. Miller (Eds.), The handbook of
interpersonal communication (pp. 205–262). Beverly Hills, CA: Sage.
Giles, H., & Watson, B. M. (Eds.). (2013). The social meanings of languages, dialects, and
accents: An international perspective. New York: Peter Lang.
Giles, H., Williams, A., Mackie, D. M., & Rosselli, F. (1995). Reactions to Anglo- and His
panic-American accented speakers: Affect, identity, persuasion, and the English-only con
troversy. Language and Communication, 14, 102–123.
Girard, F., Floccia, C., & Goslin, J. (2008). Perception and awareness of accents in young
children. British Journal of Developmental Psychology, 26, 409–433.
Page 23 of 29

Variation
Gluszek, A., & Dovidio, J. F. (2010a). The way they speak: A social psychological perspec
tive on the stigma of nonnative accents in communication. Personality and Social Psychol
ogy Review, 14, 214–237.
Gluszek, A., & Dovidio, J. F. (2010b). Speaking with a nonnative accent: Perceptions of
bias, communication, and belonging. Journal of Language and Social Psychology, 29, 224–
234.
Gluszek, A., & Hansen, K. (2013). Language attitudes and the Americas. In H. Giles & B.
M. Watson (Eds.), The social meanings of languages, dialects and accents: An internation
al perspective (pp. 26–44). New York: Peter Lang.
Gluszek, A., Newheiser, A. -K., & Dovidio, J. F. (2011). Social psychological orientations
and accent strength. Journal of Language and Social Psychology, 30, 28–45.
Grondelaers, S., van Hout, R., & Steegs, M. (2010). Evaluating regional accent variation
in standard Dutch. Journal of Language and Social Psychology, 29, 101–116.
Guy, G. R. (1988). Language and social class. In F. J. Newmeyer (Ed.), Language:

(p. 25)
The socio-cultural context (pp. 37–63). New York: Cambridge University Press.
Heaton, H., & Nygaard, L. C. (2011). Charm or harm: Effect of passage content on listen
er attitudes toward American English accents. Journal of Language and Social
Hewstone, M., Rubin, M., & Willis, H. (2002). Intergroup bias. Annual Review of Psycholo
gy, 53, 575–604.
Hogg, M. A., & Reid, S. A. (2006). Social identity, self-categorization, and the communica
tion of group norms. Communication Theory, 16, 7–30.
Honorof, D. N., & Whalen, D. H. (2010). Identification of speaker sex from one vowel
across a range of fundamental frequencies. Journal of the Acoustical Society of America,
128, 3095–3104.
Jones, R., R. (1973). Linguistic standardization and national development. International

Journal of Psychology, 8, 51–54.
Kinzler, K. D., Corriveau, K. H., & Harris, P. L. (2011). Children’s selective trust in native-
accented speakers. Developmental Science, 14, 106–111.
Kinzler, K. D., Dupoux, E., & Spelke, E. S. (2007). The native language of social cognition.
Proceedings of the National Academy of Science, 104, 12577–12580.
Kinzler, K. D., Shutts, K., & Correll, J. (2010). Priorities in social categories. European
Journal of Social Psychology, 40, 581–592.
Page 24 of 29

Variation
Ko, S. J., Judd, C. M., & Blair, I. V. (2006). What the voice reveals: Within- and between-
category stereotyping on the basis of voice. Personality and Social Psychology Bulletin,
32, 806–819.
Kinzler, K. D., Shutts, K., Dejesus, J., & Spelke, E. S. (2009). Accent trumps race in guid
ing children’s social preferences. Social Cognition, 27, 623–634.
Krahmer, E., & Swerts, M. (2007). The effects of visual beats on prosodic prominence:
Acoustic analyses, auditory perception and visual perception. Journal of Memory and Lan
guage, 57, 396–414.
Kristiansen, T. (2001). Two standards: One for the media and one for the school. Lan
guage Awareness, 10, 9–24.
Kurzban, R., Tooby, J., & Cosmides, L. (2001). Can race be erased? Coalitional computa
tion and social categorization. Proceedings of the National Academy of Sciences, 98,
15387–15392.
Lane, M. (March 21, 2007). Best of British. BBC News Magazine. Retrieved from http://
news.bbc.co.uk/2/hi/6470095.stm.
Lambert, W. E. (1967). A social psychology of bilingualism. Journal of Social Issues, 23,

91–109.
Lambert, W. E., Hodgson, E. R., Gardner, R. C., & Fillenbaum, S. (1960). Evaluational re
actions to spoken languages. The Journal of Abnormal and Social Psychology, 60, 44–51.
Lass, N. J., Hughes, K. R., Bowyer, M. D., Waters, L. T., & Bourne, V. T. (1976). Speaker
sex identification from voiced, whispered, and filtered isolated vowels Journal of the
Acoustical Society of America, 59, 675–678.
Lev-Ari, S., & Keysar, B. (2010). Why don’t we believe non-native speakers? The influence
of accent on credibility. Journal of Experimental Social Psychology, 46, 1093–1096.
Lindemann, S. (2002). Listening with an attitude: A model of native-speaker comprehen

sion of non-native speakers in the United States. Language in Society, 31, 419–441.
Lippi-Green, R. (1997). English with an accent: Language, ideology, and discrimination in

the United States. New York: Routledge.
Luhmann, R. (1990). Appalachian English stereotypes: Language attitudes in Kentucky.

Language in Society, 19, 331–348.
Marlow, M. L. (2010). Race, power, and language criticism: The case of Hawai’i. 70, Pro
Quest Information & Learning, US. Retrieved from http://search.ebscohost.com/
login.aspx?direct=true&db=psyh&AN=2010-99110-459&site=ehost-live Available
from EBSCOhost psyh database.
Page 25 of 29

Variation
Marlow, M. L., & Giles, H. (2008). Who you tink you, talkin propah? Hawaiian pidgin de
marginalised. Journal of Multicultural Discourses, 3, 53–68.
Marlow, M. L., & Giles, H. (2010). “We won’t get ahead speaking like that!” Expressing
and managing language criticism in Hawai’i. Journal of Multilingual and Multicultural De
velopment, 31, 237–251.
McGlone, M., & Giles, H. (2011). Language and interpersonal communication. In M. L.

Knapp & J. A. Daly (Eds.), Handbook of interpersonal communication (4th ed., pp. 201–
237). Thousand Oaks, CA: Sage.
McIntosh, I., Sim, D., & Robertson, D. (2004). “We hate the English, but not because your
our pal”: A study of anti-English sentiment in Scotland. Sociology, 38, 43–59.
Mehler, J., Bertoncini, J., Barrière, M., & Jassik-Gerschenfeld, D. (1978). Infant recogni
tion of mother’s voice. Perception, 7, 491–497.
Mulac, A., & Giles, H. (1996). “You’re only as old as you sound”: Perceived vocal age and
social meanings. Health Communication, 8, 199–215.
Munro, M. J., & Derwing, T. M. (1995). Processing time, accent, and comprehensibility in
the perception of native and foreign-accented speech. Language and Speech, 38, 289–
306.
Nazarian, A., Khooshabeh, P., & Dehghani, M. (May 2012). The interplay between accent
and warmth on person perception. Poster at the Annual Convention of the Association for
Psychological Science, Chicago.
Nazzi, T., Jusczyk, P. W., & Johnson, E. K. (2000). Language discrimination by English-
learning 5-month-olds: Effects of rhythm and familiarity. Journal of Memory and
Language, 43, 1–19.
Painter, D., & Dixon, J. A. (2013). Language attitudes in Southern Africa. In H. Giles & B.
M. Watson (Eds.), The social meanings of language, dialect and accent: An international
perspective (pp. 105–124). New York: Peter Lang.
Paladino, M. P., Poddesu, L., Rauzi, M., Vaes, J., Cadinu, M., & Forer, D. (2009). Second
language competence in the Italian-speaking population of Alto Adige/Südtirol. Journal of
Language and Social Psychology, 28, 222–243.
Pantos, A. J., & Perkins, A. W. (2013). Measuring implicit and explicit attitudes toward for
eign accented speech. Journal of Language and Social Psychology, 32, 3–20.
Preston, D. (1999). A language attitude approach to the perception of regional variety. In

D. Preston (Ed.), Handbook of perceptual dialectology (pp. 359–373). Amsterdam: John
Benjamins.
Page 26 of 29

Variation
Price, S., Fluck, M., & Giles, H. (1983). The effects of language of testing on bilingual
preadolescents’ attitudes towards Welsh and varieties of English. Journal of Multilingual
and Multicultural Development, 4, 149–161.
Purkiss, S. L. S., Perrewé, P. L., Gillespie, T. L., Mayes, B. T., & Ferris, G. R. (2006). Implic
it sources of bias in employment interview judgments and decisions. Organizational Be
havior and Human Decision Processes, 101, 152–167.
Rakić, T., Steffens, M. C., & Mummendey, A. (2011a). When it matters how you pronounce
it: The influence of regional (p. 26) accents on job interview outcome. British Journal of
Rakić, T., Steffens, M. C., & Mummendey, A. (2011b). Blinded by the accent! The minor
role of looks in ethnic categorization. Journal of Personality and Social Psychology, 100,
16–29.
Ray, G. B. (2009). Language and interracial communication in the United States. New
York: Peter Lang.
Reid, S. A., Zhang, J., Anderson, G. L., Gasiorek, J., Bonilla, D., & Peinado, S. (2012). Para
site primes make foreign-accented English sound more distant to people who are disgust
ed by pathogens (but not by sex or morality). Evolution and Human Behavior, 33, 471–
478.
Riches, P., & Foddy, M. (1989). Ethnic accent as a status cue. Social Psychology Quarterly,
52, 197–206.
Rubin, D. L. (1992). Nonlanguage factors affecting undergraduates’’ judgments of nonna

tive Englisch-speaking teaching assistants. Research in Higher Education, 33, 511–531.
Rubin, D. L., & Smith, K. A. (1990). Effects of accent, ethnicity, and lecture topic on un
dergraduates’ perceptions of nonnative English-speaking teaching assistants. Internation
al Journal of Intercultural Relations, 14, 337–353.
Ryan, E. B. (1979). Why no low-prestige language varieties persist? In H. Giles & R. N. St.
Clair (Eds.), Language and social psychology (pp. 145–157). Oxford, UK: Blackwell.
Ryan, E. B., Giles, H., & Hewstone, M. (1988). The measurement of language attitudes. In
U. Ammon, N. Dittmar & K. J. Mattheier (Eds.), Sociolinguistics: An international hand
book of the science of language (vol. 2, pp. 1068–1081). Berlin: DeGruyter.
Ryan, E. B., Giles, H., & Sebastian, R. J. (1982). An integrative perspective for the study of
attitudes toward language variation. In E. B. Ryan & H. Giles (Eds.), Attitudes towards
language variation: Social and applied contexts (pp. 1–19). London: Edward Arnold.
Ryan, E. B., Hewstone, M., & Giles, H. (1984). Language and intergroup attitudes. In J. R.
Eiser (Ed.), Attitudinal judgment (pp. 135–160). New York: Springer Verlag.
Page 27 of 29

Variation
Scherer, K. R. (1979). Personality markers in Speech. In K. R. Scherer & H. Giles (Eds.),
Social markers in speech (pp. 147–210). Cambridge, UK: Cambridge University Press.
Schoel, C., Roessel, J., Eck, J., Janssen, J., Petrovic, B., Rothe, A., Rudert, S. C., &
Stahlberg, D. (2013). “Attitudes Towards Languages” (AToL) scale: A global instrument.
Sebastian, R. J., Ryan, E. B., Keogh, T. F., & Schmidt, A. C. (1979). Reactions to accents. In
Giles, H., Robinson, W. P., Smith, P. M. (Eds.), Language: Social psychological perspectives
(pp. 203–213). Oxford, UK: Pergamon.
Smith, E. R. (1993). Social identity and social emotions: Toward new conceptualizations of
prejudice. In D. M. Mackie & D. L. Hamilton (Eds.), Affect, cognition, and stereotyping:
Interactive processes in group perception. (pp. 297–315). San Diego, CA: Academic Press.
Steele, C. M. (1997). A threat in the air: How stereotypes shape intellectual identity and
performance. American Psychologist, 52, 613–629.
Stewart. C. M. (2012). Mapping language ideologies in multi-ethnic urban Europe: The

case of Parisian French. Journal of Multilingual and Multicultural Development, 33, 187–
202.
Stewart, M. A., Ryan, E. B., & Giles, H. (1985). Accent and social class effects on status
and solidarity evaluations. Personality & Social Psychology Bulletin, 11, 98–105.
Street, R. L., Jr., & Hopper, R. (1982). A model of speech style evaluation. In E. B. Ryan &
H. Giles (Eds.), Attitudes toward language: Social and applied contexts (pp. 175–188).
London: Academic Press.
Tajfel, H., & Turner, J. C. (1979). An integrative theory of intergroup conflict. In W. G.

Austin & S. Worchel (Eds.), The social psychology of intergroup relations (pp. 33–47).
Monterey, CA: Brooks-Cole.
Tsurutani, C. (2012). Evaluation of speakers with foreign-accented speech in Japan: The

effect of accent produced by English native speakers. Journal of Multilingual and Multi
cultural Development, 33, 589–603.
Turner, J. C., Hogg, M. A., Oakes, P. J., Reicher, S. D., & Wetherell, M. (1987). Rediscover
ing the social group: A self-categorization theory. Cambridge, MA: Basil Blackwell.
Ufkes, E. G., Otten, S., van der Zee, K. I., Giebels, E., & Dovidio, J. F. (2012). The effect of
stereotype content on anger versus contempt in “day-to-day” conflicts. Group Processes
and Intergroup Relations, 15, 55–72.
van Bezooijen, R. (1994). Aesthetic evaluation of Dutch language varieties. Language and
Communication, 14, 253–263.
Page 28 of 29

Variation
von Hippel, C., Wiryakusuma, C., Bowden, J., & Shochet, M. (2011). Stereotype Threat
and Female Communication Styles. Personality and Social Psychology Bulletin, 37, 1312–
1324.
Wappler, M. (February 14, 2012). Nicki Minaj and five other weird Grammy moments.
Santa Barbara NewsPress, D3.
Weick, K. (1993). The Collapse of Sensemaking in Organizations: The Mann Gulch Disas
ter. Administrative Science Quarterly, 38, 628–305.
Williams, F. (1976). Explorations of the linguistic attitudes of teachers. Rowley, MA: New
bury House.
Woolard, K. A., & Gahg, T.-J. (1990). Changing language policies and attitudes in au
tonomous Catalonia. Language in Society, 19, 311–330.
Wright, S. C., & Bougie, È. (2007). Intergroup contact and minority-language education.
Yzerbyt, V., Leyens, J.-P., & Bellour, F. (1995). The ingroup overexclusion effect: Identity
concerns in decisions about group membership. European Journal of Social Psychology,
25, 1–16.
Zahn, C. J., & Hopper, R. (1985). Measuring language attitudes: The speech evaluation in
strument. Journal of Language and Social Psychology, 4, 113–123.
Zuckerman, M., Miyake, K., & Hodgins, H. S. (1991). Cross-channel effects of vocal and
physical attractiveness and their implications for interpersonal perception. Journal of Per
sonality and Social Psychology, 60, 545–554.
Howard Giles
Howard Giles is Professor at the Department of Communication, University of Cali

fornia, Santa Barbara.
Tamara Rakić
Tamara Rakić is a Lecturer at Lancaster University.
Page 29 of 29

Language, Identity, and Culture: Multiple Identity-Based Perspectives
Language, Identity, and Culture: Multiple Identity-

Based Perspectives
Stella Ting-
Toomey and Tenzin Dorjee

Online Publication Date: Apr 2014 DOI: 10.1093/oxfordhb/9780199838639.013.013
Drawing from the intercultural and intergroup communication research literature, the ob
jective of this chapter is to unpack the relationship among language, identity, and culture
from multiple identity-based frameworks. Three theoretical frameworks—identity negotia
tion theory, intergroup communication accommodation theory, and face-negotiation theo
ry—are used to illuminate the interdependent relationship between sociocultural mem
bership issues and language/verbal interaction styles. Overall, a process competence per
spective to the understanding of language, identity, and culture is emphasized. An inter
cultural-intergroup process competence perspective contains two key ideas: being super-
mindful of the symbolic message exchange process between the two intercultural commu
nicators, and being super-mindful in understanding the sociocultural identity dynamics in
language/verbal style enactment in a multilayered cultural system. Throughout the chap
ter, ample examples are used to illustrate the dynamic interdependence among language
variations, sociocultural memberships, and shifting cultural boundary encounter issues.
Keywords: Communication Accommodation, Face Negotiation, Identity Negotiation, Intercultural-Intergroup Com

petence, Mindful
While language is the key to the heart of a culture, nonverbal communication is the heart
beat of a culture. Taken together as a package, individuals can become culturally mindful
communicators by paying responsive attention to the use of language, verbal, and also
nonverbal communication styles in particular cultural situations. Language is defined in
this chapter as an arbitrary, symbolic system that labels and categorizes objects, events,
groups, people, ideas, feelings, experiences, and many other phenomena. Language is al
so governed by the multilayered rules developed by members of a particular sociocultural
community.
In our everyday interaction, we use language to communicate, to agree or disagree with

others, to make or deny requests, and to assert or negotiate our multiple identities. Lan
guage represents a significant identity maker. Identity is conceptualized as having multi
Page 1 of 32

ple social identity, relational role, and personal identity facets (Tajfel, 1978; Tajfel & Turn
er, 1979). Social (or socio-cultural) identities can include ethnic membership identity, so
cial class identity, and professional role identity, to name a few examples. Relational role
identities can include ingroup/outgroup membership identity issues to family role identity
expectations (Ting-Toomey, 1999). Personal identities can include any unique attributes
that we associate with our individuated self in comparison to those of others.
In this chapter, we discuss various identity-based frameworks in the intercultural and in
tergroup communication domains to enrich our understanding of the intricate relation
ship between language variations and sociocultural membership communities. Individuals
mostly acquired their composite identities through socio-cultural conditioning process,
(p. 28) relational development events, individual lived experiences, and repeated commu
nication practices with surrounding others. In every intercultural or intergroup encounter

process, each individual’s multilayered identities such as group membership, relational
role, and personal identity would come into negotiation prominence pending on the con
versation topics, communication channels, relationship types, and situational frames. We
need intercultural sensitive skills to negotiate these identity issues competently.
Drawing from the intercultural and intergroup communication research literature, the ob
jective of this chapter is to unpack the relationship among language, identity, and culture
from multiple identity-based frameworks. Since there are several chapters (e.g., Howard
Giles, Anne Maass, James Pennebaker, and Klaus Fiedler) in this handbook that focus on
and discuss the psychological processes of language use, this chapter will focus on the so
ciocultural identity dimensions that impact on language and verbal styles within a variety
of cultural situations. In particular, three theoretical frameworks: identity negotiation the
ory, intergroup communication accommodation theory (CAT), and face-negotiation theory
will be used to illuminate the interdependent relationship between sociocultural member
ship issues and language/verbal interaction styles. The intergroup CAT is used as a bridge
that can enhance the identity negotiation theory and the face negotiation theory in fur
thering our understanding of the intersections of various identity-based issues and lan
guage choice.
The chapter is organized in five sections: First, a process competence lens to the under
standing of language, verbal interaction, identity, and culture is introduced as a back
drop. Second, the identity negotiation theory and its accompanying assumptions will be
explained. Third, the focal constructs of intergroup perspective and CAT and their rele
vance to language issues will be discussed and specific language issues in a distinctive di
aspora cultural community, the Tibetan speech community, will be probed. Fourth, the
face negotiation theory will be explained and the further connection between CAT and the
culture-based facework behaviors will be proffered. Fifth, directions for future research
concerning the intersection of language, verbal interaction, and sociocultural member
ship identity issues will be recommended.
Page 2 of 32

Language, Verbal Interaction, and Identity: A

Process Competence Lens
An intercultural process competence perspective emphasizes the mindful attention need
ed to understand the underlying identity issues that frame the use of language and ver
bal/nonverbal style variations in particular situations and in a particular identity member
ship community. A particular identity membership community can refer to an immigrant
group, a diaspora group, a co-culture ethnic group, or a dominant-mainstream group. A
process competence lens could shed light on the important role of sociocultural identity
dimensions that shape strategic verbal and nonverbal communication accommodation
process on the macro-level (e.g., cultural ideologies’ influence), exo-level (e.g., institution
al policies’ influence), meso-level (e.g., community values’ influence), and micro-level
(e.g., individual values) interactions (Ting-Toomey & Oetzel, 2013).
An intercultural process competence lens contains two key ideas: (1) being super-mindful
of the symbolic message exchange process and coordinated meaning construction
process via the use of appropriate, effective, and adaptive verbal and nonverbal communi
cation styles, and (2) being super-mindful in understanding more deeply the sociocultural
identity and personal identity issues in conjunction with the role of language usage in a
particular cultural system. We will explicate these two ideas in the following discussion
and a follow-up elaboration on the second idea when discussing the identity negotiation
theory, communication accommodation theory, and face negotiation theory.
A Behavioral Process Awareness Viewpoint
In discussing intercultural communication competence in general, Ting-Toomey (2009)

argues that the criteria of communication appropriateness, effectiveness, and adaptabili
ty can serve as evaluative yardsticks of whether an intercultural communicator has been
perceived as behaving competently or incompetently in an interaction episode (see also,
Spitzberg, Canary, & Cupach, 1994; Spitzberg & Changnon, 2009). The criterion of “com
munication appropriateness” refers to the degree to which the exchanged verbal and non
verbal behaviors are regarded as proper and match the expectations generated by the in
siders of the culture. To behave “properly” in any given cultural situation, competent
communicators need to have the relevant value knowledge schema of the larger situation
al norms that guide the interaction episode. They also need to acquire the specific knowl
edge schema of what constitutes appropriate or inappropriate language usage, word
choice, and verbal/nonverbal style rhythms that can achieve the desired outcome goal.
On the other hand, the criterion of “communication effectiveness” refers to the de
(p. 29)
gree to which communicators achieve mutually shared meaning and integrative goal-re
lated outcomes in an intercultural or intergroup interaction episode. More importantly, ef
fectiveness and appropriateness criteria are positively interdependent. When one man
ages a problematic communication episode appropriately with proper language usage
and with mutual-face sensitivity, the “good faith” proper behaviors can induce communi
Page 3 of 32

cation effectiveness. Likewise, when one promotes an effective mutual-interest goal-di

rected outcome, the conjoint interest goal direction can induce appropriate interaction
behaviors from the other interactional party. Deardorff (2004), in interviewing 23 schol
ars and trainers in the intercultural communication field, identifies the most preferred de
finition of intercultural competence as “the ability to communicate effectively and appro
priately in intercultural situations based on one’s intercultural knowledge, skills and
attitudes” (p. 249).
Last, the criterion of “communication adaptability” refers to our ability to change our in
teraction behaviors and goals to meet the specific needs of the situation. It implies men
tal, affective, and behavioral flexibility in dealing with the intercultural communication
situation. It signals our attunement to the other party’s perspectives, interests, goals, and
verbal/nonverbal communication approach. It also conjures willingness to modify our own
behaviors and goals to adapt to the emergent communication situation. Communication
adaptability connotes dynamic code-switching ability in a problematic interaction scene.
Dynamic cross-cultural code switching refers to the intentional learning and moving be
tween culturally ingrained systems of behavior in particular situation (Molinsky, 2007).
An Identity Orientation Awareness Viewpoint
Individuals from contrasting cultural or group membership communities often bring with
them different value patterns, perceptual biases, and interaction scripts that influence
their interpretations of competent versus incompetent communication behavior in a par
ticular situation. Sharpening the situated content knowledge and communication/lan
guage skills of the intercultural negotiators can enhance their pragmatic competencies.
According to the identity negotiation theory (Ting-Toomey, 1999, 2005a), culture-sensitive
knowledge, mindfulness, and constructive communication skills constitute the key fea
tures of the intercultural identity competence components.
While incorporating “culture-sensitive knowledge,” for example, such as value content

spectrums of individualism-collectivism and small-large power distance issues (Hofstede,
2001; E. S. Kashima & Y. Kashima, 1998; Y. Kashima, E. S. Kashima, Kim, & Gelfand,
2006; Triandis, 1995; Ting-Toomey, 2010a) or cultural tightness/looseness structural is
sues (Gelfand et al., 2011; Kashima & Gelfand, 2012), communicators can learn to uncov
er the implicit “ethnocentric lenses” they use to evaluate the ‘bizarre or unfamiliar” be
haviors in an intercultural interaction scene. Culture-sensitive knowledge also includes
the importance of developing a keen sense of awareness of self-identity issues and other-
identity issues in the communication process, and also the willingness to move beyond
the actual communication encounter process and taking into consideration the larger im
migrants’ acculturation identity change factors and host country’s institutional receptivi
ty factors.
“Mindfulness,” from the intercultural competence framework, means the willingness to

attend to one’s internal cultural and personal communication assumptions, cognitions,
and emotions and, at the same time, becoming exquisitely attuned to the other’s commu
Page 4 of 32

nication assumptions, cognitions, and emotions (Ting-Toomey, 1999). Mindfulness is about

focused meta-cognition, meta-communication, and meta-sensation attunement work
(Ting-Toomey, 2010b, 2013; Siegel, 2007; see also, Ang, Van Dyne, & Tan, 2011). Being
mindful of self-identity and other-identity complexity issues can serve as an intermediary
link between the integration of “culture-sensitive knowledge” and the execution of “con
structive communication skills.”
“Constructive communication skills” refer to our operational abilities to manage a com

munication situation flexibly via pragmatic verbal and nonverbal adaptive behaviors with
an eye to contextual sensitivity. Thus, the development of an intercultural process compe
tence lens includes the intentional integration of culture-sensitive and linguistic-sensitive
knowledge, mindful self-reflexivity and other identity-reflectivity, and the enactment of
constructive communication skills to connect and relate to culturally dissimilar others.
Understanding the role of sociocultural identity membership issues is a major starting
step in enhancing verbal and nonverbal interactional competence and developing optimal
intercultural and intergroup communication performance.
Identity Negotiation Theory: A Boundary-

(p. 30)
Crossing Perspective
To develop pragmatic communication competence in intergroup contexts, individuals
need to understand deeply sociocultural membership identity struggle issues and person
al identity adaptations as they unfold in the communication process and the larger cultur
al system. In this section, we introduce a theoretical framework, the Identity Negotiation
Theory (INT) (Ting-Toomey, 1999, 2005a) to guide us systematically in connecting the re
lationship among language/verbal style, identity, and culture in the context of minority
and immigrants’ acculturation process. In the next section, we forge connection between
the Communication Accommodation Theory (CAT) (Gallois, Ogay, & Giles, 2005; Giles,
Reid, & Harwood, 2010), and INT to further explain the critical role of appreciating lan
guage via an intergroup identity lens.
Identity Negotiation Theory: A Synopsis
By understanding how individuals define themselves and how others define them on mul
tiple grounds, persons can communicate with culturally different others appropriately, ef
fectively, and adaptively. Let us first take a look at the following case story (Ting-Toomey
& Chung, 2012, p. 158) which involved Pauline, a University Assistant Dean and her en
counter with a group of strangers in a posh restaurant on campus.
Mistaken Identity: Innocent or Guilty? I was having lunch at the university restau
rant with my work colleagues when I glanced over at the other table. The table
was beautifully decorated with rose petals and fancy packages. The women that
were going to be seated were immaculately dressed. I could see the Couture,
Chanel and Gucci. I was curious and walked over to their table. “Excuse me, your
Page 5 of 32

table is so beautiful. I was wondering what the special occasion was?” One woman
smiled and replied, “We are celebrating friendship day. We do this every year. By
the way, may I have a glass of ice tea, no cubes please?” I was totally stunned. “I
am so sorry, I did not introduce myself. I am an Assistant Dean in the College of
Arts and Sciences.” The White woman apologized and ended with, “I thought you
were the Maître D—I mean, the Head Maître D.”
As an African American woman who has worked on this campus for over a decade,
I am still disappointed and somewhat dismayed, that after all of these years, color
matters. It is a daily reminder that I am different…
In reviewing the “Mistaken Identity” real-life case story, it is obvious that multiple identi
ty clashes come into play. The story involves ethnic/racial identity clash, particular identi
ty lens interpretation, historical backdrop, professional identity issues, gender role expec
tations, and personal identity sentiments, and much more. It also provides us with a com
pelling argument for why understanding the unfamiliar other’s identity conception is so
critical in promoting competent intercultural or intergroup communication. In an identity
misalignment communication process, individuals can unintentionally or intentionally in
sult someone’s sense of social or personal identity self (see next section) as portrayed in
the dialogue line in which the White woman ended with— “I thought you were the Maître
D—I mean, the Head Maître D.” –thus, inflicting one more verbal insult to the emotional
pain that Dean Pauline has just experienced.
According to Ting-Toomey’s (2005a) INT, human beings in all cultures desire identity re
spect in the communication process. However, what constitutes the proper way to show
identity respect and consideration varies from one culture to the next. The INT perspec
tive emphasizes particular identity domains in influencing individuals’ everyday interac
tions. To illustrate, for example, individuals acquire their sense of cultural/ethnic group
membership images through their primary caretakers, peer associations, schools and me
dia influence during their formative years. Furthermore, physical appearance, racial
traits, skin color, language usage, self-appraisal, and other-perception factors all enter in
to the cultural identity construction equation.
The INT has been researched and applied primarily in immigrants’ acculturation contexts
and intergroup majority-minority interaction contexts (see, for example, Collie, Kindoh, &
Podisadlowski, 2010; Jackson, 1999, 2002) to international adjustment and re-adjustment
situations (see, for example, Hotta & Ting-Toomey, 2013; Molinksy, 2007; Onwumechili,
Nwosu, Jackson, & James-Hughes, 2003). More recently, it has been applied to bicultural/
biracial identity meaning construction and its impact on intergroup communication
strategies in various daily situations (Toomey, Dorjee, & Ting-Toomey, 2013).
Page 6 of 32

Identity Negotiation Theory: Core Assumptions
The 2005 INT version consists of the following ten core assumptions, which explain the
antecedent, (p. 31) process, and outcome components of intercultural identity-based com
munication competence:
1. The core dynamics of people’s group membership identities (e.g., cultural and eth
nic memberships) and personal identities (e.g., unique attributes) are formed via
symbolic communication with others.
2. Individuals in all cultures or ethnic groups have the basic motivation needs for
identity security, inclusion, predictability, connection, and consistency on both group-
based and person-based identity levels. However, too much emotional security will
lead to tight ethnocentrism, and, on the converse side, too much emotional insecurity
(or vulnerability) will lead to fear of outgroups or strangers. The same underlying
principle applies to identity inclusion, predictability, connection, and consistency.
Thus, an optimal range exists on the various identity negotiation dialectical spec
trums.
3. Individuals tend to experience identity emotional security in a culturally familiar
environment and experience identity emotional vulnerability in a culturally unfamil
iar environment.
4. Individuals tend to feel included when their desired group membership identities
are positively endorsed (e.g., in positive in-group contact situations) and experience
differentiation when their desired group membership identities are stigmatized (e.g.,
in hostile out-group contact situations).
5. Persons tend to experience interaction predictability when communicating with
culturally familiar others and interaction unpredictability when communicating with
culturally unfamiliar others. Interaction predictability tends to lead to either further
trust (i.e., within the optimal level) or rigidified stereotyped categories (i.e., beyond
the optimal level). Constant interaction unpredictability tends to lead to either mis
trust or haphazard expectancy surprises.
6. Persons tend to desire interpersonal connection via meaningful close relationships
(e.g., in close friendship support situations) and experience identity autonomy when
they experience relationship separations—meaningful intercultural-interpersonal re
lationships can create additional emotional security and trust in the cultural
strangers.
7. Persons tend to experience identity consistency in repeated cultural routines in a
familiar cultural environment and they tend to experience identity change (or to the
extreme, identity chaos and turmoil) and transformation in a new or unfamiliar cul
tural environment.
8. Cultural-ethnic, personal, and situational variability dimensions influence the
meanings, interpretations, and evaluations of these identity-related themes.
9. Competent identity-negotiation process emphasizes the importance of integrating
the necessary intercultural identity-based knowledge, mindfulness, and interaction
skills to communicate appropriately, effectively, and adaptively with culturally dissim
ilar others.
Page 7 of 32

10. Satisfactory identity negotiation outcomes include the feeling of being under
stood, respected, and affirmatively valued.
Cultural and Ethnic Identity Intersecting Issues
In an accelerated multicultural/multiracial identity formation society—race, ethnicity, reli

gion, social class, and culture—will become an increasingly integrative or fragmented fo
cal point for identity negotiation and re-negotiation. Cultural identity salience can be de
fined as the emotional significance that members attach to their sense of belonging or af
filiation with the larger majority culture (e.g., the larger Australian or U.S. culture). To il
lustrate, we can talk about the larger Australian or Brazilian cultural identity, or the larg
er U.S. cultural identity on the macro sociocultural membership analytical level. On the
other hand, ethnic identity salience is linked closely with the intergroup boundary mainte
nance issue across generations (e.g., third-generation Cuban Americans in the U.S.). Eth
nic identity salience can be defined as the subjective allegiance and loyalty to a group—
large or small, socially dominant or subordinate, with which one has ancestral links. Eth
nic identity can be sustained by shared objective characteristics such as shared language
or religion. It is also a subjective sense of “ingroupness” whereby individuals perceive of
themselves and each other as belonging to the same ingroup by shared historical and
emotional ties. Thus, following Berry, Kim, and Boski’s (1987) 2 X 2 cultural-ethnic identi
ty typological model, an immigrant’s identity can be classified as either an ethnic identity
maintenance type (weak on cultural identity salience, strong on ethnic identity salience),
a bicultural identity type (strong on cultural identity salience, and strong on ethnic identi
ty salience), an assimilated identity type (strong on cultural identity salience, weak on
ethnic identity salience), or (p. 32) a marginal identity type (weak on cultural identity
salience, weak on ethnic identity salience). To practice pragmatic linguistic competence
in an intercultural or intergroup communication situation, one has to move beyond look
ing at the racial physical traits or dialects/accents of the interactional partner, and learn
to observe and listen deeply to the identity stories in the particular encounter scene.
Understanding various sociocultural membership identity issues especially in a pluralistic

immigrant society is like beholding a multilayered, marble cake. Cultural-ethnic member
ship identity issues often inter-mingle or intersect with other group-based and personal-
based identity issues to create an ongoing identity kaleidoscope that impacts on language
enactment and nonverbal behavioral choices. Whether one is communicating with an in
group or outgroup member may highly influence the linguistic code-switching process. In
a nutshell, the INT assumes that human beings in all cultures desire both positive group-
based and positive person-based identities in any type of communicative situation. How
individuals can enhance identity understanding, respect, and mutual affirmative valuation
via mindful verbal and nonverbal competence practice is the essential concern of this ap
proach. Newly-arrived immigrants, minority members, and biracial/multiracial individuals
also often need to learn to use creative and adaptive language code-switching practice to
Page 8 of 32

assert their complex identity shifts in the fluctuating intergroup encounter situations
(Collie et al., 2010; Toomey et al., 2013).
Identity Negotiation Complexity: An Inter

group Perspective
Intergroup perspective augments INT’s perspective on understanding the relationships
among language, communication, identity, and sociocultural membership issues. Identity
negotiation primarily takes place in intergroup and interpersonal communication con
texts in everyday life. Intergroup perspective offers a rich theoretical extension to under
stand identity negotiation in various contexts.
Intergroup Backdrop: Two Identity Types
According to social identity theory (Tajfel, 1978; Tajfel & Turner, 1979), every individual
has two types of identity: social identity and personal identity. Social identity is based on
his or her membership/s in a group or groups, and personal identity is based on individ
ual idiosyncrasies and unique traits (see Giles, Reid, & Harwood, 2010). For example,
while social identity or group role membership identity (e.g., Stella as a full professor, fe
male, mother, and Chinese, and Tenzin as an assistant professor, male, unmarried, and Ti
betan) shapes intergroup professional interactions, personal identity (e.g., Stella being an
optimistic, creative person, and Tenzin being a considerate, empathetic person) shapes
personalized relationship development interactions.
Mounting evidence, however, indicates how both social and personal identity can assert
simultaneous effect on the evolving dynamics of communication between two or more
communicators (see Giles et al., 2010) from distinctive identity groups. Theoretically,
while these interactional contexts could be clearly distinguished, in actuality, intergroup
and interpersonal communication processes fluctuate from one moment to the next—de
pending on the conjoint social identification and differentiation processes, the conversa
tion topics, interactional goals, and critical interaction turning points that pervade in the
ongoing conversation episode.
Given the focus of this chapter, we discuss in this section intriguing relationships be
tween languages and social identity and proffer two theoretical models—one describing
the relationship between bilingualism and linguistic identity, and the other understanding
verbal/nonverbal convergence and divergence relating to intergroup and interpersonal
encounters. While differentiating these two theoretical contexts, we also show how com
municative situations can involve high intergroup salience and high interpersonal
salience. These situations involve complex negotiation of social and personal identities as
well as communication accommodation.
Page 9 of 32

Language and Social Identity Issues
Language and communication issues received early attention from Tajfel and his col
leagues (e.g., Bourhis, Giles, & Tajfel, 1973; Giles, 1978) even though social identity origi
nated in the field of social psychology (Tajfel, 1978; Tajfel & Turner, 1979). Giles and col
leagues (see Giles and Rakic, this volume) produced substantial work on language,
speech, and dialects as markers of social identity and their effect on intergroup relations
(e.g., Giles, 1973, 1978; Giles & Johnson, 1981; 1987). According to Ethnolinguistic Vitali
ty Theory (EthnoVT) (Giles, Bourhis, & Taylor, 1977; Giles, Taylor, & Bourhis, 1977), sta
tus, demographics, and institutional support not only influence intergroup relations be
tween members of different (p. 33) language groups, but they also play a significant role
in language maintenance.
In other words, group vitality—objective vitality or perceived vitality—impact intergroup

interaction, particularly communication accommodation. Group vitality can be defined as
the strength of a group measured along these dimensions: status, demographics, and in
stitutional support. Status includes social and economic status, demographics include
birth and death rates and immigration, and institutional support includes social, econom
ic, and political support from local and central governments as well as other agencies.
Group vitality provides a useful theoretical explication for why some languages thrive
(e.g., English, Spanish, and Chinese) and many others die every day (e.g., Laghu of the
Solomon Island and Old Kentish Sign Language (Tobin, 2011).
Many intergroup scholars (e.g., Giles et al., 1977; Giles & Johnson, 1981; Sachdev &
Bourhis, 1990) have suggested that languages, accents, and dialects can be the most
salient dimensions of social identity, and “the literature on the centrality of language to
group identity is substantial…” (Sachdev & Bourhis, 2005, p. 66). Tibetan diaspora in In
dia, for example, presents an illustrative case. Most of the younger generation Tibetans
educated in India could speak and write in multiple languages: Tibetan, English, and Hin
di, but a recent study (Dorjee, Giles & Barker, 2011) showed that in particular Tibetan
language is regarded as a key dimension of the Tibetan social-cultural identity. Recogniz
ing the centrality of Tibetan language to Tibetan identity and maintaining Tibetan culture
(see Bernstorff & von Welck, 2004a; Cabezon & Jackson, 1996; Dorjee, 2006; Gyatso,
1999; Shakabpa, 1967) are critical concerns in the ingroup Tibetan community. In recent
years, the Central Tibetan Administration (CTA) in India has formulated two education
policies.
One, CTA implemented Tibetan Education Policy that required most Tibetan schools in In
dia and Nepal to teach all the modern subjects, except English, in Tibetan language from
elementary level to middle school (Dorjee & Giles, 2005). Two, CTA has initiated a new
policy to design curriculum to teach Tibetan language to the Tibetan children in the West
ern diasporas (e.g., Tibetan diasporas in the U.S.A. and Canada) as a second language
(www.sherig.org). This policy is aimed at bilingual education of the Tibetan children in
the West so that they do not become linguistically assimilated into speaking just host lan
guages. His Holiness the Dalai Lama, CTA, and most Tibetans emphasize the need to
Page 10 of 32

teach, speak, and maintain Tibetan language as a key dimension of Tibetan social-cultural
identity from historical and cultural perspectives (see Cabezon & Jackson, 1996; Dorjee,
2013; Dorjee & Giles, 2009; Shakabpa, 1967; Stein, 1972).
It should be noted, however, that some groups (e.g., Arab Americans in the U.S., see
Sawaie, 1986; Chinese Canadian students, see Park, Dion, & Dion, 1985; see also Ed
wards, 1985) actually may not attach the same importance to their language as Tibetans
do. Interestingly, linguistic identity is not necessarily based on speaking the language flu
ently per se. Linguistic identity can be defined as “that part of an individual’s self-concept
that derives from his (or her) knowledge of his (or her) membership in a language (added
and italicized) group together with the value and emotional significance attached to that
membership” (Tajfel, 1978, p. 63). In other words, there is no absolute requirement for in
dividuals to speak their ethnic language, for example, to claim membership in that lan
guage group. It can be simply based on these two elements: linguistic knowledge recogni
tion and the keen emotional attachment to that language.
A Relational Perspective on Multilingualism and Social Identity
More recently, intergroup scholars have explored the complex relationship among lan
guage, social identity, and communication issues (e.g., Clement, Shulman, & Rubenfeld,
2010; Kalbfleisch, 2010; Reid & Anderson, 2010; Sutton, 2010). Interestingly, Clement et
al (2010) explored bilingual and multilingual issues between diversity and globalization.
EthnoVT can provide distinctive understanding of different types of bilingualism such as
balanced bilingualism (due to contextual factors favoring the equal status of each lan
guage and so on), additive bilingualism (situations favoring majority or minority group
members to acquire a second language in addition to their native language), and subtrac
tive bilingualism (situations pressuring minority group members to lose their ethnic/na
tive language and assimilate swiftly to adopt the dominant/host language) (Clement et al.,
2010). Illustratively stated, Tibetan diasporas in India and the U.S. provide contrasting in
sights into bilingualism.
Due to institutional support from both the Central Government of India and the State
Governments (see Dorjee, 2006), Tibetan settlers in India have been able to teach young
Tibetans (p. 34) to speak and write in Tibetan, English, Hindi or Regional Indian Lan
guage (e.g., Kanada). Even though the ethnolinguistic vitality of the Tibetan diaspora in
India cannot be compared at all to its host country’s ethnolinguistic vitality, Tibetan lan
guage and culture have thrived in India as compared to their status in their home envi
ronment—Tibet. Importantly, this case illustrates that minority group members can have
balanced or additive bilingualism or even multilingualism (as in the case of Tibetan dias
pora in India) provided their systematic language maintenance efforts are aided by host
country’s institutional support. This means the minority group members’ acquiring addi
tional second or third languages do not have to be necessarily classified as subtractive
bilinguals as some intergroup studies have argued (see Clement et al, 2010).
Page 11 of 32

On the other hand, the Tibetan diaspora in the U.S. does illustrate subtractive bilingual
ism. In the absence of institutional support and low group vitality (about 15,000 Tibetans
here), Tibetan parents struggle to teach everything Tibetan—Tibetan traditions, beliefs,
and values— and including the Tibetan language to their children. Many of the parents’
themselves, however, speak Tibetan poorly or not at all. This subtractive bilingualism may
have dire social and cultural consequences such as the perceived disconnection between
the Tibetan culture and language. Some studies (e.g., Rumbaut, 1994) suggest language
proficiency determines individual’s social identity.
Based on recent intergroup language-based research studies, we can proffer a descrip

tive model reflecting the complex relationships between bilingualism and linguistic identi
ty. We will use Tibetan diaspora to illustrate different scenarios. Scenario one can be ad
ditive bilinguals identifying strongly with their ethnic language identity. Dorjee, Giles, and
Barker’s (2011) study indicated many of their participants rating high on their language
competencies in Tibetan, English, and/or Hindi, but almost all of them strongly identified
with Tibetan identity and Tibetan language as a key dimension of that identity. Scenario
two can be subtractive bilinguals identifying with their ethnic heritage or native language
identity. While this seems counterintuitive, many young Tibetans in the U.S.A., who either
have difficulty speaking Tibetan or do not speak Tibetan due to assimilation into the
American English language, tend to hold strong Tibetan ethnic identity including Tibetan
language identity attachment. Kulyk (2011) writes, “Moreover, the identification with a
language does not always result from its use; people can feel attached to a language they
never speak and are hardly able to speak….This pertains in particular to languages of mi
grants, minorities or otherwise” (p. 629).
In other words, assimilated individuals in a host environment could still claim ethnic lin
guistic identity attachment. This implied that linguistic identity attachment is different
from actual language use and language proficiency (Libebkind, 1999). In the U.S., we find
examples among various immigrants (e.g., Mexican Americans, Tibetan Americans) who
may not speak their ethnic/native languages fluently but still claim strong ethnic linguis
tic identity awareness. Many Basque Americans who do not speak the Basque language
claim Basque language identity on a group membership level (Lasagabaster, 2008). Possi
bly, these bicultural U.S. Americans claiming their respective ethnic linguistic identity
was informed by the perception that ethnic heritage languages are essential parts of their
sociocultural identities as with Irish identity in Northern Ireland (see McMonagle, 2009).
Lastly, it is also possible that bicultural individuals may not identify (e.g., Yiddish and
English) with either of these languages as critical for two plausible reasons: (1) Language
does not occupy a strong facet of their social identity, and/or (2) other facets of their eth
nic or cultural community such as religious beliefs or rituals, customs, food, cultural arti
facts, and distinctive values shape their sociocultural membership contents. While ethnic,
racial, or religious group membership can be important to many individuals and offers a
sense of meaning, belonging, and a source of pride to these individuals (Verkuyten, 2010),
other membership types and role identities (e.g., gender identity, age identity, or profes
sional role) are important to some other individuals in particular situations (see Giles et
al., 2010). Thus, to be a competent intergroup communicator, we need to be attuned to
Page 12 of 32

the individuals’ beliefs and values concerning their perceptions of the degree of interde
pendence between their cultural/ethnic identity issues and whether they acknowledge the
vitality of their native language as part of their group membership identity, and also
whether they are capable to code-switch into another language due to their immigration
or diaspora experience.
Communication Accommodation Theory: Identity-Based Communica

tive Strategies
Given the above discussion concerning the relationships between languages and social
identity, (p. 35) communication accommodation theory offers useful insight into identity
negotiation and linguistic identity issues. In this section we provide a theoretical model
that explicates communication convergence (accommodation) and communication diver
gence (non-accommodation) in interpersonal and intergroup encounters.
Fig. 2.1 Intergroup identity negotiation: Linguistic

convergence and divergence issues.
Communication accommodation theory (CAT) has a history of over 30 years (see Gallois,
Ogay, & Giles, 2006), and it has primarily been applied to intergroup contexts including
intercultural encounters (Giles et al., 2010; Harwood & Giles, 2005). It has delineated the
conditions under which communicative convergence and divergence occur in interperson
al and intergroup encounters as well as their social consequences (Gallois et al., 2006;
Shepard, Giles, & Le Poire, 2001). Convergence can be defined as communicatively ac
commodating or adjusting to each other’s interests or needs in the encounter. For exam
ple, matching each other’s language, accent, nonverbal expressions, and communication
styles represent accommodative behaviors. In contrast, divergence can be defined as
communicatively non-accommodating or distancing from each other’s interests or needs
in the encounter. Specific examples include disengaging from the other via avoidance, si
lence, and nonverbal facial expression, and intentional dialect or language code switching
to exclude the other. We can put the two types of identity: social identity and personal
Page 13 of 32

identity on a vertical bar and cross it with communicative convergence and divergence on
a horizontal bar. These social identity-personal identity spectrum and convergence-diver
gence spectrum result in a dynamic model with four quadrants (see Figure 2.1). While the
model clearly distinguishes intergroup and interpersonal communicative strategies, they
can also manifest together in different degrees towards the center of the grid.
The first quadrant (Quadrant I) on the upper left hand corner represents intergroup mem
bership convergence or accommodative behaviors based on intergroup social approval
and other group-based motivations. According to CAT, individuals may initially orient
themselves to each other based on group membership such as culture, race, sex, ethnici
ty, age, sexual orientation, religion, professional role, to name a few. Intergroup conver
gence is observed in situations where both young and old adapt to each other’s communi
cation style, where females and males accommodate to each other’s relational expecta
tions, and members of interfaith code switch on topics of interest for the sake of inter
group harmony. Perhaps an excellent case of (p. 36) intergroup convergence is people
around the world communicating with each other in English in face to face interaction as
well as through social media such as facebook, twitter, skype, and email connection. The
importance of sociocultural context in this encounter is high regardless of the shared
medium of communication.
The second quadrant (Quadrant II) on the upper right hand corner represents intergroup
membership divergence due to motivations such as social disapproval and to highlight so
cial identity distinctiveness. Intergroup research provides evidence that members of dif
ferent social groups use a wide range of non-accommodative strategies such as avoidance
and dialect code-switching to distance themselves from each other. From a motivational
point of view, while some group members seek positive social identity in intergroup set
ting, others like to preserve their group membership identity distinctiveness. Therefore,
in situations where associating with certain group members undermine their positive so
cial identity they could use divergence strategies to uphold their positive identity distinc
tiveness such as “Vegans Save Lives!” or “Don’t be Mean, Go Green!” to assert positive
distinctiveness for their group.
The third quadrant (Quadrant III) on the bottom of the left hand corner represents per
sonal identity or interpersonal communication convergence. According to CAT, individuals
from different sociocultural backgrounds could have an initial orientation towards each
other based on their personal identity. For example, individuals accommodate to each
other’s needs based on dimensions such as personal appeals, physical attractiveness, lik
ing, and perceived like-mindedness. These dimensions constitute their personal identity.
Individuals from different social-cultural backgrounds form interpersonal relationships
based upon interpersonal attraction. They accommodate to each other’s needs both ver
bally (such as code switch, discuss common interests, self-disclosure, and share personal
stories) and nonverbally (such as reciprocal smiles and appropriate haptic/touch commu
nication). Interpersonal approval and personal willingness to communicate drive interper
Page 14 of 32

sonal convergence in communication. Thus, the importance of individuals’ sociocultural

background is minimized or not salient in this particular situation.
Finally, the fourth quadrant (Quadrant IV) on the bottom of the right hand corner repre
sents interpersonal communication divergence. In this divergent communicative situa
tion, individuals either do not seek relational approval for reasons such as personality dis
agreement, and/or have other competing needs which motivate them to use divergence or
non-accommodative verbal and nonverbal styles. Choosing not to interact with some peo
ple using a variety of strategies such as code switching, not introducing themselves, pre
tending to answer nature’s call, and changing their seats are examples of interpersonal
divergence. Basically, individuals may use any verbal and nonverbal symbols to communi
cate social distance in interpersonal encounters.
This model provides a meaningful heuristic insight into the distinction between interper
sonal and intergroup communicative strategies. That said, we envision the fluidity of com
munication across the quadrants and that the communication process can be high or low
on both intergroup and interpersonal dimensions (Giles & Hewstone, 1982; Gudykunst &
Lim, 1986). Intergroup convergence and divergence occur in situations where intergroup
salience is high (e.g., police-civilian encounter or intergenerational encounter), but inter
personal salience is low. On the other hand, interpersonal convergence and divergence
occur in situations where interpersonal salience is high (e.g., friendship encounters or in
timate relationship interactions or twitter interactions), but intergroup salience is low. Of
course, the framing or the interpretation of the actual intergroup or interpersonal en
counter process is highly contingent on each communicator’s perceptions, evaluative
schema, and cultural-racial embodied experiences. This four-quadrant model would as
sert a profound impact on the linguistic or dialect variations, nonverbal accommodation
or non-accommodation postures that individual use in a variety of intergroup interaction
situations. Perhaps negotiating identity and communicative interaction is most challeng
ing in situations where both interpersonal salience and intergroup salience are high. For
example, Jasmine, a Korean-Irish biracial American, 26, said the following about her
White boyfriend:
My boyfriend who was White had all White friends and we would hang out togeth
er sometimes. Many of his friends would make slight racial remarks against Asians
but in a joking way around me. While I took this as them trying to be funny, it ac
tually really started to bother me. When I told my boyfriend that I did not like his
friends making racial jokes against Asians he told me it was no big deal. One time
one of his friends told him that since he was dating me he could now say he “had
an Asian before” as if it (p. 37) was an accomplishment or a trophy should he ever
break up with me. A while after that I broke up with him because he would not
stand up for me, and, for hanging around such ignorant people.
(Toomey, Dorjee, & Ting-Toomey, 2013, p. 123)
Page 15 of 32

The above reaction clearly indicates the identity challenges in an intimate-interpersonal

dating relationship. Given the history of their interpersonal relationship, Jasmine told her
boyfriend that she did not appreciate his friends making racial jokes about her Asian so
ciocultural background. Lacking sensitivity, her White boyfriend simply told her “…it was
no big deal.” From an intergroup perspective, minority group members face such identity
challenges in their relationship with majority group members and what Jasmine’s
boyfriend said is a common reaction from someone in a dominant power-privileged posi
tion (see Orbe, 1998). Lastly, the situations in which both interpersonal and intergroup
salience can be low include tight intragroup interaction and mindless interactions under
the influence of alcohol and drugs. These insights from intergroup perspective along with
group membership-based communication accommodation and face saving foci can pro
vide further understanding concerning the intricate relationship among language, identi
ty, and sociocultural membership issues.
Problematic Face-Negotiation Dynamics: Inter

cultural and Intergroup Perspectives
Intercultural Face Clash
The following critical incident is an adapted story from Brislin, Cushner, Cherri, and Yong
(1986, pp. 157–158).
Who’s in Charge?
The President of XYZ Golf Club Company asked Masako Takai, the 36-year old
Chief Executive of the Marketing Division, and her staff (two male MBAs) to go to
Japan and close an important contract deal with the Nippon Company. He thought
his choice is especially effective as Masako (a third generation Japanese American
from California) knows the industry well and could also speak fluent Japanese.
Mr. Yamamoto, the 56-year old CEO of the Nippon Company was awaiting them in
his office. As Masako and her staff were being introduced, she noticed a quizzical
look on Mr. Yamamoto’s face and heard him mumbled “chief executive” to his as
sistant in an unsure manner. However, they both bowed politely to each other al
though Mr. Yamamoto felt that Masako should have bowed deeper since she
looked so young. They proceeded to the conference room and a female staff
poured them all tea. They then started their business talk. After Masako had pre
sented the merits of the marketing strategy in Japanese, referring to notes provid
ed by her staff, she asked Mr. Yamamoto what he thought. He responded by saying
that he needed to discuss some things further with the head of her department.
Masako explained that was why she was in Japan— to close the contract deal.
Smiling politely, Mr. Yamamoto replied that Masako had done a good job of ex
Page 16 of 32

plaining the marketing campaign strategy, but that he wanted to talk things over
with the person in charge.
Beginning to be frustrated, Masako stated that she had complete authority from
her company to sign the contract. At this point, Mr. Yamamoto appeared to be
quite confused and glanced at his assistant. Continuing to smile politely, however,
Mr. Yamamoto wanted to schedule another meeting with Masako and talk further.
Masako was at a boiling point at that stage. Wasted time means wasted money.
What went wrong here? Why did Mr. Yamamoto keep postponing signing the con
tract and wanting to schedule another meeting with Masako?
To analyze the story of “Who’s in Charge?’ and to answer the question of what went
wrong—everything went wrong on a culture-level analysis. Mr. Yamamoto was interpret
ing the role of Masako as either a junior staff or a translator from the XYZ Golf Club Com
pany. He was also not used to dealing with a young, female executive to represent a ma
jor firm. Furthermore, since Masako could speak Japanese fluently, Mr. Yamamoto did not
take her seriously as an American representative from the U.S. firm. He also did not de
code the term “Chief Executive” accurately early on, and he did not have faith that
Masako could actually go ahead and sign the contract on behalf of her company. While
Mr. Yamamoto came from a large power distance value orientation, Masako came from a
small power distance value dimension.
Intercultural miscommunication often involves different face-saving and face-recouping

behaviors. Face-saving and face-honoring behaviors are situated discourse. In order to
practice competent facework communication in a problematic intercultural situation, the
conflict face negotiation theory perspective may help to guide our understanding of the
interdependent nature of communicative identity, language, and culture.
(p. 38) Face Negotiation Theory Perspective: A Synopsis
In the context of the conflict face negotiation theory (FNT), face refers to a claimed sense
of desired social self-image in a relational or international setting (Ting-Toomey 1988,
2005b). The roots of FNT were inspired by the writing of Goffman’s (1955, 1959, 1967)
work on sociological facework, and Brown and Levinson’s (1987) discourse work on po
liteness theory. For a brief history that covers the origin and the evolution of the FNT, see
Ting-Toomey (1994, 2009c). Briefly stated, in any problematic discourse situation, face
loss occurs when we are being treated in such a way that our identity claims are being di
rectly or indirectly challenged or ignored. Face loss can occur on the individual level, the
group membership level, or both. Repeated face loss and face threat often lead to escala
tory conflict spirals or an impasse in the conflict negotiation process. Face gain, in con
trast, means an enhanced self-image, other-image, or both.
In fact, Spencer-Oatey (2005; van Meurs & Spencer-Oatey, 2010) proposes that the study
of facework can be understood via four relational interaction categories: a rapport-en
hancement orientation (a desire to strengthen or enhance harmonious relations between
Page 17 of 32

interlocutors), a rapport maintenance orientation (a desire to maintain or protect harmo

nious relations), a rapport-neglect orientation (a lack of concern for the quality of inter
personal relations perhaps because of a focus on the self), and a rapport-challenge orien
tation (a desire to challenge or hamper harmonious relations between the interlocutors).
In particular, when a second language is involved in the facework clash process, the tone
of voice, the nonverbal nuances, the situated linguistic codes, and the conflict assump
tions that are being enacted in the social disagreement episode can further derail the in
teraction process.
The FNT was developed as a response to the Western-biased perspective in the study of
conflict communication styles. In response to the heavy reliance on the individualistic
Western assumptions in framing various conflict approaches, Ting-Toomey (1988)
developed a cross-cultural conflict theory, namely, the conflict face negotiation theory
(FNT). The conflict FNT included an emphasis on a collectivistic, Asian-orientation per
spective to the understanding of intercultural conflict and was intended to expand the
theorizing process of existing, individualistic Western-based conflict approaches (Ting-
Toomey & Kurogi, 1998).
Core theoretical assumptions.
In sum, Ting- Toomey’s (2005a) conflict face negotiation theory assumes that (a) people in
all cultures try to maintain and negotiate face in all communication situations; (b) the
concept of face is especially problematic in emotionally threatening or identity-vulnerable
situations when the situated identities of the communicators are called into question; (c)
the cultural value spectrums of individualism-collectivism (Ting-Toomey, 2010; Triandis,
1995, 2002) and small-large power distance (Hofstede, 2001; House, Hanges, and Javidan
et al., 2004) shape facework concerns and styles; (d) individualism and collectivism value
patterns shape members’ preferences for self-oriented facework versus other-oriented
facework; (e) small and large power distance value patterns shape members’ preferences
for horizontal-based facework versus vertical-based facework; (f) the value dimensions, in
conjunction with individual, relational, and situational factors, influence the use of partic
ular facework behaviors in particular cultural scenes; and (g) intercultural facework com
petence refers to the optimal integration of knowledge, mindfulness, and communication
skills in managing vulnerable identity-based conflict situations appropriately, effectively,
and adaptively.
Cultural membership and face concerns.
Self-face concern is the protective concern for one’s own identity image when one’s own
face is threatened in the conflict episode. Other-face concern is the concern for accommo
dating the other conflict party’s identity image in the conflict situation. Mutual-face con
cern is the concern for both parties’ images and the image of the relationship. Whether
we choose to engage in self-face protection or mutual-face protection often depends on
our ingrained cultural socialization process, individual trait tendencies, and embedded
situational factors.
Page 18 of 32

More specifically, for example, in a direct empirical test of the theory by Oetzel and Ting-
Toomey (2003), the research program tested the underlying assumption of the FN conflict
theory that face is an explanatory mechanism for cultural membership’s influence on con
flict behavior. A questionnaire was administered to 768 participants in four national cul
tures: China, Germany, Japan, and the U.S. in their respective languages asking them to
recall and describe a recent interpersonal conflict. The major results of the study are as
follows: First, cultural individualism-collectivism had direct effects on conflict styles, as
well as mediated (p. 39) effects through self-construal and face concerns. Second, self-
face concern was associated positively with dominating style and other-face concern was
associated positively with avoiding and integrating styles. Third, German respondents re
ported the frequent use of direct-confrontational facework strategies; Japanese reported
the use of different pretending and accommodating strategies and minimize the severity
of the conflict situation; Chinese engaged in a variety of avoiding, accommodating, pas
sive aggressive, and third-party appeals’ tactics; and U.S. Americans reported the use of
upfront expression of feelings and remaining calm as conflict facework tactics.
Within the pluralistic U.S. sample, multiethnic research by Ting-Toomey and co-re
searchers (2000) has also uncovered distinctive conflict interaction styles in relationship
to particular ethnic identity salience issues. While previous research studies have focused
on testing the relationship between individualism-collectivism value dimensions and face
work strategies, recent research effort has focused on examining the relationship be
tween small/large power distance values and particular facework practice in the work
place (Merkin & Ramadan, 2010). Beyond broad-based cultural value pattern dimensions,
individuals do develop their unique personality attributes due to distinctive family social
ization processes and particular lived experiences.
Individual personality factors and facework.
The term, “self-construal,” was coined by Hazel Markus and Shinobu Kitayama (1991),
and is concerned with one’s personalized self-image as emphasizing either an indepen
dent or an interdependent self. In individualistic cultural communities, there may be more
situations that evoke the need for independent-based actions. In collectivistic communi
ties, there may be more situations that demand the sensitivity for interdependent-based
decisions.
The manner in which individuals conceive of their self-images should have a profound in
fluence on the expectancies of what constitute appropriate and effective responses in di
verse facework situations. Both dimensions of self also exist within each individual, re
gardless of cultural membership identity (Ting-Toomey, Oetzel, & Yee-Jung, 2001). For ex
ample, Oetzel and Ting-Toomey (2003) and Oetzel, Garcia, and Ting-Toomey (2008) found
that independent self-construal is associated positively with self-face concern and the use
of dominating/competing conflict strategies. Interdependent self-construal, on the other
hand, is associated positively with other-face concern and the use of avoiding and inte
grating conflict tactics. It would appear that independent self-construal fosters the use of
Page 19 of 32

upfront and low-context demanding interaction responses, while interdependent self-con

strual emphasizes circumspective and high-context yielding interaction patterns.
Situational factors and facework.
Two other possible factors that moderate the activation of an independent versus an in
terdependent self are situational role appraisal and ingroup/outgroup distance factors. Si
tuational role appraisal factors can include the degree of formality of the conflict setting,
the interaction climate of the situation, the role relationship between the conflict partici
pants, and the perceived goals of the facework negotiation process. To illustrate, the role
appraisal process can include an assessment of the role expectancies between the con
flict parties such as professional role identities and other salient group membership and
personal identity concerns.
For example, Merkin (2006; see also, Merkin & Ramadan, 2010) has integrated small/
large power distance value dimension to the individualism-collectivism value dimension in
explaining face-threatening response messages and conflict styles in multiple cultures.
She found that high-status individuals from large power distance cultures tend to use
both direct and indirect facework strategies to deal with face-threatening situations—de
pending on whether they were delivering positive or negative messages. Thus, an accu
rate assessment of the culture-based situational factors that frame facework strategy us
age can be critical in promoting competent conflict management outcome. Integrating
the intergroup perspective with the face-negotiation theoretical frame can yield addition
al insights among identity, face-sensitive concern, and sociocultural membership issues.
Intergroup Face Negotiation Perspective
According to social identity theory, social categories (group memberships) are mecha
nisms by which individuals relate to each other (Tajfel & Turner, 1979). Even in casual
conversations when social identity (group membership) of any conversational partner be
comes salient it changes the dynamics of conversations from interpersonal to intergroup
interaction (see Giles et al, 2010). In this regard, individuals have to constantly negotiate
intergroup boundary issues (the intertwined nature of personal and social identity) with
others and even in close friendship setting. For example, Gideon, a (p. 40) Chinese/French
American bicultural-biracial individual (Male, 25) articulated:
I feel like I have to introduce myself constantly to my friends based on my dual

heritage, rather than other deep qualities I have. It is like people only see that mix
when first meeting me and they do not really want to take the time to understand
the real, inner me. I have to talk about all the cultural background things before
we can have a further, deeper discussion. It can get annoying and frustrating at
times but I sort of understand why people or even my friends are in a way “fasci
nated” by it.
(Toomey et al., 2013, p. 122).
Page 20 of 32

The above statements clearly reflect the salience of social identity/group membership and
its impact on communication with others. Notably, this particular bicultural-biracial indi
vidual struggles with group membership-based face negotiation—his claim to positive im
age in the context of social interaction. Importantly, his face and facework are based on
social categorization or group membership. Notably, he could not simply bypassed his
group membership in encounters because others including his friends de-individualize
him with ascribed social identity based on his mixed-feature appearance.
In essence, traditional intergroup scholars have not paid much attention to face negotia
tion and the relationship between group membership-based facework and communication
convergence/divergence issues (Harwood & Giles, 2005). These integrative theoretical
concepts can enrich intergroup understanding of identity, language/verbal styles, and
communication issues. FNT explicates three types of face concern—self-face concern, oth
er-face concern, and mutual face concern to manage identity-sensitive communication is
sues in a variety of problematic interactional situations within and across cultures.
From an intergroup perspective, there could be four group membership face concerns: in
group membership face concern (IGMFC), outgroup membership face concern (OGMFC),
intergroup membership face concern (ITMFC), and community membership face concern
(CMMFC). IGMFC is the degree of protective concern for the positive communicative im
age of one’s own ingroup especially when that group-based image is threatened in social
interaction. OGMFC is the degree of concern for accommodating the outgroup members’
communicative image in social interaction from being further insulted or challenged.
ITMFC is the degree of mutual back-and-forth diplomacy concern for preserving the posi
tive images of both ingroup and outgroup in social interaction. Lastly, CMMFC is the de
gree of face-identity concern about the larger world stage’s reaction or the surrounding
eyewitness community’s reaction. Importantly, these face concerns can provide an expla
nation for why group members use different accommodative or divergent strategies in in
ternational or intergroup facework negotiation setting. These are intimately connected to
understanding intergroup categorization process and enacting skillful facework diploma
cy tactics.
Additionally, many relational distance factors are important in competent intergroup-in

tercultural facework discourse practice. One critical factor is recognizing power distance
dominance-deference issues in facework negotiation process (Holtgraves, 2009; see also,
Holtgraves & Kashima, 2008). Another factor is having a good grasp of how a particular
cultural community defines in-group and out-group and what constitutes appropriate in-
group versus out-group linguistic or verbal exchange processes. Take, for example, from
the Japanese communication lens, Midooka (1990), who categorized four groups of Japan
ese relationships, noted that the Japanese in-group consisting of kino-okenai-kankei and
nakama and the Japanese out-group consisting of najimi-no-tanin and muen-no-kankei.
Kino-okenai-kankei (“intimate ingroups”) consists of very intimate or equal-status rela

tionships in which communication is causal, open, and direct. Examples of such relation
ships are best friends, family/siblings, close relatives, childhood buddies, and dating rela
Page 21 of 32

tionships. In these relationships, differences in age or seniority are superseded by intima

cy, and no hierarchical rituals, especially in the “best friends” category, are heeded. Thus,
in Japanese “best friends” conflict situations, the process can involve more heart-to-heart
talks to direct conflict self-disclosure. Nakama (“familiar interactive ingroups”), on the
other hand, are close-contact in-group relations, especially in terms of everyday familiari
ty, yet not so much as to override status differences. These typically include everyday col
leagues in the same workplace, and here, maximum care must be taken to observe inter
personal rituals and preserving relational harmony even under stressful conflict condi
tion. A certain level of decorum or formality is expected to be maintained in this particu
lar relationship category.
On the other hand, najimi-no-tanin (“acquaintance interactive outgroups”) refers to a less

intimate, acquaintance relationship, characterized more as an outgroup rather than as an
ingroup (p. 41) relationship. For example, outgroup relationships in this instance could
mean acquaintance colleagues in other universities or a friend of a close friend who
needs a favor. While being tanin, communication behaviors toward this “familiar” out-
group member would differ greatly depending on the perceived value or reward/cost ap
praisal process of the relationship. However, since Japan is considered as an overall
group-oriented society, social ties have interlocking importance and wider interdependent
implications from one spectrum of the society to the next (Ting-Toomey & Takai, 2006). If
the relationship poses a threat to one’s public face, one is still careful to observe appro
priate interaction formality and diplomatic conflict rituals. Cautious formality is exercised
in the tanin situation more so than the nakama situation—as one misstep can be costly
and can ruin one’s reputation or face beyond just the out-group circle. Finally, muen-no-
kankei (“stranger outgroups”) indicates a purely outgroup, stranger relationship, also re
ferred to as aka-no-tanin. Since strangers are way beyond the bounds of accepted social
or personalized ties, oftentimes, no form of considerate behavior needs to be extended
between the stranger-pair as no preexisting emotional sentiments bind the two people to
gether. Indifference can be part of the conflict ritual in this peripheral outgroup category.
In sum, intergroup identity factors and relational distance parameters have a strong im
pact on what constitute appropriate and effective communication styles and adaptive
facework behaviors in particular situations and in different cultural communities. A cul
turally competent communicator would need to increase his or her awareness concerning
self and other’s social and personal identity issues and the group membership facework
issues that are being experienced and displayed in the social discourse situation. The next
section will conclude with some suggested directions for future research in the areas of
intergroup identity negotiation and intergroup facework negotiation.
Page 22 of 32

Future Research Directions

Researching Intergroup Identity Language Issues
In terms of testing the interdependent relationship among language, identity, and socio
cultural membership issues, here are some suggestions along the lines of language and
convergence/divergence issues, and language and intergroup communication competence
or incompetence issues. In responding to the language and convergence/divergence is
sues, here are some researchable questions: Under what conditions do different sociocul
tural identity members seek linguistic convergence or divergence? How do they decode
whether the other identity member is capable of code-switching in the intergroup identity
negotiation process? What paralinguistic or micro-nonverbal signals do individuals ex
change that they would actually interpret as identity approval or liking, or identity insult
ing or patronizing? How could we better study nonverbal nuances that create a strong im
pact on the language and verbal style variations in intergroup communication process?
In addressing the language and intergroup communication competence issues, here are
some research questions that need some urgent attention: What are some appropriate
and effective discourse strategies that can be used to instill super-ordinate identities and
interdependent fates among separate cultural/ethnic group circles? What is the role of a
competent translator or interpreter in the diplomatic multi-track peace-building process?
How can bilingual or multilingual mediators create a secure “third space” through the
artful use of language to promote mutual intergroup respect among diverse identity
groups?
Researching Intergroup Face Negotiation Issues
The following two research areas hold promise and need future research attention: inter
group facework situations, and intergroup facework competence. The study of face-nego
tiation in everyday discourse would definitely benefit by examining the relationship
among situations, face concerns, and facework verbal and nonverbal codes’ usage. Ques
tions such as the following need more systematic research investigations: Under what
specific situational conditions would intergroup communicators be more interested in in
tergroup mutual-face protection versus ingroup face protection? Under what identity
threat conditions would intergroup negotiators be more concerned with mutual-face pro
tection versus mutual-face annihilation? What do the language codes of honor, dignity, re
spect, insult, and vengeance in different speech communities sound like? How could
these codes be translated with optimal fidelity in correspondence to the original speech
community?
In connecting the relationship among language, verbal styles, and intergroup facework
competence, (p. 42) here are some researchable directions: What particular facework
strategies can intergroup members use to promote mutual face respect? How can bicul
tural-biracial identity members display optimal facework competence in intergroup com
munication settings? How do they make strategic choices to foster ingroup connection
Page 23 of 32

and solidarity with one group without alienating the other group? Do we need to develop
particular linguistic process competence theory for each sociocultural group or can we
engage in a cross-cultural and cross-situational theorizing process? Beyond communica
tion appropriateness, effectiveness, and adaptability, what are other competence yard
sticks we need to incorporate in order for intergroup communicators to reach optimal
competence level? What does optimal communication competence look like from a lan
guage or verbal communication competence standpoint?
Conclusions
This chapter advocates the importance of understanding complex identity issues and lan
guage variation issues from the three theoretical frameworks of identity negotiation theo
ry, intergroup perspective including communication accommodation theory, and face ne
gotiation theory. International bilingual and multilingual researchers are needed to work
more collaboratively and systematically to uncover the process competence perspective
in the study of language, communication, and sociocultural membership identity issues.
From the narrative approach to the functional- quantitative approach, more theoretical
and research efforts are needed for us to truly understand the multiplicity of identity voic
es and lenses of individuals from diverse sociocultural communities. Dynamic language
communicators are individuals who practice culture-sensitive verbal and nonverbal styles
and can code-switch fluidly with the strategic use of artful language enactment. At the
same time that they adhere to the criteria of communication appropriateness and effec
tiveness in the intergroup interaction setting, they are also highly attuning to the identity
dynamics of their fellow intergroup conversation partners.
In sum, competent communicators are highly creative individuals who can use the art of
language and strategic communication styles to convey intergroup membership identity
support. They can also display great linguistic reframing skills by de-polarizing inter
group membership tensions. They are the mindful communicators who have a secure and
grounded sense of identity and, simultaneously, lending this sense of attunement to core
identity issues that are implicitly or explicitly expressed by their intergroup conversation
partners. Toward this end, in this chapter, we have unpacked and discussed the complex
relationship among language variations, identity, and sociocultural membership from mul
tiple identity-based perspectives. Drawing from the intercultural and intergroup commu
nication literature, we are able to extend some of the existing identity-based frameworks
forward and incorporating a wide-angle lens in exploring the intersecting paths of the
ever fascinating phenomenon of language, identity, and culture.
References
Ang, S., Van Dyne, L., & Tan, M. L. (2011). Cultural intelligence. In R. J. Sternberg & S. B.
Kaufman (Eds.), Cambridge handbook on intelligence (pp. 582–602). New York: Cam
bridge Press.
Page 24 of 32

Bernstorff, D., & von Welck, H. (2004). An interview with Tenzin Gyatso, the 14th Dalai
Lama. In D. Bernstorff & H. von Welck (Eds.), In exile as challenge: The Tibetan diaspora
(pp. 107–123). New Delhi: Orient Longman.
Berry, J., Kim, U., & Boski, P. (1987). Psychological acculturation of immigrants. In Y. Y.
Kim & W. B. Gudykunst (Eds.), Cross-cultural adaptation: Current approaches (pp. 62–89).
Newbury Park, CA: Sage.
Bourhis, R. Y., Giles, H., & Tajfel, H. (1973). Language as a determinant of Welsh identity.
European Journal of Social Psychology, 3, 447–460.
Brown, P., & Levinson, S. C. (1987). Politeness: Some universals in language usage. Cam
bridge, UK: Cambridge University Press.
Brislin, R., Cushner, K., Cherrie, C., & Yong, M. (1986). Intercultural interactions: A prac
tical guide. Newbury Park, CA: Sage.
Cabezon, J. I., & Jackson, R. R. (Eds.) (1996). Tibetan literature: Studies in genre. Ithaca,
NY: Snow Lion.
Collie, P., Kindoh, S., Liu, J., & Podsiadlowski, A. (2010). Mindful identity negotiation: The
acculturation of young Assyrian women in New Zealand. International Journal of Intercul
tural Relations, 34, 208–220.
Clement, R., Shulman, J., & Rubenfeld, S. (2010). Bi- and multilingualism between diversi
ty and globalization. In H. Giles, S. A. Reid, & J. Harwood (Eds.), Dynamics of intergroup
communication (pp. 117–128). New York: Peter Lang.
Deardorff, D. (2004). The identification and assessment of intercultural competence as a

student outcome of internationalization at institutions of higher education in the United
States. Unpublished doctoral dissertation, North Carolina State University, Raleigh.
Dorjee, T. (2006). Transmitting cultural identity from generation to generation in Tibetan

diaspora. In W. Leeds-Hurwitz (Ed.), From generation to generation: Maintaining cultural
identity over time (pp. 227–254). Cresskill, NJ: Hampton Press.
Dorjee, T. (2013). Intercultural and intergroup conflict resolution: Nonviolence and mid
dle way approaches. In J. G. Oetzel & S. Ting-Toomey (Eds.), The Sage handbook of con
flict communication (2nd ed., pp. 687–711). Los Angeles, CA: Sage.
Dorjee, T., Giles, H., & Barker, V. (2011). Diasporic communication: Cultural de
(p. 43)
viance and accommodation among Tibetan exiles in India. Journal of Multilingual and
Multicultural Development, 32, 343–359.
Dorjee, T., & Giles, H. (2009). Accommodating deviance: Tibetan exiles in India. In S. K.
Singh (Ed.), Rethinking multilingualism: Issues and problems (pp. 27–60). Guwahati, In
dia: Eastern House Publications.
Page 25 of 32

Dorjee, T., & Giles, H. (2005). Cultural identity in Tibetan diasporas, Journal of Multilin
gual and Multicultural Development, 26, 118–137.
Edwards, J. (1985). Language, society, and identity. Oxford: Blackwell.
Gallois, C., Giles, H., Jones, E., Cargile, A. C., & Ota, H. (1995). Accommodating intercul
tural encounters: Elaborations and extensions. In R. Wiseman (Ed.), Intercultural commu
nication theory (pp. 115–147). Thousand Oaks, CA: Sage.
Gallois, C., Ogay, T., & Giles, H. (2006). Communication accommodation theory. In W. B
Gudykunst (Ed.), Theorizing about intercultural communication (pp. 121–148). Thousand
Oaks, CA: Sage.
Gelfand, M., Raver, J. L., Nishii, L., Leslie, L. M., Lun, J., Lim, B. C. et al. (2011). Differ
ences between tight and loose cultures: A 33-nation study. Science 332, 1100–1104.
Giles, H. (1973). Accent mobility: A model and data. Anthropological Linguistics, 15, 87–
109.
Giles, H. (1978). Linguistic differentiation between ethnic groups. In H. Tajfel (Ed.), Dif
ferentiation between social groups (pp. 361–393). London: Academic Press.
Giles, H., Bourhis, R. Y., & Taylor, D. M. (1977). Towards a theory of language in ethnic
group relations. In H. Giles (Ed.), Language, ethnicity and intergroup relations (pp. 307–
343). London: Academic Press.
Giles, H., Coupland, N., & Coupland, J. (1991). Accommodating theory: Communication,
contexts, and consequences. In H. Giles, J. Coupland, & N. Coupland (Eds.), Contexts of
accommodation: Developments in applied sociolinguistics (pp. 1–68). Cambridge, MA:
Cambridge University Press.
Giles, H., & Gasiorek, J. (2011). Intergenerational communication practices. In K. W.

Schaie & S. Willis (Eds.), Handbook of psychology of aging (7th ed., pp. 231–245). New
York: Elsevier.
Giles, H., & Hewstone, M. (1982). Cognitive structures, speech, and social situations.
Language Sciences, 4, 187–219.
Giles, H., & Johnson, P. (1987). Ethnolinguistic identity theory: A social psychological ap
proach to language maintenance. International Journal of the Sociology of Language, 68,
66–99.
Giles, H., & Johnson, P. (1981). The role of language in ethnic group relations. In J. C.
Turner & H. Giles (Eds.), Intergroup behavior (pp. 199–243). Oxford: Blackwell.
Giles, H., Taylor, D. M., & Bourhis, R. Y. (1973). Toward a theory of interpersonal accom
modation through speech: Some Canadian data. Language in Society, 2, 177–192.
Page 26 of 32

Giles, H., Wilemyns, M., Gallois, C., & Johnson, M. C. (2007). Accommodating a new fron
tier: The context of law enforcement. In K. Fielder (Ed.), Social communication (pp. 129–
162). New York: Psychology Press.
Goffman, E. (1955). On face-work: An analysis of ritual elements in social interaction. Psy

chiatry: Journal of Interpersonal Relations, 18(3), 213–231.
Goffman, E. (1959). The presentation of self in everyday life. Garden City, NY: Anchor/
Doubleday.
Goffman, E. (1967). Interaction ritual: Essay on face-to-face behavior. New York: Double
day/Anchor Book.
Gudykunst, W. B., & Lim, T. S. (1986). A perspective for the study of intergroup communi
cation. In W. B. Gudykunst (Ed.), Intergroup communication (pp. 1–9). London: Edward
Arnold.
Gyatso, T. (1999). Opening the eye of new awareness. Trans. D. S. Lopez, Jr. and J. Hop
kins. Boston, MA: Wisdom Publications.
Hofstede, G. (2001). Culture’s consequences: Comparing values, behaviors, institutions,

and organizations across cultures (2nd ed.). Thousand Oaks, CA: Sage.
Holtgraves, T. (2009). Face, politeness and interpersonal variables: Implications for lan
guage production and comprehension. In F. Bargiela-Chiappini & M. Haugh (Eds.), Face,
communication and social interaction (pp. 192–207). London: Equinox.
Hotta, J., & Ting-Toomey, S. (2013). Intercultural adjustment and friendship dialectics in
international students: A qualitative study. International Journal of Intercultural
Relations, 37, 550–566.
House, R., Hanges, P., Javidan, M., Dorfman, P., & Gupta, V. (Eds.). (2004). Culture, lead
ership, and organizations: The GLOBE study of 62 societies. Thousand Oaks, CA: Sage
Publications.
Harwood, J., & Giles, H. (Eds.). (2005). Intergroup communication: Multiple perspectives.
New York, NY: Peter Lang.
Howard, G., Reid, S. A., & Harwood, J. (Eds.). (2010). The dynamics of intergroup commu
nication. New York, NY: Peter Lang.
Kalbfleisch, P. (2010). Gendered language as a dynamic intergroup process. In H. Giles, S.

A. Reid, & J. Harwood (Eds.), Dynamics of intergroup communication (pp. 29–40). New
York: Peter Lang.
Page 27 of 32

Kashima, E. S., & Kashima, Y. (1998). Culture and language: The case of cultural dimen
sions and personal pronoun use. Journal of Cross-Cultural Psychology, 29, 461–486.
Kashima, Y., & Gelfand, M. (2012). A history of culture in psychology. In A. Kruglanski &
Strobe (Eds.), Handbook of the history of social psychology. New York: Taylor & Francis
Group.
Kashima, Y., Kashima, E. S., Kim, E., & Gelfand, M. (2006). Describing the social world:
How is a person, a group, and a relationship described in the East and the West. Journal
of Experimental Social Psychology, 31, 111–124.
Kulyk, V. (2011). Language identity, linguistic diversity and political cleavages: Evidence
from Ukraine. Nations and Nationalism, 17 (3), 627–648.
Lasagabaster, D. (2008). Basque diaspora in the USA and language maintenance. Journal
of Multilingual and Multicultural Development, 29 (1), 66–90.
Liebkind, K. (1999). Social psychology. In J. A. Fishman (Ed.), Handbook of language and

ethnic identity. New York: Oxford University Press.
Markus, H., & Kitayama, S. (1991). Culture and the self: Implications for cognition, emo
tion, and motivation. Psychological Review, 2, 224–253.
McMonagle, S. (2009). Linguistic diversity in Northern Ireland: From conflict to multicul

turalism. Paper presented at the Annual Convention of the Association for the Study of
Nationalities, New York.
Merkin, R. (2006). Power distance and facework strategies. Journal of Intercultural

(p. 44)
Communication Research, 35, 139–160.
Merkin, R., & Ramadan, R. (2010). Facework in Syria and the United States: A cross-cul
tural comparison. International Journal of Intercultural Relations. 34, 661–669.
Midooka, K. (1990). Characteristics of Japanese-style communication. Media, Culture and

Society, 12, 477–489.
Molinsky, A. (2007). Cross-cultural code-switching: The psychological challenges of adapt

ing behavior in foreign cultural interactions. Academy of Management Review, 32, 622–
640.
Jackson, R. L. (1999). The negotiation of cutlural identity. Westport, CT: Praeger.
Jackson, R. L. (2002). Cultural contracts theory: Toward an understanding of identity ne

gotiation. Communication Quarterly, 50 (3 & 4), 359–367.
Oetzel, J. G., Garcia, A., & Ting-Toomey, S. (2008). An analysis of the relationships among
face concerns and facework behaviors in perceived conflict situations: A four-culture in
vestigation. International Journal of Conflict Management, 19, 382–403.
Page 28 of 32

Oetzel, J. G., & Ting-Toomey, S. (2003). Face concerns in interpersonal conflict: A cross-
cultural empirical test of the face-negotiation theory. Communication Research, 30, 99–
624.
Oetzel, J. G., & Ting-Toomey, S. (Eds.). (2006). The Sage handbook of conflict communica
tion: Integrating theory, research, and practice. Thousand Oaks, CA: Sage.
Oetzel, J. G., Ting-Toomey, S., Yokochi, Y., Masumoto, T., & Takai, J. (2000). A typology of
facework behaviors in conflicts with best friends and relative strangers. Communication
Quarterly, 48, 397–419.
Onwumechili, C., Nwosu, P., Jackson, R. II & James-Hughes, J. (2002). In the deep valley
with mountains to climb: Exploring identity and multiple reacculturation. International
Journal of Intercultural Relations, 27, 41–62.
Orbe, M. P. (1998). Constructing co-cultural theory: An explication of culture, power, and

communication. Thousand Oaks, CA: Sage.
Park, A. W. P., Dion, K. L., & Dion, K. K. (1985). Correlates of self-confidence with English
among Chinese students in Toronto. Canadian Journal of Behavioral Science, 17, 369–378.
Rumbaut, R. G. (1994). The crucible within: Ethnic identity, self-esteem, and segmented
assimilation among children of immigrants. International Migration Review, 28, 748–794.
Reid, S. A., & Andersen, G. L. (2010). Language, social identity, and stereotyping. In H.
Giles, S. A. Reid & J. Harwood (Eds.), Dynamics of intergroup communication (pp. 91–
104). New York: Peter Lang.
Sachdev, I., & Bourhis, R. Y. (2005). Multilingual communication and social identification.
In J Harwood & H. Giles (Eds.), Intergroup communication: Multiple perspectives (pp. 65–
91). New York: Peter Lang.
Sachdev, L., & Bourhis, R. Y. (1990). Language and social interaction. In D. Abrams & M.
Hogg (Eds.), Social identity theory: Constructive and critical advances (pp. 35–51). Hemel
Hempstead, UK: Harvester Wheatsheaf.
Sawai, M. (1996). The present and future of a minority language: The case of Arabic in
the United States. Journal of Multilingual and Multicultural Development, 7, 379–392.
Shakabpa, W. D. (1967). Tibet: A political history. New Haven and London: Yale University
Press.
Shepard, C. A., Giles, H., & Le Poire, B. A. (2001). Communication accommodation theory.
In W. P. Robinson & H. Giles (Eds.), The new handbook of language and social psychology
(pp. 33–56). USA: John Wiley & Sons Ltd.
Spencer-Oatey, H. (2005) (Im)politeness, face and perceptions of rapport: Unpacking

their bases and interrelationships. Journal of Politeness Research 1, 95–119.
Page 29 of 32

Spitzberg, B., Canary, D. J., & Cupach, W. (1994). A competence-based approach to the
study of interpersonal conflict. In D. Cahn (Ed.), Conflict in personal relationships (pp.
183–202). Hillsdale, NJ: Lawrence Erlbaum.
Spitzberg, B. H., & Changnon, G. (2009). Conceptualizing intercultural competence. In D.

K. Deardorff (Ed.), The Sage handbook of intercultural competence (pp. 2–52). Thousand
Oaks, CA: Sage.
Stein, R. (1972). Tibetan civilization. Stanford, CA: Stanford University Press.
Sutton, R. M. (2010). The creative power of language in social cognition and intergroup
relations. In H. Giles, S. A. Reid, & J. Harwood (Eds.), Dynamics of intergroup communica
tion (pp. 105–116). New York: Peter Lang.
Tajfel, H. (1978). The achievement of group differentiation. In H. Tajfel (Ed.), Differentia

tion between groups: Studies in the social psychology of intergroup relations (pp. 77–
100). London: Academic Press.
Tajfel, H., & Turner, J. C. (1979). An integrative theory of intergroup conflict. In S.

Worchel & W. G. Austin (Eds.), The psychology of intergroup relations (pp. 33–47). Mon
terey, CA: Brooks-Cole.
Ting-Toomey, S. (1988). Intercultural conflicts: A face- negotiation theory. In Y. Y. Kim &

W. B. Gudykunst (Eds.), Theories in intercultural communication (pp. 213–235). Newbury
Park, CA: Sage.
Ting-Toomey, S. (Ed.). (1994). The challenge of facework: Cross-cultural and interpersonal

issues. New York: State University of New York-Albany Press.
Ting-Toomey, S. (1999). Communicating across cultures. New York: The Guilford Press.
Ting-Toomey, S. (2004). Translating conflict face-negotiation theory into practice. In D.

Landis, J. Bennett, & M. Bennett (Eds.), Handbook of intercultural training (3rd ed., pp.
217–248). Thousand Oaks, CA: Sage Publications.
Ting-Toomey, S. (2005a). Identity negotiation theory: Crossing cultural boundaries. In W.

B. Gudykunst (Ed.), Theorizing about intercultural communication (pp. 211–233). Thou
sand Oaks, CA: Sage.
Ting-Toomey, S. (2005b). The matrix of face: An updated face-negotiation theory. In W. B.

Gudykunst (Ed.), Theorizing about intercultural communication (pp. 71–92). Thousand
Oaks, CA: Sage Publications.
Ting-Toomey, S. (2009a). Intercultural conflict competence as a facet of intercultural com

petence development: Multiple conceptual approaches. In D. K. Deardorff (Ed.), The Sage
handbook of intercultural competence (pp. 100–120). Thousand Oaks, CA: Sage.
Page 30 of 32

Ting-Toomey, S. (2009b). Facework collision in intercultural communication. In F.

Bargiela-Chiappini & M. Haugh (Eds.), Face, communication and social interaction (pp.
227–249). London: Equinox.
Ting-Toomey, S. (2009c). Face negotiation theory. In S. Littlejohn & K. Foss (Eds.), Ency
clopedia of Communication Theory (pp. 371–374). Thousand Oaks, CA: Sage Publications.
Ting-Toomey, S. (2010a). Applying dimensional values in understanding intercul

(p. 45)
tural communication. Communication Monographs, 77, 169–180.
Ting-Toomey, S. (2010b). Mindfulness. In R. Jackson (Ed.), Sage encyclopedia of identity,

Volume 1 (pp. 455–458). Thousand Oaks, CA: Sage.
Ting-Toomey, S. (2014). Managing identity issues in intercultural conflict communication:

Developing a multicultural identity attunement lens. In V. Benet-Martinez & Y. -Y. Hong
(Eds.), Oxford handbook of multicultural identity (pp. XX–YY). New York: Oxford Universi
ty Press.
Ting-Toomey, S., & Chung, L. C. (2012). Understanding intercultural communication (2nd

ed.). New York: Oxford University Press.
Ting-Toomey, S., & Kurogi, A. (1998). Facework competence in intercultural conflict: An

updated face-negotiation theory. International Journal of Intercultural Relations 22, 187–
225.
Ting-Toomey, S., & Oetzel, J. G. (2001). Managing intercultural conflict effectively. Thou
Ting-Toomey, S., & Oetzel, J. G. (2013). Culture-based situational conflict model: An up
date and expansion. In J. G. Oetzel & S. Ting-Toomey (Eds.), The Sage handbook of con
flict communication (2nd ed., pp. 763–789). Los Angeles, CA: Sage.
Ting-Toomey, S., Oetzel, J. G., & Yee-Jung, K. (2001). Self-construal types and conflict
management styles. Communication Reports, 14, 87–104.
Ting-Toomey, S., & Takai, J. (2006). Explaining intercultural conflict: Promising approach
es and directions. In J. G. Oetzel & S. Ting-Toomey (Eds.), The Sage handbook of conflict
communication (pp. 691–723). Thousand Oaks, CA: Sage Publications.
Ting-Toomey, S., Yee-Jung, K., Shapiro, R., Garcia, W., Wright, T., & Oetzel, J. G. (2000).
Cultural/ethnic identity salience and conflict styles in four U.S. ethnic groups. Internation
al Journal of Intercultural Relations, 24, 47–81.
Tobin, L. (2011, February 21). Half of living languages face extinction. Retrieved from
http://www.guardian.co.uk/education/2011/feb/21/endangered-languages-re
search-project.
Page 31 of 32

Toomey, A., Dorjee, T., & Ting-Toomey, S. (2013). Bicultural identity negotiation, conflicts,
and intergroup communication strategies. Journal of Intercultural Communication Re
search, 42, 112–134.
Triandis, H. (1995). Individualism and collectivism. Boulder, CO: Westview Press.
Triandis, H. (2002). Individualism and collectivism. In M. Gannon & K. Newman (Eds.),

Handbook of cross-cultural management (pp. 16–45). New York: Lawrence Erlbaum.
Verkuyten, M. (2010). Ethnic communication and identity performance. In H. Giles, S. A.

Reid, & J. Harwood (Eds.), Dynamics of intergroup communication (pp. 17–28). New York:
Peter Lang.
Stella Ting-Toomey
Stella Ting-Toomey is a Professor of Human Communication Studies at California

State University, Fullerton.
Tenzin Dorjee
Department of Human Communication Studies California State University at Fuller

ton Fullerton, CA 92834 USA
Page 32 of 32

Language and Culture

Yoshihisa Kashima , Emiko Kashima , and Evan Kidd

This chapter argues that language and culture have a bidirectional causal link: linguistic
practices emphasizing context (contextualizing practices) may encourage people to
process a visual stimulus as a bounded figure-ground configuration (field dependent),
whereas linguistic practices emphasizing the separation of an object from its context (de
contextualizing practices) may encourage people to abstract figure from ground (field in
dependent). Once a certain cognitive style is in place, people are likely to use their lan
guage in the way congruent with their cognition. Linguistic practices thus act as an im
portant medium through which culture is transmitted and maintained. In support of this
analysis, the authors show that the geographical distribution of linguistic practices over
laps with that of cognitive style, with contextualizing linguistic practices/field-dependent
style in East Asia and decontextualizing linguistic practices/field-independent style in
Western European-derived cultures. Bilinguals show default cultural accommodation, ex
hibiting a cultural pattern congruent with the norm of the linguistic community/language
they are using.
Keywords: culture, linguistic practice, contextualizing, decontextualizing, field independent, field dependent,
holistic cognitive style, analytic cognitive style, bilingualism, biculturalism
Culture consists of a set of nongenetically transmitted information that can potentially in
fluence human action. In this chapter, the term meaning is used to refer to information
that can potentially influence human action. So, culture can be understood as a set of
nongenetically transmitted meaning. When such meaning is widely shared within a popu
lation and transmitted from one generation to another, it is said to be part of culture. In
many ways, language is an integral part of culture thus broadly defined. Language is pre
sumably unnecessary to transmit cultural information if nonhuman primates and other
animals are capable of having a culture (e.g., Whiten & van Schaik, 2007); however, the
type of cultures that humans have is inconceivable without language. Imagine a world
without a language, then try to build a parliamentary democracy, a legal system, a stock
exchange, a banking system, and myriad other infrastructures of everyday life that we
have come to take for granted. Given that these institutions are cultural products, it is
Page 1 of 28

easy to see the role that language must have played in the constitution of human culture.
In short, humans use language to form, maintain, and transform their cultures. Language
is a critical semiotic tool with which humans construct and exchange meaning (Holt
graves & Kashima, 2008).
Without a doubt, humans use language to produce desired social effects and to avoid un
desirable ones. As speech act theorists have suggested (e.g., Austin, 1964; Searle, 1969;
see Holtgraves, 2002), language use is a social action. That is to say, language is used
with an intention to produce desired end results. Nonetheless, using a tool has (p. 47)
consequences that the tool was not designed to produce. When cutting a tree with a saw,
we do not typically intend to make sawdust. Sawdust is a side effect or an unintended
consequence of using a saw. Likewise, using a language can also have its unintended con
sequences, consequences that can make up those aspects of cultures of which people are
not necessarily aware.
Such unintended consequences may stem from two sources. First, we typically use lan
guage with an audience in mind. In other words, language is used not necessarily for the
speaker, but for the listener. The way in which the speaker uses language—choice of
words, grammatical construction, perspective, and the like—is bound to influence the way
the recipient of the description understands what is described. This is because the listen
er usually constructs a mental or “situation model” of the object of description (e.g.,
Zwaan, 2004; Zwaan & Radvansky, 1998). Language can be used to describe an object or
event from a variety of perspectives, each of which will typically invite the listener to con
struct a certain mental model. Consequently, language use may end up shaping the listen
ers’ cognition as an unintended consequence of language as a semiotic tool. Second,
when the speaker uses language for communication, the process of linguistic encoding it
self can help shape the mental model that the speaker ends up making for him- or herself.
There is plenty of evidence of the cognitive impact of language use on the speaker as well
(for a review, see Holtgraves & Kashima, 2008). As we take turns in conversation, the lis
tener and speaker effects of language use can compound and reinforce each other. Once
a certain cognitive style is in place, this can drive language use. Thus, language and cul
ture are tightly connected.
In this chapter, we argue that the ways in which people use their language, which we call
linguistic practices, may act as one of the important mediums of cultural transmission;
that is, language serves as a medium through which cultural information about how to
use one’s mind is transmitted. In other words, we argue that by learning to use a lan
guage in a certain way, we may also learn to use our minds in a certain way. We first de
fine what we mean by linguistic practice and then review the literature to show that there
are a number of linguistic practices that direct listeners’ attention differently. Some lin
guistic practices direct listeners’ attention to the focal object at the expense of the con
text in which it is embedded (decontextualizing), whereas others tend to direct listeners’
attention to the context in which the object is the figure against the contextual ground
(contextualizing). Then, we suggest that the geographical distributions of linguistic prac
tices and cognitive styles appear to overlap. In particular, linguistic practices that tend to
Page 2 of 28

contextualize (contextualizing linguistic practices) are more prevalent in East Asian coun
tries. In contrast, linguistic practices that tend to decontextualize (decontextualizing lin
guistic practices) are more prevalent in Western European-based societies. Interestingly,
the geographical distribution of what Nisbett, Peng, Choi, and Norenzayan (2001) called
analytic and holistic cognitive processing styles appear to coincide with the distribution of
linguistic practices: holistic and analytic styles appear to be prevalent in Western Euro
pean-based and East Asian societies, respectively. We examine the association between
linguistic practices and social perception in monolingual populations, and then we investi
gate their links further by reviewing the literature on bilinguals’ cognition and behavior
when they use different languages.
Language Use and Linguistic Practice

Let us first illustrate what is meant by a linguistic practice. Suppose that you hear the fol
lowing utterance:
(1) The ham sandwich has just spilled the beer all over himself.
What do you make of it? Probably this doesn’t make much sense. Let us then put some
context around it. So, imagine an ordinary diner. A customer orders a ham sandwich and
a beer. The waitress delivers the order. The customer begins to attack his supper…and an
accident! She runs back to the kitchen and makes this utterance to her colleagues. Now,
does it make sense?
The utterance was in fact concocted by a cognitive linguist, George Lakoff (1987, p. 77)
as an example of metonymy, a form of figurative speech in which a part is used to refer to
the whole. So, the ham sandwich is used to refer to the man who ordered it at the diner.
Other examples of metonymy include The White House for the US government, The Krem
lin for the government of Russian Republic, and so on. Nonetheless, the point we wish to
make is not so much about metonymies, but that an utterance like this is possible as an
exemplar of language use, even though it is an unusual nonstandard English usage. We
call such a particular, and possibly one-off, instance of language use a token (p. 48) lan
guage use. However, this type of language use, or a type language use—using a
customer’s order to refer to the customer—could become a prevalent mode of reference
in this diner (e.g., The Milkshake on table 7). If most waiters and waitresses used these
metonymic extensions frequently, then the next generations of waiters and waitresses
might do so as well and subsequently pass on this mode of speech to further generations
who work at the diner. We call such repeated and widespread employment of type lan
guage use a linguistic practice (Holtgraves & Kashima, 2008; Kashima, Kashima, Kim, &
Gelfand, 2006). Whereas a token language use clearly belongs to Saussure’s (1966) pa
role, a linguistic practice is arguably part of language as a system (Saussure’s langue).
Page 3 of 28

Contextualizing Versus Decontextualizing Lin

guistic Practices
The social psychological literature has identified several linguistic practices that can be
summarized under the general dimension of contextualizing versus decontextualizing. To
explain this dimension, imagine a scene in a corporate board room. A middle-aged man in
a business suit is enthusiastically talking to a group of men and women who are intently
listening and taking notes. In verbally describing a concrete scene of this sort, social psy
chology has identified at least three general categories of information that people use: ac
tor, action, and context. Roughly corresponding to these elements, there are three ele
ments in a sentence: subject, predicate, and context. For instance, the event can be de
scribed by a sentence as:
(2) He enthusiastically talks to people in the board room.
“He” is the subject of the sentence, “enthusiastically talks to people” is a predicate, and
“in the board room” describes a physical context in which the action is carried out. Fol
lowing Langacker’s (1987, 1991) cognitive linguistic analysis, we can say that the sen
tence profiles the event as a figure against its background, bringing out the actor-action-
context configuration as a focal object of construal. The actor, “he,” is the focal object
about which this sentence invites a reader to construct a mental model (also see Langack
er, 1991). Nevertheless, one could use different linguistic devices to contextualize or de
contextualize the actor. We will discuss how different linguistic constructions of context,
predicate, and subject can contextualize or decontextualize the focal object and review
empirical studies that examined them.
Context
A first set of devices is probably most obvious and has to do with the linguistic treatment
of the context. Compare the following three sentences:
(3a) He enthusiastically talks to people.

(3b) He enthusiastically talks to people in the board room.
(3c) In the board room, he enthusiastically talks to people.
Relative to (3b), which is the same as (2), (3a) de-emphasizes the context by dropping the
reference to the context. In contrast, (3c) emphasizes the context more than (3b) by plac
ing the context at the beginning. The sentence in (3c) is somewhat awkward in English,
but it is a very common construction in other languages, for instance, in Japanese. Ar
guably, (3a) is a more decontextualizing, but (3c) is a more contextualizing linguistic prac
tice than (3b).
Page 4 of 28

Distribution of Contextual Qualifications
Fig. 3.1 Mean percentages of contextual qualifica

tions among English and Korean speakers.
From figure 4 in Kashima, Y., Kashima, E. S., Kim, U.,

& Gelfand, M. (2006). Describing the social world:
Object-centered versus process-centered descrip
tions. Journal of Experimental Social Psychology, 42,
388–396.
The linguistic practices of using contextual qualifications appear to be distributed differ

ently in different regions of the world. Kashima, Kashima, Kim, and Gelfand (2006)
conducted a cross-cultural, cross-linguistic study of verbal descriptions of social objects
in Australia and Korea and examined the likelihood of English and Korean speakers using
contextual qualifications in their written descriptions. In their study, participants were
asked to verbally describe themselves and a friend in the individual condition, their own
and their friend’s relationships in the relationship condition, and their own and their
friend’s families in the group condition. They were asked to make up to 10 open-ended
statements to describe each target. In describing the self (regardless of whether in the in
dividual, relationship, or family condition), there was no difference in the use of contextu
al qualifications between English and Korean speakers. However, in describing the other,
Koreans used contextual qualifications more than did English speakers (Figure 3.1).
Although contextual qualifications may be more prevalent in East Asian languages than in
English specifically in the context of describing others, there may be an even subtler dif
ference in the way contextual qualifications are used in East Asia. Masuda and Nisbett
(2001) presented American and Japanese students with graphical clips of swimming fish.
After viewing each clip twice, they were asked to orally describe what they saw. Verbal
protocols were carefully coded in terms of salient focal (p. 49) objects (e.g., swimming
fish) and peripheral background objects (e.g., seaweed, rocks). Although there were no
differences in the likelihood of mentioning the salient focal objects, Japanese described
more peripheral objects and mentioned time more than did Americans. Furthermore,
Japanese participants were more likely to start their first utterance with a description of
the scene by mentioning the peripheral objects, whereas Americans were more likely to
start their first utterance with a description of the focal object. Here, (4) is an exemplar
English utterance; (5), Japanese (Masuda & Nisbett, 2001, p. 928):
(4) I saw three big fish swimming from right to left.
Page 5 of 28

(5) At the beginning, a big fish was swimming towards the green seaweed.
In oral language use too, Japanese appear to use more contextualizing linguistic practices
than do Americans by referring to the context of a focal object first.
Implications for Cognition

The linguistic practice of not adding contextual qualifications may be linked to a funda
mental attribution error (Ross, 1977) or a correspondence bias (Gilbert & Malone, 1995)
—a well-known tendency for those with European-based cultural backgrounds to under
use contextual information in making a judgment about an individual. It was Jones and
Davis’s (1967) classic study on attitude attribution that identified this tendency among
American students. When they read an essay expressing a positive (vs. negative) attitude
toward Castro’s Cuba, they tended to attribute the corresponding attitude to the essay
writer even when they were told that the essay writer had no choice about the type of es
say he or she wrote. Although East Asians (Japanese in Kashima, Siegal, Tanaka, &
Kashima, 1992; Koreans in Choi & Nisbett, 1998) show a similar tendency, relative to
their American counterparts, they were more likely to take the contextual information in
to consideration when the context was made salient (Choi & Nisbett, 1998). Context ap
pears to be more cognitively accessible for East Asians than for Westerners.
Predicate
Another linguistic device used to contextualize or decontextualize the actor is the predi
cate of an utterance. According to Semin and Fiedler’s (1988, 1991; Fiedler & Kruger,
this volume) linguistic category model (LCM), different types of predicates (e.g., nouns,
adjectives, verbs) have different degrees of contextual information. Take the following
four examples:
(6a) He is a chatter box.

(6b) He is talkative.
(6c) He likes to talk.
(6d) He talks a lot.
In order to refer to the focal actor’s action, (6a) uses a noun phrase, “chatter box”; (6b)
uses an adjective, “talkative”; (6c) adds a state verb “likes”; and (6d) uses an action verb,
“talks.” LCM suggests that nouns and adjectives are most abstract and do not imply the
context in which an actor’s action is carried out. A noun phrase tends to essentialize the
referent (e.g., Gelman & Heyman, 1999), and adjectives tend to abstract specific behav
iors into a broader dispositional category. Some noun phrases have been shown to imply
an even more stable and generalizing characteristic than adjectives (e.g., Gelman & Co
ley, 1990; Hall & Moore, 1997). In contrast, verbs tend to be more contextualizing. Action
verbs describe concrete actions such as “talk,” “run,” and “walk”; state verbs describe
the state of the actor such as “like,” “abhor,” and “envy.” Action verbs are more contextu
alizing than state verbs because they typically imply a concrete action in context, where
Page 6 of 28

as state verbs imply that the action reflects the state the actor is in, which is presumably
more enduring than a dynamic action. Thus, (6a) to (6d) order the utterances from the
most decontextualizing to the most contextualizing practice.
Distribution of Predicate Use
Fig. 3.2 Percentages of different types of predicates

used by Australians and Koreans.
From figure 1 in Kashima, Y., Kashima, E. S., Kim, U.,

& Gelfand, M. (2006). Describing the social world:
Object-centered versus process-centered descrip
tions. Journal of Experimental Social Psychology, 42,
388–396.
The prevalence of different types of linguistic practices for predicates was examined by
Kashima (p. 50) et al. (2006), which we described earlier. Figure 3.2 shows the percent
ages of the four types of predicates used by English and Korean speakers in Australia and
Korea, respectively. There was an obvious and very large difference. Basically, English
speakers (top panel) used adjectives most to characterize any social object, regardless of
whether it was an individual, relationship, or group. In contrast, Korean (lower panel)
speakers used state verbs most and noun phrases somewhat. Overall, English speakers
used adjectives more and state verbs less than did Korean speakers. Kashima et al. com
puted an objectification index, which indicated the extent to which the target social ob
ject was objectified (proportion of nouns and adjectives—proportion of state and action
verbs). Australians objectified the target more (M =.54) than did Koreans (M = –.51). Al
though there was some contextual variations (Australians objectified individuals and
groups more than relationships), this overall language difference was significant through
out.
Maass, Karasawa, Politi, and Suga (2006) also found similar differences in their compar
isons of Italian and Japanese speakers. In their study 1, they had their participants de
scribe 10 aspects of their acquaintance (male and female) or gender groups (men and
women in general). Consistent with Kashima et al. (2006), Italians used adjectives more,
but verbs less, than did Japanese. In study 2, they compared Italian and Japanese individ
ual and group descriptions (again, acquaintance and gender groups) in general and in
specific contexts (at home and at school/work). Again, they found Italians used adjectives
more, but verbs less, than did Japanese for both individuals and groups. Thus, the tenden
Page 7 of 28

cy to use decontextualizing predicates was observable regardless of the type of social ob
jects.
The difference in prevalence of verb use in describing objects may be responsible for an
analogous difference in patterns of children’s first language acquisition. The vocabulary
of children learning East Asian languages such as Korean (Choi & Gopnik, 1995) and Chi
nese (Tardif, 1996; Tardif, Fletcher, Liang, Zhang, Kaciroti, & Marchman, 2008; Tardif,
Gelman, & Xu, 1999) tend to have a higher proportion of verbs than the vocabularies of
children learning European languages such as English, whose vocabularies tend to be
dominated by nouns. Indeed, Chan, Brandone, and Tardif (2009) found a greater preva
lence of verbs in Chinese mothers’ speech in their interaction with their children when
compared to English-speaking mothers’ speech.
Not only social objects, but also social events appear to follow a similar pattern. Semin,
Gőrts, Nandram, and Semin-Goossens (2002) examined Hindustani, Turkish, and Dutch.
Hindustani and Turkish appear to be similar to East Asian languages, and Dutch, to other
European languages, in linguistic (p. 51) practice surrounding predicate use. In study 1,
they asked Dutch and Hindustani (common parts of Hindi and Urdu languages spoken
widely in North India and Pakistan) speakers living in Amsterdam to generate as many
emotion words that came to their minds first and then to describe five “critical events” in
one condition or “critical experiences” in the other condition. First, Dutch speakers were
found to use fewer state verbs to name emotions than were Hindustani speakers, al
though there was no difference in adjective use, and Dutch speakers used more nouns
than did Hindustani speakers. Second, Dutch speakers used more nouns and fewer verbs
to describe critical events or experiences than did Hindustani speakers. In study 2, Semin
et al. compared Dutch and Turkish speakers’ descriptions of their own or friend’s positive
and negative life events. In this study, level of abstractness was indexed using the linguis
tic categories—adjective and noun were coded as most abstract and action verbs as least
abstract. Dutch emotion terms and event descriptions were more abstract than Turkish
ones.
All in all, there appears to be differences in prevalence of contextualizing and decontextu

alizing linguistic practices between European-based and East Asian or South Asian cultur
al groups. Dutch, English, and Italian speakers appear to use more decontextualizing
predicates of adjectives for social objects and events and nouns for emotions (but not for
social objects). Chinese (mainly Mandarin), Hindustani, Korean, Japanese, and Turkish
speakers appear to use more contextualizing, verb-based predicates, but tend not to use
decontextualizing, adjective-based predicates.
Implications for Cognition

Prevalence of adjective as opposed to verb use in social description may have a signifi
cant implication for cultural differences in social cognitive processes. In particular, this
may be closely linked to a correspondence bias (Gilbert & Malone, 1995), a tendency for
those with European-based cultural backgrounds to use personality traits (e.g., generous,
kind) in explaining social behavior. A large number of studies have shown that Americans
Page 8 of 28

often automatically encode a person’s concrete behavior in terms of personality traits

(e.g., Uleman, Winborne, Winter, & Shechter, 1986; Winter & Uleman, 1984), whereas it
has been shown that East or South Asians have a weaker tendency for characterizing peo
ple by their dispositions. This has been most clearly documented in open-ended descrip
tions of an individual (see Choi, Nisbett, & Norenzayan, 1999, for a review). Shweder and
Bourne’s (1984) pioneering research showed that North American English speakers tend
to use personality traits, whereas people in Orissa, India, tend not to use personality
traits in describing an acquaintance. Similarly, examining self-descriptions of European
Americans and Koreans, Rhee, Uleman, Lee, and Roman (1995) showed that Americans
were more likely to use traits than were Koreans. Analogous cultural differences have
been replicated in studies of person descriptions regardless of whether the self or other
was described (e.g., Bochner, 1984; Cousins, 1989; Dhawan, Roseman, Naidu, Thapa, &
Rettek, 1995; Trafimow, Triandis, & Goto, 1991), although some researchers examined
the underlying psychological meaning more closely than others (e.g., Miller, 1984, 1987).
The correspondence bias goes with the preference for adjectives in the following way.
Personality traits are concepts that are used to characterize individuals and their underly
ing dispositions. Someone who is said to be deceptive is presumed to perform deceptive
acts (e.g., lying), often in multiple contexts. It is interesting to note, however, that person
ality traits are often verbally described by using adjectives (e.g., “deceptive”), although
this is not always the case (e.g., liar). This observation suggests that the Western tenden
cy to use personality traits in person description may be explained by the Western linguis
tic practice of using adjectives in object descriptions (see Kashima et al., 2006; Maass et
al., 2006). Evidence consistent with this line of reasoning comes from Kashima et al.
(2006). They replicated the previous findings to show that English-speaking Australians
tended to use personality traits more than did Koreans, but when participants’ linguistic
practices (likelihoods of using adjectives and verbs) were statistically controlled, the cul
tural difference in personality trait ascription became nonsignificant.
The habitual tendency to use adjectives to describe a social object, and consequently to
habitually attribute personality traits to an individual, may have judgmental and percep
tual consequences. There is evidence to suggest that, given the same information about a
social action, Westerners are more likely to attribute its causes to internal, dispositional
characteristics and less likely to attribute them to external, situational factors. Miller
(1984) showed that American children and students were more likely to explain social
events by references to internal, dispositional characteristics than were their counter
parts from Orissa, India. Kashima et al. (1992) found that Australian students tended
(p. 52) to attribute the attitude expressed by an essay more strongly to the essay writer
than did Japanese students (also see Choi & Nisbett, 1998; Krull et al., 1999; Miyamoto &
Kitayama, 2002). Morris and Peng (1994) described an American and a Chinese murder
case to physics graduate students from the United States and Chinese-speaking countries
(Mainland, Hong Kong, and Taiwan) and asked them to rate the importance of personal
dispositional factors and situational factors as causes of the murder. In both cases, Chi
nese students rated situational causes as more important than did American students,
suggesting the tendency to make stronger attributions to context. For the Chinese mur
Page 9 of 28

der case, however, Americans showed a stronger tendency to attribute it to the

murderer’s dispositions than did their Chinese counterparts.
Morris and Peng (1994) extended this phenomenon to visual perception. They presented
visual information about movement patterns of animate objects (a school of fish) and
inanimate objects (circles and squares) to high school students and graduate students
from the United States and from areas with Chinese cultural backgrounds (e.g., Hong
Kong, Mainland, Taiwan) and asked them to explain the movements. American high
school students tended to make stronger dispositional and weaker contextual attributions
for animate objects than did their Chinese counterparts, although little differences were
found among graduate students or for explanations of inanimate objects.
Finally, Maass et al. (2006) reported that Italians and Japanese showed a memory bias
congruent with the suggestion that Westerners are more likely to encode social informa
tion with adjectives than are Easterners. In their study 3, they presented Italian and
Japanese students with two versions of letters of recommendation, one that used adjec
tives to describe the person and the other that used verbs. After a 5-minute distractor
task, they were instructed to list all the information verbatim. Italians correctly recalled
adjectives more than verbs, whereas the Japanese recalled more verbs than adjectives.
Maass et al. replicated this with more controlled stimulus material in study 4. Sixteen
verb-adjective pairs that share the same word stem were constructed in Italian and
Japanese (e.g., “Marco/Hideo dominates others” vs. “Marco/Hideo is dominant”). Partici
pants received a list of person descriptions, half of which included adjectives and the oth
er half verbs. After a 10-minute distractor task, they performed a recognition task in
which they had to indicate whether they had read the exact same word earlier or not.
Again, Italians recognized adjectives more accurately than verbs, whereas Japanese rec
ognized verbs more accurately than adjectives. In all, linguistic practices may facilitate
the processing of practice-congruent stimuli.
All in all, the linguistic practice of adjective use appears to be prevalent in European-
based linguistic communities, whereas that of verb use seems to be distributed widely
across Asia and the Middle East. Although there is no direct evidence for linking this with
correspondence biases in cognition, there is circumstantial evidence for this possibility.
Given that there is a correlation between the likelihood of using an adjective and the like
lihood of not using a contextual qualification (Kashima et al., 2006), some aspects of cor
respondence bias may be partly due to linguistic practices of adjective use.
Subject
A third set of linguistic devices for contextualization and decontextualization has to do

with the treatment of the subject of a sentence. Although the explicit mention of a subject
is obligatory in some languages (e.g., English, French, German), it is possible to drop a
subject in a sentence in other languages (e.g., Chinese, Japanese, Korean). Consider the
following two sentences:
(7a) He is talking to people in the board room.

Page 10 of 28

(7b) Is talking to people in the board room.
Although (7b) is an ungrammatical sentence in English, something like this is grammati

cal in other languages, such as Japanese. Relatively speaking, (7a) profiles the actor more
than (7b) by explicitly mentioning the actor. In contrast, (7b) does not even mention the
actor, suggesting the low prominence placed on the actor; instead, it highlights the action
and the process of “talking” in situ, rather than the actor that carries out the action.
This phenomenon is called a pronoun drop (e.g., Chomsky, 1981),1 the situation in which a
pronoun that may be used as the subject of a sentence is dropped (e.g., Kashima &
Kashima, 1998). To illustrate, consider an utterance used by a Japanese housewife who
was talking about Hilary Clinton in a study conducted by Kashima and Kashima (1997):
Page 11 of 28

(8) *E (watashi-wa)
Sonna, atta koto mo hanashita koto
mo nai kara
like met NOM also spoke NOM
also NEG-BE because
‘Because (I’ve) never met (her) nor spoken (to her)’
Page 12 of 28
PRINTED FROM OXFORD HANDBOOKS ONLINE (www.oxfordhandbooks.com). © Oxford University Press, 2018. All Rights Reserved. Under the terms of the licence agreement, an individ
ual user may print out a PDF of a single chapter of a title in Oxford Handbooks Online for personal use (for details see Privacy Policy and Legal Notice).

In the original Japanese utterance, the first person singular pronoun, “watashi,” is
(p. 53)
dropped as indicated by *E in front of the parentheses, which represents ellipsis. In fact,

there are three places where a pronoun could have been used had it been an English sen
tence (indicated by the parentheses in the English translation); however, no pronoun was
used at all in the Japanese original. There are some languages (e.g., Spanish) in which
verb forms change depending on the subject, thereby providing information about the
subject of the sentence; however, verbs do not change forms in other languages, such as
Japanese. Still, pronoun drops can occur. For instance, Kashima and Kashima (1997)
reported that Japanese housewives dropped the pronominal subject 87 percent of the
time when they conversed with their friends and 54 percent of the time when they con
versed with a stranger. In contrast, English-speaking Australians never dropped the
pronominal subject regardless of whom they conversed with.
More recently, a systematic research on language typology distinguished five types of

pronoun use (Dryer, 2011). Of Kashima and Kashima’s (1998) non–pronoun-drop lan
guages, there appear to be two subtypes: one in which a pronominal subject is a separate
word used in the same sentential position as a nominal subject (e.g., English), and the
other in which a distinct pronominal word is used in a location different from a nominal
subject (e.g., Hausa). Of pronoun-drop languages, Dryer distinguished three types: one
that adds verb affixes to disambiguate the subject (e.g., Spanish), one that uses clitics as
sociated with various words in a sentence (not necessarily verbs) to do so (e.g., Polish),
and one that does neither (e.g., Japanese). In the last type, the subject needs to be dis
cerned from the context of an utterance. It was also noted that there are some languages
that have a mixture of these different types.
Distribution of Pronoun Drop

Pronoun drop is a widespread linguistic practice in many parts of the world. The World
Atlas of Language Structure Online (WALS Online; Dryer & Haspelmath, 2011) shows
that a large number of languages—some spoken by small numbers of people, others by
very large numbers—indeed show a pronoun drop. In fact, many languages in East Asia,
Southeast Asia, and South Asia, as well as the Spanish and Portuguese spoken in Latin
and Central America, drop pronominal subjects.
Psychological Implications of Pronoun Drop

E. S. Kashima and Y. Kashima (1998) suggested that this type of pronoun drop tends to
contextualize the actor by reducing the prominence of the actor within the context of ac
tion. In addition, they suggested that Hofstede’s (1980, 1991) dimension of individualism
may be related to pronoun drop. Individualism is one of the dimensions that Hofstede
used to characterize different cultures in the world. Using surveys of work values at a
multinational corporation (IBM), Hofstede found that workers from different countries or
regions of the world tend to value different aspects of the work. Individualism was mea
sured by items that tapped the extent to which workers valued their own work or leisure
time as opposed to the physical and social condition of their work. Hofstede interpreted it
as an opposition between the valuing of the individual and that of the organization. Gen
Page 13 of 28

eralizing from this, he argued that countries with high individualism scores tend to em
phasize personal identity rather than collective identity, and, in those countries, it is nor
mative to look after oneself and one’s own nuclear family, rather than one’s extended fam
ily. The most individualistic countries included the United States, Australia, and other An
glophone countries, as well as mainly Western European countries; the least individualis
tic countries were from East and South Asia (e.g., Hong Kong, India), as well as Latin
America (e.g., Venezuela, Colombia). Hofstede also found that his individualism index cor
related with per capita gross national product (GNP) in 1970 at. 82; in other words, rich
er countries tend to be more individualistic.
Because Hofstede’s individualism index highlights the individual worker’s wishes and de
sires as opposed to the context in which the worker finds him- or herself as an important
aspect of work, Kashima and Kashima (1998) conjectured that the absence of pronoun
drop, or high prominence of the person, may be related to individualism. To examine this,
the language spoken by the majority in a country was used to code each country for its
language. Consistent with this reasoning, non–pronoun drop correlated with individual
ism at. 75 across 60 countries. An analogous analysis was conducted using language as a
unit of analysis. For this purpose, each language was given a score of individualism by av
eraging the individualism scores of countries whose major languages were the same. The
correlation of. 64 was obtained across 30 languages. After a correction (Kashima &
Kashima, (p. 54) 2005) partialling out other linguistic characteristics (i.e., number of first-
person and second-person pronouns available in a language) did not alter these correla
tions appreciably. It is also important to note that correlations of non–pronoun drop with
other dimensions of cultural values (e.g., Schwartz, 1994) became nonsignificant when in
dividualism was statistically controlled for, thus suggesting the centrality of individualism
as a correlate of linguistic practice of pronoun drop.
Y. Kashima and E. S. Kashima (2003) extended this analysis to examine whether pronoun
drop can account for country-level individualism in conjunction with societal factors
(1970 GNP per capita; base 10 log transformed due to skewness) and environmental fac
tor (climate as indexed by the latitude of a country’s capital city). It was surmised that
these societal and environmental factors represented material aspects of the environment
in which a cultural group is embedded, whereas pronoun drop reflects a symbolic aspect
of the culture. Gross national product, climate, and pronoun drop were all significant pre
dictors of individualism in a multiple regression analysis. In addition, pronoun drop was a
moderator of the relationships of GNP and climate with individualism. In pronoun-drop
countries, climate was the only correlate of individualism, whereas in non–pronoun-drop
countries, GNP was the only significant correlate. Overall, this pattern of findings sug
gests that material and symbolic aspects of culture can interact to have a significant im
pact on cultural values.
In a similar vein, Tabellini (2008) examined the implications of linguistic practices for cul
tural values and governance indicators. Pronoun drop may be related to the valorization
of a particularistic interpersonal relationship (i.e., drawing a strong distinction between
those who are close to the speaker and those who are not) rather than a universalistic
Page 14 of 28

stance to people. Similarly, distinction between formal and informal second-person pro
nouns (e.g., tu and vu as in French) may also contribute to the persistence of a particular
istic value orientation. His analysis indeed showed higher levels of generalized trust
(“most people can be trusted”) and respect (“high tolerance and respect for other peo
ple”; universally treating all humans as trustworthy or respectable, rather than only close
others), in countries where their languages did not drop pronouns or did not have differ
entiated second-person pronouns. This was, in fact, generalized to individual-level analy
ses, showing that speakers of languages without pronoun drop and with no differentiated
second-person pronouns were also found to value trust and respect more highly. Further
more, the quality of governance (effectiveness of law enforcement and quality of bureau
cracy) was higher in those countries whose languages do not permit pronoun drop and do
not differentiate second-person pronouns. Licht, Goldschmidt, and Schwartz (2007) also
found that norms of governance that emphasize the rule of law, lower levels of corrup
tion, and democratic accountability were more prevalent in countries where pronouns are
not dropped.
Intriguingly, Fincher, Thornhill, Murray, and Schaller (2008) examined prevalence of

pathogens as an ecological cause of cultural practices that emphasize the maintenance of
a tight ingroup boundary (i.e., collectivism). They surmised that if an ecological environ
ment has widespread human infectious diseases that are detrimental to human reproduc
tive fitness (e.g., malaria, leprosy, dengue, typhus, tuberculosis), a culture that restricts
outsiders’ contact with an ingroup (for fear of unknown infection) may evolve as an adap
tation to this environment. Fincher et al. used a variety of indices of individualism and
collectivism to index these cultural practices. Using pronoun drop as one such index, they
reported that historical prevalence of pathogens predicted the linguistic practice of pro
noun drop (r = .63). Consistent with their reasoning, contemporary pathogen prevalences
correlated with pronoun drop (r =. 44) at a lower level than did the past levels of
pathogens, suggesting that it was past ecology that set a stage for cultural adaptation. All
in all, the linguistic practice of pronoun drop—de-emphasizing the actor in a linguistic
construction of a social event—may be related to cultural values. Only time can tell
whether the empirical relationship between pronoun drop and individualism is a timeless
phenomenon that transcends the historical context or one that holds only at this point in
human history.
Summary
Different linguistic practices can make use of different elements in utterances to contex
tualize or decontextualize the actor. By adding a contextual qualification, using a verb
rather than an adjective or dropping the pronominal subject that refers to the actor, the
prominence of the actor can be reduced and the actor’s action and therefore agency can
be embedded within the context of the activity. The foregoing discussion suggests that
many of the contextualizing linguistic practices are more widely distributed in East Asia,
and more decontextualizing (p. 55) linguistic practices are more prevalent in Western Eu
ropean linguistic communities.
Page 15 of 28

The notion of contextualizing versus decontextualizing linguistic practices is related to

but different from Hall’s (1976) high- versus low-context communication. According to
Hall, a low-context communication codes as much explicit information as possible in the
message or the linguistic utterance. In contrast, a high-context communication leaves
much of the information in the physical context or the communicators’ minds, expecting
that the listener can decode its meaning without explicit references. In some sense, pro
noun drop can be thought of as a kind of high-context communication; however, verb and
context uses are not. In the latter, more explicit information is coded in the utterance.
Conceptually, our analysis of linguistic practices focuses on whether the surface form of
the linguistic construction encodes or emphasizes the context, whereas high- versus low-
context communication is more to do with whether the context of communication is used
to decode it. Nonetheless, it is important to recognize that low-context communication
and decontextualizing linguistic practices are related phenomena in that they both de-em
phasize the importance of context. Given that Hall suggests many East Asian countries
(e.g., Japan, China) are high-context cultures, it would be useful to further explore the re
lationship between the two.
Linguistic Practice and Cognitive Style

The geographical distribution of contextualizing and decontextualizing linguistic prac
tices appears to overlap with that of holistic and analytic cognitive styles, as shown by
Nisbett and his colleagues (Nisbett, 2003; Nisbett et al., 2001). Holism considers an ob
ject to be embedded within a whole; analyticism takes objects to be isolates with distinct
properties. They include as significant aspects what Witkin and his colleagues called
field-dependent and field-independent perceptual and cognitive styles (e.g., Berry, 1979;
Witkin & Goodenough, 1977). Field dependents perceive an object as embedded in a
background, whereas field independents perceive an object as a standalone. Arguably, the
correspondence bias in person perception discussed earlier is an instance of analytic,
field-independent cognitive style. Because there is an obvious similarity between contex
tualizing versus decontextualizing linguistic practices and field-dependent and field-inde
pendent cognitive styles, it raises a question about the psychological relationship be
tween the two. Do analytic field independents (vs. holistic field dependents) use decontex
tualizing (vs. contextualizing) linguistic practices, and vice versa?
There is some evidence to link the two. Klein, Ventura, Fernandes, Garcia Marques, Lica
ta, and Semin (2010) examined Portuguese-speaking literate, illiterate, and ex-illiterate
(people who have grown up illiterate, but learned written language as adults) people’s lin
guistic practice (particularly predicate use) and cognitive style. To measure linguistic
practice, they had their participants describe drawings of people performing personal
and interpersonal behaviors and coded the predicates in terms of the LCM; then, an in
dex of abstractness was computed. Cognitive style was measured by a modified version of
the framed-line task (Kitayama, Duffy, Kawamura, & Larsen, 2003). In the original ver
sion, a square frame with a vertical line was presented, and then another square frame of
a different size was shown. Participants’ task was to draw a vertical line of the same
Page 16 of 28

length (absolute task) or a line whose proportional length relative to the square equals
the original line length relative to the original square (relative task). Amount of error for
absolute tasks indexed analytic style, whereas that for relative tasks indexed holistic
style. Klein et al. modified this to make it into a reaction time (RT) task. In this version, an
original framed line is presented first, and a second framed line with a different size is
presented next. Participants are asked to respond as quickly as possible whether the sec
ond line had the same length or not (absolute task) and whether the second line relative
to its frame was proportionally equal to the first relative to its frame or not (relative task).
The average RT for relative tasks was subtracted from the average RT for absolute tasks
to index analytic (vs. holistic) cognitive style. They reported substantial correlations be
tween abstract predicate use and analytic cognitive style: .67 for literates,.83 for illiter
ates, and .75 for ex-illiterates. Regardless of literacy, linguistic practice and cognitive
style were highly related. In addition, they reported that literates used more abstract
predicates and more analytic cognitive style than did illiterates or ex-illiterates. Although
this study shows an intriguing correlation between cognitive style and linguistic practice,
future research is needed to see if there is any causal link from one to the other.
Bilinguals, Biculturals, and Language-Culture

Dynamics
If, in fact, language and culture are linked together in some sense, what happens to peo
ple (p. 56) who can speak two (or more) languages and know the cultures of those linguis
tic groups (bilingual/bicultural individuals)? Hereafter, we call these individuals bilin
guals; however, it is important to keep in mind that these individuals not only speak two
languages, but also are bicultural; that is, they can think and act in ways that are accept
able to both linguistic communities. So, the question is, for instance, do Chinese-English
bilinguals think and act more like English speakers when they use English relative to
when they use Chinese or vice versa? Cultural accommodation (Chen & Bond, 2010) is a
phenomenon in which bilinguals think and act more like a typical person of the linguistic
community that speaks the language they are using at the time (e.g., Chinese-English
bilinguals acting like English speakers when speaking English; cf. Giles, Coupland, &
Coupland, 1991). In contrast, the opposite phenomenon may be called cultural affirmation
(cf. Yang & Bond, 1980), in which bilinguals think and act more like a typical person of
the linguistic community that speaks the language they are not using at the time (e.g.,
Chinese-English bilinguals acting like Chinese when speaking English).
There is evidence of both cultural accommodation and cultural affirmation. Early evi
dence of cultural accommodation came from Ervin’s (1964) study of French-English bilin
guals who lived in the United States. She had the same English-French bilingual inter
viewers administer thematic apperception tests for each participant and recorded their
oral responses for content analysis. Based on previous ethnographic work, cultural differ
ences were hypothesized, and the bilinguals’ narratives were examined relative to the
culturally typical responses. Although the bilinguals did not show language-congruent
Page 17 of 28

narrative differences for all dimensions hypothesized, they did show cultural accommoda
tion for three of the eight dimensions examined. Narratives were found to be less
achievement oriented for women, contain more themes of withdrawal and autonomy, and
were more verbally aggressive to peers in French than in English. Together with Ervin’s
(1961) earlier research on color naming by Navaho-English bilinguals, these pioneering
studies showed the importance of language in bilinguals’ communicative activities (nar
rating and naming).
Similarly, there is evidence of cultural accommodation in the area of values and self-cog
nitions. Ralston, Cunniff, and Gustafson (1995) had Chinese-English bilingual managers in
Hong Kong respond to Schwartz’s value survey using Chinese and English and compared
their responses to monolingual US managers. They reported cultural accommodation in
four of the ten dimensions, including achievement, hedonism, tradition, and security
when the bilinguals used English. Trafimow, Silverman, Fan, and Fun Law (1997) found
that Hong Kong Chinese-English bilinguals generated more collectivist and less individu
alist self-cognitions when they described themselves in Chinese than in English. Likewise,
Ross, Xun, and Wilson (2002) also found evidence for cultural accommodation. Chinese-
born Chinese Canadians tended to endorse more Chinese worldviews, described them
selves less positively, and reported lower levels of self-esteem in Chinese than in English.
Furthermore, they did not find accommodation on interdependence. Ramirez-Esparza,
Gosling, Benet-Martinez, Potter, and Pennebaker (2006) compared English-Spanish bilin
guals’ self-report of Big Five personality characteristics, again finding a cultural accom
modation pattern in extraversion, agreeableness, and conscientiousness.
Other researchers have reported more complex patterns of findings. Watkins and Gerong
(1999) compared Filipino English-Cebuano bilingual high school students’ verbal descrip
tions of themselves in the two languages. Male bilinguals were influenced more by the
school context when using English than Cebuano, their local language. However, females
did not show any effect of language. In contrast, Kemmelmeier and Cheng (2004)
reported that female Hong Kong Chinese-English bilinguals showed a cultural accommo
dation in terms of independent self-construal; namely, more independent in English than
in Chinese. However, there was no evidence of accommodation for men.
Cultural accommodation appears to occur for other cognitive domains. Alvarado and
Jameson (2011) compared Vietnamese-English bilinguals’ judgments of similarity in emo
tion terms, especially those of shame and anguish, within the context of other emotion
terms such as excitement, happy, sad, fear, anger, and the like. They reported that their
judgments shifted in line with the monolingual speakers of the language that they were
using at the time. Lechuga and Wiebe (2011) found greater overconfidence in probability
judgment among Mexicans than Americans and reported an analogous difference in over
confidence when Mexican-American (Spanish-English) bilinguals were asked to perform
probability judgments in Spanish and English. Nonetheless, Ji, Zhang, and (p. 57) Nisbett
(2004) found that bilinguals from Mainland China and Taiwan showed cultural accommo
Page 18 of 28

dation in the domain of holistic cognitive style; however, Hong Kong and Singaporean
bilinguals did not show any effect of language.
Finally, cultural accommodation was observed in behavior. Sussman and Rosenfeld (1982)
examined interpersonal distance when Japanese and Venezuelan participants used Eng
lish or their native language in their conversations. Here, too, they found cultural accom
modation—Venezuelan bilinguals increased their distance when using English than when
using Spanish, whereas Japanese bilinguals decreased their distance when speaking Eng
lish than when speaking Japanese.
Chen and Bond (2010) conducted an elaborate study in which Hong Kong Chinese-Eng
lish female bilinguals were interviewed by Caucasian and Chinese male interviewers in
English and Cantonese. Note that this enabled them to examine the joint effects of inter
viewer identity and language. The participants’ own self-perceptions and observers’ rat
ings of behavioral patterns were examined in terms of extraversion, openness, and as
sertiveness, the dimensions on which native speakers of English and Cantonese were per
ceived to differ significantly, with English native speakers rated higher on all three. They
reported a cultural accommodation for behavioral patterns as rated by observers only
when the interviewer was Chinese; there was no accommodation with Caucasian inter
viewers, where the participants’ behaviors were rated as more extraverted, open, and as
sertive than when interacting with the Chinese interviewers. They did not find a system
atic accommodation or affirmation on self-perceptions, however.
In contrast, evidence of cultural affirmation was reported by Yang and Bond (1980; Bond
& Yang, 1982). Hong Kong Chinese bilinguals were asked to respond to a questionnaire
that was designed to tap respondents’ endorsement of more Chinese or Western atti
tudes. Each item had two options, one of which was stereotypically Chinese and, the oth
er, stereotypically Western. English and Chinese versions of the questionnaire were con
structed and administered by Chinese Cantonese and Caucasian American English speak
ers. The responses were more Chinese for the English than for the Chinese version. Bond
and Yang (1982) replicated this finding using the same instrument, as well as another
questionnaire that measured people’s endorsement of Chinese traditional sayings. Again,
Hong Kong Chinese bilinguals responded in a more Chinese way when the questionnaire
was in English than in Chinese. Interestingly, however, they found that this tendency was
reversed in some items. Some quarter-century later, Chen and Bond (2007) similarly
found cultural affirmation among Cantonese speakers in Hong Kong and Mandarin speak
ers in Beijing. In both samples, they found that the bilinguals exhibited a more Chinese
normative pattern when responding to the English than Chinese version of a question
naire, namely, a higher level of Chinese identification, a lower level of Western identifica
tion, and a lower level of self-esteem.
All in all, the language used by bilinguals does not appear to be the only determinant of
psychological processes. Indeed, Oyserman and Lee’s (2008) meta-analysis of studies that
examined the effect of language priming on participants’ endorsement of cultural atti
tudes and values reported that, of the 10 studies they examined, the 95 percent confi
Page 19 of 28

dence interval of their effect sizes was –.02 and .21 with the mean of .10, suggesting that
there was fairly large variability and there was not clear evidence of cultural accommoda
tion overall. It is likely that bilinguals who are also bicultural have default tendencies to
go along with the cultural pattern associated with the language; however, under some so
ciocultural circumstances, they may form communicative intentions to emphasize the cul
ture associated with the other language that they are not using at the time, so as to ex
hibit cultural affirmation.
A number of potential factors can trigger affirmation. One is the type of bilinguals. Com
pound bilinguals have learned two languages (or potentially more than two) in the same
social settings, mixing them at the same time; coordinate bilinguals have learned two or
more languages in distinct social settings, using one in one setting while using the other
language in the other setting (Ervin & Osgood, 1954). The latter type of bilinguals may
show clearer accommodation as a function of language used (e.g., Ervin’s studies). Anoth
er factor is an interactant’s identity. As Chen and Bond (2010) found, language may have
an effect when bilinguals interact with an actor whose identity is associated with one of
the languages (in their case, Chinese). The interactant’s identity may exert an effect sepa
rate from language itself. A third factor is the content domain. Bond and Yang (1982)
reported that Hong Kong Chinese bilinguals showed accommodation and affirmation de
pending on the item content of questionnaires. Marin, Triandis, Betancourt, and Kashima
(1983) found some evidence of more socially desirable (p. 58) response styles in some cas
es. Apparently, bilinguals’ psychological processes at any one time are a result of complex
interactions among different social factors. Future research needs to investigate these
complexities.
Conclusion
Language is a critical tool with which humans construct their cultures. The infrastructure
of everyday life that we have come to take for granted has largely been enabled by the
human ability to use language to construct and exchange meaning. Nonetheless, lan
guage as a tool can have psychological consequences that language users do not neces
sarily intend to produce. Linguistic practices are types of language well established and
in widespread use in a population, so much so that they may be considered an aspect of
language as a system. They can have unintended consequences as well, which make up
those aspects of human culture that their makers are not always aware of.
We argued that some linguistic practices are contextualizing, tending to contextualize an

object of description within its surroundings, whereas others are decontextualizing, tend
ing to abstract an object from its background and treating it like a transferable entity. We
pointed out that the geographical distribution of these linguistic practices seems to over
lap with the distribution of different cognitive styles. In processing the figure-ground
compound stimuli, the field-independent style tends to focus attention on the figure of
perception and construal at the expense of its ground, whereas the field-dependent style
processes it while binding the figure and ground together as a configuration. Contextual
Page 20 of 28

izing (decontextualizing) linguistic practices appear to go with field dependence (field in
dependence). Those with one cognitive style may use the corresponding linguistic prac
tice, which in turn may reinforce the corresponding cognitive style.
To investigate further the connection between language and culturally informed social
perception, we reviewed the literature on bilinguals’ (multilinguals’) social cognitive
processes as a function of the language that they are using. Do bilinguals exhibit a social
cognitive pattern that is congruent with the culture associated with the language they are
using at the time (cultural accommodation)? Although there is a significant amount of evi
dence for this, there are some occasions in which bilinguals, under some circumstances,
show the opposite pattern, namely, exhibiting a tendency to express the style that is con
gruent with the culture and language that they are not using at the time (cultural affirma
tion). This suggests that bilingual individuals are not at the complete mercy of the lan
guage that they use—they are capable of regulating their own thoughts and actions if
they so wish due to the circumstances at hand.
Future Directions
One of the obvious areas of future research is to explore the link between language and
culturally informed social cognition. In particular, do contextualizing and decontextualiz
ing linguistic practices influence field-dependent and -independent cognition, or does so
cial cognition drive linguistic practice? There may very well be bidirectional influences.
On the one hand, there is plenty of evidence to suggest that processing a narrative that
describes group-based (vs. individual) activities written using first-person plural (vs. sin
gular) pronouns can lead readers to act more collectivistically (vs. individualistically) not
only in European-based cultures (e.g., Brewer & Gardner, 1996; Kühnen, Hannover, &
Schubert, 2001), but also in East Asian cultures (e.g., Oyserman et al., 2009; also see Oy
serman & Lee, 2008, for a meta-analytic review). On the other hand, there is evidence
that cultural orientations drive language use. Na and Choi (2009) showed that Korean
participants who have more collectivist inclinations tend to show a greater tendency to
use a first-person plural, rather than a first-person singular pronoun, so that they may
call their brother wuri brother, for instance. Although there is correlational evidence for a
relationship between linguistic practices and cognitive style, this requires much further
investigation.
Another area of future exploration concerns research on bilinguals. Multilingualism is the

rule rather than the exception worldwide. Even in traditionally monolingual cultures (e.g.,
United States, Australia), more individuals are now operating in a multilingual environ
ment or using multiple languages in their everyday life. At work, they may use their na
tional language to communicate with their colleagues; yet, in interacting with clients,
they may use another language, such as English, as a common language. However, when
they go home, they use their ethnic language or dialect, which is distinct from their na
tional tongue. How do people manage this complex sociocultural mosaic? As we conclud
ed in the brief literature review on bilingual cognition and action, although their default
Page 21 of 28

may be cultural accommodation—to go along with the norm of (p. 59) the linguistic com
munity—they flexibly adapt their style under some circumstances. Under what circum
stances can they regulate their thoughts and actions? Such intentional adjustments would
surely require their knowledge about their own tendencies and their ability and motiva
tion to control their thoughts and actions. How are such self-regulations performed?
These represent at least two major areas of research that await further exploration.
Language is one of the most versatile psychological tools that humans have acquired dur
ing the course of their evolution. Its social and cultural implications are enormous; the
nexus of language, culture, and social behavior is an important area of study. Yet, surpris
ingly little attention has been paid to this critical issue in the area of culture and psychol
ogy. Given that there is some fertile ground for future research, we hope to see a further
growth in the area of language and culture in the near future.
References
Alvarado, N., & Jameson, K. A. (2011). Shared knowledge about emotion among Viet
namese and English bilingual and monolingual speakers. Journal of Cross-Cultural Psy
chology, 42, 963–982.
Austin, J. L. (1962). How to do things with words. Cambridge, MA: Harvard University
Press.
Berry, J. W. (1979). A cultural ecology of social behaviour. Advances in Experimental So

cial Psychology, 12, 177–206.
Bochner, S. (1984). Cross-cultural differences in the self-concept: A test of Hofstede’s in

dividualism/collectivism distinction. Journal of Cross-Cultural Psychology, 25, 273–283.
Bond, M. H., & Yang, K. -S. (1982). Ethnic affirmation versus cross-cultural accommoda
tion: The variable impact of questionnaire language on Chinese bilinguals from Hong
Kong. Journal of Cross-Cultural Psychology 13, 169–185.
Brewer, M. B., & Gardner, W. (1996). Who is this “we”? Levels of collective identity and
self-representations. Journal of Personality and Social Psychology, 71, 83–93.
Chan, C. C. Y., Brandone, A. C., & Tardif, T. (2009). Culture, context, or behavioral con
trol?: English- and Mandarin-speaking mothers’ use of nouns and verbs in joint book
reading. Journal of Cross-Cultural Psychology, 40, 584–602.
Chen, S. X., & Bond, M. H. (2007). Explaining language priming effects: Further evidence
for ethnic affirmation among Chinese-English bilinguals. Journal of Language and Social
Chen, S. X., & Bond, M. H. (2010). Two languages, two personalities? Examining lan
guage effects on the expression of personality in a bilingual context. Personality and So
cial Psychology Bulletin, 36, 1514–1528.
Page 22 of 28

Choi, S., & Gopnik, A. (1995). Early acquisition of verbs in Korean: A cross-linguistic
study. Journal of Child Language, 22, 497–529.
Choi, I., & Nisbett, R. E. (1998). Situational salience and cultural differences in the corre
spondence bias and actor-observer bias. Personality and Social Psychology Bulletin, 24,
949–960.
Choi, I., Nisbett, R. E., & Norenzayan, A. (1999). Causal attribution across cultures: Varia
tion and universality. Psychological Bulletin, 125, 47–63.
Chomsky, N. (1981). Lectures on government and binding. Dortrecht, The Netherlands:

Foris.
Cousins, S. D. (1989). Culture and self-perception in Japan and the United States. Journal
of Personality and Social Psychology, 56, 124–131.
Dryer, M. S. (2011). Expression of pronominal subjects. In M. S. Dryer & M. Haspelmath

(Eds.), The world atlas of language structures online. Munich: Max Planck Digital Library,
chapter 101. Accessed August 3, 2012. http://wals.info/chapter/101
Dryer, M. S., & Haspelmath, M. (Eds.). (2011). The world atlas of language structures on
line. Munich: Max Planck Digital Library. Accessed August 3, 2012. http://wals.info/
Dhawan, N., Roseman, I. J., Naidu, R. K., Thapa, K., & Rettek, S. I. (1995). Self-concepts
across two cultures: India and the United States. Journal of Cross-Cultural Psychology,
26, 606–621.
Ervin, S. M. (1961). Semantic shift in bilingualism. American Journal of Psychology, 74,

233–241.
Ervin, S. M. (1964). Language and TAT content in bilinguals. Journal of Abnormal and So
Ervin, S. M., & Osgood, C. E. (1954). Second language learning and bilingualism. Journal
of Abnormal and Social Psychology, 49, 139–146.
Fiedler, K., & Kruger. (this volume). Language and attribution: Implicit causal and dispo
sitional information contained in words. Handbook of language and social psychology.
Fincher, C. L., Thornhill, R., Murray, D. R., & Schaller, M. (2008). Pathogen prevalence
predicts human cross-cultural variability in individualism/collectivism. Proceedings of
Royal Society, B, 275, 1279–1285.
Gelman, S. A., & Coley, J. D. (1990). The importance of knowing a dodo is a bird: Cate
gories and inferences in 2-year-old children. Developmental Psychology, 26, 796–804.
Gelman, S. A., & Heyman, G. D. (1999). Carrot-eaters and creature-believers: The effects
of lexicalization on children’s inferences about social categories. Psychological Science,
10, 489–493.
Page 23 of 28

Gilbert, D. T., & Malone, P. S. (1995). The correspondence bias. Psychological Bulletin,
117, 21–38.
Giles, H., Coupland, J., & Coupland, N. (1991). Contexts of accommodation. Cambridge,
UK: Cambridge University Press.
Hall, D. G., & Moore, C. E. (1997). Red bluebirds and black greenflies: Preschoolers’ un
derstanding of the semantics of adjectives and count nouns. Journal of Experimental Child
Hall, E. T. (1976). Beyond culture. New York: Anchor Press/Doubleday.
Hofstede, G. (1980). Culture’s consequences. Thousand Oaks, CA: Sage.
(p. 60) Hofstede, G. (2001). Culture’s consequences (2nd ed.). Thousand Oaks, CA: Sage.
Holtgraves, T. (2002). Language as social action. Mahwah, NJ: Erlbaum.
Ji, L. -J., Zhang, Z., & Nisbett, R. E. (2004). Is it culture or is it language? Examination of
language effects in cross-cultural research on categorization. Journal of Personality and
Social Psychology, 87, 57–65.
Jones, E. E. & Harris, V. A. (1967). The attribution of attitudes. Journal of Experimental

Kashima, E. S., & Kashima, Y. (1997). Practice of the self in conversations: Pronoun drop,
sentence co-production and contextualization of the self. In K. Leung, U. Kim, S. Yam
aguchi, & Y. Kashima (Eds.), Progress in Asian social psychology (vol. 1, pp. 165–179. Sin
gapore: Wiley.
Kashima, E. S., & Kashima, Y. (2005). Erratum to Kashima and Kashima (1998) and reiter
ation. Journal of Cross-Cultural Psychology, 36, 396–400.
Kashima, Y., & Kashima, E. S. (2003). Individualism, GNP, climate, and pronoun drop: Is
individualism determined by affluence and climate, or does language use play a role?
Journal of Cross-Cultural Psychology, 34, 125–134.
Kashima, Y., Kashima, E. S., Kim, U., & Gelfand, M. (2006). Describing the social world:
Object-centered versus process-centered descriptions. Journal of Experimental Social Psy
chology, 42, 388–396.
Page 24 of 28

Kashima, Y., Siegal, M., Tanaka, K., & Kashima, E. S. (1992). Do people believe behav
iours are consistent with attitudes? Towards a cultural psychology of attribution process
es. British Journal of Social Psychology, 31, 111–124.
Kemmelmeier, M., & Cheng, B. Y.-M. (2004). Language and self-construal priming: A
replication and extension in a Hong Kong sample. Journal of Cross-Cultural Psychology,
35, 705–712.
Kitayama, S., Duffy, S. A., Kawamura, T., & Larsen, J. T. (2003). Perceiving an object and
its context in different cultures: A cultural look at new look. Psychological Science, 14,
201–206.
Klein, O., Ventura, P., Fernandes, T., Garcia Marques, L., Licata, L., & Semin, G. R. (2010).
Effects of schooling and literacy on linguistic abstraction: The role of holistic vs. analytic
processing styles. European Journal of Social Psychology, 40, 1095–1102.
Krull, D. S., Loy, M. H.-M., Lin, J., Wang, C.-F., Chen, S. X., & Zhao, X. (1999). The funda
mental attribution error: Correspondence bias in individualist and collectivist cultures.
Personality and Social Psychology Bulletin, 25, 1208–1219.
Kühnen, U., Hannover, B., & Schubert, B. (2001). The semantic- procedural-interface
model of the self: the role of self-knowledge for context-dependent versus context- inde
pendent modes of thinking. Journal of Personality and Social Psychology, 80, 397–409.
Lakoff, G. (1987). Women, fire, and dangerous things. Chicago, IL: University of Chicago
Press.
Langacker, R. W. (1987). Foundations of cognitive grammar (vol. 1): Theoretical prerequi

sites. Stanford, CA: Stanford University Press.
Langacker, R. W. (1991). Foundations of cognitive grammar (vol. 2): Descriptive applica

tion. Stanford, CA: Stanford University Press.
Lechuga, J., & Wiebe, J. S. (2011). Culture and probability judgment accuracy: The influ
ence of holistic reasoning. Journal of Cross-Cultural Psychology, 42, 1054–1065.
Licht, A. N., Goldschmidt, C., & Schwartz, S. H. (2007). Culture rules: The foundations of
the rule of law and other norms of governance. Journal of Comparative Economics, 35,
659–688.
Maass, A., Karasawa, M., Politi, F., & Suga, S. (2006). Do verbs and adjectives play differ
ent roles in different cultures? A cross-linguistic analysis of person representation. Jour
nal of Personality and Social Psychology, 90, 734–750.
Marin, G., Triandis, H., Betancourt, H., & Kashima, Y. (1983). Ethnic affirmation versus
social desirability: Explaining discrepancies in bilinguals’ responses to a questionnaire.
Page 25 of 28

Masuda, T., & Nisbett, R. E. (2001). Attending holistically versus analytically: Comparing
the context sensitivity of Japanese and Americans. Journal of Personality and Social Psy
chology, 81, 922–934.
Miller, J. G. (1984). Culture and the development of everyday social explanation. Journal
Miller, J. G. (1987). Cultural influences on the development of conceptual differentiation

in person description. British Journal of Developmental Psychology, 5, 309–319.
Miyamoto, Y., & Kitayama, S. (2002). Cultural variation in correspondence bias: The criti
cal role of attitude diagnosticity of socially constrained behaviour. Journal of Personality
and Social Psychology, 83, 1239–1248.
Morris, M. W., & Peng, K. (1994). Culture and cause: American and Chinese attributions
for social and physical events. Journal of Personality and Social Psychology, 67, 949–971.
Na, J., & Choi, I. (2009). Culture and first-person pronouns. Personality and Social Psy
chology Bulletin, 35, 1492–1499.
Nisbett, R. E. (2003). The geography of thought. and why. New York: Free Press.
Nisbett, R. E., Peng, K. P., Choi, I., & Norenzayan, A. (2001). Culture and systems of
thought: Holistic versus analytic cognition. Psychological Review, 108, 291–310.
Oyserman, D., & Lee, S. W. S. (2008). Does culture influence what and how we think? Ef
fects of priming individualism and collectivism. Psychological Bulletin, 134, 311–342.
Oyserman, D., Sorensen, N., Reber, R., & Chen, S. X. (2009). Connecting and separating
mind-sets: Culture as situated cognition. Journal of Personality and Social Psychology, 97,
217–235.
Ralston, D. A., Cunniff, M. K., & Gustafson, D. J. (1995). Cultural accommodation: The ef
fect of language on the responses of bilingual Hong Kong Chinese managers. Journal of
Cross-Cultural Psychology, 26, 714–727.
Ramírez-Esparza, N., Gosling, S. D., Benet-Martínez, V., Potter, J. P., Pennebaker, J. W.

(2006). Do bilinguals have two personalities? A special case of cultural frame switching.
Journal of Research in Personality, 40, 99–120.
Rhee, E., Uleman, J. S., Lee, H. K., & Roman, R. J. (1995). Spontaneous self-descriptions
and ethnic identities in individualistic and collectivistic cultures. Journal of Personality
Ross, L. D. (1977). The intuitive psychologist and his shortcomings: Distortions in

(p. 61)
the attribution process. Advances in Experimental Social Psychology, 10, 173–220.
Ross, M., Xun, W. Q. E., & Wilson, A. (2002). Language and the bicultural self. Personality
and Social Psychology Bulletin, 28, 1040–1050.
Page 26 of 28

Saussure, F. D. (1966). Course in general linguistics. New York: McGraw-Hill.
Schwartz, S. H. (1994). Beyond individualism/collectivism: New cultural dimensions of

values. In U. Kim, H. Triandis, Ç. Kağitçibaşi, S. C. Choi, & G. Yoon (Eds.), Individualism
and collectivism: Theory, method, and applications (pp. 85–122). London: Sage.
Searle, J. R. (1969). Speech acts. Cambridge, UK: Cambridge University Press.
Semin, G. R., & Fiedler, K. (1988). The cognitive functions of linguistic categories in de
scribing persons: Social cognition and language. Journal of Personality and Social Psy
chology, 54, 558–567.
Semin, G. R., & Fiedler, K. (1991). The linguistic category model, its bases, applications
and range. European Review of Social Psychology, 2, 1–30.
Semin, G. R., Gőrts, C. A., Nandram, S., & Semin-Goossens, A. (2002). Cultural perspec
tives on the linguistic representation of emotion and emotion events. Cognition and Emo
tion, 16, 11–28.
Shweder, R. A., & Bourne, E. J. (1984). Does the concept of the person vary cross-cultural
ly? In R. A. Shweder & R. A. LeVine (Eds.), Culture theory (pp. 158–199). Cambridge, UK:
Sussman, N. M., & Rosenfeld, H. M. (1982). Influence of culture, language, and sex on
conversational distance. Journal of Personality and Social Psychology, 42, 66–74.
Tabellini, G. (2008). Institutions and culture: Presidential address. Journal of the Euro
pean Economic Association, 6, 255–294.
Tardif, T. (1996). Nouns are not always learned before verbs: Evidence from Mandarin
speakers’ early vocabularies. Developmental Psychology, 32, 492–504.
Tardif, T., Fletcher, P., Liang, W., Zhang, Z. Kaciroti, N., & Marchman, V. A. (2008). Baby’s
first 10 words. Developmental Psychology, 44, 929–938.
Tardif, T, Gelman, S. A., & Xu, F. (1999). Putting the “noun bias” in context. Child Develop
ment, 70, 620–635.
Trafimow, D., Triandis, H., & Goto, S. G. (1991). Some tests of the distinction between the
private self and the collective self. Journal of Personality and Social Psychology, 60, 649–
655.
Trafimow, D., Silverman, E. S., Fan, R. M. -T., & Fun Law, J. S. (1997). The effects of lan
guage and priming on the relative accessibility of the private self and the collective self.
Uleman, J. S., Winborne, W. C., Winter, L., & Schecter, D. (1986). Personality differences
in spontaneous personality inferences at encoding. Journal of Personality and Social Psy
chology, 51, 396–403.
Page 27 of 28

Watkins, D., & Gerong, A. (1999). Language of response and the spontaneous self-con
cept: A test of the cultural accommodation hypothesis. Journal of Cross-Cultural Psycholo
gy, 30, 115–121.
Whiten, A., & van Schaik, C. P. (2007). The evolution of animal ‘cultures’ and social intelli
gence. Philosophical Transaction of Royal Society B, 362, 603–620. doi:10.1098/rstb.
2006.1998
Winter, L., & Uleman, J. S. (1984). When are social judgments made? Evidence for the
spontaneousness of trait inferences. Journal of Personality and Social Psychology, 47,
237–252.
Witkin, H. A., & Goodenough, D. R. (1977). Field dependence and interpersonal behavior.
Psychological Bulletin, 84, 661–689.
Yang, K. -S., & Bond, M. H. (1980). Ethnic affirmation by Chinese bilinguals. Journal of
Cross-Cultural Psychology, 11, 411–425.
Zwaan, R. A. (2004). The immersed experiencer: Toward an embodied theory of language

comprehension. The Psychology of Learning and Motivation, 44, 35–62.
Zwaan, R. A., & Radvansky, G. A. (1998). Situation models in language comprehension

and memory. Psychological Bulletin, 123, 162–185.
Notes:
(1) . To say that a pronoun is “dropped” implies that the “normal form” is to include a pro
noun, and, for this reason, some have argued that it is an English-centric nomenclature
(e.g., Dryer, 2011). We tend to agree with this judgment; however, we decided to use this
terminology for its ease of exposition.
Yoshihisa Kashima
Yoshihisa Kashima, School of Psychological Sciences, The University of Melbourne
Emiko Kashima
Emiko Kashima, School of Psychological Science, La Trobe University
Evan Kidd
Evan Kidd, Research School of Psychology, Australian National University
Page 28 of 28

Gender Similarities and Differences in Language

Campbell Leaper

In this review, gender-related variations in language are examined. It is argued that lan
guage both reflects and maintains gender divisions in society. English and other lan
guages have inherent gender biases (e.g., grammatical gender, generic use of masculine
forms, gendered titles, family surnames). In addition, average gender differences are indi
cated in children’s and adults’ uses of language. These occur in talkativeness, conversa
tion topic preferences, turn-taking and interruption, and the coordination of affiliative
and assertive functions in speech acts. However, when statistically significant average
gender differences are indicated, the magnitude of the difference is usually negligible or
small on most measures. Furthermore, the magnitude and direction of gender differences
in the use of language are moderated by contextual variables (e.g., partner familiarity,
gender composition, activity).
Keywords: human sex differences, sex roles, sexism, language, speech, conversation, communication, social inter
action, aggression, intimacy
Research on gender-related variations in language has been conducted in psychology

(e.g., Aries, 1996), linguistics (e.g., Tannen, 1994), communications (e.g., Dindia & Ca
nary, 2006), sociology (e.g., West & Zimmerman, 1991), anthropology (e.g., Maltz & Bork
er, 1982), and feminist studies (e.g., Crawford, 1995), among other fields. Several edited
volumes (e.g., Dindia & Canary, 2006; Holmes & Meyerhoff, 2003) and textbooks (e.g.,
Coates, 2004; Eckert & McConnell-Ginet, 2003; Wood, 2013) on the topic have been pub
lished. In addition, best-selling books, magazine articles, and television programs have
addressed the general public’s curiosity about presumed gender differences in language
and communication (e.g., Gray, 1992; Tannen, 1990). As reviewed here, empirical re
search calls into question some of the common assumptions that many laypersons may
have about the incidence and magnitude of gender differences in language.
Page 1 of 37

Narrowing the Scope of the Present Review

Research on language and gender has accelerated in the past three decades. In 1975,
Thorne and Henley were able to compile an annotated bibliography of nearly all known
studies on the topic. This would be a nearly impossible task today. A PsycInfo search (con
ducted September 15, 2013) using “subject = (gender OR human sex differences) AND
subject = (language OR conversation OR speech)” returned 3,414 results. Given the limit
ed length of the present review of gender similarities and differences in language, it will
necessarily be selective and incomplete.
To narrow the scope of the material reviewed, a few decisions were made. First, when
possible, evidence based on observations of people’s conversation was favored over those
based on self-report measures. Self-reports of language style and other social behaviors
can be suspect due either to inaccurate self-views or to concerns with self-presentation
(e.g., Rock, 1981).
Second, the review is based on studies employing quantitative methods. Ethno

(p. 63)
graphic and other qualitative approaches are helpful during the discovery phase of re
search. These methods can highlight some of the ways that gender-related variations in
language may depend on particular contexts or vary across individuals’ sociocultural
backgrounds (e.g., Goodwin, 2002; Kyratzis & Guo, 2001; Thorne, 1993; also see Holmes
& Meyerhoff [2003] for reviews). Although qualitative methods can provide a rich picture
of interactive context, they are correspondingly limited because they do not indicate the
reliability (statistical significance) of the observed patterns and the degree to which dif
ferent factors might be related.
Finally, when relevant meta-analyses were available, they were used to summarize aver
age gender similarities and differences for particular aspects of language. Meta-analysis
has the advantage of providing an aggregate summary of statistical significance and aver
age effect size for a gender difference across studies. This method also has the benefit of
testing various moderators that may predict if and when particular differences are most
likely. At the same time, meta-analysis is limited to detecting patterns only among those
studies that have been conducted on a topic.
When evaluating findings from studies comparing females and males, both the effect size
and the statistical significance should be considered. Even when a statistically significant
gender difference in behavior occurs, there is often much overlap between females and
males (Barnett & Rivers, 2004; Hyde, 2005). The magnitude of difference between two
groups is known as the effect size. One metric is Cohen’s d, which reflects the difference
between group means in standard deviation (SD) units. It also reflects the amount of
overlap in the distributions of the two groups’ scores. According to Cohen’s (1988)
guidelines, differences are considered large when the effect size d = .8 or greater (less
than 53 percent overlap), medium when d = .5 (66 percent overlap), or small when d = .2
(85 percent overlap). An effect size below. 2 is therefore seen as negligible (i.e., trivial).
The effect sizes associated with various average gender differences will be regularly indi
Page 2 of 37

cated throughout this review. The reader is advised to keep in mind these guidelines for
interpreting their magnitude.
Organization of the Review

The rest of the review is divided into the following sections. First, gender biases in Eng
lish and other languages are reviewed. Second, the evidence for gender similarities and
differences in people’s use of language in conversation is examined. Finally, some conclu
sions and ideas for future directions are offered.
Gender Biases in Languages

Feminist scholars, sociolinguists, and psychologists have highlighted ways that English
and other languages can be construed as sexist or gender-biased (see Graddol & Swann,
1989; Henley, 1989; Leaper & Bigler, 2004; Stahlberg, Braun, Irmen, & Sczensy, 2007, for
reviews). Forms of gender bias in language reviewed below include the regular marking
of gender in words, the generic use of masculine nouns and pronouns, gendered titles,
family surnames, and essentialist language.
Marking Gender in Languages
There are three variations in the ways that gender is marked in languages across the
world (Stahlberg et al., 2007). First, in gendered languages, nouns and pronouns are gen
dered (e.g., Spanish, Russian, Hebrew, Hindi). Because each noun is gendered, there are
no gender-neutral terms that can be used to refer to persons. When referring to both fe
males and males, the masculine form is commonly used. For example, to refer generally
to “adults” in Spanish, the likely word would be “adultos” (masculine form) rather than
“adultas” (feminine form). Second, in natural gender languages, only pronouns are gen
dered (e.g., English, Swedish). For example, in English, singular third-person pronouns
(he/his and she/her) are gendered but not the nouns. Finally, in genderless languages,
gender is not grammatically marked in either nouns or pronouns (e.g., Finnish, Hungari
an, Turkish, Mandarin); however, genderless languages do have words that can be used
to distinguish gender (e.g., words for female and male). For example, in Turkish, the
third-person singular pronoun “o” is used to refer to either a female or a male. But the
words for “woman” and “man” in Turkish are “kadin” and “adam,” respectively.
Overcoming gender bias in language may be more challenging in languages with gram
matical gender, such as Spanish. Each noun is marked as feminine or masculine. Accord
ingly, virtually every sentence reifies the importance of gender distinctions in people’s
minds. Increasing the salience of gender, in turn, may increase the likelihood that people
use gender to categorize and stereotype others (Bigler & Liben, 2007). By extension, one
might expect that languages with a grammatical gender (p. 64) may be more likely than
English (and other natural gender languages) to elicit gender-biased views (see Prewitt-
Freilino, Caswell, & Laakso, 2012). To test this hypothesis, Wasserman and Weseley
Page 3 of 37

(2009) randomly assigned high school students to complete a measure of gender attitudes
written in either English or a language with a grammatical gender (Spanish or French).
Students were more likely to endorse gender-stereotypical attitudes after completing the
survey in Spanish or French than in English. Thus, there is tentative evidence that the
manner gender is marked in a language may have some influence on individuals’ gender-
related thinking.
Optional Marking of People’s Gender
As previously explained, in natural gender languages such as English, it is possible to use

gender-neutral nouns to index people (e.g., “student,” “parent,” “person”). This means
that, to some extent, a choice is made whether to linguistically mark other people’s gen
der in a statement. There are some contexts when it is clearly pertinent to do so in the
conversation. For example, a speaker would likely distinguish between females and males
when discussing certain physical or medical topics that affect the sexes differently. How
ever, there are many times when speakers unnecessarily mark people’s gender in conver
sation. For example, many teachers commonly welcome their students with expressions
such as “Good morning, girls and boys” (as opposed to “Good morning, students”). In this
situation, the gender of the persons is not pertinent. (Why is it necessary for a teacher to
separately greet the girls and the boys?) By way of contrast, it would seem strange if the
teacher greeted children by their race, eye color, or handedness (Leaper & Bigler, 2004).
The linguistic practice of unnecessarily addressing people by their gender both reflects
and reinforces the organization of social life along gender lines. That is, the repeated use
of gender labels in discourse increases the salience of gender in children’s (and adults’)
daily lives—and thereby increases the likelihood of gender stereotyping (Bigler & Liben,
2007).
Generic Use of Masculine Nouns and Pronouns
As noted above, one form of linguistic gender bias in English and many other languages is
the generic use of masculine nouns and pronouns. Examples in English include terms
such as “mankind” or “chairman,” as well as the use of “he” to refer generally to all per
sons (e.g., “When a student studies for an exam, he is likely to do well”). In many lan
guages with grammatical gender, the masculine form is used to refer to both genders. For
example, in Spanish, reference to “the children” would commonly be made using “los
niños” (the masculine form) rather than “las niñas” (the feminine form).
The use of masculine linguistic forms to refer to general cases is considered problematic
for three interrelated reasons. First, studies with children (e.g., Hyde, 1984; Liben, Bigler,
& Krogh, 2002) and adults (e.g., Hamilton, 1988; Martyna, 1980) indicate that people
tend to imagine male characters when reading material written using the masculine
generic. Thus, girls and women who read material using the masculine generic may be
less likely to find it self-relevant and less likely to identify with the characters (Conkright,
Flannagan, & Dykes, 2000; Stout & Dasgupta, 2011). Second, when the masculine lin
guistic form is the default, it may reinforce the tendency to view gender-neutral cases as
Page 4 of 37

male. For example, researchers have observed that parents and teachers were over
whelmingly more likely to label gender-neutral animal characters in stories as male than
female (DeLoache, Cassidy, & Carptenter, 1987; Gelb, 1989; Lambdin, Greer, Jibotian,
Wood, & Hamilton, 2003). Finally, the pervasive use of the masculine generic systemati
cally reinforces the higher status of males relative to females. In this regard, MacKay
(1980) noted that the masculine generic meets the criteria for effective propaganda: fre
quent occurrence, early age of exposure, and association with high-prestige sources (e.g.,
teachers, books).
In the United States and many other English-speaking countries, the more inclusive use
of “he or she” rather than “he” to refer to generic cases has increased in prevalence over
the years. Nonetheless, some people resist making these changes. Perhaps it is not sur
prising that researchers found gender-inclusive language was less likely to be favored by
individuals with sexist than nonsexist attitudes (Parks & Roberton, 2004; Swim, Mallett,
& Stangor, 2004).
Gendered Titles
Henley (1989) noted that linguistic practices in English and other languages often define
women as subordinate to men. One such practice is the use of titles preceding surnames.
In English and other languages, there is a single salutation for men, which is Mr. Howev
er, until recent decades, (p. 65) a woman was addressed variously as Miss if she was un
married and Mrs. if she was married. There is obvious inequity inherent in these naming
practices. A woman’s marital status is immediately signaled by her title, whereas a man’s
status is not. Thus, through usage of these titles, a woman is defined by her relationship
to a man—but not the reverse.
Coinciding with the feminist movement in the 1970s, the term Ms. was introduced into
the English language. Ms. and Mr. are parallel titles whereby neither signals the
individual’s marital status. Whereas Ms. was often viewed as radical when first intro
duced, it has gained wide acceptance over the years (Crawford, Stark, & Renner, 1998).
Today, Ms. is the standard used in business and journalism. Furthermore, research gener
ally suggests that the Ms. title has positive connotations (Crawford et al., 1998) and may
possibly reduce gender-biased perceptions of women’s agency and competence in profes
sional settings (Dion & Schuller, 1990; Malcomson & Sinclair, 2007).
Despite the mainstream acceptance of Ms. as a title, many speakers and writers continue
to prefer Miss and Mrs. For example, in an online survey, Lawton, Blakemore, and Vartan
ian (2003) found that married women were much more likely to favor Mrs. over Ms. The
same researchers also found that many young women (below age 20 years) misconstrued
the meaning of Ms. to signify an unmarried woman. The researchers inferred that some
younger women might consider Ms. as a title for single women who are too old for Miss.
Paradoxically, if the latter view became a common linguistic practice, it would define
women even more than the traditional Miss/Mrs. distinction has done. That is, not only
Page 5 of 37

would women be defined by marital status (Miss vs. Mrs.), they would additionally be de
fined by age (Miss for single/younger and Ms. for single/older).
Family Surnames
Another common gender inequity in naming occurs in the selection of a surname when a
woman and a man marry. The traditional practice in many cultures has been for the
woman to take the man’s family name. Although most women take their husband’s last
name upon marriage, it is now common in many Western societies for women not to
change their name. Also, for some couples, the solution is to hyphenate both last names
(e.g., Brown-Davis). Still another option is for the couple to create a novel surname that
each partner adopts (e.g., see Rapacon, 2011).
Surveys indicate that women and men commonly cite tradition as a justification for why
the wife should take the husband’s family surname (Robnett & Leaper, 2013; Twenge,
1997). When women cite tradition, however, they may not realize its historical basis. The
practice of a woman taking her husband’s family name goes back to patriarchal traditions
whereby the woman was considered property of the man (Boxer & Gritsenko, 2005). A re
lated tradition is paternalism, which refers to the notion that the man is responsible for
protecting the woman in exchange for her childbearing and housekeeping. Attitudes in
support of paternalistic practices are known as benevolent sexism (Glick & Fiske, 1996).
Robnett and Leaper (2013) observed in a sample of heterosexual undergraduate women
and men that benevolent sexism was positively related to favoring the adoption of the
man’s family name in marriage.
Hoffnung (2006) examined correlates of married women’s name choices in the United
States. Factors related to nontraditional name choices (e.g., not changing name, hyphen
ation) included higher educational level, older age at marriage, higher career commit
ment, and stronger feminist attitudes. Also, women of color were more likely than white
women to favor nontraditional name choices.
The increasing acceptance of hyphenated names was implicated in a study of American

college students (Forbes, Adams-Curtis, White, & Hamm, 2002). Participants were asked
to rate hypothetical married women and men with either hyphenated or nonhyphenated
married names. The students were more likely to rate married men and women higher in
both instrumental/assertive and expressive/affiliative qualities if the targets had hyphen
ated names.
Even when a woman retains her own family name after marriage to a man, another dilem
ma that typically remains is what surname to give any offspring. The traditional practice
in many cultures, of course, has been for offspring to take the father’s last name. Chil
dren are rarely given the mother’s last name. (In many Spanish-speaking countries, how
ever, children are commonly given both parents’ last names.) One contemporary alterna
tive for many Western couples is to hyphenate both (p. 66) parents’ family names. As a
way to retain family names, however, this strategy will prove challenging in succeeding
generations when two individuals with hyphenated names get married (e.g., Jane Garcia-
Page 6 of 37

Harris and Hugo Cheng-Stein become married). Still another strategy mentioned earlier
is to create a novel family name for the child, as well as for both parents.
Essentialist Language
Essentialism is another means by which gender bias may occur in language (Gelman, Tay
lor, & Nguyen, 2004). Essentialist language occurs when males and females are each de
scribed as having inherent qualities. Essentialist statements can be either prescriptive or
descriptive. Examples of prescriptive essentialist statements include “Only girls can play
with dolls” or “Only boys can play with footballs.” These sentences explicitly state that
certain qualities are supposed to be associated with one gender rather than the other. Al
ternatively, examples of descriptive essentialist statements include “Girls play with dolls”
or “Boys like football.” Unlike the prescriptive statements, descriptive statements do not
assert that only girls can play with dolls or that only boys can play football. However, de
scriptive essentialist statements offer broad generalizations about each gender. They do
not acknowledge the possibilities of within-gender variability (e.g., “Some—but not all—
girls play with dolls”) or overlap between the genders (e.g., “There are more boys than
girls who like football”).
Researchers have observed mothers commonly using descriptive essentialist language—

but not prescriptive essentialist language—in their speech to children (Friedman, Leaper,
& Bigler, 2007; Gelman et al., 2004). Furthermore, in these studies, descriptive essential
ist statements were commonly expressed even among mothers who endorsed gender-
egalitarian attitudes. Whether they intend to do so or not, adults who use essentialist
statements may perpetuate gender stereotypes in children (Bigler & Liben, 2007; Leaper
& Bigler, 2004).
Another way that language can essentialize gender is through using the terms “feminine”
and “masculine” to describe social behaviors (Leaper, 1995; Lott, 1981). This usage be
came popular with the advent of the Bem Sex-Role Inventory (Bem, 1974) and the Person
al Attributes Questionnaire (Spence & Helmreich, 1978) to measure self-perceived attrib
utes. Both measures include separate Femininity and Masculinity scales comprised of
items assessing self-perceived affiliative attributes (e.g., understanding, emotional) and
assertive attributes (e.g., self-confident, independent), respectively. However, a problem
with labeling behaviors as “feminine” and “masculine” is that it perpetuates the notion
that certain behaviors are inherently either female or male (Lott, 1981; also see Leaper,
1995). Thus, for a man to show affection implies that he is somehow not masculine; con
versely, for a woman to act confident implies that she is somehow not feminine.
Alternatives to describing behavior clusters as “feminine” and “masculine” include the

terms affiliative and assertive (e.g., Leaper & Ayres, 2007; Leaper & Smith, 2004), com
munal and agentic (Bakan, 1966), or socioemotional and instrumental (Bales, 1950,
1970), respectively. These options allow us to consider these behaviors as human
attributes that occur in both women and men (Lott, 1981). If one wishes to specifically
discuss how these behaviors are differentially socialized in males and females, the com
Page 7 of 37

pound adjectives “masculine-stereotyped” and “feminine-stereotyped” can be used to

characterize them. The latter usage of the term “stereotyped” underscores the point that
these behaviors are socially prescribed for (rather than inherent to) persons based on
their gender.
Gender-Related Variations in Language Use

Having previously reviewed some of the ways that English and other languages can re
flect gender bias, the next section addresses ways that males and females sometimes dif
fer in their uses of language in social interaction. That is, gender differences and similari
ties in the areas of talkativeness, turn-taking and interruption, conversation topic prefer
ences, and speech acts are explored.
The average effect size associated with a particular gender difference will be reported
when a relevant meta-analysis has been conducted. The reader is advised to review the
conventions for interpreting effect sizes described in the section “Narrowing the Scope of
the Present Review.” As noted later, most significant average gender differences in lan
guage are typically in the negligible or small range. However, meta-analyses often point
to various moderators affecting the likelihood and the magnitude of gender differences.
These include individual characteristics (e.g., age), activity settings (e.g., conversation
topic), and the relationship among interaction partners (e.g., familiarity, gender composi
tion). Cultural and social-structural factors may also moderate (p. 67) some gender differ
ences; unfortunately, however, the numbers of studies comparing individuals from diverse
backgrounds have been insufficient to consider these moderators in meta-analyses.
Talkativeness
Many people stereotype women as talkative relative to men (see James & Drakich, 1993).
In support of this view, some studies found that women were more likely than men to cite
talking as a favorite activity in their same-gender friendships (see Aries, 1996). However,
stereotypes are often inaccurate. Also, stating a preference for a particular activity such
as talking does not necessarily mean that one performs that behavior more often than
others. To help shed some light on the extent that gender differences in talkativeness
might occur, two separate meta-analyses were conducted of studies of children (Leaper &
Smith, 2004) and adults (Leaper & Ayres, 2007). (Positive effect sizes reflect higher aver
age talkativeness for women than men.) The meta-analysis of child studies indicated that
girls were significantly more talkative than were boys, although the average effect size
was negligible in magnitude (d = .11). In the meta-analysis of adult studies, there was al
so a significant average gender difference in talkativeness with a negligible effect size;
and it was men rather than women who tended to score higher in average talkativeness
(d = –.14).
A study published in the same year as Leaper and Ayres’ (2007) meta-analysis is also re
vealing. Mehl, Vazire, Ramirez-Esparza, Slatcher, and Pennebaker (2007) recorded the
speech of 396 adults aged 17–29 years in the United States and Mexico over 4–10 days.
Page 8 of 37

They found no significant gender difference in talkativeness (d = .07). This study is no
table, given the researchers were sampling natural conversations over the course of sev
eral days. Thus, the overall trend—from Mehl and colleagues’ (2007) recent study and the
two previously cited meta-analyses—is the absence of a meaningful average gender dif
ference in talkativeness.
There may be certain conditions when more substantive gender differences in talkative
ness are more likely to occur. The aforementioned meta-analyses indicated significant
moderator effects on average gender differences in talkativeness. Among studies of chil
dren (Leaper & Smith, 2004), one significant moderator was the children’s age. A signifi
cantly larger effect size was seen in studies of children below 3 years (d = .32) than stud
ies of children 3–16 years of age (d = .08). The greater talkativeness among girls than
boys during the toddler years may possibly reflect girls’ faster average rate of language
development—with boys eventually catching up (Gleason & Ely, 2002).
Another moderator of gender differences in children’s talkativeness was the conversation

partner (Leaper & Smith, 2004). There was not a significant average gender difference
across studies of children’s interactions with peers (d = .03). However, a significant dif
ference in talkativeness—with girls higher than boys—was seen in children’s interactions
with adults (d = .19). Most of the child–adult studies were with mothers. This is pertinent
because in a different meta-analysis Leaper, Anderson, and Sanders (1998) found that, on
average, mothers were more talkative with daughters than with sons (d = .29). (There
were not enough studies to consider child gender effects on fathers’ talkativeness.) The
direction of possible influence is not clear, however. The child average gender difference
in talkativeness may result because some mothers talk more to daughters than sons. A
second possibility is that girls are more likely than boys to initiate talk with mothers, and
this leads to greater mother talkativeness with daughters than sons. A third alternative is
that there is a reciprocal influence whereby, on average, mothers and daughters each ini
tiate talk more to one another than do mothers and sons.
Significant moderators of gender differences in talkativeness were also indicated for

adults (Leaper & Ayres, 2007). One was the gender composition of the interaction part
ners. Men’s greater talkativeness was more likely during mixed-gender interactions (d =
–.28) than same-gender interactions (d = –.08). This pattern suggests that greater male
talkativeness may reflect some men’s attempts to establish dominance in mixed-gender
interactions (Carli, 1990; Ridgeway & Smith-Lovin, 1999). During mixed-gender interac
tions, some men may wish to establish their authority through talking; at the same time,
some women may wish to affirm their support through deference.
Another contextual moderator of gender differences in adult talkativeness was the activi
ty setting. This factor influenced the magnitude as well as the direction of average gen
der differences. Men were significantly more talkative during discussions of impersonal
topics (d = –.79) or assigned decision-making tasks (d = –.13). In contrast, women were
significantly more talkative during socioemotional contexts that involved either interac
tions with children or self-disclosure with another adult (d = .39). The conversational top
Page 9 of 37

ics in which gender differences occurred are relatively (p. 68) gender-typed: Self-disclo
sure or discussions with children reflect the emphases on nurturance and expressiveness
associated with women’s traditional gender role; in contrast, decision-making tasks and
discussing current events are compatible with the emphases on instrumentality and self-
assertion associated with men’s traditional gender role. Thus, gender-typed settings may
increase the salience of gender-typed norms for some women and men and lead to corre
spondingly different levels of conversational engagement.
Turn-Taking and Interruption
Conversations involve different speakers taking turns to talk. Conversation analysts have
been interested in how these turns are negotiated. One approach has been to consider
asymmetries in the length of speakers’ conversational turns. In Leaper and Ayres’ (2007)
meta-analysis, different operational definitions of talkativeness were considered. One was
the duration of talking. For this particular measure, significantly higher average talking
time was indicated for men than women (d = –.24). This suggests that men were more
likely to control the conversational floor for longer periods of time. In support of this
dominance interpretation, Mast’s (2002) meta-analysis indicated a significant average as
sociation (r = .51) between speaking time and measures of dominance. Moreover, the as
sociation was significantly stronger for men (r = .56) than women (r = .44).
In addition to talk time, another way that some speakers seek to control a conversation is
through interruption. Anderson and Leaper (1998) conducted a meta-analysis to test for
average gender differences in adults’ use of interruption. On average, men were signifi
cantly more likely than women to interrupt during conversation—but the effect size was
negligible (d = .12). The researchers further differentiated between intrusive interrup
tions (i.e., breaking into the other speaker’s utterance) and other types of interruption
(e.g., short acknowledgements). A more meaningful effect size (d = .23) was indicated
among studies specifically looking at intrusive interruptions. Gender composition (same-
vs. mixed-gender interactions) was not a significant moderator, which may contradict
some earlier proposals that some men use interruption to dominate women in conversa
tion (see Aries, 1996, for a review). Instead, it appears that intrusive interruptions may be
more likely for men than women—regardless of the partner’s gender.
Conversation Topics
People discuss a variety of topics in their conversations. Some examples include relation
ships, personal problems, work, sports, entertainment, fashion, health, finances, philo
sophical ideas, and current events. There have been no meta-analyses testing for average
gender differences in conversation topic preferences; however, there are two observation
al studies of conversation topics among university students. First, Bischoping (1993)
found women were more likely than men to discuss people and relationships (32 vs. 15
percent, respectively). In addition, Newman, Groom, Handelman, and Pennebaker (2008)
similarly found that women were significantly more likely than men to reference social
processes (d = .14). They also noted higher average incidences among men than women
Page 10 of 37

in discussions of money (d = –.39) and sports (d = –.39). The magnitudes of these differ
ences are in the negligible to small range. Also, the samples are limited to college stu
dents. Therefore, the generalizability of these findings needs to be tested in future re
search.
As reviewed next, there are other studies that have considered possible gender-related
variations in talk about particular kinds of topics. These include discussions about person
al disclosures and gossip about other people.
Self-Disclosure in Conversation
Self-disclosure refers to discussions of personal thoughts or feelings. The research litera

ture points to greater average disclosure among females than males. First, this was indi
cated in a narrative review of studies examining self-disclosure during childhood and ado
lescence (Buhrmester & Prager, 1995). Half (49 percent) of studies indicated self-disclo
sure was higher among girls than boys, but 13 percent of studies found boys were higher
and 38 percent found no significant difference. When age was taken into account, greater
female disclosure was reported in 59 percent of studies of children aged 9–11 years and
in 89 percent of studies of youth aged 12–20 years. Thus, existing reports suggest that av
erage gender differences in self-disclosure may increase from childhood into adolescence.
A meta-analysis of studies on self-disclosure in child and adolescent samples would help
to clarify the magnitude of the average gender difference as well as the possible impact
of moderator variables.
A meta-analysis has been conducted of studies testing for gender differences in adults’
self-disclosure (Dindia & Allen, 1992). Across studies, there was a significant difference
with a negligible to small effect (p. 69) size, whereby greater self-disclosure was slightly
more likely for women than men (d = .18). The gender composition of the interaction
partners was a moderator. Although significant average gender differences occurred in
both contexts, the effect size was more meaningful during same-gender interactions (d
= .37) than mixed-gender interactions (d = .13). The partners’ relationship and familiarity
also moderated the effects. Average differences were larger during interactions with fa
miliar partners (friends: d = .28; spouses/partners: d = .22) compared to interactions with
strangers (d = .07).
Average gender differences in self-disclosure behavior reflect the gender-typed norms

that often emerge in same-gender friendships. A meta-analysis of friendship expectations
in children and adults indicated that females were more likely than males (d = .39) to ex
pect self-disclosure, understanding, and emotional support in their same-gender friend
ships (Hall, 2011). Many males may actively avoid expressions of intimacy in their same-
gender friendships. Traditional masculinity norms dictate that vulnerable emotions
should be controlled and not shared (Levant & Richmond, 2007; Oransky & Fisher, 2009).
For many traditional boys and men, self-disclosure is viewed as a sign of weakness. This
socialization may begin in childhood. Some studies have found that parents were more
likely to discuss emotions with daughters than sons (e.g., Aldrich & Tenenbaum, 2006;
Page 11 of 37

Fivush, Brotman, Buckner, & Goodman, 2000). Later in childhood, male peers may tease
boys who discuss their feelings as being weak, feminine, or gay.
Over time, the demands of conforming to traditional masculinity may lead to skill deficits.
Higher rates of alexithymia (the inability to put emotions into words) are positively relat
ed to hypermasculinity among boys and men (Levant & Richmond, 2007). Also, peer pres
sures toward emotional restrictiveness in adolescents’ same-gender friendships may con
tribute to difficulties in later heterosexual relationships for men and women (see Leaper
& Anderson, 1997, for a review). In Dindia and Allen’s (1992) meta-analysis, there was a
small but meaningful average gender difference in self-disclosure between heterosexual
spouses (d = .22), with women higher than men in disclosures. If men are reluctant to
share their feelings to their partners, it can undermine the quality of a relationship. Stud
ies indicate that self-disclosure is related to relationship satisfaction in heterosexual cou
ples (e.g., Finkenauer & Hazam, 2000; Sprecher & Hendrick, 2004) and in lesbian and
gay couples (Kurdek, 2005).
Online Self-Disclosure
The advent of electronic social media (e.g., e-mail, text messaging, Facebook, etc.) has
broadened the means by which individuals can communicate with one another. As with
other forms of communication, electronic media can be used to pursue intimacy or ag
gressive goals (see the later section, “Aggressive Controlling Speech Acts”). With regards
to potentially positive outcomes, online communication can stimulate self-disclosure and
enhance relationship quality (see Valkenburg & Peter, 2009). The benefits of online com
munication with friends may be especially beneficial for adolescent boys, given the reluc
tance of many boys to disclose with one another in face-to-face interactions (Schouten,
Valkenburg, & Peter, 2007). A question to test in future studies is whether online disclo
sures spill over to face-to-face interactions.
Gossip
Gossip refers to conversations about persons who are not present. Some gossip can be
neutral, inasmuch that the speaker only describes something about another person in a
nonevaluative manner (e.g., “I heard Jamal got a new job”). Much of the time, however,
gossip involves offering some kind of positive or negative evaluation about a third party
who is not present.
Several affiliative functions are associated with gossip. They include entertainment, es
tablishing shared norms regarding desirable (and undesirable) behavior, and creating a
sense of solidarity (Foster, 2004; Eckert, 1990; Leaper & Holliday, 1995). Negative gossip
about a third party may also be a means to assert power in relationships by negotiating
norms for acceptable behavior with friends. That is, when two speakers evaluate another
person’s behavior, they define what are considered appropriate social norms.
Page 12 of 37

Despite a common conception of women as gossips (see Levin & Arluke, 1985), there has
been surprisingly little research specifically testing for gender differences in gossip dur
ing childhood or adulthood. (Some researchers studying indirect aggression have includ
ed negative gossip in composite measures, but they usually do not separately report find
ings for gossip.) Among the few studies testing for gender differences specifically in gos
sip, the observed pattern has been of higher rates among females than males (Coyne,
Archer, & Eslea, 2006; Leaper & Holliday, 1995; Levin & Arluke, 1985).
Traditional gender differences in power and status may be related to average gender dif
ferences in (p. 70) gossip (Eckert, 1990; Leaper & Holliday, 1995). Men’s status and self-
worth have traditionally been defined through instrumental pursuits such as work and
sports, whereas women’s status and self-worth have been derived through their close re
lationships. Also, aggression and overt competition have traditionally been encouraged
more in males than females (see later section “Aggressive Controlling Speech Acts”).
Thus, girls and women may be more likely to negotiate their status within their friendship
networks in indirect ways through gossip, whereas boys and men may be more apt to
compete directly for dominance with their peers. Consistent with this interpretation, one
recent study indicated that adults tend to perceive individuals who often gossip as less
powerful than those who rarely gossip (Farley, 2011).
Speech Acts
One common approach in the analysis of language style has been to examine the social
functions of utterances. These social functions are known as speech acts (e.g., Searle,
1969). Many researchers distinguish between self-assertive and affiliative functions. As
sertion refers to acts that exert the individual’s agency and influence, whereas affiliation
refers to acts that affirm and join with the other person. Other constructs that are similar
to assertion and affiliation include instrumentality and expressiveness (Bales, 1970), task
and socioemotional orientations (Bales, 1950, 1970), and agency and communion (Bakan,
1966).
Whereas some researchers have framed assertive and affiliative functions as mutually ex
clusive (dichotomous approach), others have conceptualized them as orthogonal dimen
sions (multidimensional approach). The latter approach allows for the possibilities of
speech acts that can range from low to high in affiliation and simultaneously range from
low to high in assertion. Earlier studies of language and communication were based on
the dichotomous approach (e.g., Anderson & Blanchard, 1982). But an increasing number
of studies are applying a multidimensional approach, whereby high assertion and high af
filiation are not viewed as mutually exclusive (e.g., Leaper, 1991; Leaper, Tenenbaum, &
Shaffer, 1999; Leman, Ahmed, & Ozarow, 2005; Penman, 1980; Strough & Berg, 2000;
Tenenbaum, Ford, & Alkhedairy, 2011). Collaborative speech acts refer to statements that
are high in both affiliation and assertion. Examples include making initiatives for joint ac
tion, elaborating on the other person’s comment in a relevant manner, and expressing
support. Obliging speech acts reflect high affiliation and low assertion. Examples include
simple expressions of agreement, conceding to the other person’s view, and requests for
Page 13 of 37

help. Conversely, controlling speech reflect high assertion and low affiliation. Examples
include hostile comments, directive statements, or expressions of disagreement.
The subsequent sections address findings regarding gender-related variations in collabo

rative speech, obliging speech, and controlling speech. When reporting average effect
sizes from meta-analyses, the direction will be positive if females scored higher than
males in collaborative or obliging (i.e., high affiliation) speech acts. In contrast, the direc
tion will be positive if males scored higher than females in controlling speech (i.e., high
assertion and low affiliation). After reviewing the evidence for average gender differences
in speech acts, some of the factors that moderate some of these differences will be con
sidered.
Overall Assertive and Affiliative Language

Meta-analyses point to average gender differences in the relative emphases on affiliation
and assertion in the speech acts of children (Leaper & Smith, 2004) and adults (Leaper &
Ayres, 2007). Findings from these meta-analyses are summarized in Table 4.1. On aver
age, studies indicate that females were somewhat more likely than males to use speech
acts high in affiliation (collaborative or obliging speech acts combined). However, the
magnitudes of the average effects were small with children (d = .26) and negligible with
adults (d = .12). In the same set of meta-analyses, males were slightly more likely than fe
males to use speech acts low in affiliation and high in assertion (controlling speech acts).
Although the average effect sizes were significant, the magnitudes were negligible in
studies of children (d = .11), as well as of adults (d = .09). More meaningful effect sizes
were often indicated when particular speech acts were examined (reviewed next).
Page 14 of 37

Table 4.1 Effect sizes and statistical significance for tests of average gender differences in speech in children and
adults
Language k Combined N d 95% Confi Direction

variable dence interval
(CI)
Overall Affiliative Speech
Children 46 2,694 0.26* 0.16/0.28 F>M
Adults 54 2,781 0.12* 0.06/0.18 F>M
Elaboration/Active Understanding
Children 24 1,251 0.41* 0.13/0.57 F>M
Adults 3 152 0.41* 0.10/0.72 F>M
Support/
Praise
Children 7 811 0.13 –0.01/0.26 F=M
Adults 10 554 0.16* 0.01/0.32 F>M
Agreements
Page 15 of 37

Children 9 492 0.05 –0.11/0.22 F=M
Adults 13 818 0.08 –0.03/0.18 F=M
Simple Acknowledgments
Children 8 517 0.09 –0.08/0.26 F=M
Adults 26 1,577 –0.01 –0.09/0.08 F=M
Overall Assertive Speech
Children 75 3,495 0.11* 0.05/0.17 M>F
Adults 50 2,541 0.09* 0.02/0.15 M>F
Suggestions
Children 13 1,092 0.13* 0.02/0.24 M>F
Adults 13 934 0.27* 0.16/0.38 M>F
General Task Comments
Adults 9 386 0.38* 0.20/0.56 M>F
Directives
Page 16 of 37

Children 38 1,806 0.23* 0.15/0.32 M>F
Adults 11 467 –0.06 –0.22/.09 M=F
Disagree
ments
Children 11 477 0.10 –0.07/0.27 M=F
Adults 10 499 –0.06 –0.22/0.09 M=F
Criticism
Children 16 918 –0.01 –0.12/0.10 M=F
Adults 12 672 –0.13* –0.25/–0.02 F>M
The above findings are from meta-analyses of gender differences in language among children (Leaper & Smith,
2004) and adults (Leaper & Ayres, 2007). “Overall affiliative language” refers to a composite of all speech acts that
reflect high affiliation. “Overall assertive language” refers to a composite of all speech acts that reflect unmitigated
assertion (high assertion and low affiliation). General task comments were only included in the meta-analysis of
adults’ speech.
(*) p < .05.
Page 17 of 37

Collaborative Language
Collaborative speech acts are high in both affiliation and assertion (Leaper, 1991; Pen
man, 1980). These acts include elaboration and verbal support. Elaboration occurs when
a person builds on the prior speaker’s statement in a relevant manner. For example,
among children, this might include advancing the plot in a fantasy-play scenario that a
friend has introduced; among adults, it includes (p. 71) (p. 72) building on a partner’s pri
or comment in ways that demonstrate active understanding. Verbal support, another type
of collaborative speech, is expressed through the use of reassuring comments or praise.
These speech acts affirm the other person and therefore reflect high affiliation; at the
same time, supportive statements are moderately assertive because they are aimed at di
rectly influencing the recipient (i.e., make the person feel better).
Average gender differences in collaborative speech have been indicated in reviews. In

Leaper and Smith’s (2004) meta-analysis of children’s speech, elaborations (or responsive
statements) were significantly more likely among girls than boys (d = .41). No significant
gender difference in supportive statements was found, although the trend was toward a
higher average among girls than boys (d = .13). A similar pattern was indicated in Leaper
and Ayres’ (2007) meta-analysis of adults’ speech. Elaboration was significantly more like
ly among women than men (d = .41). Supportive statements were also significantly more
common among women than men, although the effect size was negligible (d = .16).
Average gender differences in collaborative speech have implications for close relation
ships. Active understanding and verbal support are associated with satisfaction in hetero
sexual couples (Pasch & Bradbury, 1998), as well as in gay and lesbian couples (Kurdek,
2005). Traditional gender roles, however, may disadvantage women in cross-gender rela
tionships. Some studies have found, on average, that men were less likely than women to
provide supportive responses to partners in cross-gender friendships (Leaper, Carson,
Baker, Holliday, & Myers, 1995) and heterosexual relationships (Steil, 2000). This aver
age difference has been observed in studies of heterosexual couples in America and other
parts of the world including China (Xu & Burleson, 2001) and Costa Rica (Schwarzer &
Gutiérrez-Doña, 2005).
One potential limitation of some of the research on spousal satisfaction and support is
that many studies have been based on individuals’ self-reports of perceived support in
their relationships—which may not always be accurate. Neff and Karney (2005) sought to
address this concern by collecting both observational and diary data of heterosexual cou
ples’ supportive responses. They found no average difference between women and men in
providing verbal support. However, there were average gender differences in the timing
of support. Women’s supportive responses were tied to the severity of their spouse’s
stress. If their husband’s stress was more severe, then women tended to increase their
positive support. In contrast, men’s provision of emotional support was not related to the
severity of their wives’ stress.
Page 18 of 37

Co-Rumination
Although verbal support is generally considered a positive sign of support, recent re
search suggests it is possible to have too much of a good thing. Sometimes, a personal
problem can be discussed excessively in a relationship. This is known as co-rumination
(Rose, Carlson, & Waller, 2007). Although actively discussing a friend’s or a partner’s per
sonal problems can be perceived as supportive, it may inadvertently reinforce the other
person’s negative feelings about the event. A 6-month longitudinal study of children indi
cated that frequent co-rumination was related to increased positive friendship quality in
girls as well as boys. Only for girls, however, was co-rumination additionally associated
with increased anxiety and depression (Rose et al., 2007). Co-rumination may have in
creased these symptoms in girls but not boys because girls are at greater risk for inter
nalizing symptoms (anxiety and depression). Also, because self-disclosure tends to be
more common among girls than boys, perhaps co-rumination more readily becomes a tip
ping point toward distress for girls than boys.
Research on co-rumination suggests that, as with many things in life, finding the right
balance is important when it comes to providing verbal support in close relationships. In
addition to friendships, this is indicated for marital relationships. Brock and Lawrence
(2009) examined degrees of reported partner support in relation to marital satisfaction
over a 5-year period. They considered overprovision as well as underprovision of support
(i.e., more or less support received than desired, respectively). Consistent with traditional
gender roles, wives were more likely to report underprovision of support from their hus
bands than the reverse. There was no average gender difference in reported overprovi
sion of support. However, both under- and overprovision of partner support were related
to declines in marital satisfaction for women and men. Thus, it is in both women’s and
men’s interest to provide and receive optimal support in close relationships.
Obliging Language
Obliging speech acts are high in affiliation but relatively low in assertion (Leaper, 1991).
They reflect an emphasis on going along with the other (p. 73) person without directly in
fluencing her or him. Two forms of obliging speech acts are agreement and simple ac
knowledgments. An agreement is high in affiliation and low in assertion because it af
firms another speaker’s proposal without asserting anything new (e.g., “Sure, let’s do
that”). Simple acknowledgments include short statements (“That’s too bad”) or minimal
responses (“Um-hm”) that demonstrate the person is listening. As summarized in Table
4.1, meta-analyses do not indicate significant gender differences in agreements or simple
acknowledgments among either children or adults.
According to characterizations of the traditional female role as emphasizing high affilia

tion and low assertion (e.g., see Spence & Helmreich, 1978), one might have expected
obliging speech acts would be more common among females than males. The more likely
pattern is that females are somewhat more likely than males to coordinate affiliative and
assertive goals (collaboration) in their social interactions (e.g., Leaper, 1991). Indeed, as
noted in the previous section, significant average gender differences (with modest effect
Page 19 of 37

sizes) in collaborative speech acts were indicated for children and adults in meta-analy
ses.
Politeness: Affiliation With or Without Assertion?

Politeness terms include phrases expressing civility such as “please” and “thank you.”
These speech forms are construed as affiliative language because they demonstrate inter
personal sensitivity. Timmerman’s (2002) meta-analysis indicated a significantly higher
likelihood of politeness terms among women than men (d = .27). This pattern was indicat
ed during same-gender and mixed-gender contexts.
Researchers have differed regarding whether or not politeness should be seen as power
less (unassertive) (see Timmerman, 2002). In some situations, politeness terms may re
flect high affiliation and low assertion. An example might be saying “thank you” in re
sponse to receiving a compliment. Speakers also may be more likely to use politeness
terms to signal deference to higher-status persons. In other contexts, however, politeness
terms may be high in both assertion and affiliation. An example might be requesting
someone to do something (e.g., “Please bring me a glass of water”). These are potentially
interesting distinctions to explore more fully in future research.
Tentative Speech: Interpersonal Sensitivity or Powerlessness?

In a frequently cited monograph, Lakoff (1973, 1977) proposed that women were more
likely than men to use various forms of tentative speech, including expressions of uncer
tainty (disclaimers [e.g., “I may be mistaken, but…”] or qualifiers [e.g., “It was somewhat
unclear”]), hedges (phrases that downplay the force of one’s assertion [e.g., “I guess…”]),
and tag questions (a confirmation-seeking question following a declarative statement
[e.g., “That seems interesting, doesn’t it?”]). Leaper and Robnett (2011b) recently con
ducted a meta-analysis of studies testing for average gender differences in these speech
forms. On average, all three forms were significantly more likely among women than
men, with negligible to small effect sizes (expressions of uncertainty: d = .33, hedges: d
= .15, tag questions: d = .23).
Lakoff (1973, 1977) proposed that women were more likely than men to use tentative lan
guage due to power asymmetries between men and women. If so, then one might find
larger average differences during mixed-gender than same-gender interactions. However,
Leaper and Robnett’s (2011b) meta-analysis did not point toward this pattern; gender
composition was not a significant moderator.
Rather than characterize tentative speech forms as unassertive, it may be more meaning
ful to consider their potential affiliative function. Speakers may sometimes use tentative
language to soften their assertion and thereby increase the other person’s involvement.
That is, they may reflect politeness and interpersonal sensitivity. For example, consider a
speaker who prefaces a statement of opinion with a disclaimer (“I may be wrong but I
think…”). Rather than directly asserting one’s point of view, this strategy acknowledges
the listener may have a different view (see Leaper & Robnett, 2011b). Thus, these speech
Page 20 of 37

forms can be both assertive (stating one’s view) and affiliative (sensitivity to other’s per
spective).
Controlling Language
Controlling speech acts are high in self-assertion and relatively low in affiliation. They
emphasize the speaker’s agency while negating the other person. That is, they reflect un
mitigated assertion. If one thinks of the affiliation dimension as ranging from highly posi
tive to highly negative toward the other person, a distinction can be made between con
trolling acts that are nonaggressive (i.e., moderately low in affiliation) or aggressive (i.e.,
very low in affiliation). Evidence for average gender differences in each type of control
ling speech is described below. Afterward, research on gender-related variations in verbal
sexual harassment and sexist jokes is reviewed.
(p. 74) Nonaggressive Controlling Speech Acts

Nonaggressive forms of controlling speech include making suggestions or other task-ori
ented comments, directives, disagreements, and criticism. As summarized in Table 4.1,
males were significantly more likely than females to use some nonaggressive speech acts
in studies of children and adults. In Leaper and Smith’s (2004) meta-analysis of studies
examining children’s speech, there were no significant average gender differences in ex
pressions of disagreement (d = .10) or criticism (d = –.01). However, boys were signifi
cantly more likely than girls to make suggestions (d = .13) and directives (d = .23). In
Leaper and Ayres’ (2007) meta-analysis of gender and adults’ speech, no significant gen
der differences occurred in disagreements (d = –.06) or directives (d = .00). However,
men were significantly more likely than women to offer suggestions (d = .27) and to use
general task comments (d = .38). The significant average differences among children and
adults are consistent with prior conceptualizations of the traditional male role as empha
sizing task orientation (see Anderson & Blanchard, 1982).
Leaper and Ayres (2007) additionally noted a significant but negligible average difference
in criticism—with women slightly higher than men (d = –.13). Perhaps the pattern reflect
ed a greater average willingness among women than men to express feelings. In the con
text of heterosexual relationships, women may be more likely than men to make emotion
al demands (e.g., “I wish you were more affectionate” or “You aren’t listening”) (Holley,
Sturm, & Levenson, 2010).
Aggressive Controlling Speech Acts

Archer’s (2004) meta-analysis indicated that verbal aggression was more likely among
males than females across studies of children and adults. (Age was not a significant mod
erator.) Average effect sizes were reported separately for studies based on either obser
vations or peer reports. Among observational studies, the effect size was significant but
negligible in magnitude (d = .14). However, when studies employed peer report mea
sures, the average effect size was considerably larger (d = .51).
Page 21 of 37

Additional research is needed to clarify the reasons for the difference in effect size due to
methodology. One possibility is that peer reports were subject to people’s gender stereo
types and led to exaggerated perceptions of gender differences in aggression. An alterna
tive possibility is that peer reports were more accurate than observational studies be
cause peers can take into account how people act across times and contexts.
Verbal Sexual Harassment

Sexual harassment occurs when an individual directs unwanted sexual conduct toward
someone. As with other forms of aggressive behavior, sexual harassment can be physical
(e.g., unwanted touching) or verbal (e.g., unwelcome sexual comments). The American
Association of University Women (AAUW, 2011) recently conducted a large-scale study of
sexual harassment during adolescence. They surveyed a nationally representative sample
in the United States of 1,965 students in grades 7–11. Verbal forms of sexual harassment
included unwelcome sexual comments or jokes, as well as being called lesbian or gay in a
negative way. Girls were twice as likely as boys to report being targets of unwelcome sex
ual comments/jokes (46 percent of girls vs. 22 percent of boys). Girls and boys reported
experiencing anti-gay or anti-lesbian insults at similar rates (18 percent of girls vs. 19
percent of boys). Other surveys of adolescents’ experiences with sexual harassment have
indicated similar patterns, although the percentages vary somewhat (e.g., Lacasse, Purdy,
& Mendelson, 2003; Leaper & Brown, 2008; Pepler et al., 2006).
Ilies, Hauserman, Schwochau, and Stibal (2003) conducted a meta-analytic review of the
incidence of workplace sexual harassment for women in the United States. Across sam
ples, the researchers found 24 percent of women reported having experienced sexual ha
rassment at work. Reports tended to be higher for women in the military compared to pri
vate sector, government, or academic work settings. The review did not differentiate be
tween verbal and nonverbal forms of sexual harassment. Also, it did consider men’s expe
riences with workplace sexual harassment. According to other sources, at least 40 per
cent of women and 13 percent of men in the United States experienced sexual harass
ment in the workplace (Aggarwal & Gupta, 2000). Furthermore, studies from other coun
tries indicate that sexual harassment is a problem across the world for many adolescents
and adults (see Leaper & Robnett, 2011a; Willness, Steel, & Lee, 2007, for reviews).
Studies generally indicate that males are more likely than females to be perpetrators of
sexual harassment during adolescence in school settings (see AAUW, 2011; Leaper & Rob
nett, 2011a, for reviews) and during adulthood in the workplace (see Aggarwal & Gupta,
2000). In addition to being somewhat more liable to be targets of sexual harassment, fe
males are more likely than males to be negatively affected by sexual harassment during
(p. 75) adolescence (AAUW, 2011; Fineran & Bolen, 2006). Sexual harassment also tends
to have a more negative impact on sexual-minority boys than on heterosexual boys

(Kosciw, Greytak, Diaz, & Bartkiewicz, 2010). Workplace sexual harassment appears to
have similar effects on women and men according to a recent meta-analysis (Chan, Lam,
Chow, & Cheung, 2008). For both adolescents and adults, negative reactions to sexual ha
rassment include internalizing symptoms (e.g., anxiety, depression) and declines in acade
Page 22 of 37

mic or work performance (see AAUW, 2011; Leaper & Robnett, 2011a; Chan et al., 2008;
Willness et al., 2007, for reviews).
Online Sexual Harassment

In addition to face-to-face encounters, verbal sexual harassment can be instigated thor
ough the use of electronic media such as text messages, e-mail, and web postings. Ac
cording to the AAUW (2011) survey, 36 percent of girls and 24 percent of boys experi
enced online sexual harassment. Forms of online harassment included being sent unwel
come sexual messages (26 percent of girls and 13 percent of boys), being the target of
unfavorable sexual rumors (17 percent of girls and 8 percent of boys), and being called
gay or lesbian in a negative manner (12 percent of girls vs. 12 percent of boys). The use
of electronic media for harassment is an increasing concern among researchers, policy
makers, teachers, and parents.
Sexist Jokes
As previously noted, sexist jokes are one form of verbal sexual harassment. Studies of
adolescents indicate that girls are more likely than boys to report feeling distressed when
told sexist jokes (AAUW, 2011). The latter finding may reflect the fact that most sexist
jokes are disparaging of females. Sexist jokes can also undermine women’s status in the
workplace (Boxer & Ford, 2010).
Unlike outright hostile comments (e.g., “Women make terrible managers”), sexist jokes
are a more subtle form of sexism. The joke-tellers’ typical rebuttal to charges of poor
taste is that they are “just joking” and the offended party lacks a sense of humor (Boxer &
Ford, 2010). However, men’s enjoyment of sexist (female-disparaging) humor is correlat
ed with sexist attitudes (Thomas & Esses, 2004) and acceptance of rape myths (Ryan &
Kanjorski, 1998). Women’s enjoyment of sexist humor is associated with a belief in adver
sarial heterosexual relationships and tolerance of sexual violence (Ryan & Kanjorski,
1998).
Sexist humor may activate sexist thinking in men who already hold sexist attitudes. In ex
perimental studies, Ford and colleagues have found that exposure to sexist (vs. neutral)
jokes increased tolerance of sexist events (Ford, Wentzel, & Lorion, 2001) or decreased
support of women’s organizations (Ford, Boxer, Armstrong, & Edel, 2008) among men
with hostile sexist attitudes but not among men with nonsexist attitudes. This work and
the previously cited studies point to the pernicious effects that sexist jokes can have on
women and men.
Moderators of Average Gender Differences in Speech Acts
In light of the exaggerated views of male–female communication differences (see Barnett

& Rivers, 2004; Cameron, 2007; MacGeorge, Graves, Feng, Gillihan, & Burleson, 2004), it
is revealing to note that average gender differences in communication style across stud
ies are typically negligible to small when they occur. To understand the kinds of factors
that may influence if and when average gender differences in language do occur, modera
Page 23 of 37

tor analyses can be helpful. To illustrate, some findings associated with three moderators
of gender differences in affiliative and assertive language are reviewed: gender composi
tion, familiarity, and activity setting. The reader is also referred to the cited meta-analy
ses for examples of other moderators.
Gender Composition
Prior research indicates that the gender composition of a dyad or a group can have a sig
nificant impact on people’s behavior. During same-gender interactions, individuals are
more likely to share similar gender-related norms for social behavior. In contrast, during
mixed-gender interactions, gender may become a status characteristic that differentiates
individuals from one another (Deaux & Major, 1987; Wood & Karten, 1986). Thus, some
researchers (e.g., Carli, 1990; Ridgeway & Smith-Lovin, 1999) have proposed that gender
composition effects point to the relative influences of gender-related ingroup norms ver
sus intergroup dominance on observed gender differences in social behavior.
As noted earlier, support of the dominance and status explanation was indicated for gen
der differences in adults’ talkativeness. That is, differences were greater during mixed-
gender than same-gender interactions. In contrast, support of the social (p. 76) norms ex
planation was indicated for average gender differences in assertive and affiliative speech.
Among studies of children, this was seen regarding gender differences in controlling-as
sertive speech in Leaper and Smith’s (2004) meta-analysis. Across all forms of controlling
speech, the average incidence was significantly higher for boys than girls among studies
of same-gender interactions (d = .29); in contrast, there was no difference among studies
of mixed-gender interactions (d = .01). Among studies of adults, Leaper and Ayres’ (2007)
meta-analysis similarly indicated that gender differences were more pronounced during
same-gender than mixed-gender interactions across measures of affiliative speech (same-
gender: d = .33, mixed-gender: d = .01) and controlling speech (same-gender: d = .29,
mixed-gender: d = .03).
The social norms explanation is consistent with intergroup theory and other approaches
emphasizing how ingroup members tend to share similar behavioral norms (Tajfel &
Turner, 1979). Also, the social norms interpretation does not have to be seen as contra
dictory to the male dominance explanation (Leaper, 2000b). Gender inequalities in society
shape the ways that girls and boys are commonly socialized (Eagly, Wood, & Diekman,
2000; Leaper, 2000b).
Asymmetrical Accommodation During Mixed-Gender Interactions?

One question that the meta-analyses did not address was whether the absence of gender
difference in speech acts within mixed-gender interactions was due to one gender or both
genders modifying the communication styles that they might favor during same-gender
interactions. Individual studies addressing this point suggest that females may be most
likely to modify their preferred style. For example, Miller, Danaher, and Forbes (1986)
considered gender composition effects in relation to children’s uses of conflict strategies.
During conflicts with same-gender peers, boys tended to use fewer conflict-mitigation
strategies (e.g., compromise) and more power-assertive strategies (e.g., demands) than
Page 24 of 37

did girls. During conflicts between cross-gender peers, there were no average differences
in strategies because girls, on average, increased their use of power-assertive strategies
and decreased their use of conflict-mitigation strategies. In contrast, boys, on average,
used similar rates of power-assertive and conflict-mitigation strategies during same-gen
der or cross-gender conflicts. Thus, from an early age, many girls may find that they must
play by the boys’ rules to get their way in society.
Familiarity
People tend to be more likely to enact some gender-typed behaviors with unfamiliar than
familiar partners (Deaux & Major, 1987). This may occur for at least two reasons. First,
when encountering strangers, individuals may be more apt to invoke their gender
schemas to form expectations about them. A second reason is that people tend to be more
concerned about self-presentation with strangers than with familiar persons. For many in
dividuals, appearing in a gender-stereotypical manner may be important in these situa
tions. Although familiarity was not a significant moderator in the meta-analysis of
children’s language (Leaper & Smith, 2004), it was a significant factor in the meta-analy
sis of adults’ language (Leaper & Ayres, 2007). Consistent with the self-presentation mod
el, gender differences were significantly more likely among strangers than familiar part
ners (e.g., friends, spouses/partners, family). This was seen with regards to affiliative
speech (strangers: d = .18, close relationships: d = .02) and controlling-assertive speech
(strangers: d = .22, close relationships: d = –.04).
Activity Setting
As emphasized in social cognitive theory (Bussey & Bandura, 1999), the likelihood of en
acting particular behaviors depends partly on having the opportunity to practice them.
Gender-typed play activities during childhood create correspondingly different occasions
for girls and boys. Gender-typed activities for boys tend to stress competition (e.g.,
sports), aggression (e.g., playing war), and task orientation (e.g., playing with construc
tion toys). Conversely, gender-typed activities for girls tend to focus more on nurturance
(e.g., caring for baby dolls) and close relationships (e.g., playing house). Different activi
ties, in turn, tend to emphasize different styles of talk (Leaper, 2000a; Maltz & Borker,
1982). In support of this argument, average gender differences in affiliative speech (with
girls higher than boys) were moderated by the activity setting in Leaper and Smith’s
(2004) meta-analysis. The average gender difference was larger during unstructured ac
tivities (d = .65) than structured (i.e., assigned) activities (d = .20). In a similar manner,
Leaper and colleagues’ (1998) meta-analysis comparing mothers’ speech to daughters
versus sons indicated that differences in affiliative speech were mitigated when the activ
ity was assigned. The (p. 77) smaller effect sizes seen in structured activities therefore
suggests that the type of activity or conversation topic may partly mediate (a) gender dif
ferences in children’s affiliative speech as well as (b) parents’ differential speech to girls
and boys (Huston, 1985; Leaper, 2000a).
The activity or topic may have a different impact on gender differences in adults’ speech
style. Gender differences in communication during childhood may be largely a conse
Page 25 of 37

quence of the different activity contexts in which many girls and boys are engaged. When
placed in similar situations, most girls and boys may have the flexibility to modify their
behavior to fit the activity (e.g., girls and boys similarly used affiliative speech when
asked to play with toy foods and plates [Leaper, 2000a]). Over time, however, these expe
riences may shape individuals’ expectations and preferred styles (see Leaper, 2000b).
When placed in feminine- or masculine- stereotyped settings as adults, gender-typed

women and men interpret and respond differently to these situations. This idea is consis
tent with the patterns indicated in Leaper and Ayres’ (2007) meta-analysis of gender dif
ferences in adults’ speech. In unstructured settings, there were no significant gender dif
ferences in affiliative or assertive speech. But there were significant differences indicated
during particular activities. (This is the opposite of the pattern seen among children.) For
example, during discussions of assigned nonpersonal topics, men’s averages were signifi
cantly higher than those of women for controlling-assertive speech (d = .48). Conversely,
during discussions of assigned personal topics, women’s average was significantly higher
for affiliative speech (d = .20). These patterns may reflect the different kinds of situations
in which gender-typed women and men (a) expect to enact gender-typed behaviors (as
emphasized in expectation states theory) and/or (b) feel comfortable and have a stronger
sense of self-efficacy (as emphasized in social cognitive theory).
Conclusion and Future Directions

Three important conclusions can be drawn from the foregoing review of research on gen
der and language. First, languages tend to be gender biased. For example, masculine
nouns and pronouns are commonly used to refer generically to people. Second, when av
erage gender differences in language style occur, they tend to be small in magnitude with
a great deal of overlap in the distributions of females and males. And third, gender-relat
ed variations in language behavior tend to vary across different situations. The latter two
points are striking in light of popular accounts of females and males as being so different
in their respective communication styles that they might be from different cultures or
planets (see Barnett & Rivers, 2004; Cameron, 2007; MacGeorge et al., 2004, for discus
sions of this issue).
In a recent review of various meta-analyses testing for gender differences in various so
cial behaviors, Hyde (2005) advocated for the gender similarities hypothesis. She noted
that for many social behaviors, the degree of similarity between males and females is
much greater than the amount of difference. Yet, why do popular myths about gender dif
ferences in language style persist? As with other kinds of stereotyping, people tend to pay
more attention to cases that support their views and to ignore instances that contradict
them (Rudman & Glick, 2008). Also, people tend to associate themselves with like-minded
people who endorse and practice similar gender roles. In these ways, stereotyped beliefs
are perpetuated.
Page 26 of 37

Although the average effect sizes for most language behaviors tend to be small or negligi
ble, there are some consistent patterns when statistically significant gender differences
were indicated. One pattern was females tend to be more likely to use language that coor
dinated affiliative and assertive goals (collaborative speech). Conversely, males tend to be
more likely than females to use language that emphasizes assertion over affiliation (con
trolling speech). There was no evidence from meta-analyses that females were more like
ly than males to use language that emphasizes affiliation while downplaying assertion
(obliging speech). With all of these average patterns, however, situational factors tended
to moderate the incidence and magnitude of gender differences. Some of these included
familiarity among interaction partners, the gender composition of the dyad or group, and
the activity setting or conversation topic.
Understanding how contextual factors influence gender-related variations in language

should be a continued focus for research. In addition to facets of the interactive context
(e.g., familiarity, gender composition, activity), it is also important to consider how
people’s sociocultural backgrounds may play a role. One of the shortcomings of the
present review is that cultural and socioeconomic variations in gender-related differences
in language were not addressed in more (p. 78) depth. Unfortunately, there have been rel
atively few quantitative studies taking into account culture and other social-structural
factors as possible moderators of gender differences in language. Indeed, this is a limita
tion for the behavioral sciences in general (Henrich, Heine, & Norenzayan, 2010). Ethno
graphic research (e.g., see Goodwin, 2002; Kyratzis & Guo, 2001) and a few pertinent
quantitative studies (e.g., Filardo, 1996) suggest some ways that gender-related language
might vary depending on people’s sociocultural backgrounds and contexts.
Another suggested area for future research is to clarify how social-cognitive processes
underlie gender differences in language. The results have been mixed in past studies
when researchers have sought to identify possible cognitive mediators of gender differ
ences in language. It may be that certain cognitive processes (e.g., goals) are better pre
dictors than other cognitive processes (e.g., attitudes) of language style. Furthermore,
links among cognitive processes, gender, and behavior may vary across different con
texts. For example, people’s attitudes and values may have a weaker impact on their be
havior when they are interacting with strangers (Deaux & Major, 1987).
Continued research on language and gender is important for at least two reasons. First,
there is widespread interest among the general public on the topic—as seen in the ongo
ing popularity of books and other media on the topic. Unfortunately, much of the informa
tion dispensed is inaccurate and perpetuates unfounded gender stereotypes (see Barnett
& Rivers, 2004, for a review). This is all the more reason for more scientific research.
Finally, advancing our understanding of gender and language is important because it has
practical implications. Language is used to create and maintain gender divisions in soci
ety. Furthermore, the way that people use language in conversation affects the quality of
their relationships with co-workers, friends, and romantic partners. Gender imbalances in
language style are correspondingly related to problems in each of these relationship con
Page 27 of 37

texts. At the same time, as this review has highlighted, we have a good understanding of
some of the ways that language can be used to facilitate greater gender equality and to
improve people’s relationships.
References
Aggarwal, A. P., & Gupta, M. M. (2000). Sexual harassment in the workplace (3rd ed.).
Vancouver, BC: Butterworths.
Aldrich, N. J., & Tenenbaum, H. R. (2006). Sadness, anger, and frustration: Gendered pat
terns in early adolescents’ and their parents’ emotion talk. Sex Roles, 55, 775–785.
American Association of University Women (AAUW). (2011). Crossing the line: Sexual ha
rassment at school. Washington, DC: Author.
Anderson, K. J., & Leaper, C. (1998). Meta-analyses of gender effects on conversational in
terruption: Who, what, when, where, and how. Sex Roles, 39, 225–252.
Anderson, L. R., & Blanchard, P. N. (1982). Sex differences in task and social-emotional
behavior. Basic and Applied Social Psychology, 3, 109–139.
Archer, J. (2004). Sex differences in aggression in real-world settings: A meta-analytic re

view. Review of General Psychology, 8, 291–322.
Aries, E. (1996). Men and women in interaction: Reconsidering the differences. New York:
Oxford University Press.
Bakan, D. (1966). The duality of human existence. Boston: Beacon.
Bales, R. F. (1950). Interaction process analysis: A method for the study of small groups.
Reading, MA: Addison-Wesley.
Bales, R. F. (1970). Personality and social behavior. New York: Holt, Rinehart & Winston.
Barnett, R. C., & Rivers, C. (2004). Same difference: How gender myths are hurting our
relationships, our children, and our jobs. New York: Basic Books.
Bem, S. L. (1974). The measurement of psychological androgyny. Journal of Consulting &

Clinical Psychology, 42, 155–162.
Bigler, R. S., & Liben, L. S. (2007). Developmental intergroup theory: Explaining and re
ducing children’s social stereotyping and prejudice. Current Directions in Psychological
Science, 16, 162–166.
Bischoping, K. (1993). Gender differences in conversation topics, 1922–1990. Sex Roles,

28, 1–18.
Page 28 of 37

Boxer, C. F., & Ford, T. E. (2010). Sexist humor in the workplace: A case of subtle harass
ment. In J. Greenberg (Ed.), Insidious workplace behavior (pp. 175–205). New York: Rout
ledge/Taylor & Francis.
Boxer, D., & Gritsenko, E. (2005). Women and surnames across cultures: Reconstituting
identity in marriage. Women & Language, 28, 1–11.
Brock, R. L., & Lawrence, E. (2009). Too much of a good thing: Underprovision versus
overprovision of partner support. Journal of Family Psychology, 23, 181–192.
Buhrmester, D., & Prager, K. (1995). Patterns and functions of self-disclosure during
childhood and adolescence. In K. J. Rotenberg (Ed.), Disclosure processes in children and
adolescents (pp. 10–56). New York: Cambridge University Press.
Bussey, K., & Bandura, A. (1999). Social cognitive theory of gender development and dif
ferentiation. Psychological Review, 106, 676–713.
Cameron, D. (2007). The myth of Mars and Venus. New York: Oxford University.
Carli, L. L. (1990). Gender, language, and influence. Journal of Personality and Social Psy
chology, 59, 941–951.
Chan, D. K., Lam, C. B., C. S. Y., & Cheung, S. F. (2008). Examining the job-related, psy
chological, and physical outcomes of workplace sexual harassment: A meta-analytic re
view. Psychology of Women Quarterly, 32, 362–376.
Coates, J. (2004). Women, men, and language (3rd ed.). New York: Pearson Longman.
Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.). Hills
dale, NJ: Lawrence Erlbaum.
Conkright, L., Flannagan, D., & Dykes, J. (2000). Effects of pronoun type and gen
(p. 79)
der role consistency on children’s recall and interpretation of stories. Sex Roles, 43, 481–
497.
Coyne, S. M., Archer, J., & Eslea, M. (2006). “We’re not friends anymore! Unless…”: The
frequency and harmfulness of indirect, relational and social aggression. Aggressive Be
havior, 32, 294–307.
Crawford, M. E. (1995). Talking difference: On gender and language. Thousand Oaks, CA:
Sage.
Crawford, M. E., Stark, A. C., & Renner, C. H. (1998). The meaning of Ms.: Social assimi
lation of a gender concept. Psychology of Women Quarterly, 22, 197–208.
Deaux, K., & Major, B. (1987). Putting gender into context: An interactive model of gen
der-related behavior. Psychological Review, 94, 369–389.
Page 29 of 37

DeLoache, J. S., Cassidy, D. J., & Carpenter, C. J. (1987). The three bears are all boys:
Mothers’ gender labeling of neutral picture book characters. Sex Roles, 17, 163–178.
Dindia, K., & Allen, M. (1992). Sex differences in self-disclosure: A meta-analysis. Psycho
logical Bulletin, 112, 106–124.
Dindia, K., & Canary, D. J. (2006). Sex differences and similarities in communication (2nd
ed.). Mahwah, NJ: Lawrence Erlbaum.
Dion, K. L., & Schuller, R. A. (1990). Ms. and the manager: A tale of two stereotypes. Sex
Roles, 22, 569–577.
Eagly, A. H., Wood, W., & Diekman, A. B. (2000). Social role theory of sex differences and
similarities: A current appraisal. In T. Eckes & H. M. Trautner (Eds.), The developmental
social psychology of gender (pp. 123–174). Mahwah, NJ: Erlbaum.
Eckert, P. (1990). Cooperative competition in adolescent “girl talk.” Discourse Processes,

13, 91–122.
Eckert, P., & McConnell-Ginet, S. (2003). Language and gender. New York: Cambridge
University Press.
Farley, S. D. (2011). Is gossip power? The inverse relationships between gossip, power,
and likeability. European Journal of Social Psychology, 41, 574–579. doi 10.1002/ejsp.821
Filardo, E. M. (1996). Gender patterns in African American and white adolescents’ social
interactions in same-race, mixed-gender groups. Journal of Personality and Social Psy
chology, 71, 71–82.
Fineran, S., & Bolen, R. M. (2006). Risk factors for peer sexual harassment in schools.
Journal of Interpersonal Violence, 21, 1169–1190.
Finkenauer, C., & Hazam, H. (2000). Disclosure and secrecy in marriage: Do both con
tribute to marital satisfaction? Journal of Social & Personal Relationships, 17, 245–263.
Fivush, R., Brotman, M. A., Buckner, J. P., & Goodman, S. H. (2000). Gender differences in
parent-child emotion narratives. Sex Roles, 42, 233–253.
Forbes, G. B., Adams-Curtis, L. E., White, K. B., & Hamm, N. R. (2002). Perceptions of
married women and married men with hyphenated surnames. Sex Roles, 46, 167–175.
Ford, T. E., Boxer, C. F., Armstrong, J., & Edel, J. R. (2008). More than “just a joke”: The
prejudice-releasing function of sexist humor. Personality and Social Psychology Bulletin,
34, 159–170.
Ford, T. E., Wentzel, E. R., & Lorion, J. (2001). Effects of exposure to sexist humor on per
ceptions of normative tolerance of sexism. European Journal of Social Psychology, 31,
677–691.
Page 30 of 37

Foster, E. K. (2004). Research on gossip: Taxonomy, methods, and future directions. Re
view of General Psychology, 8, 78–99.
Friedman, C. K., Leaper, C., & Bigler, R. S. (2007). Do mothers’ gender-related attitudes
or comments predict young children’s gender beliefs? Parenting: Science and Practice, 7,
357–366.
Gelb, S. A. (1989). Language and the problem of male salience in early childhood class
room environments. Early Childhood Research Quarterly, 4, 205–215.
Gelman, S. A., Taylor, M. G., & Nguyen, S. P. (2004). The developmental course of gender
differentiation. Monographs of the Society for Research in Children Development, 69, 1,
vii–127.
Gleason, J. B., & Ely, R. (2002). Gender differences in language development. In A. M. De

Lisi & R. De Lisi (Eds.), Biology, society, and behavior: The development of sex differences
in cognition (pp. 127–154). Westport, CT: Ablex.
Glick, P., & Fiske, S. T. (1996). The ambivalent sexism inventory: Differentiating hostile
and benevolent sexism. Journal of Personality and Social Psychology, 70, 491–512.
Goodwin, M. H. (2002). Exclusion in girls’ peer groups: Ethnographic analysis of lan

guage practices on the playground. Human Development, 45, 392–415.
Graddol, D., & Swann, J. (1989). Gender voices. Oxford, UK: Basil Blackwell.
Gray, J. (1992). Men are from Mars, women are from Venus. New York: Harper Collins.
Hall, J. A. (2011). Sex differences in friendship expectations: A meta-analysis. Journal of

Social & Personal Relationships, 28, 723–747.
Hamilton, M. C. (1988). Using masculine generics: Does generic he increase male bias in
the user’s imagery? Sex Roles, 19, 785–799.
Henley, N. M. (1989). Molehill or mountain? What we know and don’t know about sex bias
in language. In M. Crawford & M. E. Gentry (Eds.), Gender and thought: Psychological
perspectives (pp. 59–78). New York: Springer-Verlag.
Henrich, J., Heine, S. J., & Norenzayan, A. (2010). The weirdest people in the world? Be
havioral and Brain Sciences, 33, 61–135.
Hoffnung, M. (2006). What’s in a name? Marital name choice revisited. Sex Roles, 55,
817–825.
Holley, S. R., Sturm, V. E., & Levenson, R. W. (2010). Exploring the basis for gender differ
ences in the demand-withdraw pattern. Journal of Homosexuality, 57(5), 666–684.
Holmes, J., & Meyerhoff, M. (Eds.). (2003). The handbook of language and gender. Cam
bridge, MA: Blackwell.
Page 31 of 37

Huston, A. C. (1985). The development of sex typing: Themes from recent research. De
velopmental Review, 5, 1–17.
Hyde, J. S. (1984). Children’s understanding of sexist language. Developmental Psycholo

gy, 20, 697–706.
Hyde, J. S. (2005). The gender similarities hypothesis. American Psychologist, 60, 581–
592.
Ilies, R., Hauserman, N., Schwochau, S., & Stibal, J. (2003). Reported incidence rates of
work-related sexual harassment in the United States: Using meta-analysis to explain re
ported rate disparities. Personnel Psychology, 56, 607–631.
James, D., & Drakich, J. (1993). Understanding gender differences in amount of talk: A
critical review of the research. In D. Tannen (Ed.), Gender and conversational interaction
(pp. 281–312). New York: Oxford University Press.
Kosciw, J. G., Greytak, E. A., Diaz, E. M., & Bartkiewicz, M. J. (2010). The 2009 national
school climate survey: The experiences of lesbian, gay, bisexual, and transgender youth in
our nation’s schools. New York: GLSEN.
Kurdek, L. A. (2005). What do we know about gay and lesbian couples? Current Direc
tions in Psychological Science, 14, 251–254.
Kyratzis, A., & Guo, J. (2001). Preschool girls’ and boys’ verbal conflict strategies
(p. 80)
in the U. S. and China: Cross-cultural and contextual considerations. Research on Lan

guage and Social Interaction, 34, 45–74.
Lacasse, A., Purdy, K. T., & Mendelson, M. J. (2003). The mixed company they keep: Po
tentially offensive sexual behaviors among adolescents. International Journal of Behav
ioral Development, 27, 532–540.
Lakoff, R. T. (1973). Language and woman’s place. Language in Society, 2, 45–79.
Lakoff, R. T. (1977). Women’s language. Language and Style, 10, 222–247.
Lambdin, J. R., Greer, K. M., Jibotian, K. S., Wood, K. R., & Hamilton, M. C. (2003). The
animal = male hypothesis: Children’s and adults’ beliefs about the sex of non-sex-specific
stuffed animals. Sex Roles, 48, 471–482.
Lawton, C. A., Blakemore, J. E. O., & Vartanian, L. R. (2003). The new meaning of Ms.:
Single, but too old for Miss. Psychology of Women Quarterly, 27, 215–220.
Leaper, C. (1991). Influence and involvement in children’s discourse: Age, gender, and
partner effects. Child Development, 62, 797–811.
Leaper, C. (1995). The use of masculine and feminine to describe women’s and men’s be
havior. Journal of Social Psychology, 135, 359–369.
Page 32 of 37

Leaper, C. (2000a). Gender, affiliation, assertion, and the interactive context of parent-
child play. Developmental Psychology, 36, 381–393.
Leaper, C. (2000b). The social construction and socialization of gender. In P. H. Miller &
E. K. Scholnick (Eds.), Toward a feminist developmental psychology (pp. 127–152). New
York: Routledge Press.
Leaper, C., & Anderson, K. J. (1997). Gender development and heterosexual romantic rela
tionships during adolescence. In W. Damon (Series Ed.) & S. Shulman & W. A. Collins (Is
sue Eds.), Romantic relationships in adolescence: Developmental perspectives (New Di
rections for Child Development, No. 78, pp. 85–103). San Francisco: Jossey-Bass.
Leaper, C., Anderson, K. J., & Sanders, P. (1998). Moderators of gender effects on parents’
talk to their children: A meta-analysis. Developmental Psychology, 34, 3–27.
Leaper, C., & Ayres, M. M. (2007). A meta-analytic review of gender variation in adults’
language use: Talkativesness, affiliative speech, and assertive speech. Personality and So
cial Psychology Review, 11, 328–363.
Leaper, C., & Bigler, R. S. (2004). Gendered language and sexist thought. Monographs of
the Society for Research in Child Development, 69, 1, 128–142.
Leaper, C., & Brown, C. S. (2008). Perceived experiences with sexism among adolescent
girls. Child Development, 79, 685–704.
Leaper, C., Carson, M., Baker, C., Holliday, H., & Myers, S. B. (1995). Self-disclosure and
listener verbal support in same-gender and cross-gender friends’ conversations. Sex
Roles, 33, 387–404.
Leaper, C., & Holliday, H. (1995). Gossip in same-gender and cross-gender friends’ con
versations. Personal Relationships, 2, 237–246.
Leaper, C., & Robnett, R. D. (2011a). Sexism. In J. R. Levesque (Ed.), Encyclopedia of ado
lescence (pp. 2641–2648). New York: Springer.
Leaper, C., & Robnett, R. D. (2011b). Women are more likely than men to use tentative
language; aren’t they?: A meta-analysis testing for gender differences and moderators.
Psychology of Women Quarterly, 35, 129–142.
Leaper, C., & Smith, T. E. (2004). A meta-analytic review of gender variations in children’s
language use: Talkativeness, affiliative speech, and assertive speech. Developmental Psy
chology, 40, 993–1027.
Leaper, C., Tenenbaum, H. R., & Shaffer, T. G. (1999). Communication patterns of African
American girls and boys from low-income, urban background. Child Development, 70,
1489–1503.
Page 33 of 37

Leman, P. J., Ahmed, S., & Ozarow, L. (2005). Gender, gender relations, and the social dy
namics of children’s conversations. Developmental Psychology, 41, 64–74.
Levant, R. F., & Richmond, K. (2007). A review of research on masculinity ideologies us
ing the Male Role Norms Inventory. The Journal of Men’s Studies, 15, 130–146.
Levin, J., & Arluke, A. (1985). An exploratory analysis of sex differences in gossip. Sex
Roles, 12, 281–286.
Liben, L. S., Bigler, R. S., & Krogh, H. R. (2002). Language at work: Children’s gendered
interpretations of occupational titles. Child Development, 73, 810–828.
Lott, B. (1981). A feminist critique of androgyny: Toward the elimination of gender attri
butions for learned behavior. In C. Mayo & N. M. Henley (Eds.), Gender and nonverbal be
havior (pp. 171–180). New York: Springer-Verlag.
MacGeorge, E. L., Graves, A. R., Feng, B., Gillihan, S. J., & Burleson, B. R. (2004). The
myth of gender cultures: Similarities outweigh differences in men’s and women’s provi
sion of and responses to supportive communication. Sex Roles, 50, 143–175.
MacKay, D. G. (1980). Psychology, prescriptive grammar, and the pronoun problem. Amer
ican Psychologist, 35, 444–449.
Malcomson, K. A., & Sinclair, L. (2007). The Ms. stereotype revisited: Implicit and explicit
facets. Psychology of Women Quarterly, 31, 305–310.
Maltz, D. N., & Borker, R. (1982). A cultural approach to male-female miscommunication.

In J. J. Gumperz (Ed.), Language and social identity (pp. 195–216). Cambridge, UK: Cam
bridge University Press.
Martyna, W. (1980). Beyond the “he/man” approach: The case for nonsexist language.
Signs, 5, 482–493.
Mast, M. S. (2002). Dominance as expressed and inferred through speaking time: A meta-
analysis. Human Communication Research, 28, 420–450.
Mehl, M. R., Vazire, S., Ramirez-Esparaza, N., Slatcher, R. B., & Pennebaker, J. W. (2007).
Are women really more talkative than men? Science, 317, 82.
Miller, P. M., Danaher, D. L., & Forbes, D. (1986). Sex-related strategies for coping with
interpersonal conflict in children aged five and seven. Developmental Psychology, 22,
543–548.
Neff, L. A., & Karney, B. R. (2005). Gender differences in social support: A question of
skill or responsiveness? Journal of Personality and Social Psychology, 88, 79–90.
Newman, M. L., Groom, C. J., Handleman, L. D., & Pennebaker, J. W. (2008). Gender dif
ferences in language use: An analysis of 14,000 text samples. Discourse Processes, 45,
211–236.
Page 34 of 37

Oransky, M., & Fisher, C. (2009). The development and validation of the Meanings of Ado
lescent Masculinity Scale. Psychology of Men and Masculinity, 10, 57–72.
Parks, J. B., & Roberton, M. A. (2004). Attitudes toward women mediate the gender
(p. 81)
effect on attitudes toward sexist language. Psychology of Women Quarterly, 28, 233–239.
Pasch, L. A., & Bradbury, T. N. (1998). Social support, conflict, and the development of
marital dysfunction. Journal of Consulting & Clinical Psychology, 66, 219–230.
Penman, R. A. (1980). Communication processes and relationships. London: Academic

Press.
Pepler, D. J., Craig, W. M., Connolly, J. A., Yuile, A., McMaster, L., & Jiang, D. (2006). A de
velopmental perspective on bullying. Aggressive Behavior, 32, 376–384.
Prewitt-Freilino, J. L., Caswell, T. A., & Laakso, E. K. (2012). The gendering of language: A
comparison of gender equality in countries with gendered, natural gender, and gender
less languages. Sex Roles, 66, 268–281.
Rapacon, S. (2011, September 2). Five choices for changing your name after marriage.
Kiplinger.com. Retrieved from http://www.kiplinger.com/columns/starting/archive/
5-choices-for-changing-your-name-after-marriage.html
Ridgeway, C. L., & Smith-Lovin, L. (1999). The gender system and interaction. Annual Re
view of Sociology, 25, 191–216.
Robnett, R. D., & Leaper, C. (2013). “Girls don’t propose! Ew.”: A mixed-methods exami
nation of marriage tradition preferences and benevolent sexism in emerging adults. Jour
nal of Adolescent Research, 28, 96–121.
Rock, D. L. (1981). The confounding of two self-report assertion measures with the ten
dency to give socially desirable responses in self-description. Journal of Consulting and
Clinical Psychology, 49, 743–744.
Rose, A. J., Carlson, W., & Waller, E. M. (2007). Prospective associations of co-rumination
with friendship and emotional adjustment: Considering the socioemotional trade-offs of
co-rumination. Developmental Psychology, 43, 1019–1031.
Rudman, L. A., & Glick, P. (2008). The social psychology of gender: How power and inti
macy shape gender relations. New York: Guilford.
Ryan, K. M., & Kanjorski, J. (1998). The enjoyment of sexist humor, rape attitudes, and re
lationship aggression in college students. Sex Roles, 38, 743–756.
Scheuble, L. K., Johnson, D. R., & Johnson, K. M. (2012). Marital name changing attitudes
and plans of college students: Comparing change over time and across regions. Sex
Roles, 66, 282–292.
Page 35 of 37

Schouten, A. P., Valkenburg, P. M., & Peter, J. (2007). Precursors and underlying process
es of adolescents’ online self-disclosure: Developing and testing an “Internet-attribute-
perception” model. Media Psychology, 10, 292–314.
Schwarzer, R., & Gutiérrez-Doña, B. (2005). More spousal support for men than for
women: A comparison of sources and types of support. Sex Roles, 52, 523–532.
Searle, J. R. (1969). Speech acts. New York: Oxford University Press.
Sprecher, S., & Hendrick, S. S. (2004). Self-disclosure in intimate relationships: Associa

tions with individual and relationship characteristics over time. Journal of Social & Clini
cal Psychology, 23, 857–877.
Spence, J. T., & Helmreich, R. (1978). Masculinity and femininity: Their psychological di
mensions, correlates, and antecedents. Austin: University of Texas Press.
Stahlberg, D., Braun, F., Irmen, L., & Sczesny, S. (2007). Representations of the sexes in
language. In K. Fiedler (Ed.), Social communication (pp. 163–187). New York: Psychology
Press.
Steil, J. M. (2000). Contemporary marriage: Still an unequal partnership. In C. Hendrick

& S. S. Hendrick (Eds.), Close relationships: A sourcebook (pp. 124–136). Thousand Oaks,
CA: Sage.
Strough, J., & Berg, C. A. (2000). Goals as a mediator of gender differences in high-affilia
tion dyadic conversations. Developmental Psychology, 36, 117–125.
Stout, J. G., & Dasgupta, N. (2011). When he doesn’t mean you: Gender-exclusive lan
guage as ostracism. Personality and Social Psychology Bulletin, 37, 757–769.
Swim, J. K., Mallett, R., & Stangor, C. (2004). Understanding subtle sexism: Detection and
use of sexist language. Sex Roles, 51, 117–128.
Tajfel, H., & Turner, J. C. (1979). An integrative theory of intergroup conflict. In W. G.

Austin & S. Worchel (Eds.), The social psychology of intergroup relations (pp. 94–109).
Monterey, CA: Brooks-Cole.
Tannen, D. (1990). You just don’t understand: Women and men in conversation. New York:
Morrow.
Tannen, D. (1994). Gender and discourse. New York: Oxford.
Tenenbaum, H. R., Ford, S., & Alkhedairy, B. (2011). Telling stories: Gender difference sin
peers’ emotion talk and communication style. British Journal of Developmental Psycholo
gy, 29, 707–721.
Thomas, C. A., & Esses, V. M. (2004). Individual differences in reactions to sexist humor.
Group Processes & Intergroup Relations, 7, 89–100.
Page 36 of 37

Thorne, B. (1993). Gender play: Girls and boys in school. New Brunswick, NJ: Rutgers
University Press.
Thorne, B., & Henley, N. M. (1975). Language and sex: Difference and dominance. Row
ley, MA: Newbury House.
Timmerman, L. M. (2002). Comparing the production of power in language on the basis of

sex. In M. Allen, R. W. Preiss, B. M. Gayle, & N. A. Burrell (Eds.), Interpersonal communi
cation research: Advances through meta-analysis (pp. 73–88). Mahwah, NJ: Erlbaum.
Twenge, J. M. (1997). “Mrs. His Name”: Women’s preferences for married names. Psy
chology of Women Quarterly, 21, 417–429.
Valkenburg, P. M., & Peter, J. (2009). Social consequences of the Internet for adolescents:
A decade of research. Current Directions in Psychological Science, 18, 1–5.
Wasserman, B. D., & Weseley, A. J. (2009). Qué? Quoi? Do languages with grammatical
gender promote sexist attitudes? Sex Roles, 61, 634–643.
West, C., & Zimmerman, D. H. (1991). In J. Lorber & S. A. Farrell (Eds.), The social con
struction of gender (pp. 13–37). Thousand Oaks, CA: Sage.
Willness, C. R., Steel, P., & Lee, K. (2007). A meta-analysis of the antecedents and conse
quences of workplace sexual harassment. Personnel Psychology, 60, 127–162.
Wood, J. T. (2013). Gendered lives: Communication, gender, and culture (10th ed.).
Boston, MA: Wadsworth.
Wood, W., & Karten, S. J. (1986). Sex differences in interaction style as a product of per
ceived sex differences in competence. Journal of Personality and Social Psychology, 50,
341–347.
Xu, Y., & Burleson, B. R. (2001). Effects of sex, culture, and support type on perceptions
of spousal social support: An assessment of the “support gap” hypothesis in early mar
riage. Human Communication Research, 27, 535–566.
Campbell Leaper
Campbell Leaper is Professor of Psychology, University of California at Santa Cruz.
Page 37 of 37

Working Together
Working Together
Herbert H. Clark and Tania Henetz

Online Publication Date: Aug 2014 DOI: 10.1093/oxfordhb/9780199838639.013.019
People find themselves working, or collaborating, with others at every turn in their daily
affairs—in preparing meals, playing games, transacting business, gossiping, and even
walking down crowded streets. In joint activities like these, people have shared goals
they can accomplish only by establishing agreements or joint commitments about who is
to do what when. Evidence shows that people reach these agreements in pairs of actions
in which one person proposes a joint action or position and another agrees to it. People
use these pairs both alone and in combination to reach agreements on the goals, sub
goals, and timing of their joint activities as they go along. Thus, it takes coordination to
work together, but that in itself takes communication.
Keywords: Collaboration, joint commitments, agreements, working together, communication
Almost everywhere we go, we find ourselves working with others to get things done. At
home, we join other members of the family in cooking, cleaning, watching television, and
playing games. At work, we join colleagues in planning projects, making coffee, and gos
siping. Outside the home and workplace, we debate with friends, take children to sports
events, play friends in tennis, and bargain with clerks. Even when our goals don’t require
other people, we are usually in the presence of others—in offices, corridors, streets,
sports arenas—and take them into account. In all these circumstances, we do more than
take note of others. We collaborate with them (collaborate, etymologically, means “work
with”) to accomplish our goals (Clark, 1996, 2006).
How do we do this? The fundamental way is through communication—for example, by

speaking. But to restrict ourselves to speech would distort the way we actually work to
gether. We also communicate by pointing at things, placing objects on counters, standing
in queues, gesturing with our hands, arms, faces, and bodies, and simply making manifest
to others what we are doing. In the presence of others, we cannot not communicate (Wat
zlawick, Bavelas, & Jackson, 1967).
Page 1 of 30

Working Together
To work together is to engage in joint activities, and these require the coordination of the
people engaged in them. We begin by characterizing joint activities and how the partici
pants coordinate with each other in them. We then take up the techniques people use in
coordinating on larger enterprises. Finally, we examine the way people coordinate on
communication itself. In face-to-face conversation, we speak to others while they listen;
we gesture to them while they watch; and we point at things while they register what we
are pointing at. Each of these is also a joint action, and they, too, take coordination.
Joint Activities
What is a joint activity? Compare a person doing something alone, a solo activity, with
two people doing roughly the same thing but together, a joint activity:
• Juggling. One clown juggling three balls alone versus two clowns juggling six balls
between them.
(p. 86) • Dancing. A ballerina dancing solo versus a couple waltzing.
• Playing cards. A person playing solitaire versus two people playing gin rummy.
• Playing the piano. A pianist playing a Mozart sonata versus two pianists playing a
Mozart duet.
• Passing through a door. One person versus a couple passing through a doorway.
• TV stand. One versus two people assembling a TV stand.
In solo activities, individual people work autonomously—at places, times, and speeds of
their own choosing. Not so in joint activities. A joint activity is one in which: (a) there are
two or more participants, (b) the participants have a shared set of goals, and (c) the ac
tions of each participant in achieving those goals depend in part on the actions of the oth
er participants (Clark, 1996). Participants in joint activities have to coordinate with each
other on the places, times, and speeds of their individual actions.
How do people coordinate with each other in such activities? One influential account was
proposed by philosopher Paul Grice (1975) for conversations or talk exchanges. As Grice
put it: “Our talk exchanges…are characteristically, to some degree at least, cooperative
efforts; and each participant recognizes in them, to some extent, a common purpose or
set of purposes, or at least a mutually accepted direction (p. 45).” As a result, partici
pants in talk exchanges expect the current speaker to adhere to what Grice called the co
operative principle:
Make your conversational contribution such as is required, at the stage at which it

occurs, by the accepted purpose or direction of the talk exchange in which you are
engaged (p. 45).
In Grice’s view, people take for granted that “contributions” to a talk exchange are to be
interpreted against the “accepted purpose or direction of the talk exchange.” Grice’s ac
count is remarkable for its insights into talk exchanges and how people contribute to
Page 2 of 30

Working Together
them. But what about joint activities in general, such as two people juggling, waltzing,
playing gin rummy, playing a duet, passing through a doorway, and building a TV stand?
None of these is a talk exchange, so how would Grice’s cooperative principle apply? In
the framework we describe, juggling, dancing, and assembling a TV stand together are
basic joint activities. In each of these, the participants coordinate with each other by
means of communicative exchanges, which are a close kin of Grice’s talk exchanges. But
communicative exchanges are conceptually distinct from the joint activities they coordi
nate.
Varieties of Joint Activities
Joint activities fall largely into two classes. One class is joint physical activities, which
proceed mainly by a series of joint physical actions. In sports, we find two people playing
tennis, ten women playing basketball, and twenty-two men playing soccer. In the arts, we
find two dancers doing a tango, a jazz trio playing a Dave Brubeck classic, eight singers
performing a Brahms motet, and an opera company putting on The Marriage of Figaro. In
everyday circumstances, two people pass each other in a corridor, a dozen people jostle
to get through an exit, and scores of drivers share a four-lane highway.
The second class of activities is joint informational activities, ones that proceed by the
participants establishing a series of joint positions about states of the world—“That is
mine,” “She is called Margaret,” “I’ll give you $10 for that,” and “I want to go home now.”
By joint position, we mean an intellectual position or proposition that the participants are
jointly committed to holding or assuming for current purposes.
Joint informational activities are also common. Examples: two people gossiping on the
telephone; a person ordering a book at a bookstore; senators negotiating a piece of legis
lation; and judge, prosecutors, defense lawyers, witnesses, and defendants in court carry
ing out a trial. In these, participants do most of their work with language, although in oth
er informational activities, they rely on other forms of communication. At ticket counters,
people signal that they want to buy a ticket by standing in a queue. In bookstores, cus
tomers signal that they wish to buy books by placing them on checkout counters (Clark,
2003).
Most joint activities are a mix of physical and informational actions. In soccer, the players
coordinate physically, whereas the referee coordinates informationally (with whistles and
raised cards). In conversation, people coordinate informationally with talk and gestures
but physically with the arrangement of their bodies (Clark, 2003, 2005; Kendon, 1990).
Joint physical and informational activities seem radically different, but the techniques by
which people coordinate in them are remarkably similar.
Partitioning Joint Activities
What, then, is the difference between engaging in joint versus solo activities? One day at
home, Josephine decided to play a Mozart piano sonata, (p. 87) so she sat down at her pi
ano, started playing, and continued until she reached the end. Later that day, John
Page 3 of 30

Working Together
stopped by, and the two of them decided to play a Mozart piano duet. This time, the two
of them together sat down at her piano, started playing, and continued until they reached
the end. From a recording of the two performances, it would be hard to tell the sonata
from the duet as single unified performances. So what was the difference?
It took certain actions to perform the duet that were not needed for the sonata. When
Josephine played alone, she herself chose what to play, when to start, what tempo to use,
and in what style to play each part. But when she and John played together, they had to
coordinate on what to play, when to start, what tempo to use, and in what style to play
each part. They used language to coordinate some of these features (“Let’s play the sec
ond movement”) but gestures and manner of playing to coordinate others (see Clark,
2005). The duet required the added activity of coordinating the players.
To make this concrete, compare one versus two people assembling a TV stand. In an un
published study by Julie Heiser and Barbara Tversky, singletons and pairs of Stanford
University students were brought into the lab and asked to assemble a TV stand from its
parts without written instructions. Their actions were captured in video recordings. It
was clear to everyone—singletons and pairs alike—that they couldn’t assemble the stand
without a plan, so they had to devise one. But how?
All of the participants quite spontaneously used a means-ends analysis of the task to guide
their actions (Newell & Simon, 1972). A means-ends analysis is one that begins with the
top goal, here, “assemble TV stand,” which is then decomposed recursively into subgoals
(such as “attach side 1 to the top piece”) that are to be worked on sequentially to reach
the top goal. Here is the hierarchy of goals, subgoals, and sub-sub-goals that emerged for
a singleton named Sam as he assembled the TV stand:
1. Assemble TV stand
1.1. Arrange parts
1.1.1. Put sides in one pile
1.1.2. Put screws, pegs in a second pile
1.1.3. Put wheels in a third pile
1.2. Assemble parts

1.2.1. Attach side 1 to top-piece
1.2.1.1. Insert pegs in side 1
1.2.1.2. Affix top piece to pegs in side 1
1.2.2. Attach side 2 to top-piece

1.2.3. Etc.
Sam began by arranging the parts into three piles (sides, pegs and screws, and wheels)
and went on to combine the parts bit by bit. He did not plan the entire hierarchy from the
outset. With the means-ends analysis, he started with the top goal and initial subgoals
and filled in the other subgoals as he went along.
Page 4 of 30

Working Together
Next, consider Ann and Burton as they assembled the TV stand together. They, too, did a
spontaneous means-ends analysis and proceeded through its hierarchy of goals and sub
goals in sequence. In one 20-second segment, they carried out these five pairs of actions:
Ann’s action Burton’s action
1 A gets cross-piece B holds side-piece
2 A holds cross-piece B inserts peg
3 A affixes cross-piece B holds side-piece
4 A inserts peg B holds side-, cross-piece
5 A affixes side-piece B holds side-, cross-piece
Each line represents a joint action by the two of them. In line 3, for example, Burton held
a sidepiece steady while Ann affixed the crosspiece onto it. Each of the joint actions ful
filled a subgoal of a joint action at a higher level, “attach both side-pieces to cross-piece,”
which was itself a subgoal of the top goal, “assemble TV stand.” In short, Ann and
Burton’s actions emerged as a hierarchy of actions very much as Sam’s did.
What was added by Ann and Burton was communication. Whereas Sam did the means-
ends analysis on his own, Ann and Burton did theirs together, reaching agreement on its
goals, subgoals, and timing. Here is how they did it for the 20-second segment just de
scribed:
Page 5 of 30

Working Together
1 Ann Should we put this in, this, this little like kinda cross
bar, like the T? like the I bar?
2 Bur Yeah, we can do that.

ton
3 Ann So, you wanna stick the screws in. Or wait is, is, are
these these things, or?
4 Bur That’s these things I bet. Because there’s no screws.

ton
5 Ann Yeah, you’re right. Yeah, probably. If they’ll stay in.
6 Bur I don’t know how they’ll stay in but

ton
7 Ann Right there.
8 Bur Is this one big enough?

ton
9 Ann Oh [untranscribable] I guess cause like there’s no oth

er side for it to come out.
10 Bur M-hm.
ton
11 [8.15 sec]
12 Bur Now let’s do this one

ton
13 Ann Okay
In lines 1 and 2, Ann and Burton agreed to attach the crosspiece to the top piece
(p. 88)
first, and in lines 3 through 6, they reached agreement on the first of its subgoals. And so
it went, with successive agreements about what pieces to connect when, how to orient
each piece, who was to hold, and who was to attach. Sam did all he needed to do without
uttering a word.
Page 6 of 30

Working Together
As this example illustrates, joint activities can in general be partitioned into two sets of
actions:
1. The basic joint activity is what the participants are basically trying to do. What
Ann and Burton were “really doing,” in their words, was assembling the TV stand.
Their basic joint activity consisted of the actions and positions that were needed for
their basic goal, “assembling the TV stand.”
2. The coordinating joint actions are what the participants do to coordinate their ac
tions in the basic activity. These consist of communicative acts about the basic activi
ty, as in the talk between Ann and Burton.
Assembling the TV stand took both sets of actions. The first set was the assembly proper,
which Sam did on his own, but Ann and Burton did together. The second set was needed
by Ann and Burton, but not by Sam, to coordinate their individual actions within the as
sembly proper.
Physical and informational activities contrast mainly in their basic activities. When ten
people play basketball, the basic activity proceeds largely as a series of joint physical ac
tions—players passing to each other, dribbling around each other, blocking each other,
and so on. When two people plan a vacation, in contrast, the basic joint activity proceeds
largely as a series by joint positions—establishing where to go, when, and for how long.
Despite these differences, participants coordinate with each other in both activities by
means of speech, gestures, and other signals. The basic activities may be radically differ
ent, but the coordinating joint actions are very much alike.
It takes coordination, therefore, to carry out joint activities, and it takes communicative
acts to achieve that coordination. But to agree on a joint course of action is really to es
tablish a joint commitment to that course of action. How is that done?
Agreeing to Work Together

In assembling the TV stand, Ann and Burton agreed on each goal and subgoal as they
went along. Neither could have reached these agreements on their own. Why? Because
an agreement is a joint commitment, and these can only be established by all the parties
working together (see Clark, 1996; Gilbert, 1990). It takes actions from each of the par
ties to make an agreement, so, with two parties, it takes at least one action from each
party. This leads to a remarkable prediction: spontaneous discourse should be built on
pairs of actions. And it is. Almost nothing in discourse gets done without them.
Projective Pairs
The most basic form of agreement is for one person to initiate the process and the other
to complete it. In assembling the TV stand, Ann and Burton agreed to one of their sub
goals in this way:
Page 7 of 30

Working Together
1 Ann Should we put this in, this, this little like kinda cross
bar, like the T? like the I bar?
2 Bur Yeah we can do that.

ton
In line 1, Ann proposed that the two of them put in the I bar, and in line 2, Burton took up
her proposal and agreed to it. The result was a joint commitment to put in the I bar. Even
though Ann initiated the process and Burton completed it, what they reached was a joint
agreement. Schematically:
Part 1: A proposes to B a joint action or position for the two of them

Part 2: B takes up A’s proposal and agrees to it.
This is called a projective pair (Clark, 2004; Clark & Krych, 2004). Projective pairs are
versatile in both form and function. As for form, the first part may be either spoken or
gestural, and the second part, too, may be either spoken or gestural. When both parts are
spoken, the exchange is called an adjacency pair (Schegloff, 2007; Schegloff & Sacks,
1973). Adjacency pairs (a) consist of two spoken turns in sequence, a first pair part and a
second pair part; (b) with each part produced by a different speaker; and (c), once one
party has produced the first part, it is conditionally relevant for another (p. 89) party to
produce the second part. Projective pairs are defined by the same properties, except that
each part may be either spoken or gestural.
Projective pairs with gestural parts are common but, of course, only when people can see
each other. Here is an example from assembling the TV stand:
1 Burton [extends hand with peg]
2 Ann [grasps peg]
In line 1, Burton proposes to Ann that she grasp the peg, and in line 2, she agrees to his
proposal by grasping it. Once Burton had produced the first part (line 1), it was condition
ally relevant for Ann to produce the second part (line 2). If Ann hadn’t grasped the peg,
Burton would have interpreted her as refusing his offer. For everyday examples, imagine
a hostess handing a guest a drink, a poker player handing the dealer several cards, or a
professor handing a student a syllabus.
Projective pairs are versatile in function, too. The prototype is the question–answer pair,
as in this example (LL 2.7.1232)1:
Alice: w- where are they going then, .
Brian: to Spain
Page 8 of 30

Working Together
In line 1, Alice asks Brian a question, making it conditionally relevant for him to provide
an answer in the next turn, which he does. Note that answers can be either spoken or
gestural, as illustrated by this exchange:
1 Professor [to a classroom of students] Who’d like a syllabus?
2 Student [raises her hand]
In answering the question in line 1, students could either say “I would” or raise their
hands. So, although a question projects an answer, the answer may be spoken, gestural,
or both. When asked “Is Helen around?” one could answer by saying “yes,” by nodding, or
by doing both (see Clark, 2012; Stivers et al., 2009).
Projective pairs serve many other functions as well. Here are a few examples from a sin
gle telephone call between Jane and Kate (LL 8.3d.230):2
Pair parts Example
1. summons Jane: [causes Kate’s telephone to ring]
2. answer Kate: Miss Pink’s office -
1. greeting Kate hello,
2. greeting Jane: hello,
1. request Jane: could you give her a message *for me,*
2. promise Kate: *certainly,*
1. gratitude Kate: thank you very much indeed,
2. reply Jane right,
1. farewell Kate: bye bye,
2. farewell Jane bye,
Indeed, every turn in Jane and Kate’s call was either the first or second part of such pro
jective pairs, and this is characteristic of dialogue.
Projective pairs are also used in reaching more complicated joint commitments. Three
ways they are used are in chaining, embedding, and presequencing.
Page 9 of 30

Working Together
Chaining
It is common in dialogues to link projective pairs together in chains. Linking consists of

(1) producing one projective pair, (2) construing its second part simultaneously as the
first part of a next projective pair, and (3) producing the second part of the next pair. In
the following example, Abby and Bess are on the telephone (SW 2382-72):
1 Abby Could you hold the phone for one second.
2 Bess Sure.
3 Abby Thanks.
4 [Pause]
5 Bess Okay, I’m back.
In lines 1 and 2, Abby makes a request of Betty, which Betty agrees to. These two lines
constitute a request-compliance pair. But Abby construes Betty’s “sure” in line 2 as a fa
vor, so she takes up Betty’s “sure” with “thanks.” Lines 2 and 3 constitute a favor-grati
tude pair. Betty’s “sure” in line 2 is construed both as the second part of one pair and as
the first part of a next pair. Conversations are littered with chains. Here are a few three-
part chains with stripped down examples for each:
First pair Part 1 → Part 2
Second pair Part 1 → Part 2
Question-answer-eval A: What’s pi? B: 3.1416 A: Correct

uation
Offer-agreement-com A: Want a B: Yes, A: [hands B

pliance beer? please a beer]
Favor-thanks-acknowl A: [hands B a B: Thanks A: No prob

edgment beer] lem.
Request-compliance- A: Hold on a B: Sure A: Thanks

thanks sec
Links can even emerge from parts of turns, as in this example from a lunch
(p. 90)
counter (Merritt, 1976, p. 324):
Page 10 of 30

Working Together
1 Customer Do you have the pecan Danish today?
2 Server Yes we do.
3 Would you like one of those?
4 Customer Yes, please.
5 Server [turns to get]
Lines 1 and 2 are a simple question-answer pair. But once the server has answered the
question, she goes on in line 3 to make an offer “Would you like one of those?” Her offer
initiates an offer-acceptance pair (lines 3 and 4), which is then linked to a request-compli
ance pair (lines 4 and 5). People use linked projective pairs for all manner of purposes.
Embedding
Another way to create complex agreements is to embed one projective pair inside anoth
er. In this example, the waitress is getting a customer’s order by means of a question-an
swer pair (Merritt, 1976):
1 Waitress What’ll you have girls?
2 Cus I’ll have a bowl of clam chowder and a salad with

tomer Russian dressing.
As it happened, this customer didn’t have enough information to answer the waitress’s
question, so what she actually did was first initiate the question-answer pair in boldface:
Page 11 of 30

Working Together
1 Waitress What’ll you have girls?
2 Customer What’s the soup

of the day?
3 Waitress Clam chowder
2 Customer I’ll have a bowl of clam chowder and a salad with Russian dressing.
Page 12 of 30

Working Together
Embedded pairs like this are called side sequences (Jefferson, 1972); when they are em
bedded within an adjacency pair, as here, they are also called insert sequences (Schegloff,
2007).
Embedding is remarkably useful in everyday conversation. People initiate side sequences

whenever they need more information to go on, like the customer asking for the soup of
the day. And embedding can be recursive, as in this example from a notions counter (Mer
ritt, 1976, p. 331):
Page 13 of 30

Working Together
1 Server Next
2 Customer Roast beef on rye
3 Server Mustard or mayonnaise?
4 Customer Excuse me?
5 Server What?
6 Customer Excuse me,

I didn’t
hear what
you said?
7 Server Do you want mustard or mayonnaise?
8 Customer Mustard please
9 Server [Provides sandwich with mustard]
Page 14 of 30

Working Together
The server initiated one side sequence in line 3; within that, the customer initiated anoth
er in line 4; and within that, the server initiated yet another in line 5. Recursion of side
sequences may not be common, but it is available when people need it.
Presequencing
The final strategy, called presequencing, is illustrated in an exchange between Ann and
Betty about a belt Ann wants to offer Betty (LL 7.1d.1320):
1 Ann oh there’s one thing I wanted to ask you,
2 Bet m-hm, -
ty
3– [Ann and Betty take ten turns as Ann describes a type

12 of belt]
13 Ann would you like one.
14 Bet oh I’d love one Ann –

ty
From the very start, Ann realized that she couldn’t offer Betty the belt without de
(p. 91)
scribing it, and that would take time. So she initiated the adjacency pair in lines 1 and 2
with what Schegloff (1980) called a prequestion: she asked Betty whether she could ask
Betty a question. Once Betty agreed (“m-hm” in line 2), Ann took the time she needed to
describe the belt for Betty (lines 3 through 13). Only then did she ask the question she
had set up with her prequestion, “Would you like one?” The structure is this:
A. Presequence (lines 1 and 2)

B. Preliminaries to target sequence (lines 3 through 12)
C. Target sequence (lines 13 and 14)
Presequences can be initiated not only by prequestions, but by preannouncements, pre

requests, preinvitations, prenarratives, and others, as in these stripped down examples:
Page 15 of 30

Working Together
Type of presequence Part 1 → Part 2
Preannouncement A. Do you know what? B. No, what?
Preinvitation A. Doing anything tonight? B. No, why?
Prerequest A. Do you have size C bat B. Yes, we

teries? do.
Prenarrative A. Did I tell you about my B. No, you

interview? didn’t.
What makes presequences so useful is that person A gets person B to agree to allow A to
do something that requires extra time, space, information, or other considerations. With
the prequestion, Ann was given space to establish preliminaries to the target question.
With prenarratives, A is given space to tell a complete narrative. With prerequests, A gets
information that is needed for making a request, as here:
1 Customer Do you have uh size C flashlight batteries?
2 Server Yes sir.
3 Customer I’ll have four please
4 Server [turns to get]
Presequences are also useful when B’s consent isn’t a trivial matter. Here is a preques
tion put to Monica Lewinsky during an interview before a grand jury prior to President
Bill Clinton’s impeachment hearings in 1998:
Page 16 of 30

Working Together
1 Juror Okay. I have a question that’s a bit on the deli

cate side.
2 Lewin Okay.
sky
3 Juror But this is just something that I need to know.
4 Lewin Sure.
sky
5 Juror Did you and the President ever engage in sexual re
lations using cigars?
6 Lewin Yes, we did.

sky
Hierarchies in Dialogue
To repeat, the reason that people communicate is to coordinate their actions and posi
tions in basic joint activities (such as building a TV stand, negotiating a contract, or play
ing basketball). People’s immediate goal is to establish agreements about what each is to
do when, where, and how, which they do with projective pairs—alone, or in chains, or em
bedded, or in presequences. But are these techniques all they need to carry out those ac
tivities? The answer is no.
To make this point, let us start with a complete joint activity, a call to Directory Enquiries
at British Telecom in Cambridge England, in 1974, made by a man we will call Charles
(from Clark & Schaefer, 1987):
Page 17 of 30

Working Together
1 Charles [rings telephone at “Directory Enquiries”]
2 Operator Directory Enquiries,
3 for which town, please?
4 Charles In Cambridge
5 Operator What’s the name of the people?
6 Charles It’s the Shanghai Restaurant, it’s not in my direc

tory, but I know it exists
7 Operator It’s Cambridge 67924
8 Charles 67924
9 Operator That’s right
10 Charles Thank you very much
11 Operator Thank you,
12 good bye
13 Opera [hang up telephones to break contact]

tor,
Charles
(p. 92) Charles and the operator traversed this exchange with remarkable efficiency—as if
they knew exactly what to expect next. But how did they know what to expect?
The answer at first seems obvious: Charles and the operator relied on the structure creat
ed by the projective pairs, presequences, and linked sequences that made up this ex
change. They opened with a summons-response pair, which they construed as a pre-en
quiry plus its response:
1 Sum Charles [rings telephone at “Directory En

mons quiries”]
2 Response Opera Directory Enquiries

tor
Page 18 of 30

Working Together
As a result, the operator, without other preliminaries, was able to initiate the question-an
swer pairs in lines 3 and 4 and lines 5 and 6. And she was able to tell Charles the tele
phone number in lines 7, 8, and 9 in two linked adjacency pairs: an offer plus uptake in
lines 7 and 8 and a read-back plus verification in lines 8 and 9. Once that was complete,
and again without other preliminaries, Charles was able to initiate the closing of the con
versation with a gratitude exchange (lines 10 and 11), and the operator was able to initi
ate a good-bye exchange (in line 12)—although Charles didn’t happen to take that up.
And, finally, the two of them broke contact (see Clark & French, 1981).
But the course of this dialogue—and its efficiency—cannot be accounted for by projective
pairs, presequences, and linked sequences alone. The course was ultimately determined
by the joint activity Charles and the operator were engaged in—the exchange of a name
and address for a telephone number. At the start, the two of them presupposed the top
goal (“exchanging the name and address for a number”), which led to a hierarchy of goals
and subgoals:
1 C&O enter joint activity of a directory enquiry (lines 1–2)

1.1 C&O exchange name/address for number (lines 3–8)
1.1.1 C&O exchange information needed by O (lines 3–6)
1.1.1.1 C&O exchange name of town (lines 3–4)
1.1.1.2 C&O exchange name of people (lines 5–6)
1.1.2 C&O exchange number requested by C (lines 7–9)

1.1.2.1 O tells C the number (line 7)
1.1.2.2 C&O verify the number (lines 8–9)
1.2 C&O exchange gratitude for O’s assistance (lines 10–11)
2 C&O exit from joint activity (lines 12–13)
It was the emergent hierarchy that enabled them to establish where they were at each
moment in their call. It was the top goal, for example, that enabled them to construe
“[rings telephone]” followed by “Directory Enquiries” as a pre-enquiry. Although they es
tablished each agreement locally, they established the hierarchy of agreements globally.
To summarize, joint activities have a goal hierarchy that emerges in time. When people
work with others, they have to communicate to collaborate on these goals. The basic unit
of collaboration is the projective pair. Through the artful use of chaining, embedding, and
presequencing, people deploy simple projective pairs to create complex agreements or
joint commitments, allowing them to carry out goals that are often quite complex.
Managing Communication
To communicate is, etymologically, to “make common,” to establish information as com
mon or shared. But how do people make certain that the right information becomes truly
common or shared?
Page 19 of 30

Working Together
One approach has focused on failures in communication (e.g., Schegloff, Jefferson, &
Sacks, 1977). The idea is that dialogues have “intrinsic troubles,” as when someone mis
hears, misspeaks, or becomes distracted. So participants must monitor their dialogues for
such troubles and, when they find them, get them repaired.
An alternative approach has focused, instead, on success in communication (Clark, 1996;

Clark & Schaefer, 1989). The idea here is that people monitor (p. 93) their dialogues for
success as well as failure. They try to establish, as they go along, the mutual belief that
they have understood each other well enough for current purposes. It takes more than
the repair of problems to do that. It also takes displays of success. The process of estab
lishing these beliefs is called grounding (Clark, 1996; Clark & Brennan, 1991; Clark &
Schaefer, 1989; Clark & Wilkes-Gibbs, 1986).
The idea is that people in conversation keep track of two lines of action. Their primary
goal is to reach agreements on their official business, and these make up the primary line
of their dialogue. When Burton said, “Now let’s do this one,” and Ann replied “Okay,” they
agreed on the next step in the official business of assembling the TV stand. At the same
time, people in dialogue have the on-going or collateral goal of making sure they have
succeeded in each step of their communication. Ann and Burton made sure, before they
went on, that Ann had understood Burton’s request, “Now let’s do this one,” well enough
for current purposes. Establishing that an utterance has been understood is grounding,
and it is this process that creates the second or collateral line of action in dialogue. But
just what is grounding, and how does it manifest itself?
Grounding Based on Positive Evidence
The goal in grounding an utterance is the agreement that it has been understood well
enough for current purposes. Reaching that agreement should take at least two actions,
one from each party to the agreement. But from a traditional point of view, that conclu
sion seems like nonsense. Consider Burton’s assertion in the following exchange:
1 Burton It’s starting to look like a TV stand
2 Ann He-he he-he. It’s starting to, yeah.
For the traditionalist, once Burton has produced line 1, he is done. He has made an asser
tion, and no second action is needed. But what if Ann hadn’t responded? It would be rea
sonable to conclude that she might not have heard or understood him. And, if she hadn’t,
what Burton had produced in line 1 was for nothing. To tell a person something entails
getting that person to understand that something. Without Ann’s response in line 2, Bur
ton couldn’t have been certain he had told her anything. So even the simplest assertion
requires two actions: the initial utterance (as in line 1), and a response with positive evi
dence of understanding (as in line 2).
Page 20 of 30

Working Together
Acknowledgments
The simplest type of positive evidence is acknowledgments, such as “yeah”, “yes,” “m-
hm,” “uh-huh,” or nodding (Jefferson, 1984, 2002; Schegloff, 1982). In the following, April
is relating a brief anecdote to Brigit (MaF S3.15:27):
1 April so like . my little brother I think was in . like. fourth

grade
2 Brig m-hm
it
3 April when we got . the dog so that’s fine
4 Brig ah
it
5 April an- she’s small but um.. yeah our golden retriever was
like so nice but .
6 Brig oh
it
7 April he was just like . too much work
8 Brig yeah
it
9 April so we gave him away
In the course of April’s anecdote, Brigit introduced four brief responses—“m-hm,” “ah,”
“oh,” and “yeah.” When Brigit said “m-hm” in line 2, she acknowledged April’s statement
in line 1. And with “yeah” in line 8, she acknowledged April’s statement in line 7. Ac
knowledgments tend to be brief, reduced in volume, and often in overlap with the
partner’s speech (Oreström, 1983).
But what are acknowledgments? Intuitively, they are claims of understanding-so-far.

When Brigit says “m-hm” in line 2, she is telling April, “Yes, I understand your utterance-
so-far well enough for current purposes.” Yes is a common feature of acknowledgments.
Although Brigit could have said, “go on” or “and…” asking April to continue, she used
forms of yes instead. Acknowledgments in other languages tend to be forms of yes as well:
ja in German and Dutch, oui in French, si in Spanish, da in Russian, vai in Greek, ken in
Hebrew, hai in Japanese, and shi and wei in Mandarin Chinese (Clark, 2012).
Page 21 of 30

Working Together
Assessments
A related type of positive evidence is assessments such as “oh,” “really,” “you don’t say,”
or “good God!” (Goodwin, 1986). When Brigit said “ah” in line 4 and “oh” in line 6, she
did more than acknowledge April’s last statement. She made an assessment of its con
tent. With “ah” she expressed satisfaction, and with “oh” mild surprise. With an assess
ment, (p. 94) a speaker presupposes understanding-so-far and makes a brief comment.
With both acknowledgments and assessments, then, agreeing on a common understand

ing takes two actions in sequence—two collateral actions. We will call this a closure pair.
Schematically:
Part A As I produce “[utterance]”, are you understanding this

1 well enough for current purposes?
Part B Up to this point, yes.

2
In this notation, roman print represents what A and B say explicitly, and italic print repre
sents what they mean implicitly. In line 1 of April and Brigit’s interchange, April is tacitly
asking Brigit “Are you understanding this so far well enough for current purposes?” and
in line 2, Brigit explicitly answers “m-hm,” but tacitly adds, “up to the moment I say ‘m-
hm.’”
Closure pairs are created by two people, A and B, working in collaboration. In some in
stances, A creates a space for B to add an acknowledgment or assessment. That makes
the closure pair a type of projective pair: the first part projects the second part as the
next action. In other instances, A leaves it up to B at what point to produce an acknowl
edgment or assessment.
Acknowledgments such as “m-hm,” “yeah,” and nodding seem so insignificant that it is

hard to see how they could influence dialogue, but they do. In one study (Matarazzo,
Saslow, Wiens, Weitman, & Allen, 1964), an interviewer engaged sixty policemen and fire
men in an employment interview and, for a third of each interview, he nodded slightly
throughout each of their responses. His nodding increased the length of their responses
by 50 percent. Apparently, the interviewer’s nods signaled an interest in what they were
saying and encouraged them to say more (see also Kita & Ide, 2007).
Assessments have an added effect. In one study (Bavelas, Coates, & Johnson, 2000), par
ticipants were each asked to tell another participant a story in which they had experi
enced a close call. Unbeknownst to the narrators, some of the listeners had been asked to
count, as they listened, all the narrator’s words that began with “t.” Although the dis
tracted listeners produced just as many acknowledgments as the attentive ones, they pro
duced far fewer assessments. Indeed, the distracted listeners produced almost none at
all. So, when they failed to say “good God!” or “wow!” at the mention of a close call, nar
rators were left wondering how well they were telling their stories. As a result, narrators
Page 22 of 30

Working Together
with distracted listeners found it more difficult to tell their stories (e.g., they would go on
and on) than did narrators with attentive listeners. As Bavelas and her colleagues put it,
the listeners were really co-narrators.
Projective Pairs
Probably the commonest type of positive evidence is implicit in the second parts of pro
jective pairs. Consider this question-answer pair plus follow-up (MaF S3.8:58):
1 Brigit anyway . what are your favorite movies?
2 April um . I love musicals
3 Brigit oh nice. I do too. what are your favorites?
Brigit’s question in line 1 projects an answer about the class of movies April likes best.
April’s answer in line 2, “I love musicals,” is positive evidence that she understood
Brigit’s question because it is, appropriately, about movies she likes best. Does Brigit ac
cept it as positive evidence? Yes, she does. By producing the assessment “Oh nice” in line
3, she completes the closure pair in lines 2 and 3. Second parts of projective pairs are
usually an excellent source of positive evidence for grounding.
Dealing with Negative Evidence
Grounding becomes more difficult when people run into troubles communicating. It is
easy for a conversation to go off track when a speaker misspeaks or cannot be heard, or
when a listener mishears or gets distracted. These troubles, left uncorrected, can lead to
bigger problems down the line. So, as part of grounding, people in conversation will want
to prevent these troubles or repair them as soon as circumstances allow. Suppose Amy
makes a mistake in an utterance addressed to Ben. To ground the utterance, the two of
them must: (1) reach agreement on Amy’s mistake and its correction and (2) reach agree
ment that Ben has understood Amy’s utterance as corrected well enough for current pur
poses. Negative evidence complicates the process by adding the first of these steps.
Retrospective Strategies
People regularly collaborate in repairing problems that have already occurred. As Sche
gloff et al. (1977) argued, participants have two preferences (p. 95) in these repairs. First,
speakers prefer to make their own repairs. Consider:
Dan we got my first cat . our first cat when I was in nursery (MaF
S4.6:07)
Dan’s replacement of “my first cat” with “our first cat” is a self-repair. Next consider:
Page 23 of 30

Working Together
1 Maggie you fancy it yourself do you? -
2 Julia what, the men’s doubles?
3 Maggie yeah,
4 Julia well more than the singles, yes, - (LL 7.3e.278)
When Julia didn’t understand Maggie’s “it,” she initiated the side sequence in lines 2 and
3 to clear up the trouble. “The men’s doubles” is Julia’s repair, not Maggie’s, making it an
other-repair. In conversation, self-repairs are strongly preferred over other-repairs.
The second preference is for speakers to initiate their own repairs. When Dan repaired
“my first cat” to “our first cat,” he did so without being prompted. It was a self-initiated
repair. For a contrast, consider the next example:
1 Roger well there’s no general agreement on it I should think,
2 Sam on what,.
3 Roger on uhm – on the uhm – the mixed up bits in the

play, the *uhm*
4 Sam *yes* (LL 3.5a.283)
In this case, Roger clarified the word “it” only after Sam asked for clarification (“on
what?”). Roger’s repair was other-initiated. In conversation, self-initiated repairs are
strongly preferred to other-initiated repairs.
These two preferences reflect a deep principle of communication: all else being equal,
people try to communicate with the minimum of joint or collaborative effort (Clark &
Wilkes-Gibbs, 1986). Other-repairs are, in general, very costly in effort: it took Maggie
and Julia two extra turns and five extra words to complete Maggie’s repair. The same
goes for other-initiated repairs: It took Roger and Sam two extra turns and fourteen
words to complete Roger’s repair. Self-repairs and self-initiated repairs incur much small
er costs. It took Dan only three extra words to make his self-initiated self-repair. Minimiz
ing joint effort favors both self-repairs and the self-initiation of repairs.
Prospective Strategies
When speakers anticipate problems in conversation, they often warn their partners about
the imminent problems because doing so makes it easier to ground what is said.
Page 24 of 30

Working Together
One prospective strategy is to annotate about-to-be-made repairs. Although the repairs

themselves look backward, the annotations look forward. Compare these three self-re
pairs:
Dan we got my first cat . our first cat when I was in nursery
(MaF S4.6:07)
Mike it wa- I mean it’s better than nothing I guess (MaF

S15.11:45)
Robert or do they only know about the-e practical, . excuse me

experimental aspects, of reading (LL 2.4.736)
Whereas Dan introduced his replacement (“our first cat”) without comment, Mike and
Robert annotated theirs. Mike added “I mean” to suggest that he was now going to say
what he meant, and Robert added “excuse me” to apologize for having to correct what he
said. Other common annotations include no, sorry, like, and you know. All of these are ad
dressed not to the official business of the conversations but to the wording or meaning of
what was being said. They belong to the collateral line of the conversation.
Another prospective strategy is to warn addressees of impending delays. When people

plan and produce utterances on the fly, they are often forced to introduce delays in speak
ing. They may have trouble retrieving a word, planning what to say, or keeping track of
too many thoughts at once (Garrett, 1982; Levelt, 1989). In monitoring their own speech,
speakers can apparently identify when a delay is imminent and tell addressees roughly
when it will occur.
The commonest way to signal delays in English is with the words uh and um (Clark & Fox
Tree, 2002). Fillers like these have long been associated with difficulties in production
(e.g., Boomer, 1965; Bortfeld, Leon, Bloom, & Schober et al., 2001; Christenfeld, 1994;
Goldman-Eisler, 1958; Levelt, 1989; Maclay & Osgood, 1959; Shriberg, 1996; Smith &
Clark, 1993). Recall April’s answer to “What are your favorite movies?”:
April: um . I love musicals
Apparently, April wasn’t ready to answer immediately, so she produced “um” and
(p. 96)
paused 325 msec before initiating “I love musicals.” Evidence shows that speakers use uh
and um to announce that they are initiating delays in speaking (Clark & Fox Tree, 2002).
They use uh for minor delays and um for major ones.
A final way to warn about delays is to prolong words. Take these two examples:
Page 25 of 30

Working Together
Tim an—d I don’t know what else do we do? (MaF S8.24:47)
Al I think that the—e [deep breath] uh the situation that you

Gore describe has gotten completely out of hand.
Although and would ordinarily take about 200 msec to pronounce, Tim prolonged it to
1,100 msec. Such prolongations signal ongoing delays (Clark & Fox Tree, 2002). Like
wise, when speakers pronounce words the and a to rhyme with bee and bay instead of
duh, they are telling their addressees that they are initiating a delay because of problems
in speaking (Fox Tree & Clark, 1997).
Is announcing an impending delay useful? The answer is yes. Listeners recognize that uh
and um and prolonged words signal imminent trouble, and they try to infer the source of
that trouble (Arnold, Kam, & Tanenhaus, 2007; Arnold, Tanenhaus, Altmann, & Fagnano,
2004; Barr & Seyfeddinipur, 2010; Brennan & Schober, 2001). For example, listeners as
sociate “thee-um” but not “thuh” with difficulty in referring to a new or difficult-to-name
object (Arnold et al., 2004, 2007). Listeners are also less surprised at hearing unpre
dictable words that are preceded by uh or um (Corley, MacGregor, & Donaldson, 2007).
Such is the value of prospective strategies.
Conclusion
Working together itself takes work. People cannot combine their individual actions into
something bigger—doing the tango, playing a piano duet, playing basketball, or talking—
without reaching agreements on each step of their joint activity. How do they reach these
agreements? By communicating with each other—with speech, gestures, and other ac
tions. But communication itself is also a joint activity, so it requires the same agreements.
We have to work together just to communicate about working together.
References
Arnold, J. E., Kam, C. L. H., & Tanenhaus, M. K. (2007). If you say thee uh you are de
scribing something hard: The on-line attribution of disfluency during reference compre
hension. Journal of Experimental Psychology: Learning, Memory, and Cognition, 33(5),
914–930.
Arnold, J. E., Tanenhaus, M. K., Altmann, R. J., & Fagnano, M. (2004). The old and thee,
uh, new. Psychological Science, 15(9), 578–582.
Barr, D. J., & Seyfeddinipur, M. (2010). The role of fillers in listener attributions for speak
er disfluency. Language and Cognitive Processes, 25(4), 441.
Bavelas, J. B., Coates, L., & Johnson, T. (2000). Listeners as co-narrators. Journal of Per
sonality and Social Psychology, 79(6), 941–952.
Page 26 of 30

Working Together
Boomer, D. S. (1965). Hesitation and grammatical encoding. Language & Speech, 8, 148–
158.
Bortfeld, H., Leon, S. D., Bloom, J. E., & Schober, M. F. (2001). Disfluency rates in conver
sation: Effects of age, relationship, topic, role, and gender. Language and Speech, 44(2),
123–147.
Brennan, S. E., & Schober, M. F. (2001). How listeners compensate for disfluencies in
spontaneous speech. Journal of Memory and Language, 44(2), 274–296.
Casillas, M., & Henetz, T. (unpublished). The Meet a Friend Corpus. Stanford University,
Stanford CA.
Christenfeld, N. (1994). Options and UMS. Journal of Language and Social Psychology,
13(2), 192–199.
Clark, H. H. (1996). Using language. Cambridge: Cambridge University Press.
Clark, H. H. (2003). Pointing and placing. In S. Kita (Ed.), Pointing. Where language, cul
ture, and cognition meet (pp. 243–268). Hillsdale NJ: Erlbaum.
Clark, H. H. (2004). Pragmatics of language performance. In L. R. Horn & G. Ward (Eds.),

Handbook of pragmatics (pp. 365–382). Oxford: Blackwell.
Clark, H. H. (2005). Coordinating with each other in a material world. Discourse Studies,
7(4–5), 507–525.
Clark, H. H. (2006). Social actions, social commitments. In S. C. Levinson & N. J. Enfield

(Eds.), Roots of human sociality: Culture, cognition, and human interaction (pp. 126–150).
Oxford: Berg Press.
Clark, H. H. (2012). Unspoken questions, unspoken answers. In J.-P. de Ruiter (Ed.), Ques
tions and answers (pp. 81–100) Cambridge: Cambridge University Press.

Levine, & S. D. Teasley (Eds.). Perspectives on socially shared cognition (pp. 127–149).
Washington, DC: APA Books.
Clark, H. H., & Fox Tree, J. E. (2002). Using uh and um in spontaneous speech. Cognition,
84, 73–111.
Clark, H. H., & French, J. W. (1981). Telephone goodbyes. Language in Society, 10, 1–19.
Clark, H. H., & Krych, M. A. (2004). Speaking while monitoring addressees for un
(p. 97)
derstanding. Journal of Memory and Language, 50(1), 62–81.
Clark, H. H., & Schaefer, E. F. (1987). Collaborating on contributions to conversation.

Language and Cognitive Processes, 2, 19–41.
Page 27 of 30

Working Together
Clark, H. H., & Schaefer, E. F. (1989). Contributing to discourse. Cognitive Science, 13,
259–294.
Clark, H. H., & Wilkes-Gibbs, D. (1986). Referring as a collaborative process. Cognition,

22, 1–39.
Corley, M., MacGregor, L. J., & Donaldson, D. I. (2007). It’s the way that you, er, say it:
Hesitations in speech affect language comprehension. Cognition, 105(3), 658–668.
Fox Tree, J. E., & Clark, H. H. (1997). Pronouncing “the” as “thee” to signal problems in
speaking. Cognition, 62, 151–167.
Garrett, M. F. (1982). Production of speech: Observations from normal and pathological

language use. In A. W. Ellis (Ed.), Normality and pathology in cognitive functions (pp. 19–
76). New York: Academic Press.
Gilbert, M. (1990). Walking together: A paradigmatic social phenomenon. Midwest Stud

ies in Philosophy, 15(1), 1–14.
Godfrey, J. J., Holliman, E., & McDaniel, J. (1992). SWITCHBOARD: Telephone speech cor
pus for research and development. Proceedings of the IEEE conference on acoustics,
speech and signal processing (pp. 517–520). San Francisco, CA: IEEE.
Goldman-Eisler, F. (1958). Speech production and the predictability of words in context.

Quarterly Journal of Experimental Psychology, 10, 96–106.
Goodwin, C. (1986). Between and within: Alternative sequential treatments of continuers

and assessments. Human Studies, 9, 205–217.
Grice, H. P. (1975). Logic and conversation. In P. Cole & J. L. Morgan (Eds.). Syntax and
semantics 3: Speech acts (pp. 41–58). New York: Academic Press.
Jefferson, G. (1972). Side sequences. In D. Sudnow (Ed.), Studies in social interaction (pp.
294–338). New York: Free Press
Jefferson, G. (1984). Notes on a systematic deployment of the acknowledgement tokens

“Yeah” and “Mm hm.” Tilburg Papers in Language and Literature, 20, 1–18.
Jefferson, G. (2002). Is “no” an acknowledgement token? Comparing American and British

uses of (+)/(-) tokens. Journal of Pragmatics, 34, 1345–1383.
Kendon, Adam. (1990). Conducting Interaction: Patterns of behavior in focused encoun

ters. Cambridge: Cambridge University Press.
Kita, S., & Ide, S. (2007). Nodding, aizuchi, and final particles in Japanese conversation:
How conversation reflects the ideology of communication and social relationships. Journal
of Pragmatics, 39(7), 1242–1254.
Levelt, W. J. (1989). Speaking: From intention to articulation. Cambridge, MA: MIT Press.
Page 28 of 30

Working Together
Maclay, H., & Osgood, C. E. (1959). Hesitation phenomena in spontaneous English

speech. Word, 15, 19–44.
Matarazzo, J. D., Saslow, G., Wiens, A. N., Weitman, M., & Allen, B. V. (1964). Interviewer
head nodding and interviewee speech durations. Psychotherapy: Theory, Research &
Practice, 1(2), 54–63.
Merritt, M. (1976). On questions following questions in service encounters. Language in

society, 5(3), 315–357.
Newell, A., & Simon, H. A. (1972). Human problem solving. Englewood Cliffs, NJ: Pren
tice-Hall.
Oreström, B. (1983). Turn-taking in English conversation (vol. 66). Lund, Sweden:

Gleerup.
Schegloff, E. A. (1980). Preliminaries to preliminaries: “Can I ask you a question?” Socio

logical Inquiry, 50, 104–152.
Schegloff, E. A. (1982). Discourse as an interactional achievement: Some uses of “uh huh”

and other things that come between sentences. In D. Tannen (Ed.), Analyzing discourse:
Text and talk. Georgetown University Roundtable on Languages and Linguistics 1981 (pp.
71–93). Washington, DC: Georgetown University Press.
Schegloff, E. A. (2007). Sequence organization in interaction: Volume 1: A primer in con

versation analysis. Cambridge: Cambridge University Press.
Schegloff, E. A., Jefferson, G., & Sacks, H. (1977). The preference for self-correction in
the organization of repair in conversation. Language, 53, 361–382.
Schegloff, E. A., & Sacks, H. (1973). Opening up closings. Semiotica, 8(4), 289–327.
Shriberg, E. (1994). Preliminaries to a theory of speech disfluencies. PhD thesis, Universi

ty of California at Berkeley.
Smith, V. L., & Clark, H. H. (1993). On the course of answering questions. Journal of
Memory and Language, 32, 25–38.
Stivers, T., Enfield, N. J., Brown, P., Englert, C., Hayashi, M., Heinemann, T.,…, & Levin
son, S. C. (2009). Universals and cultural variation in turn-taking in conversation. Pro
ceedings of the National Academy of Sciences, 106(26), 10587–10592.
Svartvik, J., & Quirk, R. (1980). A corpus of English conversation. Lund, Sweden: Gleerup.
Watzlawick, P., Bavelas, J. B., & Jackson, D. D. (1967). Pragmatics of human communica
tion. New York: W. W. Norton.
Page 29 of 30

Working Together
Notes:
(1) . Many examples in this chapter come from one of three corpora: London-Lund (LL;
Svartvik & Quirk, 1980), Switchboard (SW; Godfrey, Holliman, & McDaniel, 1992), and
Meet a Friend (MaF; Casillas & Henetz, unpublished). Each example is marked with a
corpus identification (e.g., LL), text number (e.g., 2.7), and beginning line or time-stamp
(e.g., 1232).
(2) . Two stretches of speech that overlap are enclosed in pairs of asterisks.
Herbert H. Clark
Herbert H. Clark is Professor of Psychology at Stanford University.
Tania Henetz
Tania Henetz, Department of Psychology, Stanford University.
Page 30 of 30

Perspective Taking and Its Impostors in Language Use: Four Patterns of De
ception
Perspective Taking and Its Impostors in Language Use:

Four Patterns of Deception
Dale J. Barr

Language users seem skilled at adapting their language processing to changes in context.
Although such adaptations may give the impression of perspective taking, other mecha
nisms could give rise to the same behavior. As a consequence, merely observing that a
speaker produced an utterance that matched a listener’s informational needs, or that a
listener interpreted an utterance in the manner a speaker intended, does not warrant the
inference that any perspective taking actually occurred. This chapter reviews four “pat
terns of deception” that yield behaviors that look like perspective taking but that, on clos
er inspection, actually reflect simpler kinds of processing. These patterns are personified
in four basic character types: the Double (egocentric projection), the Charlatan (attribute
substitution), the Conspirator (independent processing by functionally encapsulated sub
systems), and the Freeloader (interactive coordination). The existence of these alterna
tive explanations suggests a need for greater caution in interpreting evidence from stud
ies of dialogue.
Keywords: Language processing, psycholinguistics, perspective taking, pragmatics
But, of all my troubles, this was the chief: I was every day and every hour assailed
with accusations of deeds of which I was wholly ignorant; of acts of cruelty, injus
tice, defamation, and deceit; of pieces of business which I could not be made to
comprehend…
– James Hogg, The Private Memoirs and Confessions of a Justified Sinner

(1824/1997)
Intersubjectivity is one of the enduring mysteries in the study of language use: how can
different people with different beliefs use language to arrive a common understanding?
Despite pervasive ambiguity, people seem to communicate their intentions effectively;
moreover, they seem to do so with relative ease. These observations suggest the exis
Page 1 of 21

ception
tence of powerful constraints on language processing. What are these constraints, and
how are they incorporated into the production and interpretation of utterances?
One influential proposal has been that language users reduce ambiguity by producing
and interpreting language against their common ground—the set of information interlocu
tors believe to be shared (Clark & Marshall, 1981). Common ground is distinct from
shared knowledge inasmuch as it is a special kind of meta-knowledge: a given piece of
shared information is not common ground unless interlocutors are mutually aware that it
is shared. For example, you might know some salacious gossip about your friend, who
knows that people have been spreading gossip about her and knows of its nature, but
who doesn’t know whether you know it. Although shared, this gossip is not yet part of
your common ground. Typically, language users can infer that some piece of information
is part of their common ground on the basis of certain co-presence heuristics—on the ba
sis of sharing perceptual experiences (being located in the same environment (p. 99) and
thus perceiving the same objects and hearing the same speech) or on joint membership in
certain social communities (Clark & Marshall, 1981).
The proposal that people speak and interpret language against the background of their
common ground has been highly influential because it offers a possible solution to the
problem of pervasive ambiguity in language use. Speakers can tailor their utterances to
the common ground in a process known as audience design (Clark & Murphy, 1982). Like
wise, listeners can use common ground to constrain the set of possible meanings they
might assign to an utterance (Clark & Carlson, 1981). However, achieving this reduction
in ambiguity through common ground requires language users to pay the cost of continu
ally accessing and using information about their interlocutor’s perspective (Lin, Keysar, &
Epley, 2010; Rossnagel, 2000).
The strategy of processing language against common ground has obvious appeal because
it would seem to promote accurate communication. However, the psycholinguistic litera
ture suggests that language users do not always follow this strategy. In many cases,
speakers or listeners make use of privileged information—information that is known not to
be shared with the interlocutor (e.g., Brown & Dell, 1987; Ferreira & Dell, 2000; Keysar,
Barr, Balin, & Brauner, 2000; for review, see Barr & Keysar, 2006). In general, what the
literature suggests is that language users can make use of common ground in certain cir
cumstances but are not always able (or willing) to pay the cost of perspective taking. In
fact, it is not always necessary for language users to pay this cost. At least in some cases,
speakers and listeners can reduce ambiguity using other, less cognitively demanding tac
tics (Barr, 2004; Barr & Keysar, 2002; Brown & Dell, 1987; Ferreira & Dell, 2000; Horton
& Gerrig, 2005; Pickering & Garrod, 2004).
The recognition that there are multiple strategies for ambiguity resolution beyond per
spective taking has greatly enriched the theoretical landscape in the study of language
use. At the same time, it has made investigation of language processing more challenging
because a language user can produce or interpret utterances in ways that are consistent
with common ground but without any perspective taking ever having taken place. In oth
Page 2 of 21

ception
er words, there can be multiple explanations for the same behavior, and, as a result, it is
often difficult to identify or create the conditions necessary to distinguish these alterna
tive explanations from true perspective taking. My goal in this chapter is to characterize
these theoretical and empirical challenges by identifying four “patterns of deception”—
four patterns of behavior that are distinct from, but easily confused with, genuine per
spective taking.
By genuine perspective taking, I mean behavior satisfying the following criteria: (1) it
arises when there is a disparity in beliefs between a perspective-taking agent (henceforth
agent) and a target person’s perspective (henceforth target), (2) the agent adapts its pro
cessing or behavior because of this disparity, (3) the target’s beliefs are accessible and ac
curately represented by the agent, (4) the agent produces adaptations that are appropri
ate to the target’s perspective, and (5) these adaptations are undertaken spontaneously.
The first criterion is needed because in cases where there is no disparity, it is generally
not possible to distinguish perspective taking from behavior that is merely egocentric; the
second, because true perspective taking must be produced in response to a perceived
change in the common ground, as opposed to changes in the perspective of the agent
(Keysar, 1997). The third criterion rules out various uninteresting cases in which perspec
tive taking fails due to ignorance or inaccurate beliefs about the target. The fourth criteri
on indicates that we should only accept adaptations that are appropriate to the target’s
needs. Finally, the fifth criterion states that we should restrict our definition to cases in
which language users recognize the need for perspective taking and adapt their behavior
spontaneously (i.e., without prompting by an experimenter or through feedback received
from their conversational partner).
Four Patterns of Deception

Here, I highlight research that unmasks four characteristic “patterns of deception” that
violate one or more of these criteria. These patterns are personified in terms of four im
postor types: the Double, who processes language against a minimally altered copy of his
or her own knowledge; the Charlatan, who produces the illusion of perspective taking
through the process of attribute substitution; the Conspirator, in which parallel process
ing by functionally independent, informationally encapsulated subsystems gives the false
impression of information integration; and finally, the Freeloader, who offloads the com
putational burden of perspective taking to the communicative environment, usurping the
credit for accomplishments that are the result of interactive processes and not of individ
ual cognition.
(p. 100) The Double
Because communication involves exchanging information, the basic acts of speaking and
understanding would seem to require, as a prerequisite, at least some minimalistic form
of perspective taking. After all, to even have the idea of saying something requires know
ing that there are things that one’s interlocutor needs to know or would be interested in
Page 3 of 21

ception
knowing. Likewise, homing in on a speaker’s intended meaning would also seem to mini
mally require considering what it is that one’s interlocutor might want to have under
stood. But as the pattern of the Double shows, the amount of perspective taking required
to get communication off the ground may be very minimal indeed.
The Double engages in a very minimalistic form of perspective taking, in which agents
adapt their speaking or understanding not to the target, but to a minimally altered copy
of their own knowledge. Essentially, the Double is creating or interpreting messages for
him- or herself, minus some critical information. In production, the Double operates by
planning speech while suppressing the information content of the utterance. For exam
ple, in referential communication, a “director” agent may know the identity of some refer
ent and need to convey this information to the addressee. The proposal here is that the
Double accomplishes this not by designing an utterance taking into account what the ad
dressee knows and does not know, but by asking him- or herself the question “What
would I need to know to be able to identify the referent, knowing everything I now know
except for the referent’s identity?”
To empirically distinguish such “minimal” perspective taking from more sophisticated

versions requires creating situations in which information available to the agent is un
available to the target. The extent to which speaking and understanding are influenced by
this “privileged” information provides evidence for this strategy. For example, the fact
that a speaker will call an object the circle in the context of a square and the small circle
in the context of a larger circle could reflect perspective taking, or it could reflect the
minimal adaptation of the Double. To distinguish these accounts, it is critical to manipu
late the agent’s belief about whether the relevant context is also available to the target.
A classic study by Horton and Keysar (1996) found that speakers used modified expres
sions such as the small circle in cases where, from the target’s point of view, an unmodi
fied expression such as the circle would be sufficient. Furthermore, they found that this
“egocentrism” was exacerbated when speakers were put under time pressure, suggesting
that utterances are initially designed for a minimally adapted version of the speaker’s
own knowledge and that speakers adapt this description to the addressee’s perspective
given sufficient time and cognitive resources. Later research has further confirmed that
the generation of modified expressions depends to a strong degree on speaker-internal
processes: instructing speakers to conceal the contrasting privileged information makes
them even more likely to include it in their speech (Wardlow Lane, Groisman, & Ferreira,
2006); indeed, giving them financial incentives to conceal even further increases the ten
dency to leak information (Wardlow Lane & Liersch, 2012).
Operation of the Double can also be seen in studies of the comprehension of definite re
ferring expressions. When people interpret instructions to, for example, “pick up the
small candle,” they often find it difficult to realize that the candle intended by the speak
er can only be the smallest one that they and the speaker both know about. Instead, they
often consider the smallest candle they know about, even if they know the speaker can’t
see it, and indeed often (about 20 percent of cases) erroneously choose it as the intended
Page 4 of 21

ception
candle (Keysar et al., 2000). This pattern of egocentrism has been replicated over a large
number of studies (Apperly et al., 2010; Begeer, Malle, Nieuwland, & Keysar, 2010; Epley,
Keysar, van Boven, & Gilovich, 2004a; Epley, Morewedge, & Keysar, 2004b; Keysar, Lin, &
Barr, 2003; Lin et al., 2010; Nilsen & Graham, 2009). Although the basic pattern of ego
centrism can hardly be denied, there is currently vigorous debate on its implications for
theories of spoken language comprehension (see Brown-Schmidt, Gunlogson, & Tanen
haus, 2008; Hanna, Tanenhaus, & Trueswell, 2003; Heller, Grodner, & Tanenhaus, 2008;
Nadig & Sedivy, 2002; Wu & Keysar, 2007, for counterevidence and caveats; see Barr,
2014, and Brown-Schmidt & Hanna, 2011, for review and discussion; see also the section
on “The Conspirator”).
The strategy of the Double can be seen as an anchoring-and-adjustment reasoning

process (Tversky & Kahneman, 1974), wherein a preliminary model of the interlocutor is
built from one’s own knowledge (anchoring) and then incrementally adapted toward the
knowledge of the target person (adjustment) (Epley et al., 2004a; Keysar & Barr, 2002).
Critically, this adjustment process is assumed to require cognitive effort and is typically
insufficient, resulting in a model that is biased toward one’s own knowledge.
(p. 101) This strategy of substituting one’s own knowledge will be effective whenever
there is extensive overlap between the perspectives of agent and target. It is possible,
then, that language users will be more likely to adjust away from the “egocentric” anchor
when the target is perceived to be dissimilar from the self. This yields an interesting and
counterintuitive prediction: people should be more egocentric when talking to friends.
This prediction was confirmed in a study by Savitsky, Keysar, Epley, Carter, and Swanson
(2011), who found that speakers were more prone to overestimate how well they con
veyed their meanings when speaking to friends than when speaking to strangers. In addi
tion, they found that listeners experienced greater interference from privileged informa
tion when they interpreted speech from friends. In short, talking to someone familiar
causes people to drop their “cognitive guard” against their native tendency toward ego
centrism.
The Charlatan
Much of human cognition can be regarded as heuristic in nature: rather than mapping in
puts to outputs using some computationally demanding or uncertain target function, it
approximates the mapping using some alternative function that sacrifices accuracy for ef
ficiency, a characteristic known as “satisficing” (Simon, 1996). In decision making, substi
tuting an easier problem for a harder problem is known as attribute substitution (Kahne
man & Frederick, 2002). As Kahneman and Frederick put it: “judgment is mediated by a
heuristic when an individual assesses a specified target attribute of a judgment object by
substituting another property of that object—the heuristic attribute—which more readily
comes to mind.” (p. 53). For example, when assessing whether it is safer to travel from A
to B by car or by plane, one should attempt to compare two ratios: the ratio of unsafely
completed car journeys to all car journeys, compared to the ratio of unsafely completed
air journeys to all air journeys. However, one may answer this question instead by consid
Page 5 of 21

ception
ering how readily unsafe journeys by either mode come to mind. If there have recently
been significant air disasters in the news, one may be more likely to consider car travel as
safer. In such an example, the heuristic attribute of “ease of recall” is substituted for the
target attribute of relative risk.
The Charlatan uses such attribute substitution to produce behavior that appears to be
shaped by the target attribute of the target person’s perspective but that, in reality, is
based on heuristic attributes available to the self.1 Often, this reliance on heuristic attrib
utes will be sufficient for successful communication because the heuristic attributes are
strongly correlated with the target attributes. However, the correlated nature of these at
tributes means that often quite clever experimentation is required to distinguish attribute
substitution from perspective taking.
Consider, for example, decisions about how to articulate a word. A variety of evidence
suggests that the less contextually predictable a word is, the more clearly speakers will
tend to articulate it (Bell, Brenier, Gregory, Girand, & Jurafsky, 2009; Lieberman, 1963);
indeed, listeners seem to infer whether a referent is given or new in the discourse based
on the intelligibility of the referring expression (Fowler & Housum, 1987). It would there
fore seem that phonological planning in language production is closely calibrated to the
comprehender’s informational needs. The available evidence generally suggests, howev
er, that articulatory clarity is driven by availability of information to the speaker rather
than by the speaker’s model of the listener. Bard and colleagues (Bard et al., 2000) found
that speakers produced less intelligible tokens of a word when an expression was repeat
ed, regardless of whether the expression was new to the addressee. In a more recent
study (Arnold, Kahn, & Pancani, 2012), the speaker’s beliefs about the availability of a
referent to an addressee were directly manipulated by the addressee’s nonverbal behav
ior. There was no evidence that speakers spoke less intelligibly when the referent was
available to the addressee. However, Galati and Brennan (2010) found some evidence for
audience sensitivity: whereas the measured duration of words seemed only to reflect the
speaker’s knowledge, a set of participant raters who rated the intelligibility of the words
out of context gave higher ratings to words spoken to less-knowledgeable addressees.
Moving up the ladder of language production, studies of syntactic processing have, on the
whole, discredited the idea that speakers choose syntactic constructions to avoid ambigu
ity or to make their speech easier for an addressee to parse (Arnold, Wasow, Asudeh, &
Alrenga, 2004; Ferreira & Dell, 2000). There is also little evidence that, when using a syn
tactically ambiguous construction, speakers use speech prosody to disambiguate it
(Kraljic & Brennan, 2005). Instead, these choices seem to be driven by what is easier to
produce or more salient to the speaker. For example, much of a speaker’s decision re
garding whether to use the optional complementizer that in sentence complement struc
tures (p. 102) (e.g., “the coach knew (that) a good player would be hard to find”) depends
on cognitive availability of the material in the complement clause (Ferreira & Dell, 2000),
with higher availability resulting in fewer instances of that. Even when speakers are made
explicitly aware that a construction will be syntactically ambiguous to the listener, they
Page 6 of 21

ception
are drastically overconfident in their ability to use prosody to signal an intended meaning
(Keysar & Henly, 2002).
At the next-higher level of production, speakers have to make choices about how to refer
to a referent. Unlike syntactic or prosodic choices, such choices would seem to be more
accessible to consciousness and therefore more subject to beliefs about the interlocutor’s
informational needs. When referring to an entity, speakers have a large of variety of refer
ential forms at their disposal. Philosophers and language scientists have long assumed
that it is accessibility of information for the addressee that is the driving factor in speak
ers’ referential choices (Ariel, 1988; Chafe, 1976; Gundel, Hedberg, & Zacharski, 1993).
However, here is another case where careful research has unmasked the Charlatan’s
clever tricks.
Speakers produce full noun phrases when a referent is new and pronouns when a refer
ent is the salient focus of discussion. But do they do this based on what is accessible in
the addressee’s discourse model, or based on what is accessible in their own discourse
models? To distinguish these possibilities, Fukumura and van Gompel (2012) had speak
ers describe a critical action performed by one of two characters (of different genders) to
an addressee who had to recreate the action using toys. Just before the speaker de
scribed the critical action, Fukumura and van Gompel (2012) auditorily presented a con
text sentence, manipulating two factors: (1) whether the character mentioned in the con
text sentence was the same character later mentioned in the critical sentence (the “criti
cal” character) or a different character; and (2) whether the addressee or only the speak
er heard the context sentence. The question was whether the speaker, in describing the
critical action, used a pronoun (e.g., “he stood up”) or a full noun phrase (“the admiral
stood up”). If the choice is based on the accessibility of referents for the addressee, then
speakers should only use pronouns when the context sentence was about the critical
character and the addressee had heard it. Instead, Fukumura and van Gompel’s speakers
seemed influenced only by whether the context sentence mentioned the critical character,
using pronouns at a dramatically higher rate when it did. Intriguingly, speakers did not
seem altogether indifferent to whether addressees heard the context sentence; there was
a (not fully significant) trend toward speakers using fewer pronouns when the addressee
did not hear the context sentence. However, this was true even when the competitor was
mentioned. In short, it seems that speakers largely base their choices of referential form
on accessibility within their own discourse model. They may also be sensitive to the possi
bility that addressees may not be knowledgeable about a topic, but this causes them sim
ply to be more explicit overall, not just in the only case where it actually matters to the
addressee (for similar effects, see Ferreira & Dell, 2000).
Choosing the form of a referring expression is just one type of high-level planning deci
sions that speakers have to undertake when generating spoken utterances. Another type
of planning decision concerns how speakers package information along different visual
and auditory channels. In many conversational settings, speakers have visual contact with
their conversational partner, such that they can use gestures and facial expressions to
convey their meanings alongside speech. It is critical for speakers to consider what infor
Page 7 of 21

ception
mation channels are available to their interlocutor when packaging information; for ex
ample, it is likely to be ineffective to attempt to gesturally demonstrate an idea when
speaking to a visually impaired person. A number of studies support the idea that speak
ers are sensitive to addressees’ access to visual information channels; for example, speak
ers gesture more frequently and vividly when they have visual contact (Alibali, Heath, &
Myers, 2001; Bavelas, Gerwing, Sutton, & Prevost, 2008; Cohen & Harrison, 1973).
In most situations, there is parity across production and comprehension modalities in

terms of the availability of information channels. In other words, it is unusual for one par
ty in a communication to have access to a modality that is not also accessible to the other
party.2 When you can see and hear me, it is generally safe to assume that I can also see
and hear you. But this parity opens up the opportunity for attribute substitution: specifi
cally, speakers can substitute the heuristic attribute of can I see and hear my interlocu
tor? for the target attribute of can my interlocutor see and hear me? Given this, the find
ing that speakers gesture more under visibility conditions may not truly reflect adapta
tion to the fact of being seen, but rather, a learned response to the visible availability of
the interlocutor.
(p. 103) It was only recently that researchers had the idea of deconfounding the effects of
seeing from the effects of being seen on gesture production. Mol, Krahmer, Maes, and
Swerts (2011) assessed the rate of representational gesture3 production among speakers
retelling a cartoon under various communication conditions while independently varying
the factors of seeing and being seen. In one of the two experiments, they used a special
webcam setup that allowed for simulated eye contact between interlocutors. In this ex
periment, they found that the rate of gesture production was influenced only by being
seen and not by seeing. This supports the idea that the distribution of information across
channels reflects genuine perspective taking (but see the section on “The Freeloader” for
possible caveats).
Although so far the discussion of attribute substitution has focused on language produc
tion, this stratagem is also available to language comprehension. Consider how people
determine who is being referred to when they hear a person’s initial name, such as Kevin.
Initial names are highly ambiguous due to conservatism in how parents name their chil
dren; for example, about half of the approximately 2 million baby boys born in the United
States in 2011 received one of just 140 names (Barr, Jackson, & Phillips, 2013). When
your friend asks, “did you hear about Kevin?,” how do you know which Kevin she is refer
ring to? One possibility is that you access the common ground you share with your friend,
narrowing down the set of Kevins to those people named Kevin that are contextually rele
vant, that both of you know, and that both of you know that you both know. However, an
other possibility is that you have learned to associate hearing the word Kevin in your
friend’s voice with the target Kevin. Upon hearing the compound cue of the name and
your friend’s voice, you access the referent immediately, thus bypassing common ground.
To distinguish these possibilities, Barr et al. (2014) took advantage of the naturally occur
ring common ground existing between university students in the form of social networks.
Page 8 of 21

ception
Pairs of friends from a given year of study were recruited to play a communication game
along with another student (the assistant) who was an outsider to the friends’ social cir
cle. One of the friends (the addressee) heard the name of a mutually known target person
from the same year (e.g., “the next person will be Kevin”) and had to select the corre
sponding photo from a computerized display while his or her eyes were tracked. On any
given trial, addressees were led to believe that the name for the target had been chosen
either by the friend or by the outsider and that this “designer” agent had then written
down his or her choice on a slip of paper. Critically, the name on the slip of paper could be
read aloud by either the friend or the assistant; thus, in addition to friends reading mes
sages they themselves had written, they sometimes read messages that the assistant had
written and vice versa.
An analysis of eye data from the first 1.5 seconds after name onset revealed that ad
dressees gazed more quickly and more reliably at the target when the name was spoken
by the friend, regardless of whether the name was chosen by the friend or by the
outsider. This supports the idea that listeners identify persons from spoken names
through episodic associations rather than through common ground. However, it was not
the case that addressees simply ignored the identity of the designer because effects of
the designer were observed on response time. This case provides support for the general
idea that the ordinary operations of memory can serve as a proxy for computations of
common ground (Horton & Gerrig, 2005).
The Conspirator
One of the more nuanced questions that has arisen in recent years concerns not whether
an agent used information about a target perspective at all during processing, but how
such information was used. Specifically, a key question is the extent to which agents are
able to use information about a target perspective to constrain the operations at various
levels of language processing. To date, these more nuanced questions have been asked
mostly about language comprehension rather than language production and have focused
on the extent to which perspective information constrains the lexical level of processing
(recognizing words in context and accessing their meanings).
The comprehension of words is an incremental process that does not wait to resolve am
biguities until a linguistic constituent has been filled; rather, the comprehension system
generates hypotheses on-the-fly, moment-by-moment, as information accrues (Allopenna,
Magnuson, & Tanenhaus, 1998; Marslen-Wilson, 1987). For example, by the time one has
heard the first syllable buh of the word bucket, the comprehension system will have al
ready reduced the set of possible lexical candidates from a large set to a much smaller set
of words consistent with the evidence, such as buckle and bucket. But it is only in psy
cholinguistic labs that listeners hear single words in isolation, without having (p. 104) ac
cess to any prior information that makes certain words more or less likely. In general, lis
teners have access to a wide array of prior information that can constrain their expecta
tions from a wide variety of sources. A long-standing debate in spoken language compre
Page 9 of 21

ception
hension concerns how effectively this information constrains the lexical hypotheses gen
erated about the input.
Some of the earliest claims about the use of common ground in language processing fa
vored an “egocentric first” model in which comprehenders initially interpret utterances
from their own perspectives and only later consider information about speakers’ perspec
tives as part of a secondary, optional, resource-demanding stage of processing (Keysar et
al., 2000; Keysar, Barr, Balin, & Paek, 1998). However, this view has been largely discon
firmed, with visual-world eye-tracking studies showing that listeners pay less attention to
privileged information than to information in common ground. Indeed, effects of common
ground generally are present during the earliest moments of lexical processing (Brown-
Schmidt et al., 2008; Hanna et al., 2003; Nadig & Sedivy, 2002). Moreover, listeners seem
to spontaneously access information about the speaker’s perspective, doing so even when
considering the speaker’s perspective is unnecessary to resolve ambiguity (Barr, 2008b;
Heller et al., 2008).
Taken together, these findings would seem to suggest that common ground modulates
lexical processing from the earliest moments of comprehension. As such, they have been
marshalled in support of probabilistic constraint-based models of language comprehen
sion, in which various structures and interpretations are assigned probabilities through a
weighted combination of cues (Hanna et al., 2003; Nadig & Sedivy, 2002). Under this in
terpretation, whether or not a speaker knows about a particular referential alternative is
seen as just one of many cues that are used in parallel to resolve references. Jurafsky
(2003) notes that inasmuch as constraint-based models assume weighted cue combina
tion and probabilistic outputs, they can be considered a notational variant of Bayesian
models. These models have a critical requirement of cognitive penetrability—that is, that
“top-down” information available to higher level subsystems, such as information about a
speaker’s perspective, must be also available to lower level systems engaged in “bottom-
up” processing, such as those involved in lexical processing.
Although it might be tempting to see the evidence from comprehension as settling mat
ters in favor of Bayesian-type integration accounts, the problem is that it is possible to
have concurrent effects of top-down perspective taking and bottom-up lexical processing
in a system lacking in cognitive penetrability—that is, even when the lexical processing
subsystem is not influenced by the top-down information. When visual-world researchers
observe concurrent effects of a contextual constraint and of linguistic processing, they all
too readily interpret this pattern as evidence for interactive integration (Barr, 2008a).
But, as any good Bayesian would tell you, observing only the output of the belief-updating
process (the posterior) cannot possibly provide any information about whether or not the
new information (linguistic evidence) was combined with the prior information (context)
in an optimal way. Participants in perspective-taking experiments, before hearing any re
ferring expression, already know which referential alternatives are privileged and which
are shared. This creates a visual bias favoring shared over privileged referents. Effects of
common ground during the expression could simply be a continuation of these prior ef
fects. To be sure, the fact that this bias exists shows that listeners are sensitive to infor
Page 10 of 21

ception
mation about the speaker’s perspective; however, from observing that this bias continues
during processing of the referring expression, it is not valid to conclude that the informa
tion modulated lexical processing. The relevant evidence for information integration is
how the incoming linguistic information changes the probabilities of gazing at potential
referents; and particularly, whether the boost in probability is larger for shared than for
privileged alternatives, as Bayesian (or probabilistic constraint-based models) would pre
dict.
Barr (2008b) tested the predictions of information integration by examining lexical com
petition within a visual-world paradigm. Participants viewed computerized displays de
picting four referential alternatives and heard a speaker telling them to click on a target
picture (e.g., “Click on the bucket”). One of the three remaining alternatives, the critical
picture, was either a phonological competitor (i.e., shared a phonological onset with the
target, e.g., a buckle) or a baseline picture (i.e., phonologically unrelated picture, e.g., a
stepladder). Importantly, this critical object was either privileged or shared. The question
was whether the boost in activation for competitors (measured against baseline) was larg
er when the critical object was shared than when it was privileged.
Fig. 6.1 Predicted and observed posterior probabili

ty of gazing at a privileged competitor plotted
against the observed posterior for a shared competi
tor (uniform prior).
Over all three experiments, there was clear evidence that comprehenders anticipated ref
erences to (p. 105) shared referential alternatives: typically, by the onset of the referring
expression, they were three to four times more likely to be gazing at a shared referent
than a privileged referent. In other words, they had a strong prior expectation that the
speaker would refer to one of the shared alternatives. However, none of the experiments
provided any evidence that listeners were able to integrate this information with incom
ing speech. Data from one such experiment (experiment 2) appear in Figure 6.1, showing
the observed competition effect compared to the predictions of a Bayesian model. The
competition from privileged alternatives was far greater than a Bayesian information-inte
gration account would predict; indeed, in each experiment, there was no evidence that
the competition from privileged alternatives was any less than that induced by shared al
ternatives (the case of a uniform prior).
What this implies is that comprehenders represent speakers’ beliefs and spontaneously
activate them in time to constrain processing but that the neurocognitive system respon
sible for mapping incoming speech to referential alternatives has no access to this infor
Page 11 of 21

ception
mation. This is particularly surprising when considered against a further finding. In one
experiment, the top-down effects of perspective were contrasted with top-down effects of
a linguistic nature (specifically, top-down effect from verb semantics). Although the ef
fects of perspective showed the anticipation-without-integration pattern, the effects of the
linguistic constraint showed a pattern that was nearly Bayes-optimal (Barr, 2008b).
These findings suggest the intriguing possibility that at least some of the difficulties peo
ple experience in using knowledge about others may lie not in a failure to represent what
others know in an accurate or timely fashion, but as the result of constraints imposed by
neurocognitive architecture on the flow of information between autonomous (or semiau
tonomous) processing modules. The independent, parallel operations of these seemingly
autonomous subsystems conspire to give an illusion of information integration.
The Freeloader
The Freeloader, like the Conspirator, is unduly credited with more sophisticated process
ing than is actually undertaken; however, in this case, the processing in question is not di
vided across separate processing modules within the same individual but rather across
separately acting individuals. When you interact with another person, the behaviors you
and your partner produce are not independent: you will reciprocally influence one
another’s behavior (Pickering & Garrod, 2004; Pickering & Garrod, this volume). For ex
ample, addressees play an active role in determining an utterance’s informational content
through cues they provide to the speaker (Bavelas, Coates, & Johnson, 2000; Clark &
Wilkes-Gibbs, 1986; Gann & Barr, 2012; Kraut, Lewis, & Swezey, 1982; Schober & Clark,
1989), and speakers monitor these cues as they speak (Clark & Henetz, this volume;
Clark & Krych, 2004; Gann & Barr, 2012). The same might be true for comprehenders—to
the extent that different interpretations of an utterance lead to different overt actions and
to the extent that comprehenders have access to online feedback from a speaker who is
monitoring their behavior, they could modify their interpretations based on this feedback
(although to (p. 106) my knowledge this possibility has not been investigated). In close-
knit interactive situations, certain acts of speaking and understanding might be better
viewed as collaborative achievements rather than individual acts.
The Freeloader takes advantage of interactive situations to reduce his own perspective-
taking burden (Barr & Keysar, 2005; Fussell & Krauss, 1992; Gann & Barr, 2012). Model
ing the listener’s knowledge is an effortful activity fraught with uncertainty. Interactive
situations enable interlocutors to reduce uncertainty about one another’s perspectives by
acting on uncertain suppositions rather than requiring them to more effortfully do so
through cognitive computation; viewed in this way, acts of speaking and understanding in
dialogue can constitute a form of epistemic action (Kirsh & Maglio, 1994). A speaker who
is uncertain about whether an addressee is familiar with an expression (does she know
that “NHS” stands for the National Health Service in Britain?) can piece together the
likelihood that she knows it by considering other things she knows about the individual (I
think she once visited the UK, and her aunt married an Englishman). When there is no po
tential for interaction (the speaker is crafting an e-mail), the speaker would have only her
Page 12 of 21

ception
own knowledge to rely on to decide what form is appropriate. But in interaction, the
speaker can simply take a gamble by emitting the uncertain material (“I can’t imagine
something like the NHS ever existing in the US”) and observing her partner’s reaction
(she looks confused; now she is asking me what the NHS is). Given that the best possible
model of the addressee’s knowledge is the addressee herself, this strategy is necessarily
more accurate than any estimation strategy that a speaker could possibly undertake.
The fact that speakers’ and listeners’ conversational acts are intertwined in interactive di
alogue means that studying socially immersed behavior can lead to erroneously crediting
individual language users with feats of perspective taking that are more appropriately
characterized as interactively emergent. To illustrate, an influential study by Brennan and
Clark (1996) sought to test the argument for partner-specificity in referring to everyday
objects: that speakers choose referring expressions based on their history of interaction
with specific partners. They found that if you have been calling a particular shoe “the
loafer” to distinguish it from another shoe when talking to me, you would be likely to con
tinue calling it thus even in contexts in which the term is overspecific (because it is the
only shoe); more importantly, they found that you would use overspecific terms at a high
er rate when you continued talking to me as compared to when you spoke to another per
son. But why? Is it because in designing your utterances you are considering my perspec
tive, or because a new partner gave you feedback that led you to stop using the overly
specific term?
In a follow-up study, Gann and Barr (2012) recruited triads of participants and got speak
ers to entrain on very unusual ways of describing very typical objects. For instance, by
presenting a very typical candle in the context of a half-melted candle, they got speakers
to frequently call the typical candle “the unmelted candle.” The question was whether
speakers would continue using these unusual expressions when the contrasting object
(the melted candle) was removed from the context. Like Brennan and Clark’s speakers,
they continued using the overly specific terms; in this case, about 70 percent of the time.
Unlike Brennan and Clark (1996), however, speakers were only tested on each entrained
term a single time with a single addressee, giving them no opportunity to change their in
dividual descriptions in response to feedback. Under these circumstances, there was no
evidence that speakers were any less likely to use unusual entrained terms such as un
melted candle with a new addressee than with the old partner.
The danger of erroneously attributing interactively emergent effects to individual pro

cessing lurks most obviously in the practice of aggregating over multiple referring
episodes and treating the result as if it were a picture of what takes place within a single
episode. However, interactive effects can also emerge within a single referring episode,
where its operation can often be quite subtle. In the same study by Gann and Barr (2012),
speakers described abstract figures to addressees, much in the manner of the well-known
study by Clark and Wilkes-Gibbs (1986). This time, however, the images were presented
on computer monitors, and addressees selected images using a computer mouse. Speak
ers could observe the addressee’s movement of the mouse cursor and could use this as an
index of their emerging understanding. Indeed, there was evidence that speakers did
Page 13 of 21

ception
this: the longer the lag between the onset of the speaker’s expression and the onset of
movement of the addressee’s mouse cursor, the more words speakers used to describe
the target.
In sum, in close-knit, interactive dialogue, many of the adjustments that speakers make
for their (p. 107) addressees—and possibly, that addressees make for their speakers—may
emerge out of interactive processes and, as such, do not reflect spontaneous adjustments
to a perceived change in the common ground. It can be argued that these adjustments do
reflect a kind of perspective taking; after all, when a speaker notes that an addressee is
not understanding, they need to determine how to adapt their speech to improve the situ
ation. But as we saw earlier, careful research is needed to know whether adaptations to
signals of miscoordination reflect true consideration of the addressee’s perspective or the
thoughtless deployment of heuristics that have been learned through experience and
which may or may not be advantageous for the situation at hand.
Conclusion
In summary, we have considered four explanations for partner-adapted behavior in lan
guage use, personified in four impostor types: the Double, who substitutes a minimally
adapted version of its own knowledge for that of the target; the Charlatan, who employs
attribute substitution; the Conspirator, in which the parallel but independent operation of
autonomous subsystems conspires to produce an illusion of information integration; and
the Freeloader, who gets solitary credit for interactively emergent processes.
To the extent that we are interested in mechanistic explanations of speaking and under
standing, the existence of these four patterns of deception calls for greater caution in
how evidence from language use in dialogue informs psycholinguistic theories. Re
searchers currently seem to be well aware of the pattern of the Double, but less aware of
the three other personality types. The Charlatan is often overlooked and can be particu
larly tricky to unmask because this requires identifying heuristic attributes that might be
available in a given situation and varying them independently of target attributes. The
Conspirator is generally ignored, even though studies claimed to support interactive ef
fects of perspective on lexical processing generally show the same anticipation-without-
integration pattern (see Barr, 2014, for discussion).
It has recently been argued that studies of perspective taking in noninteractive contexts
(Brown-Schmidt & Hanna, 2011) or that use confederates (Kuhlen & Brennan, 2013) have
limited ecological validity. But the pattern of the Freeloader highlights the invalidity of
drawing conclusions about individual psycholinguistic processing from close-knit interac
tive situations. To be sure, studying language processing in dialogue is critical for under
standing the interactive strategies that language users bring to bear in typical face-to-
face settings. But it must be kept in mind that the behaviors produced in such circum
stances are not independent, and claims about individual processing should be appropri
ately tempered. This is currently not the case. For example, although Brennan and Clark’s
own data suggest that lexical choices emerged through interaction, researchers still
Page 14 of 21

ception
heavily cite this as a key study supporting perspective taking in production, for example:
“Extensive evidence has shown that speakers select linguistic forms on the basis of the
knowledge, intentions, or goals of the addressee (e.g., Brennan & Clark, 1996)” (Arnold et
al., 2012, p. 505). To distinguish competing psycholinguistic explanations, there can be no
substitute for the controlled experiment, and there is no escaping the tradeoff between
ecological validity and experimental control. It should also be kept in mind that settings
of language use vary in their interactivity, with even some face-to-face interactions show
ing relatively low interactivity due to social status differences or sociocultural norms re
garding nonverbal behavior and interruption (e.g., behavior in a job interview versus in
the pub). They also vary in their multimodality, from rarefied exchanges in text messaging
to full blown face-to-face interaction. To the extent that people take advantage of interac
tivity to lower their computational burden, this means that studying language in dialogue
might cause us to underestimate rather than overestimate language users’ perspective-
taking abilities.
The search for mechanisms underlying perspective taking in language use has led to the
identification of a variety of strategies that work as a proxy for perspective taking. If lan
guage use rests on an almost ad hoc collection of strategies, how do language users know
which strategies are appropriate in which situation? Perhaps this approach is merely
shifting the explanatory burden rather than resolving it. It is indeed the case that this
type of approach leaves much to be explained. However, it at least points a finger in the
direction where such explanations could be profitably sought: namely, in processes of so
cial development and acculturation. Our communicative skills are honed over a lifetime of
experience with various communicative partners in various social settings. One well-
known hallmark of expert skill is that it involves the appropriate structuring of a database
of knowledge that enables experts to see underlying patterns that are hidden (p. 108) from
the novice and to access relevant information in a timely manner. Experts are thus better
at recognizing particular classes of situations and have developed heuristics that enable
processing that is efficient and generally accurate but subject to systematic biases. When
an expert categorizes a situation as a new instance of an old problem, this results in the
obligatory retrieval of solutions and processing strategies that have proven effective in
the past (Logan, 1988). If the tricks that we learn are good enough for understanding one
another—or at least, to comfort us with the illusion of doing so—then we should not fault
ourselves for using them.
References
Alibali, M. W., Heath, D. C., & Myers, H. J. (2001). Effects of visibility between speaker
and listener on gesture production: Some gestures are meant to be seen. Journal of Mem
ory and Language, 44, 169–188.
Allopenna, P. D., Magnuson, J. S., & Tanenhaus, M. K. (1998). Tracking the time-course of
spoken word recognition using eye movements: Evidence for continuous mapping models.
Journal of Memory and Language, 38, 419–439.
Page 15 of 21

ception
Apperly, I. A., Carroll, D. J., Samson, D., Humphreys, G. W., Qureshi, A., & Moffitt, G.
(2010). Why are there limits on theory of mind use? Evidence from adults’ ability to fol
low instructions from an ignorant speaker. Quarterly Journal of Experimental Psychology,
63, 1201–1217.
Ariel, M. (1988). Referring and accessibility. Journal of Linguistics, 24, 65–87.
Arnold, J. E., Kahn, J., & Pancani, G. (2012). Audience design affects acoustic reduction
via production facilitation. Psychonomic Bulletin & Review, 19, 505–512.
Arnold, J. E., Wasow, T., Asudeh, A., & Alrenga, P. (2004). Avoiding attachment ambigui
ties: The role of constituent ordering. Journal of Memory and Language, 51, 55–70.
Bard, E. G., Anderson, A. H., Sotillo, C., Aylett, M., Doherty-Sneddon, G., & Newlands, A.
(2000). Controlling the intelligibility of referring expressions in dialogue. Journal of Mem
Barr, D. J. (2004). Establishing conventional communication systems: Is common knowl

edge necessary? Cognitive Science, 28, 937–962.
Barr, D. J. (2008a). Analyzing “visual world” eyetracking data using multilevel logistic re
gression. Journal of Memory and Language, 59, 457–474.
Barr, D. J. (2008b). Pragmatic expectations and linguistic evidence: Listeners anticipate

but do not integrate common ground. Cognition, 109, 18–40.
Barr, D. J. (2014). Visual world studies of conversational perspective taking: Similar find
ings, diverging interpretations. In P. Pyykkönen-Klauck & M. W. Crocker (Eds.), Visually
situated language comprehension. Manuscript in progress.
Barr, D. J., Jackson, L., & Phillips, I. (2014). Using a voice to put a name to a face: The
psycholinguistics of proper name comprehension. Journal of Experimental Psychology:
General, 143, 404–413.
Barr, D. J., & Keysar, B. (2002). Anchoring comprehension in linguistic precedents. Jour
nal of Memory and Language, 46, 391–418.
Barr, D. J., & Keysar, B. (2005). Making sense of how we make sense: The paradox of ego
centrism in language use. In H. L. Colston & A. N. Katz (Eds.), Figurative language com
prehension: Social and cultural influences (pp. 21–41). Mahwah, NJ: Erlbaum.
Barr, D. J., & Keysar, B. (2006). Perspective taking and the coordination of meaning in lan
guage use. In M. J. Traxler & M. A. Gernsbacher (Eds.), Handbook of psycholinguistics
(2nd ed., pp. 901–938). Amsterdam: Elsevier.
Page 16 of 21

ception
Bavelas, J. B., Gerwing, J., Sutton, C., & Prevost, D. (2008). Gesturing on the telephone:
Independent effects of dialogue and visibility. Journal of Memory and Language, 58, 495–
520.
Begeer, S., Malle, B. F., Nieuwland, M. S., & Keysar, B. (2010). Using theory of mind to
represent and take part in social interactions: Comparing individuals with high-function
ing autism and typically developing controls. European Journal of Developmental Psychol
ogy, 7, 104–122.
Bell, A., Brenier, J. M., Gregory, M., Girand, C., & Jurafsky, D. (2009). Predictability effects
on durations of content and function words in conversational English. Journal of Memory
and Language, 60, 92–111.
Brennan, S. E., & Clark, H. H. (1996). Conceptual pacts and lexical choice in conversa
tion. Journal of Experimental Psychology: Learning, Memory, & Cognition, 22, 1482–1493.
Brown, P. M., & Dell, G. S. (1987). Adapting production to comprehension: The explicit
mention of instruments. Cognitive Psychology, 19, 441–472.
Brown-Schmidt, S., Gunlogson, C., & Tanenhaus, M. K. (2008). Addressees distinguish

shared from private information when interpreting questions during conversation. Cogni
tion, 107, 1122–1134.
Brown-Schmidt, S., & Hanna, J. E. (2011). Talking in another’s shoes: Incremental per
spective-taking in language processing. Dialogue and Discourse, 2, 11–33.
Chafe, W. L. (1976). Givenness, contrastiveness, definiteness, subjects, topics, and point

of view. In C. N. Li (Ed.), Subject and topic (pp. 25–55). New York: Academic Press.
Clark, H. H., & Carlson, T. B. (1981). Context for comprehension. In J. Long & A. Badde
ley (Eds.), Attention and performance IX (pp. 313–330). Hillsdale, NJ: Erlbaum.
Clark, H. H., & Krych, M. A. (2004). Speaking while monitoring addressees for under
standing. Journal of Memory and Language, 50, 62–81.
Clark, H. H., & Marshall, C. R. (1981). Definite reference and mutual knowledge. In A. K.
Joshe, B. L. Webber, & I. A. Sag (Eds.), Elements of discourse understanding (pp. 10–61).
Cambridge, UK: Cambridge University Press.
Clark, H. H., & Murphy, G. L. (1982). Audience design in meaning and reference.
(p. 109)
J. Le Ny & W. Kintsch (Eds.), Language and comprehension (pp. 287–299). Amsterdam:

North Holland Publishing.

22, 1–39.
Page 17 of 21

ception
Cohen, A. A., & Harrison, R. P. (1973). Intentionality in the use of hand illustrators in
face-to-face communication situations. Journal of Personality and Social Psychology, 28,
276–279.
Epley, N., Keysar, B., van Boven, L., & Gilovich, T. (2004a). Perspective taking as egocen
tric anchoring and adjustment. Journal of Personality and Social Psychology, 87, 327–339.
Epley, N., Morewedge, C. K., & Keysar, B. (2004b). Perspective taking in children and
adults: Equivalent egocentrism but differential correction. Journal of Experimental Social
Ferreira, V. S., & Dell, G. S. (2000). Effect of ambiguity and lexical availability on syntac
tic and lexical production. Cognitive Psychology, 40, 296–340.
Fowler, C. A., & Housum, J. (1987). Talkers’ signaling of “new” and “old” words in speech
and listeners’ perception and use of the distinction. Journal of Memory and Language, 26,
484–504.
Fukumura, K., & van Gompel, R. P. G. (2012). Producing pronouns and definite noun
phrases: Do speakers use the addressee’s discourse model? Cognitive Science, 36, 1289–
1311.
Fussell, S. R., & Krauss, R. M. (1992). Coordination of knowledge in communication: Ef

fects of speakers’ assumptions about others’ knowledge. Journal of Personality and Social
Galati, A., & Brennan, S. E. (2010). Attenuating information in spoken communication:

For the speaker, or for the addressee? Journal of Memory and Language, 62, 35–51.
Gann, T. M., & Barr, D. J. (2012). Speaking from experience: Audience design as expert
performance. Language and Cognitive Processes. Manuscript in press.
Gundel, J. K., Hedberg, N., & Zacharski, R. (1993). Cognitive status and the form of refer
ring expressions in discourse. Language, 69, 274–307.
Hanna, J. E., Tanenhaus, M. K., & Trueswell, J. C. (2003). The effects of common ground
and perspective on domains of referential interpretation. Journal of Memory and Lan
guage, 49, 43–61.
Heller, D., Grodner, D., & Tanenhaus, M. K. (2008). The role of perspective in identifying
domains of reference. Cognition, 108, 831–836.
Hogg, J. (1824/1997). The Private Memoirs and Confessions of a Justified Sinner. Ware,
UK: Wordsworth Classics.
Horton, W. S., & Gerrig, R. J. (2005). The impact of memory demands on audience design
during language production. Cognition, 96, 127–142.
Page 18 of 21

ception
Horton, W. S., & Keysar, B. (1996). When do speakers take into account common ground?
Cognition, 59, 91–117.
Jurafsky, D. (2003). Probabilistic modeling in psycholinguistics: Linguistic comprehension

and production. In R. Bod, J. Hay, & S. Jannedy (Eds.), Probabilistic linguistics (pp. 39–95).
Cambridge, MA: MIT Press.
Kahneman, D., & Frederick, S. (2002). Representativeness revisited: Attribute substitu

tion in intuitive judgment. In T. Gilovich, D. Griffin, & D. Kahneman (Eds.), Heuristics and
biases: The psychology of intuitive judgment (pp. 49–81). Cambridge, UK: Cambridge Uni
versity Press.
Keysar, B. (1997). Unconfounding common ground. Discourse Processes, 24, 253–270.
Keysar, B., & Barr, D. (2002). Self-anchoring in conversation: Why language users do not
do what they “should”. In T. Gilovich, D. Griffin, & D. Kahneman (Eds.), Heuristics and bi
ases: The psychology of intuitive judgment (pp. 150–166). Cambridge, UK: Cambridge
University Press.
Keysar, B., Barr, D. J., Balin, J. A., & Brauner, J. S. (2000). Taking perspective in conversa
tion: The role of mutual knowledge in comprehension. Psychological Science, 11, 32–38.
Keysar, B., Barr, D. J., Balin, J. A., & Paek, T. S. (1998). Definite reference and mutual
knowledge: Process models of common ground in comprehension. Journal of Memory and
Language, 39, 1–20.
Keysar, B., & Henly, A. (2002). Speakers’ overestimation of their effectiveness. Psychologi
cal Science, 13, 207–212.
Keysar, B., Lin, S., & Barr, D. J. (2003). Limits on theory of mind use in adults. Cognition,
89, 25–41.
Kirsh, D., & Maglio, P. (1994). On distinguishing epistemic from pragmatic action. Cogni
tive Science, 18, 513–549.
Kraljic, T., & Brennan, S. E. (2005). Prosodic disambiguation of syntactic structure: For
the speaker or for the addressee? Cognitive Psychology, 50, 194–231.
Kraut, R. E., Lewis, S. H., & Swezey, L. W. (1982). Listener responsiveness and the coordi
nation of conversation. Journal of Personality and Social Psychology, 43, 718–731.
Kuhlen, A. K., & Brennan, S. E. (2013). Language in dialogue: when confederates might
be hazardous to your data. Psychonomic Bulletin & Review, 20, 54–72.
Lieberman, P. (1963). Some effects of semantic and grammatical context on the produc
tion and perception of speech. Language and Speech, 63, 172–187.
Page 19 of 21

ception
Lin, S., Keysar, B., & Epley, N. (2010). Reflexively mindblind: Using theory of mind to in
terpret behavior requires effortful attention. Journal of Experimental Social Psychology,
46, 551–556.
Logan, G. D. (1988). Toward an instance theory of automatization. Psychological Review,

95, 492–527.
Marslen-Wilson, W. D. (1987). Functional parallelism in spoken word-recognition. Cogni

tion, 25, 71–102.
McNeill, D. (1996). Hand and mind: What gestures reveal about thought. Chicago: Uni
versity of Chicago Press.
Mol, L., Krahmer, E., Maes, A., & Swerts, M. (2011). Seeing and being seen: The effects
on gesture production. Journal of Computer-Mediated Communication, 17, 77–100.
Nadig, A. S., & Sedivy, J. C. (2002). Evidence of perspective- taking constraints on

children’s on-line reference resolution. Psychological Science, 13, 329–336.
Nilsen, E., & Graham, S. (2009). The relations between children’s communicative per
spective-taking and executive functioning. Cognitive Psychology, 58, 220–249.
Pickering, M. J., & Garrod, S. (2004). Toward a mechanistic psychology of dialogue. Be
Rossnagel, C. (2000). Cognitive load and perspective taking: Applying the automatic-con
trolled distinction to verbal communication. European Journal of Social Psychology, 30,
429–445.
Savitsky, K., Keysar, B., Epley, N., Carter, T., & Swanson, A. (2011). The closeness-commu
nication bias: Increased egocentrism among friends versus strangers. Journal of Experi
mental Social Psychology, 47, 269–273.
Schober, M. F., & Clark, H. H. (1989). Understanding by addressees and overhear

(p. 110)
ers. Cognitive Psychology, 21, 211–232.
Simon, H. A. (1996). The sciences of the artificial. Cambridge, UK: MIT Press.
Tversky, A., & Kahneman, D. (1974). Judgment under uncertainty: Heuristics and biases.
Science, 185, 1124–1131.
Wardlow Lane, L., & Groisman, M., & Ferreira, V. S. (2006). Don’t talk about pink ele
phants! Psychological Science, 17, 273–277.
Wardlow Lane, L., & Liersch, M. J. (2012). Can you keep a secret? Increasing speakers’
motivation to keep information confidential yields poorer outcomes. Language and Cogni
tive Processes, 27, 462–473.
Page 20 of 21

ception
Wu, S., & Keysar, B. (2007). The effect of culture on perspective taking. Psychological
Science, 18, 600–606.
Notes:
(1) . In a sense, the Double might be seen as one of the many manifestations of the Char
latan, inasmuch as it substitutes its own knowledge for that of the interlocutor.
(2) . Arguably, disparity in access to channels has recently become somewhat more com
mon than it has been traditionally, with the rise of video conferencing software like Skype
in combination with the variation in the availability of a front-facing webcam across mo
bile devices.
(3) . Representational (iconic) gestures are those that iconically depict their referents, in
contrast to beat gestures, used for emphasis, and deictic gestures that direct a listener’s
attention (see McNeill, 1996).
Dale J. Barr
Dale J. Barr, Institute of Neuroscience and Psychology, University of Glasgow
Page 21 of 21

Hand and Facial Gestures in Conversational Interaction
Janet Bavelas , Jennifer Gerwing , and Sara Healing


Conversational hand and facial gestures are an integral part of language use in face-to-
face dialogue. Extensive research shows that conversational hand gestures are tightly
synchronized with words to demonstrate anything that can be represented (directly or
metaphorically) as size, shape, position, or action and that they are highly sensitive to the
immediate communicative context. Although research on conversational (nonemotional)
facial gestures is much more limited, they, too, are precisely timed with words and con
text to demonstrate anything that can be represented by a facial configuration—whether
in the past, present, or future and even hypothetical or metaphorical. Both hand and fa
cial gestures can also function as collateral communication (meta-communication). This
chapter includes theoretical, methodological, and technical requirements for studying
these gestures in conversational interaction.
Keywords: Conversational hand gestures, conversational facial gestures, language use, dialogue, conversational
interaction, collateral communication
In face-to-face dialogue, interlocutors spontaneously combine words, prosody, hand ges

tures, and facial gestures to produce rapidly changing integrated messages (Bavelas &
Chovil, 2000, 2006). This chapter presents a framework for studying these integrated
messages as social communication and then an overview of what experimental research
is discovering about how the visible elements of communication contribute to language
use in conversational interaction.
Looking Closely at Conversational Hand and

Facial Gestures
To appreciate the dynamic, fleeting quality of conversational hand and facial gestures, it
is helpful to see (rather than to read descriptions of) as many examples as possible. This
chapter uses sequential frame shots from our research videos, a format that may help the
reader imagine how these actions actually looked (and may lead to noticing them more of
Page 1 of 32

ten in everyday life). In his advice on how to get the most out of examples presented in
print form, McNeill (1985, p. 352) urged readers to act out the examples in order to expe
rience the single coordinated action of motor–speech synchrony that the examples are in
tended to show.
Fig. 7.1 While describing a scene from the movie

Shrek 2, the speaker made the hand gesture in frame
2 and the facial gestures in frames 2 and 3. The
three frame shots cover a 6.75-second period. (The
face of the addressee, who was sitting across a table
from the speaker, is inset at the upper right.)
Figure 7.1 presents three frames in a 6.75-second video excerpt (from Bavelas, Gerwing,
& Healing, 2014). The speaker’s task was to watch several scenes from the animated
movie Shrek 2 and then retell them to an addressee. She had just finished describing an
attack on Shrek by the cat character (Puss in Boots). In frame 1, she began to describe
the next scene, in which Shrek picked up the defeated cat by the back of the neck, lifting
him close to his face; the cat started begging for his life. By frame 2, she had raised her
hand up in front of her own face, pinching her first two fingers and thumb as if lifting and
then suspending the cat, presenting an image of Shrek holding the cat in front of his
face. This hand gesture demonstrated information that was either left ambiguous
(p. 112)
in her words (e.g., the shape of Shrek’s hand) or was missing from her words (e.g., that
Shrek suspended the cat in the air at face level).
The speaker’s face was also active. In frame 2, while saying “neck,” she demonstrated
Shrek’s slightly gloating expression as he held the cat at his mercy. In frame 3, she began
to quote the cat (“Oh I’m so sorry!”) while wrinkling her forehead and pursing her lips,
creating a worried and pleading look. Each of these facial gestures portrayed one of the
characters and illustrated her narrative for the addressee. Like her hand gestures, each
added relevant details that her words did not convey (e.g., Shrek’s expression as he held
the cat, then the cat’s apprehension).
Some Definitions and Assumptions

The close analysis of the three frames in Figure 7.1 illustrates the defining characteristics
of conversational hand and facial gestures. Each gesture complemented the speaker’s
words and was timed precisely with the relevant words. Both hand and facial gestures oc
cur as part of a spontaneous stream of speech, and they are tightly synchronized in both
meaning and timing with the accompanying words. The speaker’s hand gesture in frame 2
began and ended exactly with the words it was related to. The face is highly mobile, so
speakers can make facial gestures even more quickly than hand gestures—at the rate of a
Page 2 of 32

single word or even syllable. In fact, Figure 7.1 includes only two of the speaker’s facial
gestures in the excerpt; there were two more (one before and one after frame 3). Alto
gether, in 4 seconds, she made four different facial gestures related to what she was say
ing.
The synchrony between visible and audible elements of communication results in efficien
cies that could not be matched by words alone. Even if words could convey exactly the
same information, the speaker would be limited to presenting it in a sequential, linear
manner (e.g., “Shrek picks him up by the nape of his neck with his thumb and two fingers
and holds him up at eye level in front of his face, and looks at him in sort of a gloating
way.”) The additional words needed to convey information that had been conveyed by the
speaker’s gestures would require a much longer description than the original integrated
message.
This chapter treats hand and facial gestures—but not all nonverbal behaviors—as part of
language use in face-to-face dialogue. There are several types of nonverbal behaviors that
are not part of language use, including those that are involuntary or reflexive (e.g., blink
ing), instrumental (e.g., turning a door key), static postures (e.g., sitting with crossed
arms), or behaviors that are just as likely to occur when no one else is present (e.g.,
squinting in bright light). In addition, the hand and facial gestures discussed in detail in
this chapter are those that occur within the context of conversation, not those occurring
in nonspeaking contexts such as hand signals (e.g., the hitch-hiking sign) or still photos.
The Role of Context in Meaning
Fig. 7.2 Image of the nested contexts that succes

sively determine the meaning of words, as well as
hand and facial gestures in face-to-face dialogue.
Page 3 of 32

In Figure 7.1, the meaning of each of the speaker’s hand and facial gestures depended on
the set of nested contexts in which they occurred. Figure 7.2 presents one way to visual
ize these contexts. The setting was a university psychology experiment for class credit.
The interlocutors were unacquainted students. Their task was for the speaker to tell the
addressee the scenes from Shrek 2. Their dialogue so far consisted of their conversation
up to this point, which culminated in a particular microsocial moment in the dialogue
(Bavelas, 2007), for example, demonstrating Shrek picking up the cat. In a different set of
nested contexts, a hand and facial gesture combination similar to the one in frame 2
(p. 113) would have an entirely different meaning: imagine the meaning of these gestures
if she had been in a pub talking with friends about how much fun it would be to drop a
water balloon on her boss.
Notice that the same layers of context also determine the meaning of any word or phrase
in a dialogue. For example, in this particular context, “the cat” unambiguously meant the
Puss in Boots character rather than any other feline. A meaning-based analysis of hand
gestures, facial gestures, and words must always take into account these contextual lev
els—which is what everyone does every day, in every dialogue.
Hand and Facial Gestures as Demonstrations

We have proposed (Bavelas, Gerwing, Sutton, & Prevost, 2008; Bavelas et al., 2014) that
conversational hand and facial gestures fit into a larger view of language, communica
tion, and the general theory of signs initiated by Charles Peirce (e.g., Buchler, 1940).
Peirce distinguished between symbols (e.g., words), indices (e.g., a weather vane), and
icons (e.g., a picture). Clark and Gerrig (1990; Clark, 1996, chapter 6) put these three
categorical distinctions into a social context by emphasizing how they function in every
day language use: with symbols, a speaker can describe something to an addressee by us
ing words in their conventional, arbitrarily established meanings. Using indices, a speak
er can also indicate something for the addressee by pointing at it, either physically or
with a demonstrative pronoun (e.g., “I want that one”). Using icons, a speaker can demon
strate something by presenting a visual or auditory version of it. Demonstrations include
direct quotations (Clark & Gerrig, 1990), metaphors, conversational facial gestures, and
conversational hand gestures (Bavelas et al., 2014).
Demonstrations share several defining features. Rather than having standardized mean
ings, they “work by enabling others to experience what it is like to perceive the things de
picted” (Clark & Gerrig, 1990, p. 765). They do so by resembling their referent in some
way. Although a demonstration resembles its referent, it is also a selective transformation
of the literal or actual properties of the referent; it selects some features and leaves out
others. The gestural demonstration in frame 2 of Figure 7.1 presented the speaker’s ver
sion of the scene she had seen. Her hand resembled Shrek’s hand holding the cat, but it
was not an exact replication. She selected (and emphasized) the image of Shrek holding
the cat between his fingers in front of his face (but did not include Shrek bouncing his
hand slightly, as shown in the movie). In the same frame, her face resembled how Shrek
Page 4 of 32

looked, but she selected a mildly gloating and victorious version and left out the anger
that was clearly present in his face in the actual movie scene.
A Functional, Communicative Approach

This chapter concerns the communicative functions of hand and facial gestures in social
interaction. First, the focus is on function rather than type. That is, we will be asking
“What is this gesture doing in this dialogue?” rather than “What category does this ges
ture go in?” A functional approach leads to terms such as demonstrated, portrayed, pre
sented, complemented, and marked, all of which direct attention to the function of the
gesture in its microsocial context. Although hand gestures can be classified as iconics or
illustrators, such categorization abstracts them from their interactional context. Second,
our focus on communicative function steers away from questions such as “What does this
facial gesture tell us about what the individual is feeling?” Inferences about the cognitive
or emotional processes associated with a hand or facial gesture are distinct from examin
ing what it is doing in the interaction. Functions are not exclusive: actions that are com
municative may have cognitive or emotional functions as well. Although there is no logi
cal reason to privilege one function over another (much less to set them in opposition),
they do require quite different research methods. In (p. 114) any case, contemporary re
search has moved past the adversarial debate about whether gestures communicate to
the more interesting question of how they do so, which opens up entirely new possibilities
for research and theorizing.
Methodological Issues
As implied in the previous paragraph, experiments on the communicative functions of
conversational hand and facial gestures require a research design in which one condition
is a face-to-face dialogue. The next sections explicate this requirement and critically ex
amine some conventional barriers to meeting it.
Conversational Interaction as Face-to-face Dialogue
Conversational hand and facial gestures serve their communicative functions within face-
to-face dialogues. Clark (1996, pp. 9–10) specified the defining characteristics of face-to-
face dialogue, which stress two key principles: first, a face-to-face dialogue is unmediated,
which means that the interlocutors are in the same physical environment, able to see and
hear each other, and able to produce and receive at once and simultaneously. Second, the
interlocutors can freely interact and collaborate with each other, moment by moment. In
Clark’s terms, they are acting as themselves (e.g., not as a confederate), determining
what actions to take when (e.g., without a script), and acting extemporaneously, in real
time. These features provide essential criteria for research designs that will yield data
relevant to face-to-face dialogue.
Page 5 of 32

True Dialogues Versus Quasi-Dialogues
Traditionally, experimental social psychologists have shown a methodological preference

for what Bavelas and Healing (2013) called quasi-dialogues, in which the participant in
teracts with a confederate, the experimenter, or another participant who is acting under
instructions that constrain his or her actions. These interactions do not fit Clark’s (1996)
definition of true dialogues because one of the participants is not acting as him- or her
self, is not determining his or her own actions, and is not able to extemporize. This sec
tion explores the reasons underlying the choice of quasi-dialogues. (See also Kuhlen &
Brennan, 2013.)
First, many scholars have pointed out that social psychology has remained largely the
study of the individual rather than the dyad (e.g., Bavelas, 2005, 2007; Clark, 1985;
Danziger, 1990; Sears, 1951; Thibaut & Kelley, 1959). As Danziger pointed out, the line
between the person and the environment has historically been drawn tightly around the
individual, so that the interlocutor in a dialogue is “outside” this line—just another part of
the environment. However, when the interest is in how the speaker’s actions function in
dialogue, then the unit of study must include the dialogue itself.
Even social psychologists with an interest in communication during social interaction of
ten use quasi-dialogues, perhaps because of an erroneous application of the principle of
reductionism. Luria (2004, p. 537) described reductionism as the assumption that “the ba
sic goal of science is to reduce complex phenomenon to separate simple parts, and that
such reduction provides significant explanations of phenomena.” By this principle, a true
dialogue is a complex phenomena that can be better studied by reducing it to one individ
ual in a controlled interaction. However, Luria went on to question this assumption:
Although the philosophy of reductionism was accepted as a general principle in

the natural sciences and psychology, there are grounds to suppose it may be false.
To study a phenomenon, or an event, and to explain it, one has to preserve all its
basic features: one must be able to describe their rules and their mechanisms
without the loss of any [characteristics of the phenomenon]. It can easily be seen
that reductionism may very soon conflict with this goal. (p. 537)
In their critical examination of reductionism, Reber, Allen, and Reber (2009, p. 663) point
ed out that the problem is not with reductionism as a scientific principle per se but with
its overgeneralization to all topics in all fields. Reducing a complex phenomenon to its
smallest possible units makes sense if the smaller units are the topic of interest. But
studying the smaller units will not be informative if a more complex phenomenon of inter
est is lost in the process. For example, if one is interested in learning about salt (sodium
chloride), then studying the properties of sodium (a metal so reactive that it almost ex
plodes in water) and chlorine (a poison gas) would not be very enlightening. If the phe
nomenon of interest is social interaction in dialogue, then studying individuals is similarly
unenlightening.
Page 6 of 32

For experimental researchers, reductionism also underlies a second justification for the
use of quasi-dialogues, namely, the assumption that the complexity of a real dialogue pre
cludes experimental (p. 115) control. The fear is that if one interlocutor acts spontaneous
ly and the other is also free to react spontaneously, experimental control will be lost and
the ensuing variability will overwhelm any differences between conditions. However, face-
to-face dialogue is inherently well organized. In everyday life, even strangers regularly
accomplish complex tasks together in dialogue (e.g., giving directions, discussing pur
chases, exchanging technical information, making a diagnosis, or conducting an inter
view). The numerous experiments reviewed in this chapter demonstrate that allowing two
individuals to respond extemporaneously to each other within an assigned task does not
lead to unanalyzable data. Far from being chaotic, dialogues are intrinsically orderly;
those of us who microanalyze dialogue are constantly awed by its precision and orderli
ness.
Some Technical Issues
As a practical matter, observing the precision of face-to-face dialogue requires digitized

video with both interlocutors on screen at all times (e.g., using split-screen) plus software
for frame-by-frame analysis and annotation (e.g., ELAN, http://tla.mpi.nl/tools/tla-tools/
elan/; Wittenburg, Brugman, Russel, Klassman, & Sloetjes, 2006). Using comprehensive
operational definitions, analysts can capture the details of dialogue with high interanalyst
agreement. (Gullberg’s, 2010, informative outline of methodological concerns in gesture
research in second language acquisition is relevant to gesture research more widely, es
pecially pp. 86–91.)
Conversational Hand Gestures

Imagine you are quietly observing two people having a conversation in a restaurant. You
can’t hear their speech, but you can see that they make a variety of hand movements. Not
all of these movements are conversational hand gestures. First, some are instrumental ac
tions involving objects (e.g., cutting their food with a knife and fork) or their own body
(e.g., pushing hair behind an ear). These actions are often called object- and self-oriented
adaptors, and when they serve purely instrumental functions, they have no communica
tive meaning. Second, you may see one person point to a corner of the restaurant,
prompting the other diner to look in that direction as well. These pointing (indicating or
deictic) gestures function to establish a joint focus of attention. Third, one person at the
table catches the server’s eye and holds one palm up while pretending to write on it with
the other hand. If you are familiar with the cultural norms of the gesturer, you under
stand this movement as a request for the bill. If the server nods and then comes to the ta
ble with the bill, the gesture was effective. These hand signals, which are usually called
emblems, have stereotypic forms and culturally specified meanings so they can function
as messages outside of conversation. Aside from instrumental actions, pointing, and em
blems, you see both diners moving their hands as they talk. These improvised movements
seem to mean something, but you won’t know their meaning without hearing the diners’
Page 7 of 32

words because these are likely to be conversational (or co-speech) hand gestures, as de
fined in the opening paragraphs of this chapter. (For a thorough review of the history and
variety of gestures, see Kendon, 2004.)
Conversational hand gestures have at least three phases (Kendon, 1980). Most gestures
begin with a preparation phase, lifting or moving the hands into position. The meaningful
depictive action that follows is the actual gesture stroke. After the stroke, there is often
some retraction, as the hands move back to a resting position or prepare for the next ges
ture. Immediately before or after the stroke, there may also be a brief hold, in which the
hands pause in position. These sequences occur smoothly, in a matter of seconds or less,
and they are tightly synchronized with the stream of speech.
The following sections focus, first, on what kinds of information the gesture stroke can
contribute to a dialogue and, second, on how social factors affect them in a dialogue. The
studies referred to in this section were experiments in which at least one condition ful
filled all of the criteria for a face-to-face dialogue.
What Can Hand Gestures Demonstrate?
Hand gestures are suited to depicting some kinds of information (e.g., shape, size, or di
rection) and not others (e.g., color, flavor, or names). Still, as will be seen in this section,
the variety of possibilities is surprising. Most of the examples are from spontaneous con
versational hand gestures made during laboratory experiments using dialogic tasks de
signed to elicit gestures (e.g., retelling a narrative, describing an object, or explaining to
an addressee how to do something). They are highly varied because each was a demon
stration specific to a particular micromoment in a dialogue. Recall that demonstrations do
not have standardized conventional meanings. Instead, they resemble their referent and
require the addressee to recognize that this particular hand action, in this context, is a
selective version of the referent.
(p. 116) Hand Actions

Gestures are well suited to demonstrating hand actions that are real or hypothetical; lit
eral or metaphorical; done in the past, present, or future; and representing one’s own or
anyone’s hands. In an experiment by Gerwing and Bavelas (2004), speakers first used a
set of objects and later demonstrated their actions to an addressee (without the object).
One of these objects was a woven tube-like toy that trapped the users’ index fingers when
they tried to pull them out. When telling their addressee about it, speakers usually ges
tured their actions, often by holding their index fingers in a straight, horizontal line with
the tips touching, so the addressee could imagine a tube holding the fingers in this posi
tion. Sometimes they pulled their fingers back and forth to demonstrate how the fingers
were (or could have been) trapped in the tube. Another object was a propeller-like toy
that could be launched by spinning the upright stick between vertical palms. Speakers
later gestured how they had launched the toy by moving one vertical palm rapidly past
the other. In a study by Furuyama (2000), speakers taught addressees how they should
fold an origami figure—without any paper. The teachers used gestures to demonstrate
Page 8 of 32

making the various folds. Thus, the addressee had to see the hands as though they were
manipulating a piece of paper that was not actually there.
Speakers also demonstrate hand actions other than their own, as when the speaker in
Figure 7.1 used her hand to represent Shrek’s hand. The addressee had to see the
speaker’s hand as if it were Shrek’s. In another experiment (Bavelas et al., 2008), speak
ers had to describe a picture of a woman holding a fan in front of her chest. Speakers of
ten included a gestural demonstration of the fan, as if holding a fan themselves.
Finally, gestures of hand actions can be metaphorical. In an experiment by Rowbotham,

Holler, Lloyd, and Wearden (2011), interviewers asked participants to describe a recent
pain experience. Although pain is an intangible, subjective experience, the participants
often used gestures to demonstrate various aspects of their pain. One person said, “It
feels like they’re [sic] just sat there with like a hammer, hitting me, that’s how it feels.”
During the italicized words, the speaker gestured as if holding a hammer and hitting her
self repeatedly on her temple (p. 8). Her demonstration required the addressee not only
to “see” this action but also to appreciate that the headache felt as if the hammer were
pounding her temple.
Spatial Features
Fig. 7.3 While describing a propeller toy to her ad

dressee, the speaker made three related hand ges
tures. The three frame shots cover a 1.4-second peri
od. (The addressee was sitting directly across from
the speaker and filmed in split screen, indicated here
by a dotted line.)
Hands are not limited to portraying hand actions. Speakers can use their gestures to
demonstrate physical features of an object such as its shape, size, or orientation. For ex
ample, some speakers in the Gerwing and Bavelas (2004) experiment used their hands to
depict the propeller toy itself. Figure 7.3 presents one such pair of gestures: in frame 1,
precisely as the speaker said “a rod,” she began to bring her upright index finger straight
down, drawing a vertical line for the rod. As she said “then two” (in frame 2), she drew a
horizontal line to represent the propellers at the top of the rod. In frame 3, she held her
gesture at the end of the horizontal line while she finished saying “propellers.”
In other experiments, speakers have used gestures to demonstrate the shape of knotted
pipes, dome-shaped roofs, or bridges (Holler & Stevens, 2007; see their Figures 1–3, pp.
17–19), the evolving shape of an origami figure (Furuyama, 2000), and the shape and size
of distinctive features of an unusual dress (Bavelas et al., 2008; see their Figures 2–4, pp.
504–507).
Page 9 of 32

Spatial gestures can also represent virtual spaces, that is, when the referent does not ac
tually exist. Bavelas, Gerwing, Allison, and Sutton (2011) asked (p. 117) participants to de
sign a floor plan together without pencil and paper. They “drew” their plans on the table
with their gestures, demonstrating the location, size, and shape of virtual rooms, which
did not actually exist.
A single gesture can be dense with information, that is, demonstrating more than one fea
ture of the referent (Gerwing & Allison, 2011). In several of the examples described in
this section, gestures depicting shape also conveyed accurate information about the
referent’s size (e.g., Holler & Stevens, 2007; Bavelas et al., 2008) or other features such
as location (Gerwing & Allison, 2009, see their Figure 1, p. 313) or orientation.
Motion
Hands can move in almost any direction and are well suited to demonstrating aspects of
motion. In an experiment by Parrill (2008), participants retold a cartoon in which a cat
had swallowed a bowling ball, which then began rolling down a hill, transporting the
helpless cat along with it. Speakers often abstracted and selected aspects of this motion
to demonstrate with their gestures. For example, Parrill’s figures illustrate gestures that
depicted the direction the cat rolled (moving the hand diagonally downward; her Figure
1, p. 287), the rolling manner of the movement (tracing a circular motion in one place;
her Figure 4, p. 289), as well as both direction and manner simultaneously (tracing a cir
cular motion moving downwards; her Figure 5, p. 290). In Özyürek’s (2002) experiment,
participants used gestures to show the direction of motion as characters went into or out
of buildings or across a street (Özyürek, Appendix, p. 703). Gestures depicting motion can
also be metaphorical; for example, representing the lessening of pain by moving the hand
downward (Rowbotham et al., 2011, Table 4).
Relative Positions
Some gestures demonstrate both the referent and its relationship to something else, such
as another gesture, the gesturer’s body, or a nearby object. In discussions about floor
plans, Gerwing and Allison (2009) found two-handed gestures that demonstrated the rela
tive positions of two or more rooms. For example, one participant gestured and held the
location of the entrance with her right index finger while showing the proposed location
of the kitchen with her left hand. She then held her left hand in position while moving her
right hand over to the proposed location of the living room (Gerwing & Allison, 2009,
their Figure 1, p. 313).
The rest of a speaker’s body provides an additional resource: speakers can use the rela
tionship between their hand(s) and body to demonstrate relative positions. In Figure 7.1
of this chapter, the speaker used the relationship between her gesture and her face to de
pict where Shrek suspended the cat while listening to his pleading. In another experi
ment (Bavelas et al., 2008), speakers demonstrated the distinctive features of a dress.
Men as well as women often placed their gestures in relation to their own body. Some
outlined the bodice of the dress as a deep “V” over their own chest (Bavelas et al., 2008,
their Figure 4, p. 507) or showed the unusually wide shape of the skirt by extending their
Page 10 of 32

hands out from their own waist (Gerwing & Allison, 2011, their Figure 2, p. 311). In de
scriptions of their pain experiences, Rowbotham et al.’s (2011) participants also placed
their gestures in relation to their body; for example, one participant put her palms over
her lower stomach while leaning forward to show how she positioned herself as a reac
tion to the pain (Rowbotham et al., 2011, table 4, p. 17). Gestures can also build on the lo
cation of previous gestures, as if the previous gesture left a “residue” behind that could
be used by later gestures. For example, in Figure 7.3 of this chapter, notice that the par
ticipant gestured the propellers directly over where she had previously placed the top of
the rod. Healing and Gerwing (2012) analyzed the gestures of speakers who described a
picture of a maze-like route. They found that speakers often linked contiguous sections of
the route by gesturing one section, pausing, and then beginning the next gesture from
the same position.
Gestures and Words
The previous examples show that gestures are particularly suited for many—but not all—
kinds of information; for example, although they can demonstrate something, they cannot
name it. Words and gestures are useful for communicating different kinds of information,
and Bavelas and Chovil (2000, 2006) proposed that they work together to form integrated
(but not necessarily redundant) messages. A semantic feature approach (e.g., Beattie &
Shovelton, 1999) can analyze how information is distributed between words and gestures,
and it yields reliable quantitative evidence of how their integration works. One applica
tion of this approach examined how dyads used gestures and words when designing floor
plans (Gerwing & Allison, 2009). Gestures were significantly more likely to contribute
spatial information (i.e., size, (p. 118) shape, and relative location of the rooms) whereas
words contributed categorical information (e.g., the name of a room). One speaker said
“there’s a kitchen” while pointing to a space adjacent to where she had located the en
trance. The words named the room, and the gestures specified its location. To understand
the speaker’s message, the addressee had to integrate speech and gesture. A follow-up
study on dress descriptions (Gerwing & Allison, 2011) reported a similar pattern of inte
gration: in face-to-face dialogues, speakers’ gestures were significantly more likely than
their words to convey spatial information such as the size and shape of the skirt. When
describing their pain experiences, participants in Rowbothom et al. (2011) were more
likely to put information about the location and size of their pain in their gestures rather
than in their words, while putting information about the cause, the effect, and their
awareness of the pain significantly more often in their words.
Conversational Factors that Affect Gestures
The physical capacities of hands to demonstrate referents are by no means the only deter
minants of what gestures actually do in a dialogue. The form of a gesture is also deter
mined by its function in the dialogue at the precise moment it occurs. Most of the experi
ments cited in the preceding sections will reappear here because their primary purpose
Page 11 of 32

had been to show how factors in the social interaction can affect how and when the inter
locutors use gestures.
Mutual Visibility
A frequent experimental manipulation in gesture research consists of varying whether or
not the speaker and addressee are mutually visible (e.g., speaking face to face vs.
through a partition). Since Cohen and Harrison’s initial experiment (1973), researchers
have hypothesized that, if the rate of gesturing decreases when the addressee is not visi
ble, then gestures serve a communicative function; if not, then they serve a cognitive
function (e.g., helping the speaker formulate utterances).
However, there are several reasons to doubt the utility of this design for such a broad hy
pothesis. First, the results of numerous experiments on the effect of mutual visibility on
overall gesture rate have been mixed, with half finding that mutual visibility leads to a
significantly higher rate (Alibali, Heath, & Myers, 2001; Cohen, 1977; Cohen & Harrison,
1973; Emmorey & Casey, 2001; Krauss, Dushay, Chen, & Rauscher, 1995; Mol, Krahmer,
Maes, & Swerts 2009a, 2009b) and the other half finding no significant difference in over
all rate (Bavelas et al., 1992, 2008, 2014; de Ruiter, Bangerter, & Dings, 2012; Holler, Tut
ton, & Wilkin, 2011; Pine, Burney, & Fletcher, 2010; Rimé, 1982). Two of the latter studies
found both results, depending on the kind of gesture studied (Bavelas et al., 1992; de
Ruiter et al., 2012). The second problem is that all of the experiments reporting a signifi
cantly higher rate of gesture in the visibility condition were quasi-dialogues, whereas the
other group were all true dialogues. Clearly, the results of the former do not generalize to
the latter. (See Bavelas & Healing 2013, for a discussion of possible reasons.)
Finally, several experiments have shown that overall rate measures do not capture the ef
fects of mutual visibility on gesture use: speakers make qualitatively different gestures
when addressees can see them than when they cannot; see Bavelas and Healing (2013,
pp. 76-79, and Table 5). For example, mutual visibility affects whether or not gestures
convey essential information. When their addressee will see their gestures, speakers are
significantly more likely to convey information in them that does not appear in words.
When their addressee will not see their gestures, speakers’ gestures are less likely to add
any additional information; that is, they tend to be redundant with the words (e.g., Bave
las et al., 2008; de Ruiter et al., 2012). Bavelas et al. (2008) also found that speakers’ ges
tures in face-to-face conversations were significantly larger than those on the telephone.
Healing and Gerwing (2012) found that, when describing the maze-like route to a visible
addressee, speakers were significantly more likely to link sequential gestures into a cohe
sive whole, maintaining their relative positions. When their addressee could not see them,
gestures for sequential features were not linked but instead tended to pile up in one area.
Some specific gesture functions require visibility; for example, interlocutors who could
see each other were more likely to mimic the form and meaning of each other’s gestures
(Holler & Wilkin, 2011).
Page 12 of 32

Shared Social Space

Just as speakers can use their own body as a resource for contributing to gestures’ mean
ing, they can also use the space that they share with their addressee. Özyürek (2000,
2002) compared experimental conditions in which speakers were seated in different spa
tial configurations with their addressees. For example, they might be seated directly
across from the addressee or talking to (p. 119) two addressees in a triangular configura
tion (see examples in Özyürek’s, 2002, Figure 1, p. 692, and Figure 2, p. 698). When
demonstrating how a cartoon character moved from one place to another, the speakers
displayed sensitivity to how the relative seating positions of their addressees and them
selves would influence the meaning of their gestures. For example, they changed the di
rection of their gestures to match whether the gesture was moving into or out of the
shared space. Interlocutors can even share their gestural space. Furuyama’s (2000) study
on origami instruction reported that instructors and learners often collaborated with their
gestures in the same physical space, pointing to or acting on a single virtual origami fig
ure that was not actually in that space.
Common Ground
It is well established for words that speakers are sensitive to whether or not they share
common ground with their addressee (Haviland & Clark, 1974). For example, they pro
vide fuller, more informative verbal descriptions when they know the addressee is unfa
miliar with the information. Gerwing and Bavelas (2004) showed a similar effect for ges
tures by manipulating common ground between participants. Speakers described the
same object twice: once to an addressee who they knew had just manipulated the same
object and once to an addressee who they knew had not. The gestures that speakers used
to identify the object for the addressee who did not share common ground were judged to
convey more information, be more complex, or be more precise than those the speaker
used to identify the object for the addressee who did share common ground (e.g., their
Figures 3 and 4, pp. 169–170). Holler and Stevens (2007) found an effect of common
ground on the relationship between the gestures and words that speakers used to convey
the size of several disproportionately large objects. They found that speakers who de
scribed these objects to addressees who had not seen them used gestures significantly
more often than words. Furthermore, their gestures depicted the objects’ sizes more ac
curately than gestures used in the common ground condition. Conversely, when partici
pants shared common ground, speakers used words to refer to the objects significantly
more often than they used gestures.
As their dialogue unfolds, the interlocutors accumulate more common ground. Speakers
can introduce new information that, when they refer to it later, becomes given information
and part of their common ground. Gerwing and Bavelas (2004) found that gestures for an
object would change as the information depicted went from new to given. For example,
the gestures would change from large and precise to smaller and vaguer. These changes
in gestures parallel what happens with words: new words that become given change from
noun to pronoun or from a distinct pronunciation to a shorter and less distinct pronuncia
tion. Fowler (1988) pointed out (for words) that such attenuation is not sloppiness but
Page 13 of 32

rather serves the communicative function of marking the status of the information in the
dialogue. We propose that attenuation serves the same function for gestures.
Grounding with Gestures

The accumulation and coordination of common ground (grounding) is an interactive
process during which participants give each other evidence that they have understood
each other “well enough for current purposes” (Schober & Clark, 1989, p. 228). Several
researchers have examined the role gestures play in grounding by analyzing sequences of
immediately contiguous utterances that involved gestures. Clark and Krych (2004) varied
visibility to examine the role of gestures in addressees’ responses to their speaker’s in
structions. In each pair of participants, the speaker had a model constructed from Lego
blocks and instructed the addressee (who could not see the model) how to build it. In the
pairs who could see each other, addressees often responded with gestures that demon
strated their current understanding of the instructions. Speakers monitored these gestur
al signals of understanding, often following up by modifying their instructions, sometimes
in mid-utterance. In their study of pairs designing floor plans together, Bavelas et al.
(2011) focused on responses to the 552 gestures that had contributed information that
was missing from the words (i.e., were nonredundant). The vast majority of addressees’
responses to these gestures indicated understanding of the information. Moreover, the
majority of speakers’ follow-up responses accepted that the information in the gesture
had become common ground (e.g., speakers presented new information that presupposed
understanding of the previously presented information). These and other studies that ex
amined the relationship between interlocutors’ behaviors (e.g., Furuyama, 2000; Holler &
Wilkin, 2011) represent necessary next steps in understanding how gestures are integrat
ed sequentially into dialogue.
(p. 120) Dialogue Versus Monologue

The use of conversational hand gestures seems to be tied to dialogue, regardless of visi
bility. Bavelas et al. (2008, 2014) showed that a major determinant of the overall rate of
gesturing is whether the context is a dialogue or a monologue. They created three experi
mental conditions: a face-to-face dialogue, a dialogue on the telephone, and a monologue
into a microphone. The telephone dialogues met the criteria for a dialogue (summarized
earlier) except that the interlocutors were not in the same room or able to see each other.
In both experiments, the two dialogue conditions elicited a significantly higher rate of
gestures than the monologue condition, and these rates were not significantly different
from each other. Speaking in a monologue suppressed the gesture rate to about one-third
of the two dialogues.
Gestures as Collateral Communication

Clark (1996, pp. 241–243 and 255–257) proposed that dialogue has two parallel tracks.
The primary track is the topic of the conversation; the secondary track is about the
process of having the dialogue (e.g., “OK, I understand”), which is meta-communication
or collateral communication. We propose that hand gestures can also function as collater
al communication. For example, when speakers change the form of their gesture to mark
Page 14 of 32

it as common ground, their depiction of the referent is the primary track, and its attenu
ated form is the secondary track (in this case, marking the “given” status of this informa
tion in the dialogue). Entrainment is another form of collateral communication. Clark and
Wilkes-Gibbs (1986) and Schober and Clark (1989) have illustrated verbal entrainment:
when describing Tangram figures, interlocutors who could not see each other entrained
on the same arbitrary verbal descriptions of the figures they were working with. Using
the same task with interlocutors who could see each other, Holler and Wilkin (2011) found
gestural entrainment. For example, addressees often mimicked the speaker’s gesture to
show that they had understood which Tangram the speaker was referring to (e.g., their
Figures 2 and 3, p. 145). Their mimicry was not conveying information about the topic
(i.e., identifying the figure); it was instead collateral communication, showing mutual un
derstanding of that information.
Some gestures operate entirely as collateral communication. Bavelas et al. (1992) and
Bavelas, Chovil, Coates, and Roe (1995) identified a subset of gestures with interactive
functions. These gestures do not convey information about the topic of the conversation.
Instead, they refer directly to the addressee, in two senses: (a) The gesture is physically
oriented directly at the addressee with no other depictive features, and (b) the gesture’s
meaning refers to the addressee and to the process of dialogue itself. Examples include
moving the fingers or palms toward the addressee as if metaphorically delivering infor
mation to the addressee or flicking the index finger toward the addressee to cite a previ
ous contribution (equivalent to “as you just said…”). (See sketches of some interactive
gestures in Bavelas et al., 1992, their Figure 1, pp. 474–475, and in Bavelas et al., 1995,
their Figure 1, p. 396.) A series of experiments designed to confirm their interactive func
tion showed that these gestures occurred at a significantly higher rate in the following
conditions: (a) when speakers were talking to an addressee rather than narrating alone to
a camera (Bavelas et al., 1992), (b) when speakers were talking to a visible addressee
rather than to one who could not see them (Bavelas et al., 1992), and (c) when interlocu
tors who could see each other were working together in a back-and-forth dialogue rather
than contributing in alternating monologues (Bavelas et al., 1995). Two microanalyses
further confirmed the hypothesized functions of interactive gestures (Bavelas et al., 1992,
pp. 483–486; Bavelas et al., 1995, Study 2).
Summary of Hand Gestures
An examination of the dates of the publications just sampled reveals that experimental re
search on hand gestures in dialogue is growing rapidly. In addition, the International So
ciety for Gesture Studies with its regular conferences, the journal Gesture, and other pub
lications in a wide range of refereed outlets all suggest that the field is a vital one, with a
wide variety of research foci and established methods. Even better, the research is open
ing up new questions about hand gestures in conversational interaction.
Page 15 of 32

Conversational Facial Gestures

Let’s go back to people-watching in a restaurant. If you are paying close attention to the
diners’ faces, the first thing you would probably notice is that faces bear little resem
blance to still photographs. With their numerous independent muscle groups, human
faces are extraordinarily fast and flexible. Ekman and Friesen (1978) identified more than
forty different action units corresponding to activity in distinct facial muscles or muscle
groups, which account for the ability of the face to move flexibly (p. 121) and rapidly. You
may also notice that most of these movements seem to fit in the same three broad func
tions as hand gestures.
First, all individuals will have facial movements that are self-oriented adaptors such as
blinking, chewing, or swallowing; these actions serve instrumental functions (e.g., keep
ing the eyes moist or eating). Second, you may also notice one person directing a facial
action toward someone at a distance, even though the two of them are not engaged in
speaking with each other. For example, a diner may stare steadily across the room until
the server looks up and makes eye contact; then the diner might smile slightly and raise
his or her eyebrows expectantly to confirm a request for attention. These facial actions in
nonspeaking contexts are (like hand gestures in the same context) called emblems (Ek
man, 1977).
Finally, when you focus on two diners who are talking with each other, you will probably
see a much wider range of facial actions: smiling, frowning, tilting and moving their
heads, furrowing their eyebrows, widening their eyes, wrinkling their noses, grimacing,
and so forth. All of these actions seem to mean something, but if you can’t hear what they
are saying, you won’t know what these facial actions mean. These are likely to be conver
sational (or co-speech) facial gestures. As defined and illustrated in the opening para
graphs of this chapter, conversational facial gestures are tightly synchronized with
speech in both timing and meaning. If you’ve previously read a lot about facial expres
sions, you might assume that these are facial expressions of emotions. However, some
caution is in order because, as explained in the next section, conversational facial ges
tures are not likely to be facial expressions of emotion.
Facial Expressions Versus Facial Gestures
Kraut and Johnston (1979) introduced a distinction between two kinds of facial actions:
some that express emotion and others that display information to an interlocutor. Facial
expressions are part of an emotional, individual process, whereas conversational facial
gestures1 are part of a social, interactive process. We propose that this distinction leads
to at least four differences: first, conversational facial gestures are precisely timed with
speech, so they appear and change within seconds (as illustrated in Figure 7.1 of this
chapter). Ekman (1997) pointed out that an emotion would run its own course and would
not go on and off with a single word or phrase.
Page 16 of 32

Second, most theories of facial expressions limit them to a set of stereotypic forms corre
sponding to a small number of basic emotions (anger, happiness, disgust, etc.). Conversa
tional facial gestures, on the other hand, are part of whatever the person is saying at the
moment, so they have innumerable forms, none of which is fixed or stereotypic. For exam
ple, in Figure 7.1 there is no emotion that corresponds to the speaker’s demonstration of
how Shrek or the cat looked in the movie.
The third distinction is methodological and affects the choice of research procedures: be
cause facial expressions are often seen as stereotypic configurations that correspond to
individual emotions, a typical experiment would focus on the participant’s ability to rec
ognize these facial expressions in still photographs. In contrast, an experiment on facial
gestures in dialogue requires videorecording interlocutors in face-to-face dialogue, then
microanalyzing their facial gestures, second by second, with the accompanying words and
context.
Fourth, facial gestures include more aspects of the face than do facial expressions of emo
tion. The analysis of emotional expressions has been limited to configurations of facial
muscles, such as Ekman and Friesen’s (1978) action units. However, facial gestures often
incorporate other aspects of the facial area, especially eye movements and head posi
tions. For example, in Figure 7.1 frame 3, the tilt of the speaker’s head and the upward di
rection of her gaze portrayed the cat directing his worried expression up at Shrek. Her
head and eye positions were part of her facial gesture as a whole.
Conversational Facial Gestures as Demonstrations

These distinctions make it clear that, like hand gestures, conversational facial gestures
are demonstrations. Bavelas and Chovil (1997) first proposed that conversational facial
gestures meet the criteria of being nonliteral, selective depictions of a referent that they
resemble. In Figure 7.1, the speaker was briefly depicting some selected features of a
movie character who was gloating and then another character who was acting worried.
Although her features resembled gloating or being worried, she was not literally gloating
in frame 2 or worried in frame 3. Instead, she was creating an integrated and tightly syn
chronized demonstration consisting of her words, facial gestures, and hand gestures, all
timed together within fractions of seconds. The result allowed her addressee to perceive
directly what these characters looked like at this point in (p. 122) the movie—or at least
the speaker’s selected version of it. It is highly unlikely that an addressee would mistake
these facial gestures for the speaker’s own emotional expressions.
Although the vast majority of research on faces has been concerned with emotional ex
pressions, these expressions seem to be infrequent in conversational interactions. Frid
lund, Ekman, and Oster (1987, p. 160) reported unpublished data by Ekman and Friesen
on nearly 6,000 facial actions from interviews in which psychiatric patients with affective
disorders were mostly talking about their feelings. Even in this context, fewer than one-
third of these facial actions were classifiable as emotional expressions. In Chovil’s (1989,
1991/1992) data from ordinary conversations, the closest category to emotional expres
sions was “personal reactions” (see Personal Reactions below), which were 24 percent of
Page 17 of 32

the facial gestures she found. However, even these were timed with speech and therefore
do not fit emotion theory. Given that facial actions during dialogue are primarily conver
sational facial gestures rather than facial expressions of emotion, it is striking that there
are literally thousands of published studies on facial expressions of emotion but only a
handful of studies on conversational facial gestures. This topic is wide open for original
research.
This chapter relies on the seminal findings of Chovil (1989, 1991/1992, 2005; see also
Bavelas & Chovil, 1997), which was the first and still the only extensive description of fa
cial gestures in spontaneous face-to-face dialogues. To discover inductively what contri
butions facial gestures can make to conversation, Chovil videotaped twelve dyads (four
male, four female, four mixed) talking about a variety of assigned topics (e.g., a close-call,
an interpersonal conflict). The videos captured close-ups of both interlocutors’ faces in
split screen. Using the nested context approach illustrated in Figure 7.2, Chovil focused
on the function of each facial configuration at that particular moment in the dialogue and
developed a highly reliable systematic analysis of facial gestures (e.g., Chovil, 2005).
Smiles were excluded for practical reasons, because they are both frequent and highly
varied (Ekman, 1985, p. 150), so that work has yet to be done. Even so, she identified
more than a thousand facial actions: 880 were conversational facial gestures, 301 were
adaptors, and only 3 were not analyzable. We have maintained Chovil’s specific functions
(including applying them to some smiles) but have introduced some slightly different
groupings and terms and have illustrated them with examples from our own data (primar
ily from Bavelas et al., 2014).
The overview here follows the same format as the gesture section, first illustrating what
facial gestures are able to demonstrate and then summarizing the small experimental lit
erature on their conversational influences.
What Can Facial Gestures Demonstrate?
Faces present different possibilities and constraints than do hand gestures (or, indeed,
words) and therefore demonstrate different kinds of referents. Facial gestures cannot
portray objects, spatial relationships, motion, and the like, but they can depict what virtu
ally anyone looked like at any particular moment. Conversations are often populated with
oneself, someone else, or even imaginary characters, whether in the past, present, or fu
ture, so the capacity to depict people’s faces is very useful.
Page 18 of 32

Personal Reactions
Fig. 7.4 The speaker’s and addressee’s faces in

frame 3 showed their personal reactions as the
speaker was telling about a scene from Shrek 2. The
three frame shots cover a 6.2-second period. (The
face of the addressee, who was sitting across a table
from the speaker, is inset at the upper right.)
Both speakers and addressees can demonstrate how they are reacting in the present situ
ation, that is, their evaluation of or feeling about something in their current interaction.
For example, in Figure 7.4, the speaker was relating a scene in Shrek 2 during which the
Donkey accidentally kicked Shrek while trying to help him repel the cat’s attack. The
speaker said “and then Donkey tries to help Shrek, um, by kicking him in the crotch.” In
frame 3, as soon as she had finished saying the word “crotch,” she smiled and laughed to
demonstrate to the addressee that she thought that this was funny. We do not infer from
her smile or the smile of the addressee that they were feeling the emotion of happiness at
precisely that moment. What we can observe is that their facial reactions demonstrated
that the scene was humorous to them.
Fig. 7.5 The addressee (inset at the upper left) was

sitting directly across from the speaker, listening to
the speaker’s close-call story. In frame 2, his facial
gesture conveyed a personal reaction to a sudden
danger in the story. The three frame shots cover 2
seconds.
Figure 7.5 (from our archives) shows a different personal reaction by an addressee. The
speaker was telling a personal close-call story about walking on a wet log to cross a flood
ing river. The speaker said “and my foot slips off and I fell into the river.” Just as he fin
ished saying the word “river” (in frame 1), the addressee (in frame 2) made a shocked
face with pursed lips and said “Oooh.” The addressee’s face demonstrated specifically
and vividly his understanding of the speaker’s sudden danger at this point in the story.
This facial gesture did not mean that he was actually shocked and concerned during the .
7 second it lasted; instead, it was timed to fit the final, culminating point of the speaker’s
sentence.
Page 19 of 32

(p. 123) Facial Portrayals

Speakers and addressees can move beyond their present reactions in the dialogue and in
stead portray themselves or others in the past, the future, or even a hypothetical situa
tion. Frames 2 and 3 of Figure 7.1 show two such portrayals in which the speaker por
trayed the face of Shrek and then the cat. Another example comes from Chovil’s (1989,
1991/1992) data. When describing a minor conflict in the past, speakers often portrayed
how they had reacted to the other person—and how the other person had reacted to them
(e.g., as indignant, angry, or appeasing). One speaker said “But I was like errrrr!” Precise
ly at “errrrr,” she tightened her eyes and produced an angry face (1991/1992, p. 182).
Chovil (1989) also observed the special case of addressees’ facial motor mimicry; for ex
ample, when the speaker tells about an injury that happened to him or her, the
addressee’s facial portrayal (e.g., a wince) would be appropriate to the situation the
speaker is describing.
Facial Metaphors
Bavelas and Chovil (2000) proposed that facial gestures vary on a continuum of abstrac
tion. In Figures 7.4 and 7.5, the facial gestures literally represented the speaker’s own
face. In Figure 7.1, however, the facial gestures were somewhat more abstract because
they represented someone else’s face; that is, the addressee had to interpret the
speaker’s face as if it were Shrek’s or the cat’s face looking that way. At an even more ab
stract level, “facial metaphors convert the literal meaning of a facial reaction to an ab
stract one by using it in a metaphor” (Bavelas & Chovil, 2000, pp. 174–175). For example,
when an addressee begins to suspect that the speaker is exaggerating or teasing, he may
tilt his head and squint, as if—metaphorically—looking more closely at the story. Or an ad
dressee who is hearing a friend describe a social blunder may show a pained expression
even though the friend’s “pain” was purely social, not physical. Ultimately, facial gestures
may become even more abstract, as in many of those described in the next section.
Facial Gestures as Collateral Communication
Many facial gestures function as collateral communication (Clark, 1996, pp. 241–243,
255–257) because they demonstrate information about the status of the ongoing talk
rather than the content of the talk. For example, speakers use various facial gestures to
stress specific words or phrases, to explain pauses for a word search, to indicate a ques
tion, and to demarcate the beginning, continuation, or end of a story. Addressees use
backchannel facial gestures to demonstrate that they are attending and following. Chovil
(1989, 1991/1992) found (p. 124) that almost half of the speakers’ facial gestures in her
data did not convey topical information such as a personal reaction or a portrayal. In
stead, they marked some aspect of the discourse itself. Instead of using Chovil’s terms,
we are grouping these facial gestures under collateral communication in order to empha
size their functional similarity to collateral communication by words or gestures.
Page 20 of 32

Emphasizers, Underliners, and Question Markers
Fig. 7.6 In frame 2, the speaker’s raised eyebrows

emphasized the main word in her sentence (“AT
TACKED”). The three frame shots cover a 3.4-second
period. (The face of the addressee, who was sitting
across a table from the speaker, is inset at the upper
right.)
In Chovil’s data, the most common collateral communication with the face was the em
phasizer (e.g., Ekman, 1979), which occurred synchronously with a stressed (e.g., prosod
ically marked) word. These were typically brow movements; the facial muscles that move
the brows are rapid enough to go up and down with a single word or even syllable. Figure
7.6 shows an emphasizer in frame 2. The speaker was talking about a scene from Shrek 2
in which Shrek and Donkey were suddenly ambushed by Puss in Boots. The speaker said
“and they are stopped and ATTACKED by the cat”; precisely as she vocally stressed the
word “ATTACKED,” she raised and lowered her eyebrows. Together, the forceful prosody
and the raised eyebrows emphasized the key word in her sentence. Chovil also found
what Ekman (1979) had called underliners, which mark a whole phrase rather than a sin
gle word. Bavelas and Chovil (2000, p. 175) later speculated that the raised eyebrows
characteristic of emphasizers and underliners are a metaphor based on the startle reac
tion, although the speaker in Figure 7.6 was obviously not demonstrating that she was lit
erally startled by this word.
Fig. 7.7 The speaker’s raised eyebrows marked her

words in frame 2 as a question. The three frame
shots cover a 2.3-second period. (The face of the ad
dressee, who was sitting across a table from the
speaker, is inset at the upper right.)
The next most frequent collateral communication in Chovil’s (1989, 1991/1992) data had
a grammatical function, the question marker. Speakers frequently raised or lowered their
eyebrows when asking a question, including questions marked only with rising intona
tion. In Figure 7.7, the speaker started retelling the movie by introducing the characters.
When she came to the Donkey character, she said “You know, Donkey?” and raised her
brows exactly with “know Don-” to mark her brief phrase as a question. The addressee
nodded and smiled slightly, then the speaker continued. Chovil also found other, less fre
quent facial gestures that served as collateral communication, including ones that
Page 21 of 32

marked the beginning, continuation, or ending of a topic. For example, immediately after
frame 3 in Figure 7.1, the speaker smiled to close that part of her narrative.
Thinking Faces and Facial Shrugs
Fig. 7.8 While recalling a scene from Shrek 2, the

speaker made an extended (1.7-second) thinking face
in frames 2 and 3. (The face of the addressee, who
was sitting across a table from the speaker, is inset
at the upper right.)
The thinking (or remembering) face is a collateral communication indicating that the
speaker is thinking about what to say, trying to remember (p. 125) something, or search
ing for a word. Typically, the speaker breaks eye contact briefly by looking away (or clos
ing both eyes) and might also suddenly make either a “blank” or an “effortful” expression.
In Figure 7.8, the speaker had announced that she was about to describe the scene in
which Shrek and Donkey met Puss in Boots; instead, she paused, shook her head slightly,
and said (as if asking herself) “What happens?” while tilting her head and looking up. It
took only 1.82 seconds to move into and hold this gesture, which ended as soon as she be
gan to describe the scene. Whether or not the gesture actually helped the speaker to
think or remember, it did serve the important social function of explaining her temporary
speaking pause to the addressee. (Similarly, in Figure 7.1, as the speaker changed from
being Shrek to being the cat during the fraction of a second between frames 2 and 3, she
paused at “like— ” and made a brief thinking face, looking away with a blank face.)
Fig. 7.9 The speaker was completing her description

of a scene from Shrek 2 when, instead of continuing
on from frame 1, she made a facial shrug (in frame
2). Then she said “I guess” (in frame 3) and went on.
The three frame shots cover a 1.3-second period.
(The face of the addressee, who was sitting across a
table from the speaker, is inset at the upper right.)
Another collateral communication is the facial shrug, which (like a shoulder shrug) con
veys either not knowing or not caring about something at that moment in the dialogue
(e.g., having said enough on the topic, having said it well enough, conceding or resigning
a point). Both Ekman (1985) and Chovil (1989, 1991/1992) observed facial shrugs, which
typically involve a quick eyebrow flash and the retraction of a corner of the mouth. Figure
7.9 shows the speaker shrugging with her face by pulling the corner of her mouth inward.
Page 22 of 32

This facial shrug occurred as she came to the end of a scene in the Shrek movie and
paused between two parts of her utterance. She said “and then,” made the facial shrug,
and concluded with “I guess…they take him.” The shrug functioned to demonstrate that
the speaker was not certain whether her conclusion was actually what had happened, but
it was good enough. Thinking faces and facial shrugs function as collateral communica
tion (rather than personal reactions) because they comment on what the speaker is say
ing rather than providing topical content.
Facial Backchannels
Addressees as well as speakers make facial gestures that function as collateral communi
cation. The majority of addressees’ facial actions in Chovil’s (1989, 1991/1992) data were
backchannel responses. Chovil found that facial backchannels typically consisted of brow
raises, mouth corners turned down, or lips pressed together. Recall that Chovil did not
study smiles, but Brunner’s (1979) intensive analysis of backchannels revealed that
[Addressees’] smile beginnings showed a strong tendency to occur at the same

kinds of location as “back channel” responses (such as “yeah,” “uh-huh,” (p. 126)
and head nods). This finding indicates that the smile can function as a type of back
channel. It is argued that smiles, like other forms of back channel, make communi
cation more efficient by providing the speaker with feedback on a number of lev
els simultaneously. (p. 728)
We would add that addressees’ facial gestures are efficient for another reason as well:
they provide rapid feedback to the speaker without interrupting or taking up a turn.
Facial Gestures and Words
As many of these examples show, conversational facial gestures are meaningfully related
to the words they occur with, but they are not always redundant with them. Instead, faces
can contribute unique information to the integrated message. Within one of her major
groups, in which the speaker’s face conveyed information about the topic, Chovil (1989,
1991/1992; see also Bavelas & Chovil, 1997) analyzed how often the face provided nonre
dundant information (i.e., not at all in the accompanying words) versus how many were at
least somewhat redundant with words. A surprising 40 percent of this group were com
pletely nonredundant; that is, they conveyed information that complemented but was not
conveyed in the accompanying words.
Conversational Factors that Affect Facial Gestures
As shown in the preceding paragraphs, the interlocutors’ faces are quite active during
face-to-face dialogue, conveying both topical content and collateral communication. In
view of this ubiquity, the lack of experimental research on facial gestures in dialogue can
hardly be overemphasized (see review in Bavelas & Chovil, 2006). To our knowledge, the
published reports of dialogic data consist of the three systematic analyses cited earlier
Page 23 of 32

(Brunner, 1979; Chovil, 1989, 1991/1992; Ekman, 1979) and two experiments (Chovil,
1989, 1991; Bavelas et al. 2014).
Effects of Visibility on Addressees’ Motor Mimicry

The first controlled experiment on facial gestures in dialogue was Chovil’s (1989, 1991)
study of addressees’ motor mimicry. As described earlier, motor mimicry occurs when
someone’s reaction is not appropriate to his or her own situation; instead, it is appropri
ate to the situation of the other person. Bavelas, Black, Lemery, and Mullett (1986) had
shown that facial motor mimicry depended on eye contact with the injured person, but
their experimental setting was not a dialogue. Chovil created four experimental condi
tions: the addressee was hearing a close-call story either in a face-to-face dialogue, in a
dialogue on the telephone, in a dialogue through a partition, or when alone (listening to
an answering machine). The speakers told their own stories in the first three conditions;
the story on the answering machine was chosen from earlier data as a particularly alarm
ing one. The rate of mimicry was significantly higher in the face-to-face dialogue than in
the other three conditions, where an addressee would not see it. The rate was also signifi
cantly lower when the addressee was alone, compared to the three dialogue conditions.
Altogether, the series of experiments on motor mimicry including Bavelas et al. (1986),
Chovil (1989, 1991), Bavelas, Black, Chovil, Lemery, and Mullett (1988), and Bavelas,
Coates, and Johnson (2000) led to the conclusion that motor mimicry is not a reflexive or
emotional reaction; it functions communicatively in dialogue, specifically, as a display of
understanding. (See Bavelas, 2007, for a critical review of subsequent interpretations of
this work.)
Effects of Visibility and Dialogue on Speakers’ Facial Portrayals

To our knowledge, Bavelas et al. (2014) is the first formal experiment on speakers’ facial
gestures, specifically, speakers’ facial portrayals of characters in the movie, Shrek 2 (as il
lustrated throughout this chapter). The three experimental conditions were face-to-face
dialogue, telephone dialogue, and monologue into a tape recorder, making it possible to
assess the effects of both visibility and dialogue separately. Visibility increased the rate of
facial portrayals (i.e., it was significantly higher in the face-to-face condition than in the
telephone and tape recorder conditions combined). Dialogue also increased the rate of fa
cial portrayals (i.e., it was higher in the combined face-to-face and telephone dialogues
than in monologue). Note that Chovil’s (1989, 1991) experiment on addressees’ motor
mimicry also showed both visibility and dialogue effects.
Summary of Facial Gestures
In contrast to hand gestures, experimental research on facial gestures is rare, probably

overwhelmed by research on emotional expressions. The major descriptive studies were
published decades (p. 127) ago, but, as shown in this chapter, they fit new data just as
well. Given that faces in dialogue are lively, eloquent, efficient, infinitely varied, and ubiq
uitous, it is well past time to start investigating them in more detail.
Page 24 of 32

Conclusion
Face-to-face dialogue is the basic form of language use (e.g., Clark, 1996; Goodwin, 1981;
Levinson, 1983). The processes by which interlocutors in dialogue coordinate their ac
tions to create meaning have begun to interest scholars from many disciplines, including
linguistics, anthropology, and psychology. Research on how interlocutors use visible re
sources, such as hand and facial gestures, is intrinsic to understanding language use in
face-to-face dialogue. Regardless of the discipline, investigations of language use in dia
logue must meet, at minimum, two methodological imperatives. First, the interactions
themselves must maintain the essential features of dialogue, including interlocutors’ abili
ty to be spontaneous and to extemporize. The studies reviewed here demonstrate that
this imperative can easily be met in an experimental setting when necessary control is
achieved through careful research design rather than through reduction to a quasi-dia
logue or, as we put it elsewhere, committing “dialogicide” (Gerwing & Bavelas, 2013).
The second methodological imperative is that the original data must not only be recorded
full-face on video, but the analysis must also be conducted directly from video. Hand and
facial gestures coordinate with and complement speech, and they often provide essential
details that are not in the words. Digitized video and suitable software (e.g., ELAN, http://
tla.mpi.nl/tools/tla-tools/elan/; Wittenburg et al., 2006) permit the close observation of di
alogue as a lively synchrony of words, gestures, and faces. With frame-by-frame, repeated
viewing, analysts can observe interlocutors’ behaviors carefully and systematically. Rigor
ous procedures for analysis and detailed operational definitions, plus regular tests for in
teranalyst reliability are essential for ensuring interanalyst agreement.
The overview presented in this chapter leads to several broad conclusions: hand gestures
are not simply redundant movements nor signs of dysfluency. They contribute a wide vari
ety of information ranging from the concrete and physical to the metaphorical to meta-
communicative collateral communication. The evidence suggests that this field has ma
tured from a debate over whether gestures are communicative into solid research on how
they communicate. Nor are facial gestures a limited, static set of emotional configura
tions. They are a constantly changing accompaniment to words, adding new and nuanced
details, again ranging from the concrete to the abstract to collateral communication.
An abundant and increasing number of quantitative experiments on hand gestures are ex
ploring subtle relationships and unexpected topics. In short, this field is healthy and
growing rapidly. However, the study of facial gestures needs to emerge from the domi
nance of emotion theory in order to produce a body of research of its own. Research on
the contributions that interlocutors make with their faces cannot advance until face-to-
face dialogues are routinely recorded and microanalyzed. For young researchers, this
field is wide open for systematic description and especially for experimental research.
Every area touched on in this chapter could expand many times over. Recent research on
hand gesture provides models and methods that might apply to studies of facial gesture
as well.
Page 25 of 32

One reservation is that experimental work has too often treated gestures, faces, and
words as separable phenomena. In face-to-face dialogue, these (and other) communica
tive resources work together, and it is essential to explore the details of their division of
labor. The semantic features approach (e.g., Beattie & Shovelton, 1999; Gerwing & Alli
son, 2009) is one possible method for quantifying the contributions of each modality by
providing a systematic approach for exploring how variations in social context (e.g., mu
tual visibility, common ground) might influence the distribution of information among
hand gestures, facial gestures, and words. Another, mostly unexplored area is how all of
these come together to present a unitary multimodal configuration, such as the one em
phasized in Figure 7.1. Moreover, language use in face-to-face dialogue is more than
words, hand gestures, and facial gestures. Prosody and gaze suggest themselves as new
areas for research—provided that these studies always take place within true face-to-face
dialogues. Scraps of scripted talk and recordings of gaze at inanimate targets fit the out
moded principle of reductionism that continues to threaten the study of conversations in
face-to-face dialogue.
Other than a few exemplars reviewed here (Bavelas et al., 2011; Clark & Krych, 2004;
Holler & Wilkin, 2011; Furuyama, 2000), researchers have tended to ignore the sequen
tial relationships between interlocutors’ behaviors as their dialogue unfolds. This hesita
tion has held the (p. 128) field back from understanding how visible behaviors are sequen
tially integrated into the dialogue and how they influence subsequent utterances. One
well-established model for studying these relationships within an experimental setting is
a referential communication task (e.g., Holler & Wilkin, 2011), in which pairs of partici
pants must complete a series of steps together but one person has all of the necessary in
formation and the other person does not. Although used extensively to study the sequen
tial processes of verbal communication, this method is underused for exploring how inter
locutors use their hand and facial gestures sequentially.
On a theoretical level, our research group is pursuing a theory of demonstrations (Bave

las et al., 2008, 2014), which include not only hand and facial gestures but also direct
quotations and metaphors. We propose that demonstrations require an addressee and are
therefore uniquely linked to interacting in dialogue rather than monologue. Underpinning
this theory is the proposal that, rather than dividing the elements of dialogue along physi
cal lines such as “verbal” and “nonverbal,” it might be more useful to make functional dis
tinctions, such as between description and demonstration. Exactly how demonstrations
work alongside descriptive speech is another wide-open question.
Throughout this chapter, we have emphasized experimental research, not because it is su
perior to other methods, but because it is a useful and informative method that is most of
ten used in language and social psychology. It would be remiss not to point out that most
original insights come from everyday observations and qualitative analyses (e.g., Bavelas,
1987). These provide a springboard for research questions and hypotheses, which lead to
designing experiments that might answer or test them. There is a world of unexplored di
alogic phenomena out there, one that is not found in libraries or statistics.
Page 26 of 32

References
Alibali, M. W., Heath, D. C., & Myers, H. J. (2001). Effects of visibility between speaker
and listener on gesture production: Some gestures are meant to be seen. Journal of Mem
Bavelas, J. B. (1987). Permitting creativity in science. In D. N. Jackson & J. P. Rushton

(Eds.), Scientific excellence: Origins and assessment (pp. 307–327). Beverly Hills: Sage.
Bavelas, J. B. (2005). The two solitudes: Reconciling social psychology and language and
social interaction. In K. Fitch & R. Sanders (Eds.), Handbook of language and social inter
action (pp. 179–200). Mahwah, NJ: Erlbaum.
Bavelas, J. B. (2007). Face-to-face dialogue as a micro-social context. The example of mo

tor mimicry. In S. Duncan, E. Levy, & J. Cassell (Eds.), Gesture and the dynamic dimension
of language (pp. 127–146). Amsterdam: John Benjamins.
Bavelas, J. B., Black, A., Chovil, N., Lemery, C. R., & Mullett, J. (1988). Form and function
in motor mimicry. Topographic evidence that the primary function is communicative. Hu
man Communication Research, 14, 275–299.
Bavelas, J. B., Black, A., Lemery, C. R., & Mullett, J. (1986). “I show how you feel.” Motor
mimicry as a communicative act. Journal of Personality and Social Psychology, 50, 322–
329.
Bavelas, J. B., & Chovil, N. (1997). Faces in dialogue. In J. A. Russell & J. M. Fernandez-
Dols (Eds.), The psychology of facial expression (pp. 334–346). Cambridge, UK: Cam
Bavelas, J. B., & Chovil, N. (2000). Visible acts of meaning. An integrated message model
of language use in face-to-face dialogue. Journal of Language and Social Psychology, 19,
163–194.
Bavelas, J. B., & Chovil, N. (2006). Hand gestures and facial displays as part of language
use in face-to-face dialogue. In V. Manusov & M. Patterson (Eds.), Handbook of nonverbal
communication (pp. 97–115). Thousand Oaks, CA: Sage.
Bavelas, J. B., Chovil, N., Coates, L., and Roe, L. (1995). Gestures specialized for dialogue.
Bavelas, J. B., Chovil, N., Lawrie, D. A., & Wade, A. (1992). Interactive gestures. Discourse
Processes, 15, 469–489.
Page 27 of 32

Bavelas, J. B., Gerwing, J., Allison, M., & Sutton, C. (2011). Dyadic evidence for grounding
with abstract deictic gestures. In G. Stam, M. Ishino, & R. Ashley (Eds.), Integrating ges
tures: The interdisciplinary nature of gesture (pp. 49–60). Amsterdam: Benjamins (Ges
ture Studies series).
Bavelas, J. B., Gerwing, J., & Healing, S. (2014). The effect of dialogue on demonstrations:
Direct quotations, facial portrayals, hand gestures, and figurative references. Discourse
Processes, in press.
Bavelas, J. B., Gerwing, J., Sutton, C., & Prevost, D. (2008). Gesturing on the telephone:
Independent effects of dialogue and visibility. Journal of Memory and Language, 58, 495–
520.
Bavelas, J. B. & Healing, S. (2013). Reconciling the effects of mutual visibility on gestur
ing. Gesture 13, 63–92.
Beattie, G., & Shovelton, H. (1999). Do iconic hand gestures really contribute anything to
the semantic information conveyed by speech? An experimental investigation. Semiotica,
123, 1–30.
Brunner, L. J. (1979). Smiles can be back channels. Journal of Personality and Social Psy
chology, 37, 728–734.
Buchler, J. (Ed.). (1940). The philosophy of Peirce: Selected writings. London: Routledge
& Kegan Paul.
Chovil, N. (1989). Communicative functions of facial displays in conversation. Unpub

lished Ph.D. dissertation, University of Victoria, Victoria, BC.
Chovil, N. (1991). Social determinants of facial displays. Journal of Nonverbal Be

(p. 129)
havior, 15, 141–153.
Chovil, N. (1991/1992). Discourse-oriented facial displays in conversation. Research on

Language and Social Interaction, 25, 163–194.
Chovil, N. (2005). Measuring conversational facial displays. In V. Manusov (Ed.), The

sourcebook of nonverbal measures: Going beyond words. Hillsdale, NJ: Lawrence Erl
baum.
Clark, H. H. (1985). Language and language users. In G. Lindzey & E. Aronson (Eds.), The
handbook of social psychology (3rd ed., vol. 2, pp. 179–232). New York: Random House.
Clark, H. H. (1996). Using language. Cambridge, UK: Cambridge University Press.
Clark, H. H., & Gerrig, R. J. (1990). Quotations as demonstrations. Language, 66, 764–
805.
Clark, H. H. & Krych, M. A. (2004). Speaking while monitoring addressees for under
standing. Journal of Memory and Language, 50(1), 62–81.
Page 28 of 32


22, 1–39.
Cohen, A. A. (1977). The communicative functions of hand illustrators. Journal of Commu

nication, 27, 54–63.
Cohen, A. A., & Harrison, R. P. (1973). Intentionality in the use of hand illustrators in
face-to-face communication situations. Journal of Personality and Social Psychology, 28,
276–279.
Danziger, K. (1990). Constructing the subject. Historical origins of psychological

research. Cambridge, UK: Cambridge University Press.
de Ruiter, J. P., Bangerter, A., & Dings, P. (2012). The interplay between gesture and
speech in the production of referring expressions: Investigating the tradeoff hypothesis.
Topics in Cognitive Science 4, 232–248.
Ekman, P. (1977). Biological and cultural contributions to body and facial movement. In J.
Blacking (Ed.), Association of Social Anthropologists monograph 15, The anthropology of
the body. London: Academic Press.
Ekman, P. (1979). About brows: Emotional and conversational signals. In M. von Cranach,
K. Foppa, W. Lepenies, & D. Ploog (Eds.), Human ethology (pp. 169–249). Cambridge, UK:
Ekman, P. E. (1985). Telling lies. New York: Berkley Books.
Ekman, P. (1997). Should we call it expression or communication? European Journal of So

cial Sciences, 10, 333–359.
Ekman, P., & Friesen, W. V. (1978). The Facial Action Coding System. Palo Alto, CA: Con
sulting Psychologists Press.
Emmorey, K., & Casey, S. (2001). Gesture, thought and spatial language? Gesture, 1, 35–
50.
Fowler, C. A. (1988). Differential shortening of repeated content words produced in vari

ous communicative contexts. Language and Speech, 31, 307–319.
Fridlund, A. J., Ekman, P., & Oster, H. (1987). Facial expressions of emotion: Review of lit
erature, 1970–1983. In A. W. Siegman & S. Feldstein (Eds.), Nonverbal behavior and com
munication (2nd ed., pp. 143–224). Hillsdale, NJ: Erlbaum.
Furuyama, N. (2000). Gestural interaction between the instructor and the learner in
origami instruction. In D. McNeill (Ed.), Language and gesture (pp. 99–117). Cambridge,
UK: Cambridge University Press.
Gerwing, J., & Allison, M. (2009). The relationship between verbal and gestural contribu
tions in conversation: A comparison of three methods. Gesture, 9, 313–336.
Page 29 of 32

Gerwing, J., & Allison, M. (2011). The flexible semantic integration of gestures and words.
Comparing face-to-face and telephone dialogues. Gesture, 11, 308–329.
Gerwing, J., & Bavelas, J. B. (2004). Linguistic influences on gesture’s form. Gesture, 4,
157–195.
Gerwing, J., & Bavelas, J. (2013). The social interactive nature of gestures: Theory, as
sumptions, methods, and findings. In C. Müller, A. Cienki, E. Fricke, S.H. Ladewig, D. Mc
Neill, & S. Tessendorf (Eds.), Body-language-communication, Volume 1, Contemporary ap
proaches (Ch. 51, pp. 821-836). Berlin: Mouton De Gruyter.
Goodwin, C. (1981). Conversational organization: Interaction between speakers and hear

ers. New York: Academic Press.
Gullberg, M. (2010). Methodological reflections on gesture analysis in SLA and bilingual

ism research. Second Language Research, 26, 75-102.
Haviland, S. E., & Clark, H. H. (1974). What’s new? Acquiring new information as a
process in comprehension. Journal of Verbal Learning and Verbal Behavior, 13, 512–521.
Healing, S., & Gerwing, J. (2012 July). Spatial cohesiveness in gesture sequences. Effects
of mutual visibility. Paper presented at the meeting of the International Society for Ges
ture Studies, Lund (Sweden).
Holler, J., & Stevens, R. (2007). The effect of common ground on how speakers use ges
ture and speech to represent size information. Journal of Language and Social
Holler, J., Tutton, M., & Wilkin, K. (2011). Co-speech gestures in the process of meaning
coordination. In Proceedings of the 2nd GESPIN—Gesture in Speech and Interaction Con
ference, Bielefeld, 5–7, September 2011.
Holler, J., & Wilkin, K. (2011). Co-speech gesture mimicry in the process of collaborative
referring during face-to-face dialogue. Journal of Nonverbal Behavior, 35, 133–153.
Kendon, A. (1980). Gesticulation and speech: Two aspects of the process of utterance. In
M. R. Key (Ed.), The relationship of verbal and nonverbal communication (pp. 207–227).
The Hague: Mouton.
Kendon, A. (2004). Gesture. Visible action as utterance. Cambridge, UK: Cambridge Uni
versity Press.
Krauss, R. M., Dushay, R. A., Chen, Y., & Rauscher, F. (1995). The communicative value of
conversational hand gestures. Journal of Experimental Social Psychology, 31, 533–552.
Kraut, R. E., & Johnston, R. E. (1979). Social and emotional messages of smiling: An etho
logical approach. Journal of Personality and Social Psychology, 37, 1539–1553.
Page 30 of 32

Kuhlen, A. K., & Brennan, S. E. (2013). Language in dialogue: When confederates might
be hazardous to your data. Psychonomic Bulletin & Review, 20, 54–72.
Levinson, S. C. (1983). Pragmatics. Cambridge, UK: Cambridge University Press.
Luria, A. R. (2004). Reductionism in psychology. In R. L. Gregory (Ed.), The Oxford com

panion to the mind (2nd ed., pp. 675–676). Oxford, UK: Oxford University Press.
McNeill, D. (1985). So you think gestures are nonverbal? Psychological Review, 92, 350–
371.
Mol, L., Krahmer, E., Maes, A., & Swerts, M. (2009a). Communicative gestures and mem
ory load. In Proceedings of the 31st Annual Conference of the Cognitive Science Society,
Netherlands, 1569–1574.
Mol, L., Krahmer, E., Maes, A., & Swerts, M. (2009b). The communicative import of ges
tures. Evidence from a comparative analysis of human-human and human-machine inter
actions. Gesture, 9(1), 97–126.
Özyürek, A. (2000). The influence of addressee location on spatial language and

(p. 130)
representational gestures of direction. In D. McNeill (Ed.), Language and gesture (pp. 64–
83). Cambridge, UK: Cambridge University Press.
Özyürek, A. (2002). Do speakers design their co-speech gestures for their addressees?
The effects of addressee location on representational gestures. Journal of Memory and
Language, 46, 688–704.
Parrill, F. (2008). Subjects in the hands of the speakers: An experimental study of syntac
tic subject and speech-gesture integration. Cognitive Linguistics, 19, 283–299.
Pine, K. J., Burney, D. J., & Fletcher, B. (2010). The semantic specificity hypothesis: When
gestures do not depend upon the presence of a listener. Journal of Nonverbal Behavior,
34, 169–178.
Reber, A. S., Allen, R., & Reber, E. S. (2009). Penguin dictionary of psychology (4th ed.).
London, UK: Penguin.
Rimé, B. (1982). The elimination of visible behavior from social interactions: Effects on
verbal, nonverbal and interpersonal variables. European Journal of Social Psychology, 12,
113–129.
Rowbotham, S., Holler, J., Lloyd, D., & Wearden, A. (2011). How do we communicate
about pain? A systematic analysis of the semantic contribution of co-speech gestures in
pain-focused conversations. Journal of Nonverbal Behavior, 36, 1–21.
Schober, M. F., & Clark, H. H. (1989). Understanding by addressees and overhearers.

Cognitive Psychology, 21, 211–232.
Page 31 of 32

Sears, R. R. (1951). A theoretical framework for personality and social psychology. Ameri
can Psychologist, 6, 476–483.
Thibaut, J. W., & Kelley, H. H. (1959). The social psychology of groups. New York: Wiley.
Wittenburg, P., Brugman, H., Russel, A., Klassman, A., & Sloetjes, H. (2006). ELAN: A pro
fessional framework for multimodality research. In: Proceedings of LREC 2006, Fifth In
ternational conference on language resources and evaluation, Genoa.
Notes:
(1) . Kraut and Johnston used the term facial display, borrowed from the study of animal
behavior, where the overt behavior acts as a social display to others. Chovil also adopted
this term in her extensive work. We propose a change to facial gesture, which emphasizes
the many similarities to hand gestures.
Janet Bavelas
Janet Bavelas, Department of Psychology, University of Victoria
Jennifer Gerwing
Jennifer Gerwing, Department of Psychology, University of Victoria
Sara Healing
Sara Healing, Department of Psychology, University of Victoria
Page 32 of 32

Interactive Alignment and Language Use

Martin J. Pickering and Simon Garrod

This article discusses the challenge of dialogue for language processing. It highlights two
aspects of such processing: (1) interactive alignment, whereby speakers align their use of
language to support aligned situation models, and (2) prediction-by-simulation, whereby
speakers use comprehension processes to predict what they are about to say and listen
ers use production processes to predict what they are about to hear. It argues that these
two aspects of language processing make fluent coordinated dialogues possible. It then
discusses how alignment and prediction-by-simulation complement each other to support
efficient dialogue processing. Finally, it discusses how different communicative activity
types engage prediction-by-simulation to different degrees and argues for the importance
of the social setting in accounting for how people process language.
Keywords: Communicative activity types, comprehension, dialogue, interactive alignment, prediction, production,
simulation, situation model
What makes dialogue special? Interlocutors may initially know different things about a
topic, but if the conversation is successful, they come to think about the topic in the same
way. Unlike monologue, it is not the case that an individual works out how to describe a
situation or to interpret another individual’s description; instead, both parties work to
gether to achieve this shared understanding. Second, they achieve this alignment by man
aging their contributions so that they do not shout over one another or leave lengthy
pauses between turns; instead, their contributions are remarkably well coordinated. In
this chapter, we outline the way in which we believe interlocutors are typically able to
align their representations and coordinate their contributions. We argue that these two
properties of dialogue are closely related to each other and depend on twin mechanisms
of alignment and prediction across different levels of representation. We argue that such
a process works ideally in casual conversation and contrast this situation with other
forms of dialogue and monologue.
We first note that psychology has paid remarkably little attention to dialogue. Much of so
cial psychology is, of course, concerned with social interaction involving language, but its
focus has not primarily been on the language itself. In contrast, psycholinguistics has
Page 1 of 17

been almost entirely concerned with isolated language processing, both in the early years
(e.g., Fodor, Bever, & Garrett, 1974) and more recently (e.g., Gaskell, 2007). The vast ma
jority of studies involve an individual participant performing tasks such as deciding
whether a string of letters is a word or not, reading sentences or texts, comprehending
short segments of speech, or naming pictures. These studies have focused on the psycho
logical mechanisms involved in monologue, but it is quite possible to address the mecha
nisms involved in dialogue in a similar manner.
(p. 132) The Nature of Dialogue

People are highly adept at many joint activities, such as ballroom dancing, playing a duet,
or carrying a large object together (Sebanz, Bekkering, & Knoblich, 2006). This is re
markable, given that they have to plan their own actions (as in a solo task) but also have
to monitor their partner’s actions and make sure that their actions combine appropriately.
Dialogue is a form of joint activity involving language. It involves an extremely complicat

ed set of processes in which the interlocutors have to interweave their activities with pre
cise timing, and yet it is a skill that all speakers seem very good at (Garrod & Pickering,
2004). To understand how remarkable this is, consider this transcript of a dinner-party
conversation (Tannen, 1989), with brackets indicating overlapping speech and numbers
indicating noticeable pauses in seconds:
1— A: I shook hands with Rubenstein once? [and his hand
2—B: [Yeah we did together
3—A: That’s right. we were together. wasn’t it incredible?
4—B: (laughing) oh it was like a cushion.
5—C: What’s this?
6—A: [I (0.5) we shook] hands with Rubenstein.
7—B: [Rubenstein’s hands.]
8—D: and he had --?
9—A: his hands –
10—D: short stubby hands?
11—A: they were like (0.5) [jelly. they were like—(1.0)
12—B: [a famous concert pianist
13—A: they were like (0.5) putty. (0.5)
14—D: [really?
Page 2 of 17

15—A: [just completely soft and [limp
16—B: [mush
17—A: just mush. it was as though there was [no bone
18—B: [and warm.
19—D: and short stubby fingers?
20—A: short stubby fingers but just (0.5) totally covered with –
21—B: fat.
22—A: fat
This conversation differs greatly from formal prose (such as the rest of this chapter). In
particular, the interlocutors regularly produce elliptical and fragmentary utterances that
would make little sense on their own (e.g., 7, 10, 12, 16, 17, 20). It is jointly constructed
by all four interlocutors and involves a great deal of interruption, overlapping speech, and
disfluency. However, the interlocutors appear to be satisfied with the conversation. They
seem to understand what everyone produces, as do nonparticipants such as us. How can
this be?
Conversations like this are quite remarkable. The participants cannot be sure what con
tributions their interlocutors are going to make, so they cannot securely plan far in ad
vance. They have to construct their utterances so that they are appropriate for their ad
dressees at that particular point and therefore must pay constant attention to any feed
back (e.g., whether a particular term is understood). For example, B’s interruption at (2)
causes A to abandon (1) and produce the appropriate response (3) without delay. Inter
locutors have to decide whether to contribute to a conversation and, if so, precisely when
they should do so, and they may have to decide who to address. In addition, they have to
constantly switch between production and comprehension, even though such task switch
ing is often difficult. None of this is necessary in monologue. Yet everyone can hold a con
versation, but producing a speech (or even listening to one) is a complex skill that re
quires considerable training.
Conversation and the Alignment of Representa

tions
A few years ago, we wrote a paper asking why people find conversation (or dialogue) so
easy in comparison to monologue (Garrod & Pickering, 2004). In that paper, we argued
that conversation is easy because interlocutors tend to become aligned at different levels
of linguistic representation and therefore find it easier to perform joint activity.
Page 3 of 17

Pickering and Garrod (2004) explain the process of alignment in more detail in terms of
their interactive-alignment model. According to this account, dialogue is successful to the
extent that interlocutors come to understand the relevant aspects of the world in the
same way as each other. More specifically, interlocutors construct mental models of the
situation under discussion, and successful dialogue occurs when these models become
aligned. In accord with the text-comprehension literature, we assume that situation mod
els include information about people, time, space, causality, and intentionality (e.g.,
Zwaan & Radvansky, 1998). Alignment of situation models largely occurs as a result of
the tendency for interlocutors to repeat each other’s choices at many different linguistic
levels, such as words and grammar (e.g., Branigan, Pickering, & (p. 133) Cleland, 2000;
Brennan & Clark, 1996; Garrod & Anderson, 1987).
For example, Branigan et al. (2000) had participants take turns to describe and match pic
ture cards and found that they tended to use the form of utterance just used by their in
terlocutor. For example, they tended to use a “prepositional object” form such as the pi
rate giving the book to the swimmer following another prepositional object sentence, but
a “double object” form such as the pirate giving the swimmer the book following another
double object sentence. In other words, the study showed structural priming between in
terlocutors similar to the type of structural priming that occurs within an isolated speak
er (Bock, 1986). Similarly, Garrod and Anderson (1987) had pairs of participants play a
cooperative maze game in which they took turns to describe their positions to each other.
They tended to align on the same description scheme. For example, if one player said I’m
two along, four up, her partner tended to say I’m one along, five up; but if she said I’m at
B4, her partner tended to say I’m at A5. These players aligned on a “path” or a “coordi
nate” description scheme, rather than specific words. They also aligned on the interpreta
tion of these descriptions, for example, both treating the origin as the bottom left corner
of the maze.
Essentially, interlocutors prime each other to speak about things in the same way, and
people who speak about things in the same way are more likely to think about them in the
same way as well. Importantly, interlocutors do make use of more conscious and deliber
ate strategies on occasion, but do so as a “last resort” when automatic alignment breaks
down.
It is important to compare Pickering and Garrod’s (2004) account with Clark (1996), who
also assumes that interlocutors build up representations of their common ground, a no
tion that refers to all the information that both interlocutors believe to be shared by
themselves and their partner. Interlocutors share some common ground through being
members of the same culture (e.g., knowing about particular people, places, and events)
or through co-presence (e.g., being able to see the same objects). But they also accumu
late common ground through interaction, with one partner proposing a contribution and
the other partner accepting that contribution explicitly (e.g., saying OK) or implicitly (by
continuing the conversation without a query). Such a process of grounding presumably
Page 4 of 17

takes place alongside alignment and, indeed, involves alignment (e.g., repeating each
other’s choice of words).
We assume that the representations constructed during production and comprehensions

are tightly yoked. Pickering and Garrod (2004) assumed that the speakers and compre
henders computed the same representations (representational parity), but our recent pro
posals suggest a more nuanced account in which speakers use comprehension-based rep
resentations and comprehenders use production-based representations in an interwoven
manner (Pickering & Garrod, 2013). This account explains why linguistic repetition oc
curred in studies such as Branigan et al. (2000) and Garrod and Anderson (1987). In such
cases, the grammatical representation activated during production is yoked to the same
representation activated during comprehension.
However, such yoking may not be enough to make dialogue easy by itself. Just because
production and comprehension draw on yoked representations does not necessarily mean
that switching between production and comprehension should be straightforward. In fact,
we propose that both producers and comprehenders use yoked representations to make
predictions about upcoming contributions in a way that leads to fluency (most obviously,
timing their contributions appropriately).
Coordination of Contributions to Dialogue

We have noted that joint activities require two (or more) agents to coordinate their ac
tion, which in turn means that they are able to perceive each other’s acts and perform
their own acts together. In many of these activities, precise timing is crucial, with success
occurring only if each partner applies the right force at the right time in relation to the
other. Such success therefore requires tight interweaving of perception and action. More
over, people must predict each other’s actions because responding after they perceive ac
tions would simply be too slow. Clearly, it may also be useful to predict one’s own actions
and to integrate these predictions with predictions of others’ actions.
On the basis of motor control theory (e.g., Wolpert, 1997), Pickering and Garrod (2013)
proposed that actors construct forward models of their movements based on associations
between their intentions and the expected outcomes of these intentions. When someone
plans a hand movement, the motor command activates muscles that lead to the move
ment. In addition, an (efference) copy of the command leads to a representation of the
predicted movement (e.g., with numbers representing spatial coordinates and (p. 134)
time), and this forward model is typically ready before action execution. When that per
son perceives another actor’s hand movement, he covertly imitates the actor, derives the
actor’s upcoming intention (i.e., for the next part of the movement), and constructs a for
ward model based on his most likely action under those circumstances—and again this is
typically ready before the action occurs. In joint activity, both individuals construct for
ward models of the upcoming action and use these predictions to coordinate (e.g., in a
handshake). We refer to this form of prediction as prediction-by-simulation (in contrast to
Page 5 of 17

prediction based on perceptual experience alone, which we refer to as prediction-by-asso

ciation).
Pickering and Garrod (2013) then applied this framework to language. When someone
plans an utterance, her production command activates a series of (implemented) repre
sentations associated with speaking. Language involves representations associated with
semantics, grammar, and sound, and the production command (containing the message
that the speaker wishes to convey, including its communicative force and pragmatic con
text) activates these representations in turn, in a way that leads to articulation (e.g., Lev
elt, 1989). In addition, a copy of the command leads to a forward model consisting of se
ries of predicted linguistic representations, typically before the implemented representa
tions are constructed. When that person hears another person speak, she covertly imi
tates the speaker, derives the speaker’s upcoming production command, and constructs a
forward model based on her most likely utterance under those circumstances—again, typ
ically before speech occurs.
Pickering and Garrod (2013) discuss a range of evidence that people can predict different
aspects of other people’s utterances (meaning, grammar, sounds, etc.), that they covertly
imitate language, and that people can use covert imitation to drive predictions. For exam
ple, comprehenders predict the sounds of highly predictable words (e.g., DeLong, Ur
bach, & Kutas, 2005) as well as their grammatical properties (Van Berkum, Brown, Zwit
serlookd, Kooijman, & Hagoort, 2005), and some effects occur within 100 ms or so
(Dikker, Rabagliati, Farmer, & Pylkkänen, 2010). Evidence for covert imitation comes
from studies showing that listening to words containing particular initial sounds (e.g., /t/)
affects the production of different target phonemes (e.g., /k/ or /s/) (Yuen, Davis, Brys
baert, & Rastle, 2010). In addition, D’Ausilio, Jarmolowska, Busan, Bufalari, and
Craighero (2011) repeatedly exposed participants to a pseudoword (e.g., birro) and used
transcranial magnetic stimulation (TMS) to reveal immediate appropriate articulatory ac
tivation (associated with rr) when they heard the first part of the same item (bi, when
coarticulated with rro) compared to when they heard the first part of a different item (bi,
when coarticulated with ffo). This suggests that covert imitation facilitates speech recog
nition as it occurs and before it occurs.
Just as the accounts of action and perception can be combined into an account of joint ac
tion, so the accounts of production and comprehension can be combined into an account
of dialogue. To illustrate the potential effectiveness of prediction in dialogue, consider the
fact that interlocutors regularly finish off each other’s utterances. In the dinner-party
conversation (Tannen, 1989), speaker A described Rubenstein’s hands as Just completely
soft and limp, and speaker B uttered mush simultaneously with limp (15–16). It appears
that B predicted the meaning of what A was about to say (but not the exact word). In
terms of our proposal, B used the forward model to produce a semantic representation of
a property of Rubenstein’s hands and was therefore able to complete the process of artic
ulating mush without delay. Thus, A and B were aligned in terms of the semantic aspects
of their productions, but did not align lexically. In contrast, when A later said short stubby
fingers but just totally covered with, B completed with fat (20–21). In this case, A repeat
Page 6 of 17

ed B’s completion (22), and so it may be that both A and B were lexically aligned (i.e.,
with similar patterns of activation for particular lexical items) at the point when B spoke
(although it is also possible that A had not accessed fat until after B spoke). In such a
case, A’s behavior is complementary to B’s, rather than imitative (B completed what A
had begun, and did not repeat what A said). To do this, we assume that B covertly imitat
ed A, predicted what A would say next, and then uttered that prediction.
Although our emphasis throughout this discussion is on the use of production-based simu
lation by the addressee, it is also possible that the speaker uses such simulation to pre
dict potential responses, either to questions or to other utterances that invite a response.
For example, a parent might ask a child Have you tidied your room? and use the produc
tion system to predict the response No. The parent might even use this to prepare a fur
ther utterance (Well, do so then). In this case, the child’s response is highly predictable; it
does not matter that the parent’s prediction derives from an utterance by the parent
rather than an utterance by the child. This analysis (p. 135) is compatible with interpreta
tions of contributions to dialogue that emphasize their anticipatory quality (e.g., Linell,
1998, chapter 9).
Our account can be compared to accounts of embodied cognition (e.g., Barsalou, 1999),
which use the term simulation rather differently from us. In such accounts, mental simu
lation occurs if comprehending language (e.g., the word kick) engages the same process
es as occur when perceiving or acting in the manner described by the language (perform
ing or observing a kick; see Pulvermüller, 2005). Our account assumes that people simu
late aspects of linguistic form during comprehension and it does not commit to simulation
of content.
Effects of Prediction on Dialogue
Prediction-by-simulation is particularly important in dialogue in two senses. First, it leads

to activation of the production system during comprehension. This therefore facilitates
the regular switching between comprehension and production in dialogue. When the
speaker’s turn finishes, the addressee is likely to take the floor. In some forms of dia
logue, speaker and addressee constantly change roles. In addition, the addressee may
provide “back channel” feedback during the speaker’s turn, such as assertions (yes, go
on) or queries (eh?, who?). Even during narrative comprehension, the addressee provides
feedback to the speaker (and such feedback enhances the narrative; Bavelas, Coates, &
Johnson, 2000).
In addition, an addressee must be constantly prepared to respond. In some conversation

al settings, a response is normatively required (otherwise the addressee is flouting the
rules of turn-taking; see Sacks, Schegloff, & Jefferson, 1974). The simplest example of
this is a question directed solely (and nonrhetorically) to the addressee, which requires
some response. This is, of course, not the case in monologue, when the addressee knows
that his only job is to comprehend what he is hearing. He may use production mecha
nisms to facilitate that comprehension, but he does not have to make overt responses.
Page 7 of 17

Indeed, an addressee contrasts with other individuals who may be present in the dialogue
setting (e.g., Clark, 1996). Researchers distinguish different roles, typically speaker, ad
dressee, side-participant, and nonparticipants (licensed overhearers and eavesdroppers).
Importantly, the speaker addresses the addressee or addressees (e.g., asking him a ques
tion). The side-participant is not required to respond but is permitted to contribute during
the conversation. Thus, we might expect the addressee to predictively process to a
greater extent than side-participants or overhearers. In accord with this, Branigan, Pick
ering, McLean, and Cleland (2007) had speakers describe cards to addressees using
“prepositional object” or “double object” forms (as in Branigan et al., 2000) in the pres
ence of a side-participant. They found that addressees were more likely to repeat the
grammatical form used by the speaker than were side-participants. This suggests that ad
dressees predicted the grammatical form used by the speaker in order to prepare a re
sponse to a greater degree than did side-participants. In contrast, the tendency to repeat
was unaffected by whether the previous addressee or the previous side-participant then
spoke to the previous speaker or to someone else. This decision is not related to the de
gree of prediction required during comprehension. This account also provides an interest
ing explanation of why addressees understand dialogues better than overhearers
(Schober & Clark, 1989): the addressee is constantly preparing potential responses and is
therefore activating his or her production system, thus increasing deep processing and
hence understanding. (In contrast, Schober and Clark proposed that the speaker and ad
dressee jointly construct a common ground that is not fully accessible to the overhearer.)
In addition, the processor may emphasize prediction in dialogue rather than monologue
because dialogue is, in general, more predictable than monologue. It tends to be much
more repetitive, as Pickering and Garrod (2004) informally noted by comparing the
amount of repetition in a stretch of “maze game” dialogue (Garrod & Anderson, 1987)
with an equivalent paragraph from their paper (see also Tannen, 1989). Clearly, the more
repetitive a text, the more the advantage of emphasizing prediction over bottom-up analy
sis. Such repetitiveness occurs at many different linguistic levels, not just words. An effi
cient processor will therefore make relatively greater use of prediction.
Alignment and Prediction-by-Simulation Sup

port Each Other
We propose that alignment and prediction-by-simulation support each other in a way that
leads to communicative success. In outline, we have argued that comprehenders use
covert imitation as part of the process of predicting speakers. Such covert imitation
therefore means that the comprehender aligns his representations with the speaker. In
turn, a comprehender is more likely to predict (p. 136) the speaker correctly if they are
aligned. His predictions are based on his production system and will therefore be more
accurate if his production system more closely resembles the speaker’s production sys
tem. Alignment and prediction can be seen as opposite sides of the same coin, with align
Page 8 of 17

ment looking backward and prediction looking forward. We now expand on how this mu
tual support might work.
How Prediction Supports Alignment and Vice Versa
As we have said, the simplest way in which prediction supports alignment is that it in
volves covert imitation. In more detail, the comprehender constructs a comprehension-
based representation that is then mapped to a production-based representation, and this
representation is used to derive the speaker’s production command for the utterance so
far (e.g., for short stubby fingers but just totally covered with…) and for the upcoming
part of the utterance (fat). This means that the comprehender constructs a representation
corresponding to what the speaker has said (and in this respect aligns with the speaker)
and a representation corresponding to what she predicts the speaker will say next, which
leads to alignment if the speaker does say what has been predicted (hence, the better the
prediction, the better the alignment). In addition, the comprehender may overtly com
plete the speaker’s utterance (fat), as occurred in Tannen’s (1989) dinner party conversa
tion presented earlier. If this happens, then, of course, the speaker is more likely to align
with the comprehender.
But the relationship also works the other way. Prediction during comprehension is facili
tated when the interlocutors are well-aligned because the comprehender is more likely to
predict the speaker accurately (and the speaker is more likely to predict the
comprehender’s response, as in question-answering). One effect of this is that speaker
B’s prediction of what speaker A is going to say is more likely to accord with what B
would be likely to say if B spoke at that point. In other words, B’s prediction of B’s com
pletion becomes a good proxy for B’s prediction of A’s completion, and so there is less
likelihood of an egocentric bias.
How Prediction Helps Identify Misalignment
In the interactive alignment model, we assumed that people automatically align where
possible so that they come to use the same terms to refer to the same objects. They then
tend to make the same inferences based on those terms, so that they end up with a simi
lar understanding of the situation. But when they fail to understand the other person—for
instance, if they can’t interpret a word that they hear—they attempt to repair. This repair
may be largely automatic, for example, following a strategy of becoming more explicit
(e.g., turning pronouns into full referring expressions). But if this does not work, the
speaker must make reference to what he believes are the differences between his knowl
edge and that of his addressee.
In the prediction-based account, the comprehender monitors the speaker using her for
ward model and compares her predictions with the speaker’s behavior (here, speech). If
prediction and behavior are close enough, she tunes her forward model (predicting slight
ly differently next time) and can use this forward model to facilitate her production of her
own utterances. If not, the comprehender has to do something in order to understand.
Page 9 of 17

Speakers also use forward models to monitor their own utterances, but monitoring during
comprehension is harder because comprehenders have less access to speakers’ inten
tions than they do to their own (and also because they are different from the speaker in
other respects). The comprehender can query the speaker, and if the speaker responds,
for example by being more explicit, then understanding can result. But, alternatively, the
comprehender may be unable or unwilling to query the speaker, in which case he has to
appeal to “mentalizing” (roughly, theory of mind), and this is much harder. In other
words, he has to start the potentially difficult task of reasoning about what he thinks the
speaker is likely to mean.
The Effects of Different Social-Communicative

Settings on the Process of Interactive Align
ment
As we have noted, a few researchers have attempted to provide a psychology of language
for dialogue. But, in fact, types of dialogue differ greatly and a general theory of language
processing needs to account for dialogue as a whole. Pickering and Garrod (2004)
proposed that types of dialogue vary along a continuum. Here, we suggest that there are,
in fact, many dimensions over which dialogue varies. Importantly, some of these dimen
sions are relevant to alignment and prediction, as we shall see.
Many attempts have been made to classify types or “genres” of dialogue (or communica
tion more (p. 137) generally). Perhaps the most useful notion is that of activity type, which
is “a fuzzy category whose focal members are goal-defined, socially constituted, bounded,
events with constraints on participants, setting, and so on, but above all on the kinds of
allowable contributions” (Levinson, 1992, p. 69). Levinson gives examples such as teach
ing, a job interview, or a football game. Some of these do not (necessarily) involve com
munication, so we can restrict ourselves to communicative activity types (see Linell,
2009, pp. 201–211). They typically refer to an interaction (rather than a specific ex
change, such as a question–answer pair or a long-term “stable” relationship such as a
marriage).
Activity types vary on several dimensions (Clark, 1996, pp. 30–31; Levinson, 1992, pp. 69–
70). Perhaps most obviously, they differ in scriptedness, from marriage ceremonies in
which the participants are required to produce particular utterances at specific moments,
to chance meetings. They also differ in formality, for example ranging from some types of
interview to gossip sessions. Next, they differ in cooperativeness, with interlocutors clear
ly sometimes agreeing with each other and sometimes arguing, and on governance, with
some activity types being egalitarian and some being autocratic. Finally, communicative
activities may require different degrees of coordination. For example, in interviews, turn-
taking may be “externally managed” so that there is a prearranged ordering of speakers
and their contributions. Clearly, it may also be appropriate to decompose activities into
parts. Thus, a marriage may involve a scripted ceremony, a conventionalized but less for
Page 10 of 17

mal set of speeches, and casual conversations. Importantly, some of these dimensions and
differences may be relevant to prediction and alignment.
Prediction and Alignment

Restrictions of Content and Form.
Levinson (1992) pointed out that different activity types constrain inferences in different
ways. For example, he notes that “It’s five past twelve” indicates that a lecture should
start if it is uttered by a lecturer, with some listeners present, and the lecture is sched
uled to begin by then. In a different activity type, it would not permit this inference. Thus,
comprehenders can draw inferences from utterances when they form part of an activity.
In this case, comprehenders would predict that the lecturer was about to speak and,
specifically, to produce some form of words that are consistent with the beginning of a
lecture (e.g., Today I will talk about…).
Clearly, many activity types constrain participants’ utterances. In an exchange between

shopkeeper and customer (in the United Kingdom), both participants’ topics are highly
restricted. The shopkeeper is limited to questions such as Can I help you?, and the cus
tomer is, in turn, largely limited to asking whether the shop sells a particular product,
where it is on the shelves, or how much it costs; both participants may also make a few
generic comments (e.g., about the weather). The shopkeeper then provides little more
than answers to these questions or requests for further information (e.g., which brand the
customer prefers). Thus, both the content of the utterances and their potential forms are
much more restricted in such interactions than in casual conversation. These convention
al restrictions are particularly clear when one party violates normal “practice” by saying
something not normally associated with the activity type (e.g., the shopkeeper complains
about her unfaithful partner to the customer).
More ritualized interchanges (e.g., initiation ceremonies) or those admitting very little
variation (e.g., military commands) are, of course, even more restricted than exchanges
between shopkeepers and customers. They may require participants to utter quite specif
ic forms of words (e.g., halt, I agree to honor and obey…). In some activity types, partici
pants may respond in different but semantically equivalent ways. For example, bidders at
some auctions can use either words or hand signals (see Kuiper, 1996).
We do not seek to explain the “social” reasons underlying specific conventions. But the
implications of such conventions for language processing are considerable. Importantly,
linguistic activity types can be predictive at different levels. According to Pickering and
Garrod (2013), comprehenders construct forward models at whichever linguistic levels
are sufficiently predictable. In some contexts, the specific word is predictable (short stub
by fingers but just totally covered with…), and so the comprehender can encode this pre
diction in a forward model. In other cases, the word is not predictable but the compre
hender can predict its semantic class (e.g., something edible) or simply its grammatical
category (e.g., noun). But since predictability follows from the activity type as well as the
Page 11 of 17

linguistic context, the extent to which people use forward modeling when comprehending
dialogue (or indeed monologue) also depends on activity type.
For example, when a customer looks at the shelves in a pharmacy, he predicts that the
shopkeeper may speak, and, if she does so, that her (p. 138) utterance is very likely to be a
polite and nondirective question similar to Can I help you, sir? She predicts that the cus
tomer will respond in one of a small number of ways, including No thank you or a contex
tually relevant question. If the customer says Do you have…?, then she predicts a noun
phrase referring to a pharmaceutical product. Thus, the activity type constrains forward
modeling (by both participants) alongside the sentence context.
Thus, we can consider activity types (at any point in an interaction) as varying on the de
gree of constraint. The more restricted the activity type, the greater the degree of predic
tion possible. Clearly, if the participants conform to the conventions of the activity type
(e.g., talking exclusively about pharmaceutical products), then the conversation will be
highly predictable and will tend to be fluent. But the more restricted the activity type, the
greater the disruption if one of the participants abandons the conventions. (Of course, an
activity type can be more constraining for one than the other participant.)
Moreover, we propose that participants in an activity type can “set” their expectations
about the degree of predictability. Comprehenders normally interpret what they hear us
ing a combination of prediction and bottom-up analysis of what they hear. In a noisy but
predictable environment, they emphasize predictions; in a less noisy but unpredictable
environment, they emphasize bottom-up analysis (see Grush [2004] and Pickering & Gar
rod [2007] for discussion in terms of Kalman gain). So when participants know that the
activity type is highly constraining, they are more likely to act on their predictions.
Degree of Coupling and Coordination.
Pickering and Garrod (2004) noted that interactive activities vary according to the degree
of coupling between the interacting agents. A tightly coupled activity such as ballroom
dancing requires continuous coordination between partners, but a loosely coupled activi
ty such as golf only requires intermittent coordination (one may have to wait until one’s
partner has struck the ball, quality of play may be affected by how close the scores are,
etc.). The same is true of conversation. Whereas holding a one-to-one intimate conversa
tion may require precise and continuous coordination (e.g., interruptions, joint construc
tion of utterances, clarification requests), a diagnostic consultation between doctor and
patient might involve long contributions by the patient (e.g., descriptions of symptoms)
interrupted by extensive recommendations by the doctor (e.g., recommendations about
treatment). In general, we claim that the more tightly coupled the dialogue, the more pre
dictable it will be and the greater the extent of prediction. But what characteristics might
lead to a high degree of coupling? Here we sketch out some suggestions.
In casual conversation, interlocutors are free to contribute whenever they want, subject
to the normal rules of conversation, such as not interrupting and not responding to ques
tions addressed to someone else (Sacks et al., 1974). But this is not the case when the
Page 12 of 17

conversation is externally managed. In a job interview, each member of the appointment

panel may take a turn in asking a question of the applicant; it may not be acceptable to
speak out of turn. In fact, the order of speaking may be decided before the interview be
gins. Thus, forward modeling by members of the panel may be less relevant to their actu
al contributions (although, of course, it can be relevant for the addressee). Moreover, an
interlocutor is less likely to speak or to build on any previous speaker’s utterance, and, in
addition, such a dialogue is presumably less repetitive than free conversation. Thus, com
prehenders may de-emphasize prediction under such circumstances.
However, most dialogue is “internally managed,” so that interlocutors themselves have to

decide when it is appropriate to speak. They are remarkably good at this because inter-
turn intervals are extremely close to zero milliseconds in most exchanges (e.g., Stivers et
al., 2009). This means that the addressee knows precisely when it is appropriate to start
speaking. To do this, the addressee can use the forward model to determine (in effect)
how long an utterance the addressee would produce at that point and assumes that the
speaker will behave similarly. In accord with this, de Ruiter, Mitterer, and Enfield (2006)
found that experimental participants were very accurate at predicting the end of speak
ers’ utterances. Interestingly, their predictions remained accurate when the intonation
was removed and the words were left, but not vice versa. This follows from our account
because the phonological, syntactic, and semantic predictions are largely driven by the
words rather than by the intonational contour.
In other cases, however, constraints on the degree of coupling seem to result from institu
tional practices. For example, the formal structure of interviews (long contributions, no
interrupting, etc.) might occur so that the contribution of the candidate can be precisely
analyzed afterward or so that all candidates are treated the in same way. In the (p. 139)
case of the shopkeeper, the explanation may relate to the need to have an efficient means
of performing a transaction and the need to avoid the difficulty of raising most topics with
a stranger.
Because prediction and alignment relate to each other, the activity types in which predic
tion is made easier are likely to be those in which interlocutors are already well-aligned.
A crucial feature of activity types is that all the participants recognize them as such. In
other words, they already share a common understanding of the activity. This, of course,
means that they are already aligned with respect to an aspect of the situation model. Mis
interpretation of the activity type is, of course, possible (e.g., failure to recognize that a
dinner is actually an interview), in which case alignment is likely to break down in differ
ent ways.
So different types of conversation will differ in how predictable they are at levels relating
to word choice, grammatical form, content, or structure of turn-taking. It is therefore im
portant to categorize types or dimensions of conversation with respect to their pre
dictability. More generally, psycholinguists should not assume that data derived from one
form of language use (such as monologue or cooperative task-oriented dialogue) must be
informative about other forms of language use.
Page 13 of 17

Conclusion: Alignment and Conversational

Success
Conversational interactions are remarkably successful, given their enormous complexity.
Our explanation of how this can be has two parts. In previous publications, we proposed
that such success occurred because interlocutors align their linguistic representations
(words, grammar, meaning, etc.) and that such alignment leads to alignment of their situ
ation models in a largely automatic manner (Garrod & Pickering, 2004; Pickering & Gar
rod, 2004). But, in addition, there is good evidence that language production and compre
hension are interwoven, with producers using comprehension-based representations to
predict their own utterances and comprehenders using production-based representations
to predict their interlocutor’s utterances (Pickering & Garrod, 2013). Most of the evi
dence for such prediction comes from studies of production and comprehension in mono
logue, but, in this chapter, we argue that it plays a particularly important role in dialogue.
On the one hand, the constant switching between production and comprehension in dia
logue is likely to keep the production system active during periods of comprehension. On
the other hand, any such prediction is likely to enhance alignment in dialogue because
prediction-by-simulation involves covert imitation. The degree of such prediction and
alignment is likely to be affected by specific characteristics of the dialogue. Thus, the
study of types of dialogue (e.g., with respect to their degree of coupling) should affect
how the mechanisms of language production and comprehension are used. Most general
ly, we argue that dialogue provides a clear example of a way in which prediction facili
tates social interaction. It therefore shows how it is possible to develop an approach to an
aspect of cognition that provides a basis for interaction and thus helps begin to break
down the rigid division between the individual and the social.
References
Barsalou, L. W. (1999). Perceptual symbol systems. Behavioral and Brain Sciences, 22,
577–600.
Bock, J. K. (1986). Syntactic persistence in language production. Cognitive Psychology,

18, 355–387.
Branigan, H. P., Pickering, M. J., & Cleland, A. A. (2000). Syntactic coordination in dia
logue. Cognition, 75, B13–B25.
Branigan, H. P., Pickering, M. J., McLean, J. F., & Cleland, A. A. (2007). Syntactic align
ment and participant role in dialogue. Cognition, 104, 163–197.
Brennan, S. E., & H. H. Clark (1996). Conceptual pacts and lexical choice in conversation.
Journal of Experimental Psychology: Learning, Memory, and Cognition, 22, 1482–1493.
Page 14 of 17

D’Ausilio, A., Jarmolowska, J., Busan, P., Bufalari, & Craighero, L. (2011). Tongue corti
cospinal modulation during attended verbal stimuli: Priming and coarticulation effects.
Neuropsychologia, 49, 3670–3676.
DeLong, K. A., Urbach, T. P., & Kutas, M. (2005). Probabilistic word pre-activation during
comprehension inferred from electrical brain activity. Nature Neuroscience, 8, 1117–
1121.
De Ruiter, J. P., Mitterer, H., & Enfield, N. J. (2006). Projecting the end of a speaker’s turn:
A cognitive cornerstone of conversation. Language, 82, 515–535.
Dikker, S., Rabagliati, H., Farmer, T. A., & Pylkkänen, L. (2010). Early occipital sensitivity
to syntactic category is based on form typicality. Psychological Science, 21, 629–634.
Fodor, J. A., Bever, T. G., & Garrett, M. F. (1974). The psychology of language. New York:
McGraw-Hill.
Garrod, S., & Anderson, A. (1987). Saying what you mean in dialogue: A study in concep
tual and semantic co-ordination. Cognition, 27, 181–218.
Garrod, S., & Pickering, M. J. (2004). Why is conversation so easy? Trends in Cognitive
Sciences, 8, 8–11.
Gaskell, G. (2007). Oxford handbook of psycholinguistics. Oxford: Oxford University

Press.
Grush, R. (2004). The emulation theory of representation: Motor control, imagery, and
perception. Behavioral and Brain Sciences 27, 377–435.
(p. 140) Kuiper, K. (1996). Smooth talkers. New York: Erlbaum.
Levelt, W. J. (1989). Speaking: From intention to articulation. Cambridge, MA: MIT Press.
Levinson, S. C. (1992). Activity types and language. In P. Drew & J. C. Heritage (Eds.),
Talk at work: Interaction in institutional settings (pp. 66–100). Cambridge: Cambridge
University Press.
Linell, P. (1998). Approaching dialogue. Talk, interaction and contexts in dialogical per
spectives. Amsterdam: John Benjamins.
Linell, P. (2009). Rethinking language, mind and world dialogically: Interactional and con
textual theories of human sense-making. Charlotte, NC: Information Age Publishing.
Pickering, M. J., & Garrod, S. (2004). Toward a mechanistic psychology of dialogue. Be
Page 15 of 17

Pickering, M. J., & Garrod, S. (2007). Do people use language production to make predic
tions during comprehension? Trends in Cognitive Sciences, 11, 105–110.
Pickering, M. J., & Garrod, S. (2013). An integrated theory of language production and
comprehension. Behavioral and Brain Sciences, 36, 329–392.
Pulvermüller, F. (2005). Brain mechanisms linking language and action. Nature Reviews
Neuroscience, 6, 576–582.
Sacks, H., Schegloff, E. A., & Jefferson, G. (1974). A simplest systematics for the organiza
tion of turn-taking for conversation. Language, 50, 696–735.
Schober, M. F., & Clark, H. H. (1989). Understanding by addressees and overhearers.

Cognitive Psychology, 21, 211–232.
Sebanz, N., Bekkering, H., & Knoblich, G. (2006). Joint action: Bodies and minds moving
together. Trends in Cognitive Sciences, 10, 70–76.
Stivers, T., Enfield, N. J., Brown, P., Englert, C., Hayashi, M., Heinemann, T., Hoymann, G.,
Rossano, F., de Ruiter, J. P., Yoon, K.-E., & Levinson, S. C. (2009). Universals and cultural
variation in turn-taking in conversation. Proceedings of the National Academy of
Sciences, 106, 10587–10592.
Tannen, D. (1989). Talking voices: Repetition, dialogue, and imagery in conversational dis
course. Cambridge: Cambridge University Press.
Van Berkum, J. J. A., Brown, M. C., Zwitserlood, P., Kooijman, V., & Hagoort, P. (2005). An
ticipating upcoming words in discourse: Evidence from ERPs and reading times. Journal
of Experimental Psychology: Learning, Memory, and Cognition, 31, 443–467.
Wolpert, D. M. (1997). Computational approaches to motor control. Trends in Cognitive

Sciences, 1, 209–216.
Yuen, I., Davis, M. H., Brysbaert, M., & Rastle, K. (2010). Activation of articulatory infor
mation in speech perception. Proceedings of the National Academy of Sciences, 107, 592–
597.
Zwaan, R. A., & G. A. Radvansky (1998). Situation models in language comprehension

Martin J. Pickering
Martin J. Pickering is a professor in the Department of Psychology at the University

of Edinburgh in the UK.
Simon Garrod
Simon Garrod, Department of Psychology, University of Glasgow
Page 16 of 17

Page 17 of 17

Cognitive and Social Aspects of Coherence

Rachel Giora
Print Publication Date: Sep 2014

Subject: Psychology, Social Psychology, Cognitive Psychology
This chapter outlines the conditions for coherence of informative discourses. It reviews a
number of factors that seem to influence the degree of coherence of such non-narrative
discourses, such as degree of relevance, informativeness, and meaning salience, but not
cohesion. The requirements for discourse coherence are not linguistic but cognitive and
should apply to nonverbal discourses as well. And although verbal discourses unfold lin
early, they are represented hierarchically, in terms of prototype category, so that each in
coming proposition is weighed against the category (common) features—the discourse
topic proposition—in terms of similarity (relevance) and graded informativeness. Overt vi
olations of these requirements aim to induce pleasure, convey emotions, and challenge
norms of powerful groups while attempting to bring about social change.
Keywords: Coherence, cohesion, relevance, graded informativeness, hierarchical representation, prototype cate
gory, discourse topic, social change
In this chapter, I examine the notion of discourse coherence as it evolved from assuming
that cohesion is a necessary condition for coherence to accounting for coherence in terms
of pragmatic norms, principles of cognitive organization, and degree of meaning salience.
Given that breaching coherence is not a rare phenomenon, I review the benefits and ef
fects of such violations in terms of aesthetics, affect, and the sociolinguistics of language
change.
Cohesion
What makes a discourse coherent? What accounts for the well-formedness of a string of
utterances? Should a concept such as discourse coherence or discourse well-formedness
be reflected in discourse surface structure? Alternatively, given its conceptual nature,
might coherence be unobservable in terms of relations between adjacent sentences?
Page 1 of 23

The idea that the manifestation of linguistic competence might be observed not just with
in the sentence boundaries but also beyond the sentence level was promoted by a number
of linguists during the 1970s and 1980s. It was widely agreed then that cohesion—a sur
face structure phenomenon—both reflects and enables discourse coherence (e.g., Bellert,
1970; Daneš, 1974; Enkvist, 1978; Gutwinski, 1976; Halliday & Hasan, 1976; Reinhart,
1980; Vuchinich, 1977; inter alia; for reviews, see Reinhart, 1980; Giora, 1985a). The at
tempt was to account for discourse well-formedness in terms of the relations obtaining
between adjacent sentences. Outstanding in this respect were the Prague functionalists
who, among other things, were trying to explain discourse coherence in terms of surface-
level features of language structuring used to align linearly ordered sentential sequences
(Daneš, 1974; 1987). These kinds of surface-level structural relations constitute the cohe
sion of discourse.
Importantly, cohesion was considered a necessary condition for discourse coherence. The
claim (p. 142) was that, for a discourse to be well-formed, its sentences must be cohesive,
in that their sentence topics must be controlled by a previous mention (on the notion of
sentence topic, see Reinhart, 1981). In this framework, cohesion amounts to (sentence)
topic control. According to Daneš and Reinhart, if such topics are governed by a prior
mention in the immediate context, this would effect a cohesive string of sentences. Coher
ent discourses, then, must be made up of such cohesive strings (for a different view, see
Carrell, 1982).
For the purpose of this chapter, Daneš’s (1974) theory, outlining the ways in which a well-
formed discourse can linearly unfold—as well as Reinhart’s (1980) attempt to improve on
his, as well as Halliday and Hasan’s (1976) theory—will serve to introduce the notion of
cohesion and its contribution to discourse coherence.
In his seminal paper, Daneš (1974) delineated the various ways a given sentence topic
may be governed by a previous mention. He set out from the assumption that any nonini
tial sentence is made up of at least two information units—the theme (i.e., sentence topic)
and the rheme (i.e., comment) parts. The theme is that part of the sentence that the sen
tence is about and that conveys old or given information; the rheme is that part of the
sentence that conveys new information about the theme. Thus, in (1b) below, the theme—
She—(in bold, for convenience)1 is given or old information, referred to by a high accessi
bility marker (Ariel, 1990; 2006) since it is mentioned previously in (1a) (Jawaher Abu
Ramah); in contrast, Israel Defense Forces soldiers makes up (part of) the rheme section
of (1b), since it conveys new information about the theme. In (1c), however, the theme
These is given, referred to by a high accessibility marker, denoting facts mentioned in the
preceding sentences. In (1d), the theme (The IDF) is referred to as given—hence the defi
nite description and abbreviation/initials; Bassem, Jawaher’s brother, however, is new in
formation and makes up (part of) the rheme section of (1d). In (1e), however Bassem is re
ferred to by a high accessibility marker, denoting given information—He. As a whole,
then, the text in 1(a–e) is a cohesive discourse in that all its subsequent sentences’ topics
are governed by a previous a mention.
Page 2 of 23

(1) (a) Jawaher Abu Ramah died young. (b) She stood facing the demonstra
tors against the separation fence in her village, inhaled very large quantities
of the gas that Israel Defense Forces soldiers fired that day, collapsed and
died several hours later at a Ramallah hospital. (c) These are definitive facts.
(d) The IDF should have immediately issued a statement expressing sorrow
for the death of the demonstrator, and said it would investigate the excessive
means used for dispersing demonstrations at Bil’in, which had killed Bassem,
Jawaher’s brother, for no reason. (e) He was hit by a gas canister fired di
rectly at his chest two and a half years ago.
(Levy, 2011)
According to Daneš’s (1974) corpus-based studies, there are three ways for a text to pro
ceed cohesively from a given sentence to the next:
• Rheme–theme progression: Whereby a given sentence theme is controlled by a previ

ous rheme
• Theme–theme progression: Whereby a given sentence theme is controlled by a previ
ous theme; and
• Hypertheme–theme progression: Whereby a given sentence theme is controlled by a
previously mentioned hyper-theme.
Rheme–theme progression is exemplified by (1b–e). Theme–theme progression operates

when a given theme or sentence topic is controlled by a reference to a previous theme, as
exemplified by (2): the sentence topic of (2a) is Defense Minister Ehud Barak because the
sentence is about him; the sentence topic of (2b) is Barak, controlled by Defense Minister
Ehud Barak—the previous topic mentioned in (2a); similarly, the sentence topic of (2c)—0
—is controlled by Barak in (2b) to which it refers. Example (2) then amplifies Theme–
theme progression:
(2) (a) Defense Minister Ehud Barak spoke before an audience Tuesday as
a guest speaker in an Iran seminar held in the Tel Aviv University. (b) Barak
discussed the Islamic Republic’s nuclear program and the Middle East’s polit
ical-security state but (c) [0] was repeatedly interrupted by protestors who
waved pictures of Palestinian victims.
(Fyler, 2011)
Hypertheme–theme progression is exemplified by (3). A hypertheme is the theme of a

whole section, the topics of whose various sentences/subsections make up parts of that
theme. For instance in (3), New Jersey is the hypertheme—the topic of the whole text seg
ment:
(3) New Jersey is flat along the coast and southern portion; the northwest
ern region [of New Jersey] is mountainous. The coastal climate [of New
Jersey] is mild, but there is considerable (p. 143) cold in the mountain areas
Page 3 of 23

during the winter months. Summers [of New Jersey] are fairly hot. The
leading industrial production [of New Jersey] includes chemicals,
processed foods, coal, petroleum, metals and electrical equipment. The most
important cities [of New Jersey] are Newark, Jersey City, Paterson, Tren
ton, Camden. Vacation districts [of New Jersey] include Asbury Park, Lake
wood, Cape May, and others.
(Daneš, 1974, p. 120)
It is not the case, though, that a specific text should follow only one type of thematic pro
gression. A specific text’s progression can alternate between the various modes. For in
stance, (4) exemplifies both Theme–theme as well as rheme–theme progressions. Thus,
the repetitive topics of the first (4a) and the second (4b) sentences—they—allow for the
second theme to be controlled by the first. The third theme—he—in (4c), however, is con
trolled by the rheme of the previous sentence—the soldier—making up a rheme–theme
text-progression:
(4) (a) They managed to catch him. It was an all-out abuse. They abused him,
and I don’t think something was done about it. (b) They put him in the toilet, I
remember the soldier, I remember, he was a friend of mine, a friend from
the company. (c) And he took pride in shoving the kid’s head into the toilet.
(Anonymous [Sergeant], 2000)
Another attempt at studying the various types of cohesive devices can be found in Halli
day and Hasan’s (1976) wide-ranging research. Halliday and Hasan came up with five co
hesive devices: reference, ellipsis, substitution, lexical repetition, and connection. In
1980, Reinhart revised both views, that of Daneš’s as well as that of Halliday and
Hasan’s, subsuming them under two modes of text progression: (1) topic control via a ref
erential link—via controlling the referent of the previous topic (or scene-setting expres
sion)—rather than via mere repetition of a previous lexeme (as posited by Halliday &
Hasan, 1976); and (2) use of an explicit semantic connector when topic control fails as a
result of, for example, introducing a new sentence topic to the discourse.
The text in (5), Enkvist’s (1978, pp. 110–111) example cited in Reinhart (1980), is not co
hesive in that it lacks referential link—a reference to a referent mentioned previously. In
stead, it exhibits only lexical cohesion:
(5) I bought a Ford. The car in which President Wilson rode down the
Champs Elysees was black. Black English has been widely discussed. The
discussions between the presidents ended last week. A week has seven
days. Every day I feed my cat. Cats have four legs. The cat is on the mat.
Mat has three letters.
Although the text in (5) exhibits lexical repetition, it is not cohesive because, claims Rein
hart, it does not satisfy the requirement for topic control via a referential link.
Page 4 of 23

In (6), a semantic connector—At the same time—is used when a topic control via a refer
ential link is absent because a new sentence topic is introduced—the Jerusalem Munici
pality and the Housing and Construction Ministry:
(6) The Ir Amim organization, which encourages coexistence in Jerusalem be

tween Jews and Arabs notes that these plans, though lacking statutory status,
serve as a basis for demolition of Palestinian homes in these areas and for
scuttling building plans by Palestinians there. At the same time, the
Jerusalem Municipality and the Housing and Construction Ministry
lend a hand and even provide assurances to right-wing organizations which
declare their intention to make the Old City and the more extensive Holy
Basin more Jewish.
(Eldar, 2009)
Similarly, in (7; taken from Giora, 1990), a shift from we (the sentence topic of the first
sentence) to a new sentence topic a stamp collector, in the next sentence, is marked by
semantic connectors—A stamp collector, for example, also faces a similar problem:
(7) When we want to classify the living organisms in terms of the amount of
similarity and difference which they share, the question that arises immedi
ately is which features constitute the basis for establishing similarity and dif
ference between animals: the external shape, their place of habitat, their in
ternal structure or the activity? A stamp collector, for example, also faces a
similar problem when he wants to catalogue his stamp collection indepen
dently.
In sum, a cohesive text should exhibit (a) a referential link either via theme–theme,
rheme–theme, or hypertheme–theme progression; when these fail to be satisfied, (b) a se
mantic connector should be used to guarantee a cohesive text progression. Given that co
hesion is taken to be a necessary condition for text coherence, a coherent (p. 144) text
must unfold cohesively (see also Hoey, 1991, for a review).
To argue against the view that cohesion is a necessary condition for text coherence, one
should therefore show that a coherent text need not be cohesive (as shown in Giora,
1985a; 1985b). In the following (8), the text segment is coherent but not cohesive: adja
cent sentences are not referentially linked, nor do they connect to each other via an ex
plicit semantic connector. The topic of this discourse segment can be phrased as follow:
“Israeli politics is chauvinistic.” However, none of the consecutive sentences refers to Is
raeli politics; instead, each sentence provides support for this generalization by bringing
up all supportive evidence/instances. Even if there are some repetitions, they are lexical,
not referential—not referring to a specific previous referent:
(8) The truth is that Israeli politics continue to be frighteningly chauvinistic,

even if the current Supreme Court president and the former Knesset speaker
are women. There is not a single woman of influence and importance in the
Page 5 of 23

current Israeli cabinet. In Western Europe and the United States such a situa
tion would not be possible. The forum of seven senior cabinet ministers is an
exclusive men’s-only club, and let’s not waste our time with the defense es
tablishment. War and peace, budgets and Iran—everything is in the hands of
men. Just men.
(Levy, 2010)
Coherence
The view that cohesion is a necessary condition for coherence was undermined by later
studies into coherence (Giora, 1985a; 1985b). They showed that (a) unlike cohesion, co
herence is not a property or a relation obtaining between linearly ordered utterances;
and (b) for a text to be coherent, it need not manifest either a referential link (as posited,
for example, by Daneš [1974] and Reinhart [1980]) or an explicit semantic connector,
when the referential link fails (as posited by Reinhart, 1980). As illustrated by example
(8), cohesion is not a necessary condition for discourse coherence. What is then?
A Pragmatic Approach: The Standard Pragmatic Model
According to Grice’s (1975) model, dubbed the standard pragmatic model, any coherent
act of communication should follow the cooperative principle, which includes a tacit
agreement between interlocutors to conform to four requirements or maxims:
1. Maxim of Quantity—Be informative

• Make your contribution as informative as is required for the current purposes of
the exchange.
• Do not make your contribution more informative than is required.
2. Maxim of Quality—Be Truthful

• Only say what you believe to be true.
• Only say what you have evidence for.
3. Maxim of Relation—Be Relevant.

• Make your contribution relevant to the interaction.
4. Maxim of Manner—Be Clear

• Avoid unnecessary wordiness.
• Avoid ambiguity.
• Be brief.
• Be orderly.
Page 6 of 23

Indeed, example (8) meets all the requirements for discourse well-formedness. However,
according to Grice, for discourse well-formedness to be accomplished, one need not con
form to all the requirements. In fact, flouting one or more maxims need not impair coher
ence significantly as long as this violation is overt. Unlike a covert violation, an overt vio
lation still observes the cooperation principle, allowing addressees to assume a rational
speaker who intends them to detect the violation and derive a conversational implicature.
For illustration, consider the text in (9). Although this discourse flouts the maxim of quali
ty, stating what the speaker does not believe to be true (the “splendid job” of our fine pi
lots), this violation is overt and alerts the addressee to the ironic implicature:
(9) “Hooray to the Israeli Airforce pilots doing a splendid job” effused
Brigadier General Avi Benayahu, the IDF spokesperson, talking to Yonit Levy
—white turtleneck against a background of tanks, vis à vis hundreds of funer
als in Gaza—a token of the “splendid job” of our fine pilots.
(Levy, 2008)
As plausible as this theory is, it leaves a number of notions somewhat loosely defined. For
instance, it is not quite clear what is meant by the requirement to be informative: how
much information does one need to contribute so as not to be either overly informative or
not informative enough vis-à-vis the conversational situation? It is also not quite clear
what being relevant to the conversational interaction means.
(p. 145) A Cognitive Approach: Categorical Organization
In an attempt to better define Grice’s requirements for relevance and informativeness,

Giora (1985a; 1985b; 1988) proposed to view coherence of non-narrative discourse as ac
counted for by cognitive rather than communicative principles. Discourse organization,
she claimed, follows the same cognitive principles governing the organization of general
knowledge. According to Giora, non-narrative discourse will be coherent if it is organized
along the lines proposed to account for the structure of prototype-oriented taxonomic cat
egories (whether stable, e.g., birds, fruits, furniture, musical instruments, etc. or ad hoc;
see Barsalou, 1983).
As shown by Rosch (1973; 1975; 1978), Rosch and Mervis (1975), and Tverski (1977),
membership in a taxonomic category depends on the degree of shared features defined in
terms of family resemblance (Wittgenstein, 1953) rather than in terms of necessary and
sufficient conditions. Assuming similarity between category members results in a hierar
chical internal ordering from the least informative most redundant category member—the
prototype, sharing most of the features of the category members—to the least similar
most informative member, sharing few features with the category members while also
sharing a great number of features with noncategory members.
Based on shared features, a taxonomic category is thus organized in memory relative to

its prototypical member(s). Since (a) a prototype reflects the redundancy structure of the
Page 7 of 23

category, it is (b) the most accessible member of the set and, consequently, (c) the mem
ber that functions as the reference point relative to which inclusion of new members, in
terms of their similarity to it, is determined.
The assumption that a coherent discourse is organized along the same principles govern
ing categorical organization (Giora, 1985b; 1988) predicts that a coherent discourse
should be prototype-oriented: (a) it should boast of a prototypical member—a proposition
reflecting the redundancy structure of the text, whether derivable or made explicit (Gio
ra, 1985b), that should be most accessible and function as the reference point relative to
which inclusion of oncoming propositions is determined (the relevance requirement; Gio
ra, 1985b); and (b) its internal structure should be hierarchical, ordered linearly from the
least to the most informative proposition (the graded informativeness requirement; Giora,
1988).
In this framework, the prototypical member of a discourse (or, alternatively, the category/
discourse name/title) is termed the discourse topic (DT). Being a member of this linguistic
category, the DT must be a proposition (see also van Dijk, 1977); alternatively, it must at
least consist of an argument and a predicate, regardless of syntactic structure. For in
stance, the DT of (10) is “The weather (the argument) is going to be wet and wild for the
week ahead” (the predicate). When, for stylistic reasons, the full-fledged DT is truncated
and represented via one of its parts, such as the predicates (as in “wet and wild”), the DT
as a whole must be retrievable from the text in question:
(10) Wet and wild.2 Stormy weather returns Monday, with rain, hail and thun
derstorms. Flooding is highly likely in low-lying areas. Temperatures will drop
significantly, with strong wind. The rain will cease in the south in the after
noon. Expect more of the same through Thursday.
A DT is a generalization—a summary/list of the text utterances’ shared features. Thus,

each of the propositions in example (10) (e.g., Stormy weather returns Monday, with rain,
hail and thunderstorms) is related to the DT via its shared features (“The weather is go
ing to be wet and wild for the week ahead”). Even though, as a consequence of bearing
similarity to the DT, each proposition may also share common features with its neighbor
ing propositions, this is only a byproduct. In essence, each proposition is primarily relat
ed to a higher order proposition—the DT.
The Relevance Requirement

Assuming a categorical organization thus requires that each sentence be relevant to the
DT, that is, be related to or subsumed under a generalization that it instantiates or ex
tends in an informative manner. The relation between the various sentences/segments of
a text, then, is not linear, but hierarchical.
Since the DT is so vital to the organization of the discourse, its preferable linear position
is discourse or discourse-segment initial. Indeed, as shown in Giora (1985b), reading
Page 8 of 23

times of texts identical in every respect except for the position of the DT statement
showed that DT initial position facilitated reading times compared to DT final position.
Given this categorically based account, to be relevant, then, a contribution must be relat
ed to a DT; that is, it must be similar to or “about” a DT. When it is not, as when a new DT
or a sub-DT is (p. 146) introduced, this digression from relevance must be marked explicit
ly, whether before it is introduced or afterward.
To illustrate, consider the text in (11), which is evidently coherent. Its DT is “The possible
reasons for why most of the public [in Israel] is refusing to be vaccinated against swine
flu” stated in the first sentence. This DT utterance is followed by a list of “reasons for why
most of the public [in Israel] is refusing to be vaccinated against swine flu.” Still, not all
the text’s utterances are related to the DT. For one, the introduction of reason (c) “be
cause doctors and nurses have not been in a hurry to get vaccinated themselves” is fol
lowed by a digression to a local counterargument to (c) (“That’s not so terrible, one doc
tor said on the radio: After all, more than one doctor has advised a patient to stop smok
ing and then lit up a cigarette”). However, when the next utterance is introduced (d), it
re-establishes relevance to the DT, signaled by an explicit marker “however” (“Most influ
ential of all, however, is that the politicians, who have never turned down something they
are offered for free, are by and large refusing to be vaccinated”):
(11) Why is most of the public refusing to be vaccinated against swine flu? (a)
Perhaps because of the wide publicity given those who died, children in par
ticular, after receiving the vaccine. (b) Also because of the conflicting opin
ions about the kinds of vaccines on offer. (c) But more than either of these,
it’s because doctors and nurses have not been in a hurry to get vaccinated
themselves. That’s not so terrible, one doctor said on the radio: After all,
more than one doctor has advised a patient to stop smoking and then lit up a
cigarette. (d) Most influential of all, however, is that the politicians, who have
never turned down something they are offered for free, are by and large re
fusing to be vaccinated. (e) And the most embarrassing aspect is Netanyahu’s
refusal to be immunized. Does he know something he isn’t telling us?
(Marcus, 2010)
Although texts that include marked digressions are coherent (more coherent than texts
whose digressions are not marked as such), they are perceived as less coherent than
texts that include no digression at all. Indeed, comprehending texts with digressions, al
beit marked, is often more erroneous and effort consuming than comprehending the same
texts including no digressions (Giora, 1993).
In sum, for a discourse to be coherent, it must obey the relevance requirement (Giora,
1985b) as follows: (a) all of the discourse’s propositions or segments must be relevant/re
lated/similar to its DT—the generalization subsuming the features of the discourse seg
Page 9 of 23

ments in question; or (b) if failing to meet (a), then the digressing from the DT must be
explicitly marked, either before or after the digression.
The Informativeness Requirement

As can be predicted from the principles of categorical organization, whose internal struc
turing is hierarchical, for a discourse to be coherent, it must evolve hierarchically, as stip
ulated by the graded informativeness requirement (Giora, 1988). According to the graded
informativeness requirement, (a) each of the discourse’s propositions must not only be
relevant/similar to a DT but also informative about it, relative to its preceding proposi
tion. That is, each proposition must add new or more specific information about the DT
that is more, or at least not less informative than its preceding proposition (vis-à-vis the
DT). And, (b) if failing to meet (a), then the digression from informativeness must be ex
plicitly marked, either before or after the digression. (For digression markers signaling
givenness, such as as mentioned earlier, as you may recall, after all, see Ariel, 1985;
1988).
To illustrate the first graded informativeness requirement, consider again example (10).
In this weather forecast text, all the propositions share the category set of features (“wet”
“wild” “weather”) reflected in the DT, while also featuring specific properties (“rain,”
“hail,” “thunderstorms,” “flooding,” “low temperatures,” etc.), which provide instantia
tions for the generalization in the DT. Although each proposition adds specific information
about the DT not included in a preceding proposition, the last proposition is the most in
formative one, including all the previous features while adding a new dimension (the du
ration of this kind of weather).
To illustrate the second graded informativeness requirement, consider example (12), in

which information previously mentioned is marked as digressing from graded informa
tiveness by explicit digression markers (such as like I said earlier, I told you):
(12) …So I followed him and the officer in charge of the police station inside
the Allenby Bridge Terminal. It belongs to the Ma’ale Adumim police, like I
said earlier. We got to the station and they went into a room and I was out
side. I didn’t know why I had to stay there in the first place I didn’t. I don’t
know, something made (p. 147) me stay, I recall now. Across the closed door I
heard everything that went on inside. There was yelling: What are you doing?
How dare you behave this way? Anyway, I started hearing blows, a beating. I
didn’t know what to think. I felt quite frozen, didn’t know what to do. What,
could I break in and tell the commander to stop it? It wasn’t…You know, there
was this respect for superiors, there’s no way you’d open up and talk. Tell
them how to behave? Who, me? I was just a controller. What did you do? I
told you, I wasn’t even supposed to be there. I didn’t leave, I stayed. I want
ed to see if he really hit him or anything.
(Women Breaking the Silence, 2009)
Page 10 of 23

Degree of informativeness then regulates text ordering. Whereas the information that
category members share is redundant in that category, information they do not share is
informative in that category and determines internal ordering. The more distinctive fea
tures a category member has, the more informative it is, and hence, the further it is re
moved (in terms of representation in memory) from the prototypical, least informative
member in that set. Categorical organization then predicts graded internal structuring
from the least to the most informative message that is still relevant to the DT. Thus,
whereas degree of informativeness determines hierarchal organization, relevance to the
DT determines the boundaries of a discourse segment (for empirical evidence, see Giora,
1988).
Still, one could ask: how informative is informative? According to classical information
theories (e.g., Attneave, 1959; Shannon, 1951), a message is informative to the extent
that it reduces uncertainties. Given that a specific question/message opens up a number
of equally probable answers/options, the greater the number of possible answers a mes
sage reduces, the more informative it is. To count as informative, however, it is enough
that a message reduces uncertainties by half. In terms of categorical organization, a giv
en message may give rise to a limited number of possible answers/options—the most
probable or prototypical ones. To be informative, then, a message should reduce that
small set at least by half (Giora, 1988).
Indeed, in (10), “Wet and wild” opens up questions/uncertainties such as to what extent
wet? To what extent wild? When? Where? The next messages indeed reduce some of
these uncertainties (“Stormy weather returns Monday, with rain, hail and thunder
storms. Flooding is highly likely in low-lying areas. Temperatures will drop signifi
cantly, with strong wind. The rain will cease in the south in the afternoon. Expect
more of the same through Thursday”), thus becoming increasingly informative.
But even a question with few alternatives such as “when are you coming home? At 20:00
or at 21:00?” can be answered informatively by reducing one of these two alternatives
(e.g., “at 21:00”). A less prototypical answer in that context would be 03:00; it will be
highly informative (least probable) since it will reduce a great number of uncertainties.
In contrast, a text such as (13) is hardly informative because it reduces almost no uncer
tainty, being highly repetitive:
(13) She was gay exactly the same way. She was never tired of being gay that
way. She had learned many little ways to use in being gay. Very many were
telling about using other ways in being gay. She was gay enough, she was al
ways gay exactly the same way, she was always learning things to use in be
ing gay, she was telling about using other ways in being gay, she was telling
about learning other ways in being gay, she was learning other ways in being
gay, she would be using other ways in being gay, she would always be gay in
the same way, when Georgine Skeene was there not so long each day as when
Georgine Skeene was away.
Page 11 of 23

(Stein, 1922)
Still its most informative message (when Georgine Skeene was there not so long each day
as when Georgine Skeene was away) appears in segment final position, as would be pre
dicted by the graded informativeness requirement. The most informative message that is
still relevant to the segment DT signals segment boundary.
Discourse Segmentation
Recall that categorical organization predicts graded internal structuring from the least to
the most informative message that is still relevant to the DT. Thus, whereas degree of in
formativeness affects hierarchal organization, relevance to the DT determines the bound
aries of a discourse segment. In other words, the most informative message that is still
relevant to the given DT marks a discourse segment boundary (see Giora, 1988; on full
pronouns as segment boundary markers in Chinese, see Giora & Lee, 1996). How then
should discourse proceed?
According to the graded informativeness requirement, the introduction of a new DT moti

vates discourse progression and hence discourse segmentation. Assuming categorical or
ganization, (p. 148) there should be two preferred (i.e., coherent) ways for a text to pro
ceed from a given to a new DT. Given a specific segment (paragraph/chapter/text), a new
DT either induces segmentation and starts a new discourse segment, in which case it is
introduced initially in that new segment (14a), or it is introduced in that given segment fi
nal position so as to become the next DT and be discussed in the next segment (14b).
Both segment initial and segment final positions enjoy prominence due to the primacy
and recency effects that make them most available for future discussion (see Giora
1983a; 1983b).
Although in (14a) the DT of the first paragraph discusses a UN resolution that calls on Is
rael to open its nuclear program for inspection (bold added), the DT opening the para
graph that follows is new—providing background information about the character of such
UN resolutions (bold added). In (14b), on the other hand, the new DT of the second para
graph—the accidental discovery of penicillin—is introduced in the previous paragraph fi
nal position (Penicillin is a result of such a discovery). These two strategies thus allow for
a seamless text progression. In contrast, new information introduced in segment mid-po
sition is not intended as the next new DT (italics added). Addressing it as such (bold
added) should result in a less coherent text progression, as exemplified by the humorous
text taken from Monty Python and The Holy Grail (14c)3:
(14a) The UN general assembly has overwhelmingly approved a resolu

tion calling on Israel to open its nuclear programme for inspection.
The resolution, approved by a vote of 174 to six with six abstentions, calls on
Israel to join the Nuclear Nonproliferation Treaty (NPT) “without further de
lay” and open its nuclear facilities to inspection by the International Atomic
Energy Agency. Those voting against were Israel, the US, Canada, Marshall
Islands, Micronesia and Palau. Resolutions adopted by the 193-member
Page 12 of 23

general assembly are not legally binding but they do reflect world
opinion and carry moral and political weight. And the resolution adds to
pressure on Israel as it faces criticism over plans to increase settlement in
the West Bank, a move seen as retaliation for the assembly recognising Pales
tinian statehood.
(14b) It has often occurred in the history of science that an important discov
ery was come upon by chance. A scientist looking into one matter unexpect
edly came upon another which was far more important than the one he was
looking into…Penicillin is a result of such a discovery. In 1928 Fleming,
the British scientist, tried to grow a pure culture of bacterium…—Staphylo
coccus in Latin…(taken from Giora, 1990)
(14c) Soldier: Who goes there? King Arthur: It is I, Arthur, son of Uther Pen
dragon, from the castle of Camelot. King of the Britons, defeater of the Sax
ons, Sovereign of all England! Soldier: Pull the other one! King Arthur: I am,
and this is my trusty servant Patsy. We have ridden the length and breadth of
the land in search of knights who will join me in my court at Camelot. I must
speak with your lord and master. Soldier: What? Ridden on a horse? King
Arthur: Yes!
The relevance requirement and the graded informativeness requirement then control the
way a given discourse is segmented, reflecting the inflow of new DTs.
Coherence and Degree of Salience

Apart from obeying the relevance and graded informativeness requirements, an addition
al factor seems to affect the degree of coherence, namely, the degree of meaning
salience. According to the graded salience hypothesis (Giora, 1997; 2003), a meaning,
whether literal or figurative, is salient if it is coded in the mental lexicon; it is nonsalient
if it is not coded but relies on derivational procedures for its decoding. A salient meaning
is accessed automatically, regardless of figurativity; a nonsalient interpretation may in
volve accessing salient meanings to be followed by revision processes. According to Giora
(1997; 2003), then, degree of salience rather than degree of figurativity may affect de
gree of coherence and hence the amount of processing effort required (see Giora, this vol
ume).
By contrast, according to Grice (1975), even an overt violation, such as the use of figura
tive language, may affect degree of coherence compared to using literal language, re
gardless of degree of salience. In fact, any (overt) violation, such as any use of nonliteral
language, requires additional processes involved in deriving the conversational implica
ture. The standard pragmatic model, then, predicts that example (15a) (originally in He
brew), in which big eyes is used metaphorically, should be rated as less coherent than ex
ample (15b) (originally in Hebrew), in which the same expression is used literally. Recall
that, according to the graded salience hypothesis (Giora, 1997; 2003), both texts (p. 149)
should be rated as similarly coherent because both the figurative and literal meanings of
Page 13 of 23

big eyes (in Hebrew) are similarly salient. However, when users diverge in degree of
salience, their coherence rating will diverge too, regardless of figurativity:
(15a) Alon and Rany went to a restaurant. They ordered several appetizers,
several main dishes, and a number of deserts. Although they were still
through their first course, Alon already announced that he was completely
full. “It’s amazing,” said Rany “how I keep forgetting that you have such big
eyes.”
(15b) In his paper “The evolution of Mickey Mouse,” Stephen Jay Gould ar
gued that there are many physiological features that make us perceive cubs
as cute, such as their head being big compared to the size of their body and
the fact that they have relatively big eyes (taken from Giora, Fein, Kronrod,
Elnatan, Shuval, & Zur, 2004).
Results indeed showed that, as predicted by the graded salience hypothesis (Giora, 1997;
2003), these two texts were found to be equally coherent. In contrast, a similar pair of
texts in which the metaphoric use was novel (16a) but its literal use (16b) more familiar
(being salience-based) differed in coherence rating. The text featuring a novel use (16a)
was rated as less coherent (Giora et al., 2004):
(16a) Sharon went to sleep very late. In the morning she was supposed to
have a very important meeting. At a certain point she almost thought about
canceling it because she hates waking up in the morning, looking in the mir
ror, and seeing a geometrical abstract painting.
(16b) Sharon finished renovating her house. She put a lot of thought into de
signing the different rooms. She says she’s very pleased, but the only thing
that is still missing for the living room to look perfect is a geometrical ab
stract painting.
Indeed, in Giora et al. (2004) such novelty, whether literal (curl up and dye) or nonliteral
(weapons of mass distraction) took longer to process than salient alternatives, whether
literal (weapons of mass destruction) or nonliteral (curl up and die). It is degree of novel
ty, then, rather than nonliteralness that may hamper coherence.
Creativity: When Coherence Is Deliberately

Breached
As demonstrated earlier, coherence may often be overtly breached, thus increasing pro
cessing effort. Why then would speakers intentionally flout the requirements for coher
ence knowing the costs of such violations? Why do addressees invest in processing less
coherent discourses? There seems to be a number of factors that would offset processing
Page 14 of 23

effort (Sperber & Wilson, 1986/1995) and make processing costs worthwhile. Among
them are aesthetics, humor, affect, protest, and the establishment of a positive self-image.
Aesthetic Effects: Novelty Matters
As noted earlier, a number of studies showed that digressions, even when marked by a di
gression marker, are still experienced as less coherent and more effortful than discourses
not involving digressions. Studies show, however, that the cost of reduced coherence may
be offset by aesthetic effects, which make these texts more pleasurable than their coher
ent alternatives (Giora, 1990; 1993; Giora, Meiran, & Oref, 1996). For instance, texts in
cluding novel analogies that invite a between-domain comparison are more difficult to un
derstand than their equivalents that are free of these digressions. Notwithstanding,
processors find the texts including these analogies more aesthetic and pleasing than their
more coherent versions (Giora, 1993; Giora et al., 1996). For a different view arguing that
analogies are contributing to text comprehensibility, see, for example, Vosniadou &
Ortony, 1983; Vosniadou & Schommer, 1988).
Additionally, texts involving optimal innovations (curl up and dye)—innovations evoking

familiar stimuli (curl up and die)—whether literal, metaphoric, or sarcastic, are more ef
fortful processing-wise compared to familiar alternatives (Giora et al., 2004; Giora, Fein,
Kotler, & Shuval, forthcoming; Giora, Gazal, Goldstein, Fein, & Stringaris, 2012). Still,
they are more pleasing than the latter whether literal or nonliteral (Giora et al., 2004;
Giora et al., forthcoming; but see Bohrn, Altmann, Lubrich, Menninghaus, & Jacobs, 2012
for different findings).
Negative Affect: Indirectness Matters
Emotions and attitudes, especially negative ones, are often conveyed indirectly so as to
tone down the criticism. One such muting device is sarcastic irony (e.g., Dews & Winner,
1995; Giora, 1995; Winner, 1988, among others; for a different view, see Bowes & Katz,
2011, who show that sarcasm may also enhance the criticism and may be more offensive
and victimizing than nonsarcastic language). Sarcasm breaches discourse coherence
(p. 150) overtly and is often used to indirectly denigrate or criticize an opinion, a norm, a
thought, a policy, an individual, or a group (see Campbell & Katz, 2012). When exercised
among friends, it aims to humor others and is experienced as pleasing; among adver
saries it is not (Kotthoff, 2003). Conveying negative feelings via sarcasm, however, is per
vasive among friends compared to strangers (see Eisterhold, Attardo, & Boxer, 2006). Yet,
since it bites, it is used more often against the powerless; for example, professors teasing
students, women and men making fun of women (as shown in Eisterhold et al., 2006; on
humor, gender, and power relation, see also Kotthoff, 2006).
Conversely, given that sarcasm conveys negative affect implicitly, it often serves the pow
erless when they criticize the powerful (Hutcheon, 1994), either because it is a form of
pretense and can be denied as intended (Clark & Gerrig, 1984) or because it is a “polite
Page 15 of 23

ness strategy” (Brown, 1995; Brown & Levinson, 1978; Winner, 1988) conveying a miti
gated message compared to a direct alternative (Giora, 1995; 2003).
Perspective Taking: Whose Point of View Matters?
Optimally, innovative sarcasm is perceived as humorous and pleasing when the ironist’s
point of view is taken but not when the victim’s point of view is adopted (see van Mulken,
Burgers, & van der Plas, 2010). This has also been replicated when (self-reported) ag
gressors evaluated sarcastic comments that they perceived as more humorous and less
aggressive than did the victims of these comments (Bowes & Katz 2011; for similar find
ings regarding face-threatening questions, see Holtgraves, 2005).
In fact, taking an ingroup point of view determines one’s ability to enjoy humor. Following
Ariel and Giora (1998), Drucker, Fein, Bergerbest, and Giora (in press) show that feminist
women (scoring low on sexism) enjoy sarcasm most when it is uttered by women and
ridicules men. In contrast, they disliked it most when aimed at women by women. Femi
nist men (scoring as low on sexism) also do not enjoy sarcasm directed against women.
However, they enjoy sarcasm most when it is uttered by men against men. Although femi
nist women fully adopt their ingroup point of view, men, being the powerful group, enjoy
competing with ingroup members. At the same time, however, they also adopt the point of
view of the disempowered outgroup, exhibiting solidarity.
Taking an ingroup point of view also affected divergence from discourse norms. Following
Ariel (1986), Ariel and Giora (1998) show that what explains language differences is not a
matter of, for example, gender or ethnicity but a matter of the point of view adopted. Al
though the traditional women in our study adopt the point of view of the dominant “oth
er,” projecting in their language use a masculine point of view, feminist women adopt a
“self” (feminine) point of view. Men, however, being the powerful group, always adopt a
“self” (masculine) point of view while using language.
For underprivileged groups, challenging discourse norms may therefore be also instru
mental in developing a distinct positive self-image. For instance, for working-class individ
uals in the United Kingdom, Ireland, and Scotland belonging in closed networks, develop
ing a distinct vernacular served as a (covertly) prestigious variety, regardless of gender
(see Giles & Rakic, this volume). Among other things, this variety was used to protest
their underprivileged social status and develop their own positive self-image while pro
jecting solidarity (Coates, 1986).
Challenging discourse conventions also proved instrumental in the formation of a distinct

positive self-image among underprivileged groups in Israel. For instance, those speakers
of oriental origin who have developed self-awareness and strongly identify with their
group’s aims and perceive intergroup relations as challengeable adopted a speech diver
gence strategy. They took collective action to improve their social status by employing the
marked oriental version of Hebrew. In contrast, “mobile” Orientals, Orientals with a nega
tive self-image, who did not believe that social intergroup relations is changeable, at
Page 16 of 23

tempted to individually join the advantageous group by adopting their unmarked Euro
pean version of Hebrew (Peleg, 1991).
Challenging discourse conventions also proves effective in uncovering a covert but biased
point of view. Consider the following example (17) taken from Vonnegut (1999). Here,
Grice’s informativeness maxim is overtly breached by adding information about ethnic
origins not only when disempowered group members (black) are introduced to the dis
course, but also when members of the dominant (white) group are. Although members of
disempowered groups are usually introduced via mention of their ethnic origin (breach
ing the informativeness requirement), introducing members of the powerful groups in
vites no such introduction (p. 151) (given their unmarked social position in their eyes and
in the eyes those who adopt their point of view). Challenging this biased “norm” results in
an equal marking of members of both groups while uncovering the negative value of a
“normative” covert bias:
(17) This is a tale of a meeting of two lonesome, skinny, fairly old white men
on a planet which was dying fast. One of them was a science-fiction writer
named Kilgore Trout. He was a nobody at the time, and he supposed his life
was over. He was mistaken. As a consequence of the meeting, he became one
of the most beloved and respected human beings in history. The man he met
was an automobile dealer, a Pontiac dealer named Dwayne Hoover. Dwayne
Hoover was on the brink of going insane…. The person closest to him,
Francine Pefko, his white secretary and mistress said that Dwayne seemed to
be getting happier and happier all the time during the month before…. A
black bus boy and a black waiter discussed this singing. “Listen at him sing,”
said the bus boy.
These are just a few examples of how, when discourse coherence is violated, affective and
social goals are aimed at. Although language is not a weapon, it is an efficient tool that
might affect our mind and change both our views and our practices.
Wrapping Up
This chapter outlines the conditions for coherence of non-narrative discourse. It reviews a
number of factors that seem to influence the degree of coherence of such discourses.
Among them are degree of relevance, informativeness, and meaning salience. It further
reviews the effects of overt violations of these conditions, driven by the wish to induce
pleasure, convey emotions, or challenge norms of powerful groups while attempting to
bring about social change.
The requirements for non-narrative discourse coherence are not linguistic but cognitive.
They should therefore apply to narrative discourses as well (see Giora & Shen, 1994).
However, such discourses have additional cognitive constraints stemming from schematic
organization (see e.g., Labov, 1972; Reinhart, 1984; Shen, 1988). A subgenre of narrative
Page 17 of 23

discourse is news items. However, news structuring is different because it abides by the
rule of presenting the most informative first. Following the headlines and the lead, how
ever, the story evolves coherently (see e.g., van Dijk, 1985). Literary texts, which often
defy coherence entirely, may still abide by the requirements and be segmented in accor
dance with the predictions of the graded informativeness requirement (see e.g., Giora,
1986). And because these conditions are cognitive, they should apply to nonverbal dis
courses as well (on the cognitive-based categorical organization of the cinematic text, see
Giora & Ne’eman, 1996). These topics are not addressed in this chapter and require fur
ther research.
Acknowledgments
The research report here was supported by The Israel Science Foundation (grant No.
436/12) and the Vice President for Research and Development at Tel Aviv University En
couragement Fund.
References
[Anonymous] Sergeant. (2000). Shoving the kid’s head into the toilet. http://
www.shovrimshtika.org/testimonies_e_print.asp?id=185
Ariel, M. (1985). The discourse functions of given information. Theoretical Linguistics, 12/
2–3, 99–113.
Ariel, M. (1986). Women and men in language. Haifa: Haifa University Press. (In Hebrew).
Ariel, M. (1988). Retrieving propositions from context: Why and how. Journal of Pragmat
ics, 12/3–4, 567–600.
Ariel, M. (1990). Accessing noun phrase antecedents. London: Routledge.
Ariel, M. (2006). Accessibility theory. In K. Brown (Ed.), Encyclopedia of language and lin
guistics (2nd ed., vol. I, pp. 15–18). Oxford: Elsevier.
Ariel, M., & Giora, R. (1998). A self versus other point of view in language: Redefining
femininity and masculinity. International Journal of the Sociology of Language, 129, 59–
86.
Attneave, F. (1959). Applications of information theory to psychology: A summary of basic

concepts, methods, and results. New York: Holt.
Barsalou, L. W. (1983). Ad hoc categories. Memory and Cognition, 11, 211–227.
Bellert, I. (1970). On a condition of the coherence of texts. Semiotica, 2, 335–363.
Page 18 of 23

Bohrn, I. C., Altmann, U., Lubrich, O., Menninghaus, W., & Jacobs, A. M. (2012). Old
proverbs in new skins—an fMRI study on defamiliarization. Frontiers in Psychology,
3(204), 1–18.
Bowes, A., & Katz, A. N. (2011). When sarcasm stings. Discourse Processes, 48(4), 215–
236.
Brown, P. (1995). Politeness strategies and the attribution of intention: The case of Tzeltal
irony. In E. N. Goody (Ed.), Social intelligence and interaction (pp. 153–174). Cambridge:
Brown, P., & Levinson, S. C. (1978). Universals in language usage: Politeness phe
(p. 152)
nomena. In E. Goody (Ed.), Questions and politeness (pp. 56–311). Cambridge: Cambridge
University Press.
Campbell, J. D., & Katz, A. N. (2012). Are there necessary conditions for inducing a sense
of sarcastic irony? Discourse Processes, 49, 459–480.
Carrell, L. P. (1982). Cohesion is not coherence. TESOL Quarterly, 16, 479–488.
Clark, H. H., & Gerrig, R. J. (1984). On the pretense theory of irony. Journal of Experimen
tal Psychology: General, 113/1, 121–126.
Coates, J. (1986). Women, men and language. London: Longman.
Daneš, F. (1974). Functional sentence perspective and the organization of the text. In F.
Daneš (Ed.), Papers on functional sentence perspective (pp. 106–128). Prague: Academia/
The Hague: Mouton.
Daneš, F. (1987). On Prague school functionalism in linguistics. In R. Dirven & V. Fried

(Eds.), Functionalism in linguistics (pp. 3–38). Amsterdam: John Benjamins.
Dews, S., & Winner, E. (1995). Muting the meaning: A social function of irony. Metaphor
and Symbolic Activity, 10, 3–19.
Drucker, A., Fein, O., Bergerbest, D. & Giora, R. (in press). On sarcasm, social awareness,
and gender. Humor: International Journal of Humor Research.
Eisterhold, J., Attardo, S., & Boxer, D. (2006). Reactions to irony in discourse: Evidence
for the least disruption principle. Journal of Pragmatics, 38(8), 1239–1256.
Eldar, A. (2009, July 21). Slouching toward Sheikh Jarrah. http://www.haaretz.com/

print-edition/news/slouching-toward-sheikh-jarrah-1.280365
Enkvist, N. E. (1978). Coherence, pseudo-coherence, and non-coherence. In J. O. Östman

(Ed.), Cohesion and semantics (pp. 109–128). Åbo: Akademi.
Fyler, B. (2011, April 1). Protestors call Barak “child murderer.” http://
www.ynetnews.com/articles/0,7340,L-4009049,00.html
Page 19 of 23

Giora, R. (1983a). Segment cohesion. Text, 2, 155–182.
Giora, R. (1983b). Functional paragraph perspective. In: J. S. Petiifi & E. Siizer (Eds.), Mi
cro and macro connexity of texts (pp. 153–182). Hamburg: Buske.
Giora, R. (1985a). Towards a theory of coherence. Poetics Today, 6, 699–716.
Giora, R. (1985b). A text based analysis of non narrative texts. Theoretical Linguistics, 12,
115–135.
Giora, R. (1986). Principles of text segmentation: The case of the formally unsegmented
text. Journal of Literary Studies, 2/2, 15–28. Reprinted in 1990 in Hebrew Linguistics, 28–
30, 23–37. A special issue on Stylistics and the Literary Text.
Giora, R. (1988). On the informativeness requirement. Journal of Pragmatics, 12, 547–

565.
Giora, R. (1990). On the so called evaluative material in informative texts. Text, 10/4, 299–
320.
Giora, R. (1993). On the function of analogies in informative texts. Discourse Processes,

16, 591–611.
Giora, R. (1995). On irony and negation. Discourse Processes, 19, 239–264.
Giora, R. (1997). Understanding figurative and literal language: The graded salience hy
pothesis. Cognitive Linguistics, 8, 183–206.
Giora, R. (2003). On our mind: Salience, context, and figurative language. New York: Ox
ford University Press.
Giora, R., Fein, O., Kotler, N., & Shuval, N. (forthcoming). Know Hope: Metaphor, optimal
innovation, and pleasure. In G. Brône, K. Feyaerts, & T. Veale (Eds.), Cognitive linguistics
meet humor research. Current trends and new developments. Berlin/New York: Mouton
de Gruyter.
Giora, R., Fein, O., Kronrod, A., Elnatan, I., Shuval, N., & Zur, A. (2004). Weapons of mass
distraction: Optimal innovation and pleasure ratings. Metaphor and Symbol, 19, 115–141.
Giora, R., Gazal, O., Goldstein, I., Fein, O., & Stringaris, A. (2012). Salience and context:
Interpretation of metaphorical and literal language by young adults diagnosed with
Asperger’s syndrome. Metaphor and Symbol, 27/1, 22–54.
Giora, R., & C.- L. Lee (1996). Written discourse segmentation: The function of un
stressed pronouns in Mandarin Chinese. In J. Gundel & T. Fretheim (Eds.), Reference and
reference accessibility (pp. 113–140). Amsterdam: Benjamins Publishing Company.
Page 20 of 23

Giora, R., Meiran, N., & Oref, P. (1996). Identification of written discourse topics by struc
ture coherence and analogy strategies: General aspects and individual differences. Jour
nal of Pragmatics, 26, 455–474.
Giora, R., & Ne’eman, J. (1996). Categorical categorization in the narrative discourse: A
semantic analysis of Il conformista. Journal of Pragmatics, 26/6, 715–735.
Giora, R., & Shen, Y. (1994). Degrees of narrativity and strategies of semantic reduction.
Poetics, 22, 447–458.
Grice, H. P. (1975). Logic and conversation. In P. Cole & J. Morgan (Eds.), Syntax and se
mantics 3: Speech acts (pp. 41–58). New York: Academic Press.
Gutwinski, W. (1976). Cohesion in literary texts. The Hague: Mouton.
Halliday, M. A. K., & Hasan, R. (1976). Cohesion in English. London: Longman.
Hoey, M. (1991). Patterns of lexis in text. Oxford: Oxford University Press.
Holtgraves, T. (2005). Diverging interpretations associated with the perspectives of the

speaker and recipient in conversations. Journal of Memory and Language, 53, 551–566.
Hutcheon, L. (1994). Irony’s edge: The theory and politics of irony. London/New York:
Routledge.
Kotthoff, H. (2003). Responding to irony in different contexts: cognition and conversation.

Journal of Pragmatics, 35, 1387–1411.
Kotthoff, H. (2006). Gender and humor: The state of the art. Journal of Pragmatics, 38, 4–
25.
Labov, W. (1972). Language in the inner city (chapter 9). Philadelphia: University of Penn
sylvania Press.
Levy, G. (2008, December 29). Operation cast lead: It will end with Kleenex. http://
www.mouse.co.il/CM.television_articles_item,790,209,31362,.aspx (Hebrew, trans
lated by Elad Livnat).
Levy, G. (2010, January 24). Sexism, Mrs. Netanyahu and Mr. Livni. http://
www.haaretz.com/print-edition/opinion/sexism-mrs-netanyahu-and-mr-
livni-1.265764
Levy, G. (2011, January 9). The IDF uses propaganda like an authoritarian regime. http://
www.haaretz.com/print-edition/opinion/the-idf-uses-propaganda-like-an-authori
tarian-regime-1.335985
Marcus, Y. (2010, January 17). Five comments on the situation. http://www.haaretz.com/

print-edition/opinion/five-comments-on-the-situation-1.261510
Page 21 of 23

Peleg, O. (1991). The relationship between linguistic divergence and social identity
among Israelis of oriental origin. Unpublished MA thesis, Tel Aviv University.
Reinhart, T. (1980). Conditions for text coherence. Poetics Today, 1, 161–180.
Reinhart, T. (1981). Pragmatics and linguistics: An analysis of sentence topic.

(p. 153)
Philosophica, 27, 53–93.
Reinhart, T. (1984). Principles of gestalt perception in the temporal organization of narra

tive texts. Linguistics, 22, 779–809.
Rosch, E. H. (1973). On the internal structure of perceptual and semantic categories. In T.

E. Moore (Ed.), Cognitive development and the acquisition of language (pp. 11–144). New
York: Academic Press.
Rosch, E. H. (1975). Cognitive representation of semantic categories. Journal of Experi

mental Psychology: General, 104, 192–223.
Rosch, E. H. (1978). Principles of categorization. In E. H. Rosch & B. Lloyd (Eds.), Cogni

tion and categorization (pp. 27–48). Hillsdale, NJ: Erlbaum.
Rosch, E. H., & Mervis, C. B. (1975). Family resemblances: Studies in the internal struc
ture of categories. Cognitive Psychology, 7, 573–603.
Shannon, C. E. (1951). Prediction and entropy of printed English. The Bell System Techni
cal Journal, 30, 50–64.
Shen, Y. (1988). Schema theory and the processing of narrative texts: The X-Bar story-
grammar and the notion of discourse topic. Journal of Pragmatics, 12, 639–676.
Sperber, D., & Wilson, D. (1986/1995). Relevance: Communication and cognition. Oxford:
Blackwell.
Stein, G. (1922). Miss Furr and Miss Skeene (short story). Geography and plays (pp. 17–
22). Boston: Four Seas.
Tverski, A. (1977). Features of similarity. Psychological Review, 84, 327–352.
van Dijk, T. A. (1977). Sentence topic and discourse topic. Papers in Slavic Philology, 1,
49–61.
van Dijk, T. A. (1985). Structures of news in the press. In T. A. van Dijk (Ed.), Discourse
and communication: New approaches to the analysis of mass media discourse and com
munication (pp. 69–93). New York: Walter de Gruyter.
van Mulken, M., Burgers, C., & van der Plas, B. (2010). Wolves, confederates, and the
happy few: The influence of comprehension, agreement, and group membership on the
attitude toward irony. Discourse Processes, 48(1), 50–68.
Page 22 of 23

Vonnegut, K. (1999). Breakfast of champions. http://www.booksbooksbooks.ru/down

load/Vonnegut_Breakfast_of_Champions_pdf.pdf
Vosniadou, S., & Ortony, A. (1983). The influence of analogy in children’s acquisition of
new information from text: An exploratory study. In P. A. Niles & L. A. Harris (Eds.),
Searches for meaning in reading: Language processing and instruction. Clemson, SC: Na
tional Reading Conference.
Vosniadou, S., & Schommer, M. (1988). Explanatory analogies can help children acquire
information from expository text. Journal of Educational Psychology, 80(4), 524.
Vuchinich, S. (1977). Elements of cohesion between turns in ordinary conversation. Semi

otica, 20, 229–257.
Winner, E. (1988). The point of words: Children’s understanding of metaphor and irony.
Cambridge, MA: Harvard University Press.
Wittgenstein, L. (1953). Philosophical Investigations. Oxford Macmillan.
Women Breaking the Silence. (2009). Women Soldiers’ Testimonies, Testimony 39. http://
www.breakingthesilence.org.il/wp-content/uploads/2011/02/
Women_Soldiers_Testimonies_2009_Eng.pdf (p. 154)
Notes:
(1) . Throughout the chapter, an item in bold means emphasis is added for convenience.
Italics indicate that the italicized item is a quote and/or that it will be later referred to.
(2) . http://www.haaretz.com/hasen/spages/51551.html?item No=null (Retrieved Decem

ber 4, 2008).
(3) . http://writeonlymemory.net/python.pdf (Retrieved on December 27, 2012).
Rachel Giora
Rachel Giora, Department of Linguistics, Tel Aviv University
Page 23 of 23

Shaping Intergroup Relations Through Language

Anne Maass , Luciano Arcuri , and Caterina Suitner

This chapter addresses how intergroup relations are shaped, maintained, or modified
through language. Some of the strategies discussed are blatantly derogatory, including
homophobic epithets and metaphors, whereas others represent much more subtle linguis
tic tools, including word order, grammatical gender, language abstraction, and even
seemingly irrelevant “junk words,” such as pronouns. These language tools fulfill a wide
range of functions in intergroup relations because they drive attention to specific social
groups, guide social categorization, maintain or change stereotypes, express and protect
social identities, transmit broad cultural worldviews, and justify existing societal systems.
In its most extreme form, language serves to dehumanize minority groups, thereby creat
ing and/or justifying openly hostile intergroup relations.
Keywords: homophobic epithets, metaphors, word order, grammatical gender, language abstraction, pronouns, so
cial categorization, stereotypes, social identity, cultural worldview, system justification, dehumanization
Group stereotypes often reflect real differences in habits, beliefs, resources, privileges,
and the like, but they are also constructed, sustained, and modified by symbols and by
language well beyond actual differences. Similarly, children acquire knowledge about so
cial categories not only by observing actual differences, but also by listening to their
caregivers’ or peers’ talk to (and about) members of ingroups and outgroups. In this
chapter, we analyze the role that language plays at different stages of intergroup rela
tions. We argue that language guides our attention and helps us to chunk continuous di
mensions such as age or skin color into meaningful social categories. Language suggests
certain inferences while inhibiting others, often feeding into a self-perpetuating cycle of
stereotype maintenance. It also is an important vehicle of infrahumanization and dehu
manization, at times paving the way for treating outgroup members as less than human.
Finally, it may be used to justify unfair social systems. Together, language plays an impor
tant role in shaping intergroup relations.
Here, we analyze the different linguistic tools that are (deliberately or inadvertently) em
ployed to achieve the above aims. We start with rather blatant linguistic strategies, in
cluding racial or homophobic slurs and metaphorical figures of speech that speakers are
Page 1 of 33

likely to use in a conscious, strategic way. We then report on more subtle language tools
including word order, gendered language, and language abstraction that have powerful
effects on the listener without speakers necessarily being aware of their own language bi
ases nor of the effects they may produce. Finally, we turn to the subtle effects of appar
ently irrelevant “junk words,” such as pronouns, that seem to be nothing more than in
significant supporting actors in the discourse theater but that, as we will see, may affect
the way we envisage (p. 158) intergroup relations in a subtle but powerful way. Thus, the
chapter is organized from the most obvious and blatant to the most subtle and hidden
characteristics of language that shape intergroup relations.
Derogatory Language
The most direct and explicit form of linguistic expression of prejudice is represented by
so-called slurs or ethnophaulisms; that is, insulting, offensive, or disparaging remarks re
ferred to groups or to their members, such as fag or nigger. Although derogatory group
labels can refer to any kind of social category, such as sexual, ethnic, or national groups,
they are not equally distributed across groups but vary systematically with relative size
and status, with low-status minority groups being the prime target. Offensive terms may
allude to different aspects of the categorical membership, such as physical features (e.g.,
redneck) or the typical food of the national group, as in the case of spaghetti used to rep
resent Italian people in the United States, roastbeef to represent English people in
France, or frog to represent French people in the British or American context. Such offen
sive terms are usually created by numerical majority groups and/or by dominant groups
with a clearly hostile aim. Because such terms are largely derogatory, they are often used
to make fun of people of another group or to reinforce status and power differentials in
favor of one’s own group.
According to Allen (1983) ethnophaulisms may be divided into distinct classes, referring
to physical traits, personal traits, personal names, food habits, and group names. The
derogatory labels applied to a given group may all belong to the same class, or they may
be more complex, representing different classes. For example, four ethnophaulisms were
found to be applied to Belgian immigrants in the United States, but all of these refer to
some aspect of the group name (belgeek, belgie, blemish, flamingo), thus representing a
single class that makes the level of complexity of cognitive representation of Belgians ap
pear relatively low. In contrast, the four ethnophaulisms for Spanish immigrants in the
United States came from different classes, including group name, food, and personal
names (spanisher, spinach, Diego, José). The distribution of ethnophaulisms across differ
ent classes is interpreted as a more complex cognitive representation of the Spanish im
migrant group.
Mullen’s (2004; Mullen, Rozell, & Johnson, 2001) impressive line of research on derogato
ry labels describing immigrant groups in the United States shows that group size and fa
miliarity with the ethnic minority play a significant role in the use of ethnophaulisms. Eth
nic immigrant groups that are smaller, more salient, culturally more distant from, and
Page 2 of 33

less familiar to the host society tend to be referred to with less complex and more nega
tive slurs. A lower complexity of ethnophaulisms seems to be a sign of higher social exclu
sion, as shown by Leader, Mullen, and Rice (2009) who experimentally varied the com
plexity of the derogatory labels referring to ethnic outgroups and measured the subse
quent desire to socially exclude the members of the derogated groups. Those participants
who were exposed to simple (vs. complex) labels were more inclusive toward the out
group members (e.g., were less severe against immigration and more favorable toward
intergroup marriages). One possible interpretation of this result is that it is easier to
overcome the prejudice on one dimension than the prejudice on several dimensions; in
line with this interpretation, groups with more complex identities are better able to face
social threats (Mullen, Calogero, & Leader, 2007).
In addition to complexity, the second dimension identified by Mullen regards the evalua
tive judgment and hence the degree of offence of the label, which may vary across cultur
al and linguistic contexts. Mullen and collaborators (2001) found that the less the host so
ciety is familiar with a specific group of immigrants, the less complex and the more nega
tive the derogatory labels applied to this group tend to be. In other words, the semantic
structure and the valence of the derogatory labels can be considered reliable indices of
the prejudice against immigrant groups and, in complementary terms, of the level of fa
miliarity with those groups. Given the role of familiarity in offensive language, it may not
surprise that the number of derogatory group labels declines as the relative size and sta
tus of the minority groups increase (Mullen, 2001).
A second line of research has investigated the conditions under which derogatory group
labels are most likely to be used. In particular, Graumann (1995) has provided evidence
that derogatory labels are typically used in intragroup discourse, when majority members
communicate to each other. As suggested by Carnaghi and Maass (2008), minority group
members sometimes define their own ingroup using the derogatory label originally creat
ed by the majority group. This reappropriation of stigmatizing terms generally implies a
shift in valence; the derogatory meaning of the word changes in line with different com
municative (p. 159) goals, such as strengthening the ingroup identity by stressing its
uniqueness. This process is an interesting example of social creativity enacted by using
language as a tool of social change (for a model of reappropriation, see Galinsky, Hugen
berg, Groom, & Bodenhausen, 2003). For example, the term nigger frequently appears in
black speech as a way of referring, in an ironic or even affectionate way, to the ingroup
(Kennedy, 2002). Along the same line, homosexuals sometimes use terms such as fairy,
fag, queen, and faggot when talking about members of their own group, particularly
when talking among themselves rather than with majority members (Chauncey, 1994).
Thus, minority members often reframe group labels in a positive manner, reclaiming them
as terms of pride. At times, this redefinition may even affect the semantic meaning of the
label in majority speech. As a case in point, the German term schwul took on an increas
ingly positive connotation in the general population after gay people had started to use it
in their self-descriptions, as well as in labels of gay organizations.
Page 3 of 33

But what are the effects of derogatory labels on the listener? This question has been ad
dressed from different vantage points. Some studies have investigated the evaluative con
sequences of overhearing social slurs on perceivers’ impression formation or stereotype
application. For instance, Simon and Greenberg (1996) conducted a study in which partic
ipants (classified as pro-black, anti-black, or ambivalent) were exposed to either a deroga
tory (nigger) or to a neutral ethnic label (black) or, as a control condition, to no critical re
mark. Participants were asked to evaluate the target of the critical remark on several pos
itive and negative items. Contrary to hypotheses, there was no evidence that the deroga
tory ethnic label (nigger) would produce a more negative evaluation than the nonderoga
tory ethnic label (black), a finding that is clearly counterintuitive because it would sug
gest that ethnic or racial slurs have no negative effects on the listener. However, this re
search was not able to differentiate the stereotype activation process from the more eval
uative component of the attitude. It is possible that the derogatory labels produced nega
tive evaluative reactions but that these reactions did not reach the stage of explicit ex
pression because of deliberate inhibition of the activated knowledge (Kunda & Spencer,
2003).
To overcome these limits, subsequent studies have investigated how perceivers (including
the targets of the derogatory labels) react to derogatory group labels at an automatic lev
el, using a range of implicit measures (Carnaghi & Maass, 2007a). For instance, Carnaghi
and Maass (2007b) used a semantic priming paradigm, presenting as prime either catego
ry (i.e., gay, homosexual) or derogatory labels (i.e., fag, fairy) and as target words stereo
typical traits, counter-stereotypical traits, and traits that were irrelevant to homosexuali
ty. These target words varied also in valence, with half being positive, half negative. Re
sults showed that neutral category labels and derogatory labels activated stereotypical
content to the same degree, suggesting that the language form was irrelevant to stereo
type activation. However, there were distinct effects concerning valence. Heterosexual
participants reacted more quickly to positive targets when primed by category than by
derogatory labels, suggesting that derogatory labels inhibited the association of positive
terms related to gays. In contrast, homosexual participants were not affected by deroga
tory versus category labels but reacted more quickly overall to positive than to negative
targets. A plausible explanation for these results is that what is conceived as derogatory
in majority speech may have acquired a positive connotation in minority speech. In gener
al, the obtained results clearly demonstrate that majority and minority members have dis
tinct automatic reactions to derogatory versus category group labels. In heterosexuals,
derogatory terms inhibit positive associations with the group, whereas homosexuals react
in much the same way to derogatory labels as they react to “politically correct” labels. Al
though somewhat counterintuitive, these results suggest that the use of derogatory terms
referring to homosexuals has more severe automatic consequences for members of the
dominant (heterosexual) group than for those who are the target of such slurs.
Obviously, this does not mean that homosexual individuals do not feel offended or threat
ened by such labels, especially when used by heterosexuals (Vicars, 2006). Homophobic
slurs not only signal social exclusion (Leader et al., 2009; Mullen, 2004; Mullen & Rice,
2003), but they also frequently precede homophobic aggressions, suggesting that homo
Page 4 of 33

phobic slurs pave the way to more severe forms of prejudice. Hate speech, including ho
mophobic and ethnic slurs, also seems to be antecedent to minority suicides rates. A dra
matic example is Mullen and Smyth’s (2004) analysis of suicide rates among ethnic immi
grant groups in the United States in the 1950s that were reliably predicted by the nega
tivity of ethnophaulisms during the same time period.1 The effect of slurs on (p. 160) mi
nority members is particularly worrisome, given that ethnophaulisms and homophobic
references are extremely common, especially in adolescent speech (Plummer, 2001). For
instance, 99 percent of gay and lesbian college students in the United States report to
have heard antigay remarks on their university campuses (D’Augelli, 1992).
This also poses the question of why heterosexuals would use homophobic hate speech so
persistently. One possibility is that male heterosexuals feel threatened by gay men who
share their own sex category while at the same time violating the definition of masculini
ty (see gender inversion theory, Kite & Deaux, 1987; see also Falomir-Pichastor & Mugny,
2009). Homophobic slurs offer one possible means for heterosexual men to distance
themselves from this threatening and atypical subgroup of men and reestablish their own
masculinity. In line with this idea, male participants were found to construe their selves in
a particularly masculine way after having been exposed, either explicitly or implicitly, to
derogatory (vs. neutral category) labels (Carnaghi, Maass, & Fasoli, 2011).
Together, the research on racial or homophobic slurs suggests that derogatory language
fulfills different functions: in addition to posing an explicit threat to the minority, it in
hibits positive images about the target group in the majority listener, and it triggers iden
tity-protective strategies, thereby reinforcing the existing gap between majority and mi
nority group.
Metaphors
In the previous section on racial, national, or homophobic slurs we saw a number of ex
amples of metaphors in which food habits (frog, spaghetti, or roast beef), parts of the
body (redneck), or other attributes come to represent the entire group. However, not all
slurs are metaphors, and not all metaphors are used deliberately to insult outgroup mem
bers, which is the reason why we dedicate this section to metaphors regardless of
whether they are used as slurs.
When talking about intergroup relations, speakers often resort to metaphors, for instance
referring to immigration as tsunami (natural disaster metaphor), as invasion (war
metaphor), or as cultural contamination (disease metaphor). Also, social groups are com
monly referred to by metaphorical terms, denoting either stereotypic traits (holy roller,
crucifier) or behaviors (reproducing like rats, chickening out), or reflecting a generalized
hostile attitude (rat, parasite). Why would metaphors be so common in discourse delineat
ing intergroup relations?
Page 5 of 33

Metaphors can be defined as figures of speech in which concrete concepts (source do
main) come to represent abstract concepts (target domains; for an overview of metaphor
theory, see Landau, Robinson, & Meier, in press). According to metaphor theorists, these
figures of speech are in many ways superior to alternative linguistic forms, such as adjec
tives, denoting the same quality. First, metaphoric expressions tend to be very concrete
and tangible because they are grounded in our bodily experience or worldly knowledge
(Gibbs, 2008). As a consequence, they also tend to be very vivid and image-evoking
(Ortony, 1975). For instance, referring to immigration as a tsunami offers a very powerful
image of the phenomenon. Compared to adjectives, metaphors have the additional advan
tage in that they are very compact and efficient describing in one word, such as shark,
what would otherwise require a list of adjectives, such as merciless, tenacious, and vi
cious (Ortony, 1975). While offering concrete, vivid, and compact labels, metaphors make,
at the same time, very abstract assertions as their content generalizes to the entire cate
gory to which they are referred. For instance, the claim that “lawyers are sharks” does
not allow for many exceptions, turning an aggregate of individual professionals into a co
herent entity.
We have argued elsewhere that these features of metaphors make them particularly well-
suited for intergroup discourse, where they fulfill a number of important functions
(Maass, Suitner, & Arcuri, 2013). In intergroup relations, the prime function of metaphors
appears to be the dehumanization of outgroups, thus ultimately justifying an unjust and
degrading treatment of such groups. A large portion of the literature on the role of
metaphors in intergroup relations has focused on their capacity to dehumanize and dele
gitimize individuals or groups. Typically, metaphors are dehumanizing in one of three
ways: they denote the outgroup members as subhuman (e.g., animals such as parasites,
rats, or ants), as superhuman creatures (e.g., demons, monsters), or as automata (e.g.,
machines, robots; Bar-Tal, 1989; Haslam, 2006). In all three cases (subhuman, superhu
man, automata), the outgroup is described as lacking human essence. As argued by histo
rians and psychologists alike, such rhetorical strategies are used to justify extreme forms
of intergroup aggression, including genocide. According to Bar-Tal and Hammock (2012),
such dehumanizing language frees “human beings from the normative and moral re
straints and justify participation in violence, including the most evil actions.”
The growing literature on infrahumanization and dehumanization provides ample

(p. 161)
evidence that outgroups are not only denied typically human characteristics (Leyens, De
moulin, Vaes, Gaunt, & Paladino, 2007; Vaes, Leyens, Paladino, & Miranda, 2012), but are
also often equated with the animal world. Viki, Winchester, Titshall, Chasing, Pina, and
Russell (2006), using an Implicit Association Test (IAT), found that people associate na
tional outgroups more easily with animal- and nature-related words (and less with human-
related words) than the ingroup. Similarly, Boccato, Capozza, Falvo, and Durante (2008)
found that a regional outgroup was mainly associated with the animal kingdom, whereas
the regional ingroup was associated primarily with humanity. Some social groups are as
sociated with specific animals, such as Gypsies with wild animals (Perez, Moscovici, &
Chulvi, 2007) and Blacks with apes (Goff, Eberhart, Williams, & Jackson, 2008). Finally,
Saminaden, Loughan, and Haslam (2010) observed that indigenous (vs. industrialized)
Page 6 of 33

cultures are associated with both animals (vs. persons) and children (vs. adults), thus
putting them into an inferior status either phylogenetically (animals) or ontogenetically
(children). Importantly, intergroup metaphors may also be communicated through im
ages. For instance, an archive analysis of visual images published during the fascist peri
od in Italy revealed that Jews were often metaphorically portrayed as vultures, spiders,
parasites, microbes, and devils, whereas Blacks were more frequently associated with
apes (Volpato, Durante, Gabbiadini, Andrighetto, & Mari, 2010). Together, these studies
show a very general pattern, in which a wide range of social groups (national, racial, eth
nic, religious, cultural) are metaphorically linked to either specific animals or to the ani
mal world in general.
But how offensive are these metaphors? According to Haslam, Loughnan, and Sun (2011)
animal metaphors are particularly offensive when they either refer to disgusting or dis
liked animals (e.g., vultures, cockroaches, worms) or when they imply a high degree of
dehumanization. Interestingly, phylogenetic distance to human beings does not seem to
play a role in this picture. As the authors argue, highly offensive metaphors include both
very distant (worms, parasites) and very close species (pigs, rats, or apes).
Besides dehumanizing other groups, metaphors also serve a number of additional func
tions. In particular, they tend to guide attention to specific characteristics of the category
that come to the forefront, while other characteristics are foregone. When exposed to a
metaphor, listeners match similar features of source and target so as to maximize their
commonalities, thereby concentrating on the conceptual overlap between the two (Gen
tner & Wolff, 1997). For instance, when equating Germans with panzer, the militaristic as
pect of the German stereotype is evoked, whereas other stereotypical traits, such as pre
cision or level of organization, are overlooked. The fact that metaphors shift attention to
particular features logically implies that other features remain hidden. This hiding func
tion is easily illustrated by looking at metaphors used in intergroup conflicts such as wars
(Lakoff & Johnson, 1980). For instance, talking about such conflicts in terms of arms
races makes war appear little more than a fair competitive game, hiding its terrifying and
cruel aspects. The same is true when referring to military attacks against hospitals,
schools, and markets as combat operations or as missions that evoke relatively clean im
ages, far removed from the desperation, pain, blood, and death that people actually expe
rience in these instances. Similarly, describing the effects on the civil population as side
effects or casualties suggests that such military invasions are necessary in the service of
more important aims (such as building democracy). As these examples illustrate, in public
discourse, metaphors often serve to hide those aspects of aversive events that the gener
al public would find unacceptable if expressed otherwise.
Compared to semantically similar adjectives, metaphors also seem to provide the impres
sion that the target group is more homogeneous. There are two reasons why one may ad
vance this hypothesis. First, metaphors (shark, bradypus/sloth, tomboy) often provide
stronger labels than adjectives of similar meaning (merciless, lazy, boyish), and stronger
category labels have been shown to induce greater similarity perceptions. For instance,
Foroni and Rothbart (2011) found that pairs of women of high body weight were judged
Page 7 of 33

as more similar when described by a strong (obese) rather than neutral label (above aver
age). Second, if metaphors vividly draw attention to one aspect of the stereotype while ig
noring other features (as illustrated by the reference to Germans as panzer), then listen
ers should focus on one feature of the target group (Germans being militaristic) rather
than on multiple features, resulting an increased perceived homogeneity of the target
group. To test this hypothesis we conducted a study (Maass et al., in press) in which a
politician or a lawyer was described either by metaphors or by adjectives of similar mean
ing (e.g., lawyer: shark (p. 162) or unscrupulous) and asked participants to indicate how
similar other politicians or lawyers are to the target person. As predicted, similarity was
judged higher in the metaphor condition, suggesting that metaphors increase perceived
homogeneity.
Finally, there is first evidence that metaphors influence decision making. For instance,
Thibodeau and Borodistky (2011) have shown that framing criminality as a wild beast (an
imal metaphor) versus as a virus infecting the city (disease metaphor) led people to pro
pose different intervention strategies to counter increasing crime rates. Solutions offered
by participants were much more prevention-focused when a disease (rather than a wild
animal) metaphor was used. Similarly, Goff et al. (2008, experiment 5) report that partici
pants primed with ape words were more likely to condone violence against Blacks. To
gether, there is a small but growing body of literature showing that metaphorical lan
guage shapes, maintains, justifies, and/or changes intergroup relations.
Word Order: Who Is Mentioned First?

A basic principle of naïve causal reasoning is that the cause precedes the effect, rather
than vice versa. People are unlikely to infer causality unless this principle is respected.
This is probably the reason why word ordering is so important when conveying informa
tion about social interactions and why, in the majority of languages, subjects (agents) are
mentioned before objects (patients).2 Obviously, there are many violations of this rule, es
pecially in languages in which nouns are subject to declination, but, as a general rule,
there is a tendency to name the subject first and to attribute it the role of agent. Word or
der provides important cues for understanding and disambiguating utterances, for in
stance, allowing people to conclude that, in a phrase such as “Paolo greets Carla,” it is
Paolo (rather than Carla) who does the greeting. We argue here that word order reveals
not only causality and agency, but also other socially relevant information regarding gen
der stereotypicality, power, responsibility, and closeness, all of which play an important
role in intergroup relations.
Linguists have long argued that there are (some) semantic constraints on word order.
Cooper and Ross (1975) found that agentive, animate, and alive elements are mentioned
before passive, inanimate, and dead ones (e.g., agent and patient; speaker and listener;
subject and object; people and things; men and machines; life and death), suggesting that
agency is indeed an important determinant of word order. Recently, a more systematic
comparison of semantic and phonological ordering constraints by Mollin (2012) has re
Page 8 of 33

vealed that semantic factors are the prime predictors both of the predominant order in
which binomials are mentioned and of the frequency with which this conventional order is
violated (so-called reversibility). For instance, in binomial pairs, the first mentioned item
tends to be of greater power (adults vs. children), to constitute a cause rather than an ef
fect (cause and effect; trial and error), or to have chronological precedence (earlier or lat
er; Spring and Summer).
Given that men are stereotypically associated with greater agency and power than
women (e.g., Abele, 2003), one may suspect that these social groups are also represented
in a precise order, with males preceding females. In line with this idea, there is now con
siderable evidence that social targets are ordered systematically, rather than randomly, in
space and time. For instance, studies on graph layouts by Hegarty and collaborators
(Hegarty & Buechel’s, 2006; Hegarty, Lemieux, & McQueen, 2010; for similar findings in
drawing, see Maass, Suitner, Favaretto, & Cignacchi, 2009) have shown that both lay peo
ple and scientists tend to produce graphs so that male samples are represented first (e.g.,
on the left) compared to female samples. People also recall gender differences in a simi
lar fashion, with males being misremembered as having appeared first (e.g., on the left).
The initial interpretation offered by Hegarty and collaborators was that the normative
group that serves as standard (in this case males) is drawn before the “to be explained”
group (in this case females; Hegarty & Buechel, 2006). However, more recent work sug
gests that the men-first principle represents a specific case of the more general tendency
to represent powerful groups first (Hegarty et al., 2010). This explanation is in line with
Cooper and Ross’s (1975) and Mollin’s (2012) observations derived from language corpo
ra that more powerful elements are mentioned before less powerful ones (as in bow and
arrow; sun and moon; parent and child; cow and calf) and with McGuire and McGuire’s
(1992) argument that groups of higher sociocultural status are mentioned first.
In line with this idea, there are linguistic regularities that parallel these visual arrange
ments in graphing and that seem to reflect the same underlying principle. Verbal order
ing in binomial phrases tends to be such that males are mentioned before females (e.g.,
brothers and sisters, Romeo and Juliet, (p. 163) Tristan and Isolde), except in those
parental relations in which the woman plays a particular central role, as in parenting
(mom and dad). For example, if we Google “Romeo and Juliet,” we find about 25,400,000
results, but only 1,740,000 results for “Juliet and Romeo”; similarly “men and women”
leads to 218,000,000 results, “women and men” to about 24,800, 000 results. Such asym
metries are subject to changes over time, probably reflecting the changing gender roles
of men and women in society. A case in point is Mollin’s (2013) recent analysis of Ameri
can-English binomials over twenty decades (1810–2009) showing that the man-first prin
ciple, although still prevalent, has been declining over time (see Suitner & Maass, 2007,
for converging results on visual representation order).
Interestingly, binomial ordering is not only revealing about gender but, more generally,
about gender-related stereotypes (Hegarty, Watson, Fletcher, & McQueen, 2011). When
referring to heterosexual couples, male names tend to be mentioned before female
names, and this regularity emerges more forcefully for couples that conform to tradition
Page 9 of 33

al gender stereotypes.3 A similar pattern was found for imaginary same-sex couples for
which observers inferred that the first mentioned partner must have more masculine
traits (Hegarty et al., 2011). Thus, binomial ordering not only reflects speakers’ beliefs
about the masculinity or femininity of the persons who are being described, but it is also
used by listeners to form impressions and to make inferences.
In addition to stereotypicality, word order also reveals implicit causation and, hence, attri
bution of responsibility, although speakers may not be aware of this. It is well document
ed that events described by action verbs such as help (but not state verbs such as ad
mire) are generally attributed to the sentence subject who is seen as having caused the
event, a finding that holds both for affirmations and for questions (Semin & De Poot,
1997). Given the implicit causality attributed to the first mentioned actor, the authors pro
pose an interesting thought experiment in which a rape victim is interviewed about the
events preceding the rape by the officer asking whether the victim danced with the per
petrator (did you dance with him?) or whether the perpetrator danced with the victim
(did he dance with you?). Although the answer is likely to be affirmative in both cases (as
suming that they did dance), the two questions have very different implications, suggest
ing that either the victim or the perpetrator took the initiative.
Such word order effects, by definition, become critical when converting active (e.g., The
gang attacked Andrea) into passive phrases, given that the patient in the latter case
comes to precede the agent (e.g., Andrea was attacked by the gang). Interestingly, con
versions from active to passive voice tend to change the meaning in such a subtle way
that it easily goes unnoticed. Shifts in meaning as a function of voice have mainly been in
vestigated in the context of interpersonal violence, in which victims are assigned greater
co-responsibility when mentioned first (passive voice). A media analysis by Henley, Miller,
and Beazley (1995) found that rape cases were particularly likely to be reported in pas
sive voice. A subsequent study showed that readers are strongly affected by this language
use, showing more acceptance of violence against women and, for only male participants,
blaming the perpetrator less for the aggression when passive (vs. active) voice was used.
Bohner (2001) also showed that the use of passive voice increases as a function of the
person’s rape myth acceptance, suggesting that passive voice is indeed a way to put the
perpetrator in the background and to assign greater responsibility to the victim. Togeth
er, these and related studies suggest that word order is quite telling about who is to be
blamed. The fact that passive voice, although generally rare, is used systematically when
describing rape suggests that victims are partially blamed for their fate.
Together, these studies suggest that word order not only reflects stereotypical views of
women and men (and of other social groups), but also implies event causation. Across dif
ferent studies, the first-mentioned target was found to be perceived as more powerful,
more masculine, more agentic, and more likely to have caused the event. Although the re
lation between these implicit meanings (agency, power, masculinity, and causation) re
mains to be investigated, one may speculate that they represent a cluster of related char
Page 10 of 33

acteristics that are ultimately derived from cause–effect reasoning, assuming that events
are more likely to be caused by those who are powerful, active, and masculine.
However, word order not only reflects stereotyping (in terms of agency, power, status,
etc.) but also ingroup bias. Already, Cooper and Ross (1975) had observed that binomials
follow a “patriotic” ingroup-first rule (as in cowboys and Indians) and that, more general
ly, proximal, psychologically close, and liked items precede distal, psychological distant,
and disliked ones. Similarly, McGuire and McGuire (1992) have argued that affective
charge (p. 164) partially determines word order, with groups or individuals that are closer
to self and/or more relevant to current needs mentioned first (see also Mollin, 2012). In
line with this hypothesis, Hegarty et al. (2010, experiments 5 and 6) found that people
named closer members of couples before less familiar members and same-sex members
before other-sex members.
Together, the research reviewed here demonstrates that word order is a marker of speak
ers’ implicit stereotyping, responsibility attribution, and ingroup bias and that listeners
are able to interpret utterances along these lines. The ordering of concepts—whether dis
played in images or in utterances—plays a very subtle, difficult to detect, yet powerful
role in revealing the nature of intergroup relations and in guiding the interpretation of
human interactions. Word ordering may ultimately feed into a self-maintaining cycle, both
reflecting and creating beliefs about the relative standing of social groups.
Inclusive Versus Exclusive Language: The Case

of Gender Marking
A peculiar feature of language is its ability to draw boundaries between groups that may
implicitly include or exclude subgroups. For instance, the dated term “mother-tongue”
seems to suggest that fathers only play a subordinate role in the linguistic development of
their children. Analogously, “man–machine interaction” implies that women rarely inter
act with machines, “chairman” strongly suggests that the head of a department is male,
and “mankind” seems to exclude women from the human species altogether. The inclu
siveness or exclusiveness of language has great social relevance because it makes some
groups invisible while putting others in the limelight.
Inclusiveness of language has mainly been studied in connection with gender marking, an
ever present and unavoidable issue in grammatical gender languages, in which each noun
is gender-marked and in which many parts of speech (e.g., articles, adjectives, pronouns,
verb forms) vary as a function of the noun’s gender (Stahlberg, Braun, Irmen, & Sczesny,
2007). Whereas chances of grammatical gender bias in natural gender languages such as
English is limited to few elements (in particular, pronouns; see later section) and practi
cally absent in genderless languages such as Turkish or Finish, speakers of grammatical
gender languages face two related problems, namely, how to refer to social categories
(e.g., students) composed of both males and females and how to refer to individual
women (e.g., minister). As far as the former issue is concerned, in most grammatical gen
Page 11 of 33

der languages, the general rule is that the masculine (rather than feminine) plural is used
when referring to both male and female components of the category. Although grammati
cally correct, the literature has shown that women tend to become largely invisible when
referred to by such masculine generic forms (die Wissenschaftler, the scientists m.)
rather than by alternative forms such as dual nomination (Wissenschaftler und Wis
senschaftlerinnen), splitting form (Wissenschaftler/innen), or forms of neutralization (wis
senschaftlich Arbeitende, those active in science). A powerful illustration of this bias was
provided by Stahlberg, Sczesny, and Braun (2001) who asked German participants to
name their favorite writer, actor, or other such person. They found that the percentage of
women listed was much lower when the question was framed in the (seemingly generic)
masculine form, suggesting that female exemplars became inaccessible and hence people
simply “forgot” about women when providing their responses (for findings regarding
French, see Brauer & Landry, 2008, and Gygax & Gabriel, 2008; and for Norwegian, see
Gabriel, 2008). This (mis)interpretation of masculine generics as male-specific, rather
than as inclusive of women, is particularly strong among male participants. As is evident
from this example, the exclusive (vs. inclusive) language in which questions are formulat
ed greatly affects the cognitive accessibility of women, which in turn biases the responses
people provide.
The practical implications of generic masculine (vs. gender-fair) language forms have
been demonstrated in various social domains. Among others, masculine generics tend to
distort the results of surveys and opinion polls in favor of males (Stahlberg et al., 2001;
Stahlberg & Sczesny, 2001), they discourage women from applying for jobs advertised in
masculine form (Bem & Bem, 1973), and they lead to biased interpretations of legal infor
mation (Hamilton, Hunter, & Stuart-Smith, 1992). In contrast, nonsexist language forms
have been found to reduce male bias either partially or completely (Stahlberg et al., 2007,
for an overview). Thus, language contributes significantly to the visibility of women in so
ciety and to the way people respond to surveys, job ads, and the like.
Recently, some authors have carried this argument one step further, hypothesizing that
grammatical gender per se promotes sexist attitudes and even (p. 165) constitutes a risk
for actual socioeconomic gender equality. The underlying argument is that languages in
which gender is constantly salient tend to remind speakers and listeners of gender differ
ences, which, in practically all societies, place women in an inferior position. To test this
possibility, Wasserman and Wesley (2009, experiment 3) asked English-Spanish bilinguals
in the United States to respond to a sexism scale that was provided either in Spanish or
English. Sexism scores were found to be higher in the former than in the latter condition,
suggesting that grammatical gender languages promote sexist attitudes.4 Transposing
this argument to the institutional level, Prewitt-Freilino, Caswell, and Laakso (2012) have
shown that countries with grammatical gender languages tend to have greater economic,
political, educational, and social gender gaps than do those with natural gender or gen
derless languages, and this was true even after controlling for variables such as geo
graphic location, religious tradition, and system of government. Our archival research on
Italian word frequencies has further confirmed the association between gender stereo
types and grammar, showing that adjectives referring to agentic characteristics (e.g., de
Page 12 of 33

terminato, self-confident) are more likely to be used in their grammatically masculine

than feminine form, whereas communal adjectives (e.g., affettuosa, affectionate) appear
more frequently in their feminine form (Suitner & Maass, 2008). Together, these studies
suggest that grammatical gender is inherently damaging to gender equality because they
foster the idea that gender is a relevant dividing line.
In contrast to this claim, other authors (Braun, 2001; Stahlberg et al., 2007) have argued
that the mere presence versus absence of grammatical gender is not diagnostic of sexism
or egalitarianism. For instance, Braun’s (2001) analysis of the Turkish language, classi
fied as genderless, shows that it still contains a subtle male bias such that many neutral
terms are interpreted as referring primarily to males. Another interesting case is Norwe
gian, which underwent a major language reform in the 1970s in which the language was
“neutralized” by eliminating the feminine suffix and turning the formerly male form into a
generic meaning (similar to English). A study by Gabriel (2008), conducted 30 years after
the language reform, shows that people name fewer women in a survey framed in the
neutral way than when framed using the generic form together with the obsolete femi
nine form (dual nomination). This clearly suggests that the dual nomination is more gen
der-balanced than the neutral form. Thus, the absence of gender marking is not necessar
ily a guarantee for absence of gender bias.
The discussion of whether gender equality is better promoted by the absence of gender
marking (as in gender-free languages) or by the explicit reference to both genders (as in
the dual nomination in grammatical gender languages) remains an issue of inquiry.5 The
very capacity of grammatical gender languages to construct and explicate gender rela
tions implies that they are also able to change the conceptions of males and females. This
is illustrated by alternative language forms, such as dual nomination, that have, at times,
been found to not only reduce the male bias but even to induce a reversed, pro-female
bias (Heise, 2000; Stahlberg & Sczesny, 2001, experiment 3). According to this position,
whether gender equality is promoted or inhibited depends not so much on grammatical
gender per se, but rather on the rules governing gender-fair or unfair language use in
these languages. As a case in point, the languages spoken in the five countries with the
greatest gender equality (Iceland, Finland, Norway, Sweden, and Ireland; see 2012 Glob
al Gender Gap Index) include languages with grammatical gender (Icelandic), with natur
al gender (Swedish; English), a gender-free language (Finnish), and one where gender
has been “neutralized” by a language reform (Norwegian). This suggests that there are
indeed different routes to gender equality.
Although the question of (in)evitability of bias in grammatical gender languages is open

to debate, all of these studies argue that, in grammatical gender languages, gender-fair
language forms, in which males and females are equally salient, are preferable to generic
masculine forms. Yet, at times, linguistic visibility comes at a price. When referring to a
single individual in high-status professions, such as a female minister, surgeon, or profes
sor, speakers of grammatical gender languages at times prefer the masculine form where
as political correctness rules would suggest the use of feminized forms such as Profes
sorin in German or professoressa in Italian. The downside of the feminine version is that
Page 13 of 33

it is often associated with a status loss. For instance, Mucchi-Faina (2005) found that a
woman described as professoressa was perceived as less persuasive than one described
as a professore, and Merkel, Maass, and Frommelt (2012) observed a reliable loss in so
cial status when women professionals were referred to in the feminine form (e.g., avvo
catessa vs. avvocato). According to the latter study, such status loss is a function of asym
metrical (p. 166) language in which the feminine form is strongly marked (e.g., avvocates
sa vs. avvocato); this problem can be prevented by adopting more symmetrical forms
(e.g., avvocata vs. avvocato, ministra vs. ministro), as is typically done in Spanish. In line
with this interpretation, highly asymmetrical feminine suffices such as in Quakeress, Ac
tress, or Jewess have long been abolished from the English language (Miller & Swift,
2001).
Together, this line of research suggests that linguistic inclusion versus exclusion is a par
ticularly critical feature of grammatical gender languages. On one side, there is the risk
that masculine forms make women invisible; on the other side, there is a chance that fem
inized forms imply lower status. To overcome these difficulties, efforts have been made to
develop and promote language forms that are both inclusive and nondiscriminatory, and
these solutions necessarily vary from language to language (see gender-fair language
guidelines, Moser, Sato, Chiarini, Dmitrow-Devold, & Kuhn, 2011).
Language Abstraction
Another subtle way in which stereotypes may be perpetuated and prejudice may be trans
mitted is through language abstraction. Language abstraction refers to the linguistic cat
egory model (LCM: Semin & Fiedler, 1988; Fiedler & Kruger, this volume), a conceptual
anchor point from which several experimental paradigms, including the linguistic inter
group bias (LIB) and the linguistic expectancy bias (LEB) were derived (Maass, 1999). In
its original version, the LCM (Semin & Fiedler, 1988) distinguishes four levels of abstrac
tion of interpersonal language. The most concrete terms are descriptive action verbs
(DAV)—such A hugs B—that provide an objective description of a specific, observable
event. Interpretative action verbs (IAV) are slightly more abstract in that they describe a
larger class of behaviors (A consoles B), although they maintain a clear reference to a
specific behavior in a specific situation. The third level are state verbs (SV)—such as A
loves B—that describe stable psychological states that describe decontextualized situa
tions and behaviors but refer to a specific object (person B). The most abstract terms are
adjectives describing a general and enduring nature that is completely detached from the
specific situation, behavior, and contextual objects (e.g., A is lovely).
Interpretative implications vary systematically as one moves from the concrete to the ab
stract pole of the continuum. The more abstract the statement, the greater the amount of
information provided about the characteristics of the person whose actions are being de
scribed, the greater the implicit temporal and cross-situational stability, and the more dif
ficult it becomes to verify the veridicality of the statement and to imagine disconfirming
instances. Language abstraction has been shown to affect intergroup relations in two,
Page 14 of 33

largely independent, ways. On one side, it reflects and reinforces ingroup favoritism
(LIB); on the other, it contributes to the perpetuation of socially shared stereotypes (LEB).
Linguistic Intergroup Bias (LIB)
The LIB model predicts that language abstraction varies as a function of the group mem
bership of communicator and target and of the social desirability of the target’s actions.
In its original form, LIB predicts that ingroup members engaging in desirable behaviors
and outgroup members engaging in undesirable behaviors are described at a relatively
high level of linguistic abstraction. In contrast ingroup members engaging in the same
undesirable and outgroup members engaging in the same desirable behaviors are de
scribed at a relatively low level of linguistic abstraction. For instance, the same incidence
of maltreatment may be described as Carl hit the dog when Carl is an ingroup member,
but as Carl is aggressive when he is an outgroup member, suggesting that negative ac
tions are exceptions for ingroup members, but typical behaviors for outgroups (and vice
versa for positive actions). This bias has been shown in many different intergroup set
tings, including nations, schools, sport teams, political parties, gender and age groups,
ethnic groups, social interest groups, and the like. It even occurs for subordinate groups
that are part of larger superordinate categories such as Asian American versus European
American (Hsu, 2011).
The linguistic representation of the very same desirable or undesirable behaviors in a

more concrete or abstract manner conveys different implicit meanings. Abstract repre
sentations provide more information about the properties of the described person and
about his or her group and less information about the specific characteristics of the situa
tion. In contrast, concrete messages lead to a particularization of the event. The desirable
behavior of the ingroup member and the undesirable behavior of the outgroup member
come to be seen as typical, whereas the remaining combinations are interpreted as single
instances or as exceptions. In this way, language use has been shown to contribute in a
subtle way to the maintenance of positive ingroup and negative outgroup perceptions.
The LIB seems to be driven by ingroup protective motivation, given that it is pre
(p. 167)
dominantly found in intergroup settings in which the ingroup identity is threatened or in

which groups are in competition (Maass, Ceccarelli, & Rudin, 1996). The motivational ac
count is loosely based on social identity theory (Tajfel & Turner, 1986) and assumes that
the LIB represents one possible strategy of maintaining a positive image of one’s own
group, very much like other well-known strategies including biased reward or resource
allocation or ingroup enhancing attributions and evaluations (Wigboldus & Douglas,
2007).
The LIB is not only prevalent in interpersonal communication, but also in the mass media
(Maass, Corvino, & Arcuri, 1994) and in political communication (Menegatti & Rubini,
2007). Other applied areas in which the LIB has been shown to be influential are the
courts (Schmid & Fiedler, 1998) and hiring decisions (Rubini & Menegatti, 2008), attest
ing to the fact that language abstraction plays a critical, and often unrecognized, role in
Page 15 of 33

many settings in which the social identity of speaker and target are relevant. Interesting
ly, the LIB is also highly revealing about the speaker and about his or her relation to the
ingroup. For instance, Assilamehou and Teste (2013) have recently shown that the LIB is
the type of verbal behavior that people would expect of their fellow group members. In
fact, individuals who use the LIB are considered more desirable group members, especial
ly when their biased language use refers to the ingroup. It is therefore not surprising that
linguistic bias in favor of one’s ingroup increases as the entitativity of the ingroup in
creases (Rubini, Moscatelli, & Palmonari, 2007). Together, these studies suggest that the
magnitude of the LIB provides information not only about the speaker’s attitudes but also
about the relation between speaker and his or her group.
Linguistic Expectancy Bias (LEB)
In addition to serving ingroup protection or enhancement, language abstraction also ful

fills a second function in intergroup relations, namely, the maintenance of stereotypic be
liefs. According to the LEB (Maass, Milesi, Zabbini, & Stahlberg, 1995; Wigboldus, Semin,
& Spears, 2006), expected information is expressed in more abstract language than unex
pected information. Speakers prefer abstract language whenever they report on behav
iors that confirm prior expectancies exactly because such expectancy-congruent behav
iors are considered expressions of a stable, typical, and diagnostic behavior tendency of
the actor (Wenneker, Wigboldus, & Spears, 2005). For instance, people often endorse
stereotypes that associate blacks (rather than whites) with crime. When learning about a
crime story that confirms this expectation (involving a black rather than white suspect),
viewers are more likely to use abstract language to describe the suspect (Gorham, 2006).
The role of language abstraction in the transmission and maintenance of stereotypical im
ages is well illustrated by a recent media study in which participants read actual news ar
ticles reporting on crime committed by immigrants (Geschke, Sassenberg, Ruhrmann, &
Sommer, 2010). When exposed to news articles using abstract (vs. concrete) language,
participants reported higher estimates of immigrant’s criminal behaviors and also showed
greater prejudice against immigrants. Thus, the widely shared stereotype that immi
grants commit more crimes was reinforced by biased language use in the media. Similar
ly, Menegatti and Rubini’s (2012) archival analysis of evaluations of job applicants showed
a marked linguistic bias, especially in collective evaluations, further confirming the role
of language in reinforcing and maintaining collective worldviews.
LIB and LEB in Social Context
The main difference between the predictions advanced by LEB and LIB is that the ex
pectancy-based explanation (LEB) predicts abstract language use for expected, and con
crete language use for unexpected behaviors, regardless of valence of the behavior,
whereas the ingroup protection explanation (LIB) predicts abstract language use for posi
tive ingroup and negative outgroup behavior regardless of stereotypic expectancy. These
two mechanisms, one related to stereotypic expectancies (LEB), the other to ingroup- or
self-protective motivation (LIB), tend to be confounded in many natural real-world set
Page 16 of 33

tings, given that people generally have more negative expectations about outgroups, yet
they have been isolated in experimental situations (Maass, 1999, for an overview). Ac
cording to these studies, differential expectancies seem to be sufficient to produce differ
ential language use in intergroup settings and probably in a wide range of additional set
tings. In this case, expected events are described in more abstract language, unexpected
in more concrete language, which in turn protects existing beliefs against disconfirma
tion. At the same time, ingroup protective motives (LIB) (p. 168) come into play whenever
there is intergroup competition or when social identities are at stake. In these cases, lan
guage bias functions in much the same way as other forms of ingroup favoritism. Thus,
language abstraction is the joint function of differential expectancies and ingroup protec
tive motives. In both cases, speakers are often unaware of their biased language use and
may encounter difficulties in inhibiting these language biases (Douglas, Sutton & Wilkin,
2008).
Nouns as a Possible Fifth Level in the LCM
The LCM proposed by Semin and Fiedler distinguishes four word classes (three types of
verbs and adjectives) relevant to the interpersonal domain without considering nouns.
But impressive evidence concerning the impact of nouns (vs. adjectives) in people’s de
scriptions provided by Markman and Smith (cited in Markman, 1989) suggests that nouns
have a greater inductive potential and exert a greater effect on stereotypic representa
tions than do adjectives. In one of their studies, the authors presented two descriptions of
targets, one using a noun (e.g., “John is a sexist”) and the other an adjective (e.g., “John
is sexist”). Although pretested for semantic similarity, noun descriptions were considered
more powerful, more informative about the target, and as referring to more central and
stable features of the described person.
This is in line with linguistic research showing that nouns differ from adjectives in impor
tant ways. Nouns categorize events rather than describing event properties, they imply
multiple qualities rather than specifying a single property, they have an either-or quality
rather than allowing distinctions of degree, and they have a superordinate status in
speech production. If nouns can be considered the most abstract level of the LCM, then
they should exceed their corresponding adjectives in informativeness about the person,
perceived duration, and projection into the future. Carnaghi, Maass, Gresta, Bianchi, Cad
inu, and Arcuri (2008) tested the inductive potential of nouns, in comparison to semanti
cally similar adjectives, arguing that the stereotypes associated with the category should
become more accessible when a target person is described by a noun (Linda is a Catholic)
rather than by an adjective (Linda is catholic). In support of this idea, people exposed to
minimal phrases judged the likelihood of stereotype-congruent behaviors (e.g., going to
church) much higher in the noun than in the adjective condition, but judged the likeli
hood of incongruent behaviors (e.g., having one-night stands) greater in the adjective
than in the noun condition. At a more general level, Carnaghi et al. (2008) also observed
that nouns (compared to adjectives) are seen as providing more information about the
target person, as implying a more enduring characteristic, and as allowing better predic
tions about future behaviors. These findings suggest that nouns may constitute a logical
Page 17 of 33

extension of Semin and Fiedler’s (1988) LCM and possibly be added to the model as a
fifth level of abstraction, representing the most general type of assertion that can be
made about members of social categories. Together, there is considerable evidence that
language abstraction (including different verbs, adjectives, and possibly also nouns) plays
an important role in stereotype transmission, as well as in ingroup favoritism.
Pronouns
The final linguistic category that we consider here are pronouns, which are qualified as
seemingly unimportant function words or “junk words” (Chung & Pennebaker, 2007). Pro
nouns are linguistic markers that are subordinate to other referential elements, yet, as
we argue here, they are surprisingly important in the communication of social stereo
types since they are able to promote culturally shared worldviews and “demand a shared
understanding of their referent between the speaker and listener” (Chung & Pennebaker,
2007, p. 349; see also von Hippel, Sekaquaptewa, & Vargas, 2009; Ireland & Mehl, this
volume; Chung & Pennebaker, this volume). Pronouns have been studied in relation to in
tergroup processes under different perspectives. On one side, pronouns can refer to sin
gular or plural targets and to targets that do or do not include the self (first- vs. third-per
son). First-person singular pronouns (such as I or me) signal a focus on the self, whereas
first-person plural pronouns (such as we or us) reflect a focus on the collective. On the
other side, pronouns such as her versus his can mark the gender of the target person, ac
tivating corresponding gender stereotypes.
These features go well beyond grammatical properties, representing a subtle tool to pro
mote attentional focus and to express social membership, beliefs, and stereotypes. Given
the aims of this chapter, we report findings of three lines of research concerning (a) pro
nouns as markers of personal versus social identity, (b) pronouns reflecting and promot
ing individualistic versus collectivistic worldviews, and (c) gender-marked versus gender-
neutral pronouns.
(p. 169) Pronouns Marking Personal Versus Social Identity
First-person singular pronouns suggest a focus on the self and an increased self-aware
ness (Wegner & Giuliano, 1980) and are generally considered a marker of personal identi
ty. Pronouns such as I or me imply that the locus of attention is the self; by extension, they
are associated with greater attention to individual and personal issues. This relation has
been experimentally corroborated by various scholars, reporting an increment in the use
of first-person singular pronouns (I, me) when participants are put in a state of self-
awareness, for example, by the presence of a mirror (Dijksterhuis & Van Knippenberg,
2000). More importantly for the aims of this chapter, self-focus is also related to inter
group processes (stereotype expression and ingroup bias) and to culturally driven self-
stereotypes or self-construals. In particular, an increased self-focus leads to a momentari
ly reduced expression of stereotypes (Macrae, Bodenhausen, & Milne, 1998) and limits
the effect of stereotype activation on behaviors (Dijksterhuis & Van Knippenberg, 2000).
Page 18 of 33

Conversely, the use of the pronoun we has been interpreted as a proxy of ingroup identifi
cation and of focus on one’s group membership. As argued by Cialdini and colleagues, “by
employing the pronoun we, one is publicly able to associate oneself with another person
or group of persons. Through the use of some other designation, for example, they, one is
able to distance oneself from (i.e., to weaken the perceived association with) another per
son or persons” (Cialdini et al., 1976, p. 370). The authors confirm this strategic use of
language in intergroup relations, showing that the use of the pronoun we in relation to a
valued group is a way to maintain a positive self-esteem, and, indeed, this tendency is en
hanced when participants’ self-esteem is threatened. Similarly, Perdue, Dovidio, Gurtman,
and Tyler (1990, experiments 2 and 3) showed that first-person plural pronouns (we) are
associated with positive evaluations. Compared to plural third-person (they, them), plural
first-person pronouns (we, us) were found to facilitate the access to positive constructs.
Housley and colleagues (Housley, Claypool, Garcia-Marques, & Mackie, 2010) confirmed
this result, showing that a we prime enhanced the familiarity of a subsequent element,
presumably because such pronouns are associated with greater positivity.
It is important to note that the use of pronouns is likely to be unintentional because par
ticipants generally do not realize that they are applying a systematic linguistic strategy
that promotes social differences. Similarly, people may react to such pronouns uncon
sciously, as suggested by Perdue and colleagues (1990), in research showing that pro
nouns primed outside of participants’ awareness still affected subsequent evaluations. To
gether, these findings confirm the role of personal pronouns in intergroup processes as
linguistic tools that reflect speakers’ social (vs. personal) identity and that maintain inter
group differences.
Pronouns Reflecting and Promoting Individualistic Versus Collec

tivistic Worldviews
At the same time, pronouns are also expressions of cultural differences in individualistic
versus collectivist construals. The use of pronouns is regulated by language practices and
grammar; in turn, the use of pronouns promotes specific cultural worldviews. According
to Hofstede’s distinction between individualistic and collectivist cultures (Brewer & Chen,
2007; Hofstede, 1980), individualistic cultures (e.g., US or UK) promote self-interest and
focus on single persons as individuals, whereas collectivistic cultures (e.g., Japan or Chi
na) promote social membership and communal interests. Importantly, culture and cultur
al differences are created, transmitted, and perpetuated through communication
(Kashima, 2008; Kashima, Kashima, & Kidd, this volume), and language is the privileged
tool of human communication. Kashima and Kashima (1998; 2003) have related the use of
pronouns in grammar and natural language use to the collectivistic and individualistic dif
ferentiation proposed by Hofstede. The phenomenon they examined is called pronoun-
drop and refers to the grammatical possibility of omitting the pronominal subject in a sen
tence (e.g., went to the movies rather than we went to the movies). The subject of the dis
course remains implicit and hence has to be inferred by the listener from contextual in
formation. According to the authors, pronoun-drop is a grammatical feature of language
that has important psychological implications, given that pronouns shift the attention to
Page 19 of 33

specific targets (the self, a person, a group). Dropping pronouns means dropping the fo
cus on who is performing the action, with the action and the context remaining the focus
of the communication. Only some languages have the possibility to drop the pronoun alto
gether, such as Japanese or Korean. In other languages, such as Italian, the pronoun can
be dropped, but the information about who is performing the action is provided in the suf
fix. A third (p. 170) group of languages, such as English or German, does not allow any
pronoun-drop, and a sentence such as “went into the house” is considered grammatically
unacceptable since it is lacking the subject of the action. The authors report a correlation
of –.75 with Hofstede’s index of individualism, meaning that pronoun-drop becomes less
acceptable the more countries are classified as individualistic. Cultures that allow pro
noun-drop de-emphasize the agent of an action, whereas cultures that impede this possi
bility create a constant focus on the agent and remove the attention from the context in
which agents are embedded. This is perfectly in line with the definition of individualistic
cultures, suggesting that language is a critical factor in communicating and maintaining
this culturally shared worldview. In contrast, the requirement to continuously use contex
tual information to disambiguate the communication reminds listeners that agents are
part of the larger picture, thereby promoting a collectivistic worldview.
In line with Kashima’s findings, Marian and Kaushanskaya (2004) have shown that bilin
gual participants adjust their use of personal pronouns to the language in which they are
expressing their autobiographical memories. When participants recalled their memories
in Russian, they used more plural personal pronouns than when they recalled them in
English. This is in line with Hofstede’s definition of English and Russian cultures, the first
being more individualistic compared to the second. This finding can be interpreted as a
manifestation of self-stereotyping or of the activation of culturally promoted self-constru
als, with language working as a prime of cultural content. Kühnen and Oyserman (2002)
directly tested the focus on personal singular versus plural pronouns on the way people
elaborate information. Pronoun salience was manipulated by asking participants to circle
in a text words such as I, me versus we, us. It was found that the salience of personal sin
gular (vs. plural) pronouns promoted an attention to single elements rather than to their
relation, which is in line with the cultural differences observed among Western and East
ern cultures (Nisbett, Peng, Choi, & Norenzayan, 2001). Together, these studies illustrate
a bidirectional link between pronoun use and individualistic versus collectivist world
views.
Gender-Marked Versus Gender-Neutral Pronouns
A different approach to pronouns is represented by the literature on gender marking and

the subtle bias it produces in communication. As discussed in the section on grammatical
gender, language can convey information about the gender of the person addressed in the
discourse. This can be done in various ways: some languages provide continuous informa
tion about gender (grammatical-gender languages, e.g., Italian) by marking gender in
various linguistic elements (such as articles, adjectives, nouns), whereas, in other lan
guages (natural-gender languages, e.g., English), information about gender becomes
rarely salient. One element that is informative of gender in both grammatical-gender and
Page 20 of 33

natural-gender languages is pronouns. Personal pronouns reveal the target’s gender in

the discourse and make it accessible and salient to the listener.
Osterhout, Bersick, and McLaughlin (1997) showed that the gender communicated with
pronouns automatically activates norm-congruent stereotypes, when the stereotype does
not match the gender of the pronoun (e.g., the nurse prepared himself for the operation),
an incongruency detection that is evident with stronger event-related potentials (ERP) ac
tivation compared to a stereotype-gender match (e.g., the nurse prepared herself for the
operation). This sensitivity to pronoun gender and corresponding stereotypes is con
firmed at the behavioral level by Banaji and Hardin (1996). They primed participants with
stereotypically masculine (e.g., doctor) or feminine (e.g., nurse) roles, then presented a
subsequent masculine (he), feminine (she), or neutral pronoun (it). Participants were
faster in categorizing the pronouns when pronoun gender matched the stereotype con
tent of the prime (e.g., doctor-he) than when it did not (e.g., nurse-he). Using a similar
logic, Kennison and Trofe (2003) and Duffy and Keir (2004) showed that a mismatch be
tween stereotype content and the gender of a subsequent pronoun leads to slower read
ing times and longer fixations. Together, these and related findings suggest that a small
information unit, such as a pronoun, is able to activate an entire set of beliefs and ex
pectancies that are extremely difficult to suppress (Oakill, Garnham, & Reynolds, 2005).
In many languages, pronoun use follows the rule of generic masculine, prescribing the
use of a generic masculine pronoun when referring to a single target of unknown gender
or when referring to a group of unknown or mixed gender (even when the majority of
group members are females). A long research tradition on generically used masculine
pronouns (he) has shown that these pronouns lead to a pro-male bias (Stahlberg et al.,
2007). In addition to hampering comprehension and inhibiting the accessibility (p. 171) of
female exemplars, masculine generics are also thought to promote prejudice. For in
stance, Hyde (1984, experiment 2) asked children to estimate how qualified women are
for a fictitious profession described either in masculine generic (he), in gender-fair (they;
he or she), or in feminine forms (she). Children thought that women would do less well
when the masculine generic was used.
A critical question in terms of political correctness is whether introducing neutral pro

nouns can effectively reduce gender discrimination and make language more inclusive.
Foertsch and Gernsbacher (1997) have shown that, in the English language, the gender-
neutral pronoun they (even if used in singular form) can serve as an efficient substitute
for the generic he, since it reduced the reading time of mismatching sentences without
enhancing the reading time of matching sentences. However, the pronoun they may not
be as neutral in other languages. For instance, the French ils (= they) is masculine and
the German sie (= they), although generic when used as third-person plural, is morpho
logically identical to the feminine singular form. Comparing the three languages, Gar
nham, Gabriel, Sarrasin, Gygax, and Oakill (2012) confirmed the importance of the gram
matical morphology of generics, showing that the generic form used in German (die,
which is morphologically feminine) attenuated the masculine bias, whereas the French
form (ils, which is masculine) enhanced it. Together, these studies suggest that pronouns
Page 21 of 33

bias comprehension not only in the case of generic masculine forms (e.g., he, his), but al
so in the case of seemingly neutral forms that are morphologically masculine or feminine
(e.g., lis, sie).
Conclusion
As this brief review suggests, there are many linguistic strategies to convey smooth ver
sus hostile intergroup relations or to justify the status quo even where it is unfair. The lin
guistic toolbox is rich and varied, ranging from very obvious and easily detectable “bru
tal-force” tools such as slurs or metaphors to very subtle fine-grained “precision” tools
such as word ordering or language abstraction that may achieve the speaker’s goal very
successfully exactly because language variations are so subtle that they easily escape the
conscious control of the listener. The fact that a linguistic choice is very subtle should not
detract from the fact that it may nevertheless be very influential. One may even argue
that a linguistic tool will be impactful to the extent that it remains hidden and undetected
by the listener.
Language is undoubtedly the predominant means by which the nature of intergroup rela
tions is communicated at an interpersonal level, transmitted from generation to genera
tion, and by which the mass media create shared representations of social groups. Lin
guistic markers are not only telling about attitudes, but they also constitute an important
means to express and affirm one’s social identity.
This review illustrates the wide range of language tools involved in this process, from
very explicit and intentionally chosen rhetoric (e.g., slurs or metaphors) up to very subtle
variations in language that are likely to be used spontaneously or even unconsciously
(e.g., word ordering, active vs. passive, language abstraction, pronoun use). We have fo
cused here on those elements of language that have received the greatest attention in re
search on language and intergroup relations. However, we acknowledge that there are
many other features of language not considered here. For instance, we have not consid
ered many of the underresearched elements, such as articles (determinate vs. indetermi
nate) or verb modes (e.g., conditionals), and we have not reported on paralinguistic mark
ers of attitudes and identity, such as hedges, tag questions, or speech errors (Von Hippel,
Sekaquaptewa, & Vargas, 2009). Despite these limits, the findings presented herein clear
ly illustrate that there are multiple language tools that fulfill a wide range of functions in
intergroup relations.
In particular, they drive attention to specific social groups while pushing others into the
background (e.g., grammatical gender, active vs. passive); they guide social categoriza
tion, creating sub- and superordinate categories (e.g., grammatical gender, pronouns);
they create, maintain, or change stereotypes (e.g., metaphors, word order, language ab
straction, grammatical gender, pronouns); they express individual social identities, as
well as broad cultural worldviews (e.g., language abstraction, pronouns); they serve in
group-protective goals (e.g., language abstraction, word order); they derogate and dehu
manize outgroups (e.g., slurs, metaphors); they serve system justification (e.g., slurs,
Page 22 of 33

metaphors, word order); and they affect policy decisions (e.g., metaphors). As evidenced
by this (partial and incomplete) list, each linguistic element fulfills multiple functions, and
multiple elements are available for each function, suggesting a complex interplay of lin
guistic elements that jointly shape intergroup relations. Although relatively little is known
about the conditions that make specific elements become the favorite choice, there is evi
dence that the more (p. 172) subtle forms are preferred whenever social norms prohibit
more blatant language tools such as slurs. According to Jeffries, Hornsey, Sutton, Dou
glas, and Bain (2012), there is a general social norm, the “David and Goliath principle,”
that protects low-power and low-status groups (“David groups”) from overt criticism. This
is exactly the situation in which subtle linguistic forms of discrimination (such as lan
guage abstraction or word order) become more acceptable and more efficient tools of
communication than overt offenses (slurs or metaphors).
Finally, language plays a central, yet frequently overlooked role in psychological research
because the majority of measurement devices in social cognition involve language. Social
psychologists rarely use nonlinguistic assessments such as observational methods or
physiological measures. They also rely heavily on language when creating experimental
conditions. Thus, in large part, methodological designs imply the massive use of language
to measure and manipulate variables. Although this may be a necessity, it is problematic
that researchers rarely reflect about their specific language choices, which may guide or
bias their participants’ responses. For instance, in research conducted with grammatical
gender languages, female participants may react differently to the IAT (Greenwald,
McGhee, & Schwartz, 1998) when pronouns (I, me, my) or other stimuli (e.g., depressed)
are inclusive of their gender rather than when masculine generic forms are used (e.g.,
Italian: mio vs. mio/mia; depresso vs. depresso/a). Similarly, measures of stereotypes, prej
udice, gender roles, and the like typically ask participants to either express their degree
of agreement to a series of assertions (e.g., Women should be cherished and protected by
men) or to rate typical behaviors (e.g., acts like a leader) or traits (e.g., assertive, helpful).
We suspect that answers to these questions may vary as a function of the level of abstrac
tion at which they are expressed.6 Also, the order in which scale endpoints are labeled
(e.g., men to the left, women to the right, or vice versa) may affect responses. As these
examples show, linguistic features may not only play a critical role in stereotype transmis
sion, maintenance, and change, but also in the measurement tools that psychologists use
to assess such beliefs.
References
Abele, A. (2003). The dynamics of masculine-agentic and feminine-communal traits: Find
ings from a prospective study. Journal of Personality and Social Psychology, 85, 768–776.
Allen, L. (1983). The language of ethnic conflict: Social organization and lexical culture,
New York: Columbia University Press.
Page 23 of 33

Assilamehou, Y., & Teste, B. (2013). The effects of linguistic abstraction on evaluations of
the speaker in an intergroup context: Using the linguistic intergroup bias makes you a
good group member. Journal of Experimental Social Psychology, 49, 113–119.
Banaji, M. R., & Hardin, C. D. (1996). Automatic stereotyping. Psychological science, 7,

136–141.
Bar-Tal, D. (1989). Delegitimization: The extreme case of stereotyping and prejudice. In

D. Bar-Tal, C. Graumann, A. W. Kruglanski, & W. Stroebe (Eds.), Stereotyping and preju
dice: Changing conceptions (pp. 169–188). New York: Springer Verlag.
Bar-Tal, D., & Hammack, P. L. (2012). Conflict, delegitimization, and violence. In L. R.

Tropp (Ed.), Oxford Handbook of intergroup conflict. New York: Oxford University Press.
Bem, S. L., & Bem, D. J. (1973). Does sex-biased job advertising “aid and abet” sex dis
crimination? Journal of Applied Social Psychology, 3, 6–18.
Boccato, G., Capozza, D., Falvo, R., & Durante, F. (2008). The missing link: Ingroup, out
group and the Human speciesm, Social Cognition, 26, 224–234.
Bohner, G. (2001). Writing about rape: Use of the passive voice and other distancing text
features as an expression of perceived responsibility of the victim. British Journal of So
Brauer, M., & Landry, M. (2008). Un ministre peut-il tomber enceinte? L’impact du gener
ique masculin sur les representations mentales. [Can a Secretary of State become preg
nant? The influence of the masculine generic on mental representations]. L’Annee Psy
chologique, 108, 243–272.
Braun, F. (2001). The communication of gender in Turkish. In M. Hellinger & H. Buß

mann (Eds.), Gender across languages: The linguistic representation of women and men
(vol. 1, pp. 283–310). Amsterdam/ Philadelphia: Benjamins.
Brewer, M. B., & Chen, Y. (2007). Where (who) are collectives in collectivism? Toward
conceptual clarification of individualism and collectivism. Psychological Review, 114(1),
133–151.
Carnaghi, A., & Maass, A. (2007a). Derogatory language in intergroup context:

(p. 173)
Are “gay” and “fag” synonymous? In Y. Kashima, K. Fiedler, & P. Freytag (Eds.), Stereo
type dynamics: Language-based approaches to stereotype formation, maintenance, and
transformation. Lawrence Erlbaum Associates.
Carnaghi, A., & Maass, A. (2007b). In-group and out-group perspectives in the use of
derogatory group label: Gay vs. fag. Journal of Language and Social Psychology, 26, 142–
156.
Page 24 of 33

Carnaghi, A., & Maass, A. (2008). Derogatory language use in intergroup contexts: Are
“gay” and “fag” synonymous? In Y. Kashima, K. Fiedler, & P. Freytag (Eds.), Stereotype
dynamics: Language-based approaches to stereotype formation, maintenance, and trans
formation (pp. 117–134). New York: Erlbaum.
Carnaghi, A., Maass, A., & Fasoli, F. (2011). Enhancing masculinity by slandering homo
sexuals: The role of homophobic epithets in heterosexual gender identity. Personality and
Social Psychology Bulletin, 37, 1655–1665.
Carnaghi, A., Maass, A., Gresta, S., Bianchi, M., Cadinu, M., & Arcuri, L. (2008). Nomina
sunt omina: On the inductive potential of nouns and adjectives in person perception. Jour
Chauncey G. (1994). Gay New York: Gender, urban culture, and the making of the gay
male world, 1890–1940. New York: Basic.
Chung, C. K., & Pennebaker, J. W. (2007). The psychological function of function words. In
K. Fiedler (Ed.), Social communication: Frontiers of social psychology (pp. 343–359). New
York: Psychology Press.
Cialdini, R. B., Borden, R. J., Thorne, A., Walker, M. R., Freeman, S., & Sloan, L. R. (1976).
Basking in reflected glory: Three (football) field studies. Journal of Personality and Social
Cooper, W. E., & Ross, J. R. (1975). World order. In R. E. Grossman, L. J. San, & T. J. Vance
(Eds.), Papers from the parasession on functionalism (pp. 63–111). Chicago: Chicago Lin
guistic Society.
D’Augelli, A. R. (1992). Lesbian and gay male undergraduates’ experience of harassment

and fear on campus. Journal of Inter-personal Violence, 7, 383–395.
Dijksterhuis, A., & Van Knippenberg, A. (2000). Behavioral indecision: Effects of self-fo
cus on automatic behaviour. Social Cognition, 18(1), 55–74
Douglas, K. M., Sutton, R. M., & Wilkin, K. (2008). Could you mind your language? An in
vestigation of communicators’ ability to inhibit linguistic bias. Journal of Language and
Dryer, M. S. (2011). Order of subject, object and verb. In M. S. Dryer & M. Haspelmath
(Eds.), The world atlas of language structures. Munich: Max Planck Digital Library, chap
ter 81. Accessed October 28, 2012 at http://wals.info/chapter/81.
Duffy, S. A., & and. Keir, J. A. (2004). Violating stereotypes: Eye movements and compre
hension processes when text conflicts with world knowledge. Memory & Cognition, 32(4),
551–559.
Page 25 of 33

Falomir-Pichastor, J. M., & Mugny G. (2009). “I’m not gay…I’m a real man!”: Heterosexual
men’s gender self-esteem and sexual prejudice. Personality and Social Psychology Bulletin,
35, 1233–1243.
Foertsch, J., & Gernsbacher, M. A. (1997). In search of gender neutrality: Is singular they
a cognitively efficient substitute for generiche? Psychological Science, 8(2), 106–111.
Foroni, F., & Rothbart, M. (2011). Category boundaries and category labels: When does a
category name influence the perceived similarity of category members? Social Cognition.
29, 547–576.
Gabriel, U. (2008). Language policies and in-group favoritism: The malleability of the in
terpretation of generically intended masculine forms. Social Psychology, 39, 103–107.
Galinsky, A. D., Hugenberg, K., Groom, C. J., & Bodenhausen, G. V. (2003). The reappro
priation of stigmatizing labels: Implications for social identity. In J. Polzer, M. Neale, & E.
Mannix (Eds.), Identity issues in groups. Greenwich, CT: Elsevier Science Press.
Garnham, A., Gabriel, U., Sarrasin, O., Pascal Gygaxc, P., & Oakhilla, J. (2012). Gender
representation in different languages and grammatical marking on pronouns: When beau
ticians, musicians, and mechanics remain men. Discourse Processes, 49, 481–500.
Gentner, D., & Wolff, P. (1997). Alignment in the processing of metaphor. Journal of Mem
Geschke, D., Sassenberg, K., Ruhrmann, G., & Sommer, D. (2010). Effects of linguistic ab
stractness in the mass media: How newspaper articles shape readers’ attitudes toward
migrants. Journal of Media Psychology: Theories, Methods, and Applications. 22, 99–104.
Gibbs, R. W. Jr. (2008). The Cambridge handbook of metaphor and thought (pp. 197–211)
New York: Cambridge University Press.
Goff, P. A., Eberhart, J. L., Williams, M. J., & Jackson, M. C. (2008). Not yet human: implic
it knowledge, historical dehumanization, and contemporary consequences. Journal of Per
Gorham, B. W. (2006). News media’s relationship with stereotyping: The linguistic inter
group bias in response to crime news. Journal of Communication, 56, 289–308.
Graumann, C. F. (1995). Discriminatory discourse. Pattern of Prejudice, 29, 69–83.
Greenwald, A. G., McGhee, D. E., & Schwartz, J. L. K. (1998). Measuring individual differ
ences in implicit cognition: The Implicit Association Test. Journal of Personality and Social
Psychology, 74, 1464–1480.
Gygax, P., & Gabriel, U. (2008). Can a group of musicians be composed of women? Gener
ic interpretation of French masculine role names in the absence and presence of feminine
forms. Swiss Journal of Psychology, 67, 143–151.
Page 26 of 33

Hamilton, M. C., Hunter, B., & Stuart-Smith, S. (1992). Jury instructions worded in the
masculine generic: Can a woman claim self-defense when “he” is threatened? In J. C.
Chrisler & D. Howard (Eds.), New directions in feminist psychology: Practice, theory, and
research (pp. 169–178). New York: Springer.
Haslam, N. (2006). Dehumanization: An integrative review. Personality and Social Psy

chology Review, 10, 252–264.
Haslam, N., Loughnan, S., & Sun, P. (2011). Beastly: What makes animal metaphors offen
sive? Journal of Language and Social Psychology, 30, 311–325.
Hegarty, P., & Buechel, C. (2006). Androcentric reporting of gender differences in APA
journals: 1965–2004. Review of General Psychology, 10, 377–389.
Hegarty, P., Lemieux, A., & McQueen, G. (2010). Graphing the order of the sexes: Con
structing, recalling, interpreting, and putting the self in gender difference graphs. Journal
Hegarty, P., Watson, N., Fletcher, K., & McQueen, G. (2011). When are gentlemen first
and ladies last? Effects of gender stereotypes on the order of romantic partners’ names.
British Journal of Social Psychology, 50, 21–35.
(p. 174) Heise, E. (2000). Sind Frauen mitgemeint? Eine empirische Untersuchung zum
Verständnis des generischen Maskulinums und seiner Alternativen [Are women implied?
An empirical study on the comprehension of the generic masculine and its alternatives].
Zeitschrift für Sprache & Kognition, 19, 3–13.
Henley, N. M., Miller, M., & Beazley, J. A. (1995). Syntax, semantics, and sexual violence:
Agency and the passive voice. Journal of Language and Social Psychology, 14, 60–84.
Hofstede, G. (1980). Culture’s consequences: International differences in work-related

values, Beverly Hills, CA: Sage.
Housley, M. K., Claypool, H. M., Garcia-Marques, T., & Mackie, D. M. (2010). “We” are fa
miliar but “It” is not: Ingroup pronouns trigger feelings of familiarity. Journal of Experi
mental Social Psychology, 46, 114–119.
Hsu, L.-H. (2011). Linguistic intergroup bias tells ingroup/outgroup orientation of bicul
tural Asian Americans. International Journal of Intercultural Relations, 35, 853–866.
Hyde, J. S. (1984). Children’s understanding of sexist language. Developmental Psycholo

gy, 20, 697–706.
Jeffries, C. H., Hornsey, M. J., Sutton, R. M., Douglas, K. M., & Bain, P. G. (2012). The
David and Goliath Principle: Cultural, ideological, and attitudinal underpinnings of the
normative protection of low-status groups from criticism. Personality and Social Psycholo
gy Bulletin, 38, 1053–1065.
Page 27 of 33

Kashima, Y. (2008). A social psychology of cultural dynamics: Examining how cultures are
formed, maintained, and transformed. Social and Personality Psychology Compass, 2/1,
107–120.
Kashima, Y., & Kashima, E. S. (2003). Individualism, GNP, climate, and pronoun drop: Is
individualism determined by affluence and climate, or does language use play a role?
Kennedy, K. (2002). Nigger. The strange career of a troublesome word. New York: Pan
theon Books.
Kennison, S. M., & Trofe, J. L. (2003). Comprehending pronouns: A role for word-specific
gender stereotype information. Journal of Psycholinguistic Research, 32(3), 355–378.
Kite, M. E., & Deaux, K. (1987). Gender belief systems: Homosexuality and the implicit in
version theory, Psychology of Women Quarterly, 11, 83–96.
Kühnen, U., & Oyserman, D. (2002). Thinking about the self influences thinking in gener
al: Cognitive consequences of salient self-concept. Journal of Experimental Social Psychol
ogy, 38(5), 492–499.
Kunda, Z., & Spencer, S. J. (2003). When do stereotypes come to mind and when do they
color judgment? A goal-based theoretical framework for stereotype activation and appli
cation, Psychological Bulletin, 129, 522–544.
Lakoff, G., & Johnson, M. (1980). Conceptual metaphor in everyday language. Journal of
Philosophy, 77, 453–486.
Landau, M. J., Robinson, M. D., & Meier, B. P. (in press). Metaphorical thought in social
life, APA Press.
Leader, T. I., Mullen, B., & Rice, D. R. (2009). Complexity and valence in ethnopaulisms
and exclusion of ethnic out-groups: What puts the “hate” into Hate speech?, Journal of
Personality and Social Psychology, 96, 170–182.
Leyens, J. Ph., Demoulin, S., Vaes, J., Gaunt, R., & Paladino, M. P. (2007). Infra-humaniza
tion: The wall of group differences. Journal of Social Issues and Policy Review, 1, 139–
172.
Maass, A. (1999). Linguistic intergroup bias: Stereotype perpetuation through language.

In M. P. Zanna (Ed.), Advances in experimental social psychology (vol. 31, pp. 79–121).
San Diego, CA: Academic Press.
Maass, A., Ceccarelli, R., & Rudin, S. (1996). Linguistic intergroup bias: Evidence for in-
group protective motivation. Journal of Personality and Social Psychology, 71, 512–526.
Page 28 of 33

Maass, A., Corvino, G, & Arcuri, A. (1994). Linguistic intergroup bias and the mass media.
Revue Internationale de Psychologie Sociale, 1, 31–43.
Maass, A., Milesi, A., Zabbini, S., & Stahlberg, D. (1995). The linguistic intergroup bias:
Differential expectancies or in-group protection? Journal of Personality and Social Psy
chology, 68, 116–126.
Maass, A., Suitner, C. & Arcuri, L. (2013). The role of metaphors in intergroup relations,
In: M. J. Landau, M. D. Robinson, & B. P. Meier (eds.) The power of metaphor: examining
its influence on social life. APA.
Maass, A., Suitner, C., Favaretto, X., & Cignacchi, M. (2009). Groups in space: Stereo
types and the spatial agency bias. Journal of Experimental Social Psychology, 45, 496–
504.
Macrae, C. N., Bodenhausen, G. V. & Milne, A. B. (1998). Saying no to unwanted thoughts:

Self-focus and the regulation of mental life. Journal of Personality and Social Psychology,
74, 578–589.
Marian, V., & Kaushanskaya, M. (2004). Self-construal and emotion in bicultural bilin
guals. Journal of Memory and Language, 51, 190–201.
Markman, E. M. (1989). Categorization and naming in children: Problems of induction.

Cambridge, MA: MIT Press.
McGuire, W. J., & McGuire, C. V. (1992). Psychological significance of seemingly arbitrary

word-order regularities: The case of kin pairs. In G. R. Semin and K. Fiedler (Eds.). Lan
guage, interaction and social cognition (pp. 214–236). London: Sage.
Menegatti M., & Rubini M. (2007). A chi parliamo? Di cosa parliamo. Il ruolo
dell’astrazione linguistica nella comunicazione politica [Speaking about whom? Speaking
about what? The role of linguistic abstraction in political communication], Psicologia So
ciale, 2, 501–522.
Menegatti, M., & Rubini, M. (2012). From the individual to the group: The enhancement
of linguistic bias. European Journal of Social Psychology, 42, 36–40.
Merkel, E., Maass, A., & Frommelt, L. (2012). Shielding women against status loss: The
masculine forms and its alternatives in Italian language. Language and Social Psychology,
31, 311–320.
Miller, C. H., & Swift, K. (2001). The handbook of nonsexist writing (2nd ed.). San Jose,
CA: iUniverse, Inc.
Mollin, S. (2012). Revisiting binomial order in English: Ordering constraints and re
versibility. English Language and Linguistics 16, 79–101.
Page 29 of 33

Mollin, S. (2013). Pathways of change in the diachronic development of binomial re

versibility in Late Modern American English, Journal of English Linguistics. 41, 168–203.
Moser, F., Sato, S., Chiarini, T., Dmitrow-Devold, K., & Kuhn, E. (2011). Comparative
Analysis of Existing Guidelines for Gender-Fair Language within the ITN LCG Network:
Marie Curie Initial Training Network: Language, Cognition and Gender (ITN LCG). Re
trieved from https://ilias.unibe.ch/goto.php?
target=wiki_256402_WORK_PACKAGE_B%3A_Report_on_Milestone_1.
Mucchi-Faina, A. (2005). Visible or influential? Language reforms and gender (in)equality.

Social Science Information, 44, 189–215.
Mullen, B. (2001). Ethnophaulisms for ethnic immigrant groups. Journal of Social

(p. 175)
Issues, 57, 457–475.
Mullen, B. (2004). Sticks and stones can break my bones, but ethnophaulism can alter the
portrayal of immigrants to children. Personality and Social Psychology Bulletin, 30, 250–
260.
Mullen, B., Calogero, R. M., & Leader, T. I. (2007). A social psychological study of eth
nonyms: Cognitive representation of the in-group and intergroup hostility. Journal of Per
Mullen, B., & Rice, D. R. (2003). Ethnophaulism and exclusions: The behavioral conse
quences of cognitive representation of ethnic immigrant groups. Personality and Social
Psychology Bulletin, 29, 1056–1067.
Mullen, B., Rozell, D., & Johnson, C. (2001). Ethnophaulisms for ethnic immigrant groups:
The contributions of group size and familiarity. European Journal of Social Psychology, 31,
231–246.
Mullen, B., & Smyth, J. M. (2004). Immigrant suicide rates as a function of

ethnophaulisms: Hate speech predicts death. Psychosomatic Medicine, 66, 343–348.
Nisbett, R. E., Peng, K., Choi, I., & Norenzayan, A. (2001). Culture and systems of
thought: Holistic versus analytic cognition. Psychological Review, 108, 291–310.
Oakill, J., Garnham, A., & Reynolds, D. (2005). Immediate activation of stereotypical gen
der information. Memory & Cognition, 33(6), 972–983.
Ortony, A. (1975). Why metaphors are necessary and not just nice. Educational Theory,
25(1), 45–53.
Osterhout, L., Bersick, M., & McLaughlin, J. (1997). Brain potentials reflect violations of
gender stereotypes. Memory & Cognition, 25, 273–285.
Page 30 of 33

Perdue, C. W., Dovidio, J. F., Gurtman, M. B., & Tyler, R. B. (1990). Us and them: Social
categorization and the process of intergroup bias. Journal of Personality and Social Psy
chology, 59, 475–486.
Perez, J. A., Moscovici, S., & Chulvi, B. (2007). The taboo against group contact: Hypothe
sis of Gypsy ontologization. British Journal of Social Psychology, 46, 249–272.
Plummer, D. C. (2001). The quest for modern manhood: Masculine stereotypes, peer cul
ture and the social significance of homophobia. Journal of Adolescence, 23, 1–9.
Prewitt-Freilino, J. L., Caswell, T. A., & Laakso, E. K. (2012). The gendering of language: A
comparison of gender equality in countries with gendered, natural gender and genderless
languages. Sex Roles, 66, 268–281.
Rattan, A., & Ambady, N. (2013). Diversity ideologies and intergroup relations: An exami
nation of colorblindness and multiculturalism. European Journal of Social Psychology. 43,
12–21.
Rubini, M., & Menegatti. M. (2008). Linguistic bias in personnel selection. Journal of Lan
guage and Social Psychology, 27, 168–181.
Rubini, M., Moscatelli, S., & Palmonari, A. (2007). Increasing group entitativity: Linguis
tic intergroup discrimination in the minimal group. Group Processes and Intergroup Rela
tions, 10, 280–296.
Saminaden, A., Loughan, S., & Haslam, N. (2010). Afterimages of savages: Implicit associ
ations between “primitives,” animals and children. British Journal of Social Psychology,
49, 91–105.
Schmid, J., & Fiedler, K. (1998). The backbone of closing speeches: The impact of prose
cution versus defense language on judicial attributions. Journal of Applied Social Psychol
ogy. 28, 1140–1172.
Semin, G. R., & de Poot, J. C (1997). The question-answer paradigm: You might regret not
noticing how a question is worded. Journal of Personality and Social Psychology, 73, 472–
80.
chology, 54, 558–568.
Simon, L., & Greenberg J. (1996). Further progress in understanding the effects of
derogatory ethnic labels: The role of pre-existing attitudes toward the targeted group.
Stahlberg, D., Braun, F., Irmen, L., & Sczesny, S. (2007). Representation of the sexes in
language. In K. Fiedler (Ed.), Social communication (pp. 163–187). New York: Psychology
Press.
Page 31 of 33

Stahlberg, D., & Sczesny, S. (2001). Effekte des generischen Maskulinums und alternativ
er Sprachformen auf den gedanklichen Einbezug von Frauen [Effects of the generic mas
culine and alternative language forms on the cognitive inclusion of women] Psychologis
che Rundschau, 52(1), 131–140.
Stahlberg, D., Sczesny, S., & Braun, F. (2001). Name your favorite musician: Effects of
masculine generics and of their alternatives in German. Journal of Language and Social
Suitner, C., & Maass, A. (2007). Positioning bias in portraits and self-portraits: Do female
artists make different choices? Journal of Empirical Studies of the Arts, 25(1), 71–95.
Suitner, C., & Maass, A. (2008). The role of valence in the perception of agency and com
munion. European Journal of Social Psychology, 38, 1073–1082.
Tajfel, H., & Turner, J. C. (1986). The social identity theory of intergroup behaviour. In S.
Worchel & W. G. Austin (Eds.), Psychology of intergroup relations (2nd ed., pp. 7–24).
Chicago: Nelson-Hall.
Thibodeau, T. H., & Borodistky, L. (2011). Metaphors we think with: The role of metaphor
in reasoning. PLoS ONE, 6(2), e16782. doi:10.1371.
Vaes, J., Leyens, J., Paladino, M. P., & Miranda, M. P. (2012). We are human, they are not:
Driving forces behind outgroup dehumanization and the humanization of the ingroup. Eu
ropean Review of Social Psychology. 23, 64–106.
Vicars, M. (2006). Who are you calling queer? Sticks and stones can break my bones but
names will always hurt me. British Educational Research Journal, 32, 347–361.
Viki, G. T., Winchester, L., Titshall, L., Chasing, T., Pina, A., & Russell, R. (2006). Beyond
secondary emotions: The infrahumanization of outgroups using human-related and ani
mal-related words. Social Cognition, 24, 753–775.
Volpato, C., Durante, F., Gabbiadini, A., Andrighetto, L., & Mari, S. (2010). Picturing the
other: Targets of delegitimization across time. International Journal of Conflict and Vio
lence, 4(2), 269–287.
Von Hippel, W., Sekaquaptewa, D., & Vargas, P. T. (2009). Linguistic markers of implicit
attitudes. In R. E. Petty, R. H. Fazio, & P. Brinol (Eds.), Attitudes: Insights from the new
implicit measures. (pp. 429–458). New York: Psychology Press.
Wasserman, B. D., & Weseley, A. J. (2009).?Que? Quoi? Do languages with grammatical

gender promote sexist attitudes? Sex Roles, 61, 634–643.
Wegner, D. M., & Giuliano, T. (1980). Arousal-induced attention to self. Journal of

(p. 176)
Page 32 of 33

Wenneker, C. P. J., Wigboldus, D., & Spears, R. (2005). Biased language use in stereotype
maintenance: The role of encoding and goals. Journal of Personality and Social Psycholo
gy, 89, 504–516.
Wigboldus D. H. J, Semin G. R., & Spears R. (2006). Communicating expectancies about

others. European Journal of Social Psychology, 36, 815–824.
Wigboldus, D., & Douglas, K. M. (2007). Language, stereotypes, and intergroup relations.
In K. Fiedler (Ed.), Social communication (pp. 79–106). New York: Psychology Press.
Notes:
(1) . This holds even after controlling for the suicide rates in the respective countries of
origin.
(2) . According to the World Atlas of Language Structures only 3 percent of languages
have an inverse order, with the object preceding the subject (Dryer, 2011).
(3) . There is one important exception to this rule: when one of the pair is more familiar,
that person (whether male or female) will be mentioned first (see Hegarty et al., 2011,
study 5).
(4) . In our opinion, these findings are open to an alternative interpretation; namely, that
language may have functioned as a prime of different cultural values associated with less
er or greater gender equality.
(5) . This issue shares a number of similarities with the ongoing discussion of whether
color-blindness or multiculturalism are better suited to improve intergroup relations (see
Rattan & Ambady, 2013).
(6) . In a yet unpublished study in which we varied the lexical category of the items of a
gender stereotyping scale (verbs, adjective, nouns), we found that participants reported
greater gender differences for verbs (e.g., rationalize, help) than for adjectives (e.g., ra
tional, helpful) or nouns (e.g., rationality, helpfulness).
Anne Maass
Anne Maass, University of Padova
Luciano Arcuri
Luciano Arcuri, University of Padova
Caterina Suitner
Caterina Suitner, University of Padova
Page 33 of 33

Language, Style, and Persuasion

James Price Dillard

This chapter reviews empirical studies on language and persuasion from a message ef
fects perspective. Seven subliteratures are considered: self-referencing, rhetorical ques
tions, conclusion explicitness, vividness, gain/loss framing, powerful(less) language, and
domineeringness. For each of these, the central linguistic construct is defined and ana
lyzed prior to considering theoretical explanations for the variable’s effect on persuasion.
The state-of-the-area is evaluated and suggestions are offered to guide future efforts at
theory development.
Keywords: Persuasion, self-referencing, rhetorical questions, explicitness, vividness, framing, power, reactance,
politeness
The Latin verb suadare derives from the adjective suavem, meaning sweet or agreeable
(Chapman, 1957). When preceded by per, the verb form probably entered the English lan
guage via the French and persuadre—to persuade—came to signify talk of a certain sort.
That is, advice-giving or an effort to bring a person over to the truth. The truth, of course,
is a malleable entity, one that is often in dispute. But setting aside the common possibility
of a deceptive and self-interested message source, it is still possible to ask how individu
als give advice to others and under what circumstances listeners become enlightened.
Ancient students of rhetoric identified three aspects of a message that might move hear
ers toward the truth. The content is the substance of the message, its topic and its argu
ments. Questions of structure focus on the ways in which content can be arranged. For in
stance, a problem–solution format may be superior to an ordering in which the solution is
presented prior to the problem. Style, the third message component, is concerned with
how content is presented.
Language, which has been considered an aspect of style, is the focus of this chapter.
Page 1 of 19

Scope and Aims of This Chapter

One means of understanding the literature on language and persuasion would be to adopt
a general theory, then search for links between the central processing mechanism and
various features of language. If one of the dual-process models were chosen (Chaiken,
Liberman, & Eagly, 1989; Petty & Cacioppo, 1981), the next step would be to engage in a
focused pursuit of language variables that might influence deep versus shallow process
ing. This strategy would move in the direction of a unified view of linguistic phenomena,
but it would not accurately reflect the diverse theorizing and contested nature of the lit
erature as it currently exists. Accordingly, the approach taken here is more bottom-up
than top-down. Following a broad search of the literature, seven areas of focus were se
lected, with preference given to areas that produced (a) multiple studies of the effects of
(b) a single language variable. The result included self-referencing language, (p. 178)
rhetorical questions, conclusion explicitness, domineering language, powerful(less) lan
guage, gain-loss framing, and vivid versus pallid styles. Regrettably, space constraints
prohibited inclusion of several valuable bodies of work that also met the search criteria:
metaphor (Sopory & Dillard, 2002), language intensity (Hamilton, Hunter, & Burgoon,
1990), and agentic language (Dragojevic, Bell, & McGlone, 2013).
The first step in each of the seven sections was to articulate a language variable as it has
been used in research. Conceptual definitions were sometimes evident and, in other cas
es, implied. In both instances, they were evaluated against standard criteria for judging
theoretical constructs. To wit, constructs must be defined with precision if they are to be
a useful means of communicating among researchers. And, to the extent that a concept
possesses multiple components, those components are conceptually parallel. A common
example in persuasion research is credibility, which is often defined as having two compo
nents: expertise and trustworthiness. Both of these components are social judgments, and
they are similar in terms of level of abstraction.
A second task regarding each language variable was to express and evaluate the theoreti
cal positions employed in prior research. In keeping with the message effects tradition,
these theories were typically of the form X → Y → Z, where X is the language variable, Y is
some process, and Z is one or more persuasive outcomes. Sometimes one of the causal re
lationships were moderated by one or more other factors (M1, M2, and so on). And, com
monly, different researchers articulated different positions with respect to the nature and
functioning of Y.
To help locate the research and findings in a wider context, the language variables were
partitioned on the basis of their complexity. Next, we turn to the basis for that partition
ing.
Page 2 of 19

Features of Language in Persuasion

In ongoing efforts to classify features of language, scholars have drawn many useful lines
of analysis. For instance, any utterance might be understood in terms of phonology, syn
tax, and lexicality. The distinction made in this chapter is less nuanced but apt for the lit
erature under scrutiny. We distinguish between granular and thematic forms of language.
Granular forms are linguistic variations that exist at the level of words or phrases and
sometimes sentences. For instance, self-referencing messages are those that employ sec
ond-person pronouns (You should consider…). If the message is to be coherent, this
choice carries through the entire persuasive appeal. Indeed, inclusion of multiple forms—
first-person, second-person, and third-person—would produce shifts in perspective that
could damage message unity. But, notably, it is the same choice at every turn. For in
stance, the decision to write in active or passive voice is usually a whole-document deci
sion. Of course, deviations might occur, but typically they will be viewed as errors.
Thematic variations reflect choices made at a higher level of abstraction. Language that
is said to be vivid (vs. pallid) expresses a choice between lively and engaging, rather than
careful and reserved. To wit, Frey and Eagly (1993) write that airlines should not inform
the public of terrorist threats because the “foaming martyrs would…compete to see
whose blood-chilling threats could make headlines in the New York Times” (p. 43). Their
selection of adjectives draws from a much larger pool of lexical items and syntactic rules
than does self-referencing. Also, unlike granular features of language, thematic variations
tend not to be repeated in the same concrete form. Successful implementation of a theme
generally requires variation in concrete language elements, but consistent adherence to a
class of elements or a pivotal concept.
Granular Features of Persuasive Messages

Self-Referencing
Self-referencing language can be seen in experimental instructions such as “Try to imag

ine yourself in the situation portrayed in each of the advertisements. Think back to times
in which you have used or could have used each product” (Morrison & Shaffer, 2003, p.
272). Obviously, the researchers’ intention was to encourage participants to think about
the product with regard to themselves or, put differently, to instigate cognitive self-refer
encing. Given the very close conceptual distance between induction and desired effect, it
is not surprising that no formal definition was offered.
Other studies located self-referencing language in the appeal itself. Burnkrant and Unna
va (1989) created experimental conditions in which an advertisement for a razor con
tained either frequent use of second-person or third-person pronouns. Phrases such as
“You may recall…” were expected to prompt more cognitive self-referencing than “One
may recall…” And, indeed, the linguistic variation was found to produce effects on atti
tude and cognitive (p. 179) response that were similar to other sorts of involvement ma
Page 3 of 19

nipulations. Specifically, respondents in the self-referencing condition showed better abil

ity to discriminate variations in argument quality than did participants in the non–self-ref
erencing conditions.
Subsequent research replicated the involvement effect under other conditions including
the presence of irrelevant imagery (Burnkrant & Unnava, 1995) and the use of weak fear
appeals (Keller & Block, 1995). Importantly, it provided more direct evidence of the effect
of self-referencing on thoughts about the self (Burnkrant & Unnava, 1995; Meyers-Levy &
Peracchio, 1996). Thus, it provided more direct evidence of an X → Y → Z sort of process.
However, that same research also showed that the effect might reverse itself when com
bined with other involvement-inducing variables (Burnkrant & Unnava, 1995; Meyers-
Levy & Peracchio, 1996) or when individual differences were considered (Meyers-Levy &
Peracchio, 1996). Thus, the original X → Y → Z model was modified to include a minimum
of two moderator variables.
Rhetorical Questions
Consider three examples of a rhetorical question: Shall I compare thee to a summer’s

day? Must I argue the wrongfulness of slavery? Is the Pope a Catholic? Each of these is an
utterance expressed as an interrogative but without the expectation of an overt response
from the hearer. Aristotle may have been the first writer to systematically address the use
of rhetorical questions (Cooper, 1932). His advice for presenting questions that call for
the audience to make an obvious but silent response was to locate such questions at the
end of a speech such that they challenged an opponent’s conclusions. Subsequent theo
rizing offered speculation about the side conditions and processes for the effective use of
rhetorical questions (see also Roskos-Ewoldsen, 2003):
• Rhetorical questions enhance persuasion among audiences whose prior attitude is

positive because their subvocal response is consistent with the question (Zillmann,
1972).
• Rhetorical questions increase message processing in low-involvement audiences and
decrease it in high-involvement audiences (Petty, Cacioppo, & Heesacker, 1981).
• Rhetorical questions diminish persuasion because they are viewed as indicators of a
source’s effort to manipulate the audience (Munch & Swasy, 1985).
Individual studies exist that align with each of these three positions, but there is no pat
tern across investigations that clearly favors one position over the others or allows recon
ciliation of the three (Gayle, Preiss, & Allen, 1994). One feature of the literature that may
have impeded progress is the dearth of conceptual analysis: most work on persuasion and
rhetorical questions proceeds as if the concept requires no elaboration. This turns out to
create unforeseen problems of interpretation. For instance, the phrase “Don’t you
agree?” could be understood as a means of inducing self-referencing or, as discussed lat
er in this chapter, as an instance of powerless speech. Hence, the lay definition of rhetori
cal questions seems to lack the precision that is desired in scientific research.
Page 4 of 19

Fortunately, students of language have not been content to leave the notion of rhetorical
questions unexamined (Frank, 1990). Ilie’s (1994) explication, for one, suggests that a
more systematic treatment is possible. She details criteria for identifying rhetorical ques
tions: (a) a discrepancy between the interrogative form of the question and its commu
nicative function as a statement; (b) a polarity shift between the question and its implied
statement—that is, a question posed in the negative suggests an answer in the affirma
tive; and (c) multifunctionality, such that, apart from its function as a challenge, the ques
tion also operates as, for example, a reproach, a warning, a promise, or an exultation (cf.,
Roskos-Ewoldsen, 2003). Persuasion researchers wishing to study rhetorical questions
might find definitional problems more tractable by granting greater attention to work
such as Ilie’s, which grapples directly with explication.
Conclusion Explicitness
Explicitness is the degree to which a message source makes her or his intentions trans
parent in the message itself. Whereas explicit messages require little or no guesswork re
garding the speaker’s wants, inexplicit messages necessitate more interpretation, there
by demanding a longer cognitive route from utterance to meaning (Blum-Kulka, 1987).
This overall point is qualified by frequency of usage. “Could you pass the salt?” is so com
monly used to mean “Please pass the salt” that it seems unlikely that there is substantial
difference in cognitive effort or processing time between the two. Nonetheless, there may
still be important relational implications to be drawn from the two formulations.
There is an enormous body of work on explicitness as a property of language. However,

the research (p. 180) linking it to persuasion is more narrowly focused on explicitness as a
property of the message’s conclusion. An explicit conclusion is sometimes contrasted not
with an inexplicitly stated conclusion but with no conclusion at all. Explanatory accounts
include:
• Messages with unstated conclusions are most effective because recipients who are
required to form their own conclusions must engage the arguments to a greater de
gree (Hovland & Mandell, 1952).
• Explicit conclusions display persuasive intent that will likely stimulate reactance and
subsequent counterpersuasion (Brehm & Brehm, 1981).
• Explicit conclusions are most persuasive because they provide clear direction, even
to recipients who are uninvolved with the topic (Cruz, 1998).
Due to the fact that the overall database is small, considerable caution is needed when in
terpreting the literature. But, with that in mind, quantitative summaries of the literature
support only the third explanation (Cruz, 1998; O’Keefe, 1997): conclusion explicitness
promotes persuasion. There is no indication that conclusion-drawing influences source
credibility nor that audience involvement moderates the effect of explicitness on persua
sion. In its current form, the literature supports an unqualified X → Z model (explicitness
causes persuasion without moderators).
Page 5 of 19

Persuasion studies of explicitness seem not to have committed any gross errors in terms
of defining the concept under study. Although definitions are often casual, the opera
tionalization affords plausible precision of inference, and there is no obvious indication of
multiple components to the notion of conclusion explicitness, thereby avoiding problems
of heterogeneity among those components. If any criticism can be brought to bear it is
that the persuasion studies seem not especially ambitious. Better linkages to treatments
of explicitness in the language literature more broadly might encourage researchers to
reach further.
Thematic Features of Persuasive Messages

Vividness
Discussions of this message feature commonly note that beliefs about the power of vivid
language are widely held (e.g., Taylor & Thompson, 1982). Indeed, evidence of this im
plicit theory comes to mind quickly and easily: journalists seek the human interest angle
on a story, product promoters try to link their brand to a simple but dramatic concept,
and politicians present their version of the truth via concrete examples drawn from their
constituencies.
The single most influential piece of writing on vividness is the book chapter provided by
Nisbett and Ross (1980). At the outset, the authors note the empirical vacuum surround
ing vividness, which they say left them no alternative but to rely on “speculation…spotty
indirect evidence…[and] on anecdotal and case history evidence!” (p. 45). Despite these
caveats, the chapter exerted a profound effect on subsequent inquiry. One of its enduring
legacies is the tripartite description of the construct in which vivid information was de
fined as that which is “(a) emotionally interesting, (b) concrete and imagery-provoking,
and (c) proximate in a sensory, temporal, or spatial way” (p. 45). Each of these themes re
quires more detailed consideration.
Emotional interest was thought to result from several factors. To the extent that the infor
mation describes events that are affecting the message recipient or his or her social cir
cle, it is emotionally interesting. Emotional interest is also a function of the hedonic rele
vance of the event to the participants. In other words, information is emotionally interest
ing when it outlines circumstances that advance or conflict with the goals of the partici
pants. This aspect of vividness aligns well with appraisal theories of emotion, all of which
cast relevance and goal congruence as necessary conditions for the evocation of emotion
(e.g., Scherer, Schorr, & Johnstone, 2001). Indeed, viewing emotional interest through the
lens of appraisal theories serves to crystallize one key aspect of the definition: emotional
interest is not a feature of messages. Rather, it is an outcome of appraising the relation
ship of “the participants” (i.e., self or close others) vis á vis the environment as portrayed
in the message. It cannot be said that emotional interest is a useful definition or not,
when it is considered alone. But because it is only one of three conditions that define
Page 6 of 19

vividness, it should be expected that the remaining two would show conceptual paral
lelism.
The second major criterion for defining vividness is actually two distinguishable ideas:
concreteness and imaginability. Concrete expressions are those that provide more speci
ficity about individuals, their actions, and the context in which actions occur. For exam
ple, Smith and Shaffer’s (2000) manipulation contrasted “an increase in the probability of
an accident” (pallid condition) with “a high risk of bloody, bone-crushing accidents” (vivid
condition) (p. 777). The latter phrase is seen to be more (p. 181) vivid because it is more
specific with regard to the probability and severity of automobile accidents. In certain re
spects, concreteness (vs. abstractness) is a property of language. Concreteness provides
more information. But imaginability is different: it is an outcome rather than a feature of
language. It is the extent to which a message prompts sensory imagery, the ease with
which an image comes to mind. Indeed, imaginability has been construed as the causal
result of concreteness (Nisbett & Ross, 1980). Thus, the second major criterion is itself
two nonparallel ideas, neither of which show conceptual correspondence with the first
criterion of emotional interest.
Finally, information is considered vivid to the extent that it is close in some temporal, spa
tial, or sensory way. Events occurring in the distant past or the far future are defined as
less vivid than those taking place yesterday or tomorrow, and events that occur nearby
are more vivid than episodes that take place elsewhere. The Boston Marathon bombings
were more vivid for residents of that city than for denizens of Bombay. As with emotional
interest, proximity is not a message feature per se, but a relational concept. To judge
closeness, one must know something about the message together with something about
the message recipient. Conceivably, this final criterion—proximity—might be conjoined
with emotional interest: only events that are close and relevant and (incongruent) have
the potential to provoke emotions.
As can be seen, the three defining features of vividness do not a form a coherent whole.
Still, vividness seems to capture a set of ideas that resonate with persuasion researchers.
There are several theories of vividness effects. The early versions draw heavily from more
general theories of memory, whereas more contemporary positions emphasize cognitive
responses:
• Availability: Vivid information is more available in memory and thus better able to in
fluence judgment (e.g., Shedler & Manis, 1986).
• Differential attention: Vivid information is more likely to attract attention than is non
vivid information (e.g., Reyes, Thompson, & Bower, 1980).
• Availability-Valence: Vivid material activates associative pathways, but it is the va
lence of the information that determines the direction of vividness effects (e.g.,
Kisielius & Sternthal, 1986).
Page 7 of 19

• Figural vividness: Messages have both central (i.e., figural) and noncentral (i.e.,
ground) components, either of which can be made vivid/pallid. When the figural com
ponent is made vivid, it will enhance persuasion (Guadagno, Rhoads, & Sagarin, 2011).
Following their review of forty-seven papers, Taylor and Thompson (1982) concluded that
evidence of a vividness effect was virtually nonexistent. In terms of overall patterns in the
data, the situation in 2013 is functionally the same. Despite the existence of many careful
and creative individual studies, there is no apparent pattern in the literature as a whole.
One cannot help but think that these circumstances follow from the inchoate definition of
vividness. Overall, it is impossible to represent the literature in graphic form given that X
is not a single construct.
Gain and Loss Framing
Surely the best known example of framing is Tversky and Kahneman’s (1981) Asian dis
ease problem. In that research, participants were asked which of two programs to com
bat the disease was preferable. The gain-framed option put the decision in these terms: If
program A is adopted, 200 people will be saved; if program B is chosen, there is a one in
three probability that 600 will be saved. Participants preferred the risk-averse program A
(72 percent vs. 28 percent). The loss-framed condition put the problem somewhat differ
ently: if program C, then 400 people will die; but if program D, then there is a one in
three probability that no one will die. Here, most respondents liked program D (78 per
cent vs. 22 percent).
The simplicity of the Tversky and Kahneman’s (1981) design and the strength of the find
ings served to popularize the idea that expressing a problem in different formats could
have powerful effects on the decision itself. But, notably, the key difference within gain-
and loss-framed conditions was the presence versus absence of uncertainty (e.g., 200 will
be saved vs. a 1:3 probability that 600 will be saved). As the framing concept migrated in
to persuasion research, it also morphed into a different variable. Gain-framing came to
represent messages that highlighted the desirable outcomes associated with compliance,
whereas loss-framing was defined by an emphasis on the disadvantages of not complying
with the appeal. In other words, decision uncertainty became a constant across condi
tions, whereas framing developed into something conceptually closer to outcome valence.
The contrast can be seen in the following two statements: “If you exercise, (p. 182) you’ll
have all the benefits of a healthy heart” versus “Skipping exercise means that you’ll miss
out on the benefits of a healthy heart.”
Several theoretical positions emerged that anticipated main-effect differences between

frames:
• Loss framing is more potent than gain framing due to a basic human sensitivity to
negative information (Meyerowitz & Chaiken, 1987).
• Loss-framed messages are processed more systematically because they garner more
attention and violate expectations (Smith & Petty, 1996).
Page 8 of 19

• Gain-framed appeals are superior because they are less likely to arouse reactance
(Reinhart, Marshall, Feeley, & Tutzauer, 2007).
A series of meta-analyses by O’Keefe and Jensen (2006; O’Keefe & Wu, 2012) permit ro
bust empirical generalizations about the framing literature, three of which are presented
here in order of decreasing certainty. The first such generalization is that there is no ap
preciable difference between the two frames with regard to persuasion (O’Keefe &
Jensen, 2006). Analysis of more than 160 studies showed an effect equivalent to r = .02
with a 95 percent confidence interval of −.01 to +.04. Thus, the data are incompatible
with any of the main-effect theories of framing. In a simple frame-to-frame comparison,
neither frame is noticeably more powerful than the alternative. An important caveat is
that we know relatively little about the absolute potency of framing because only a hand
ful of studies have utilized designs that test opinion before and after message presenta
tion (e.g., Hoffner & Ye, 2009).
A second important finding concerns the effect of variation within frames. Several writers
have made the point that gain frames can be operationalized along two lines: whether
they provide the expectation of benefits or the avoidance of harmful consequences. Both
are desirable. Similarly, loss frames could convey the chances for an unwanted outcome
or for losing out on an attractive possibility. Both are undesirable consequences. Howev
er, the meta-analytic data indicate that this is a distinction without a difference. The con
fidence intervals for these comparisons include 0 (O’Keefe & Jensen, 2006).
In a third area, the findings are less clear. Gain frames appeared to be slightly more effec
tive at encouraging disease prevention behaviors, but the effect size was only r = .03.
And, that effect can largely be attributed to studies of dental hygiene behavior (O’Keefe &
Jensen, 2007). Similarly, loss frames show a very modest superiority when advocating dis
ease detection behaviors (r = −.04), but the effect seems to be carried by studies of
breast cancer self-examinations (O’Keefe & Jensen, 2009). In short, there may be some
topical domains in which one frame manifests main-effect superiority over the other. At
present, a theoretical framework that allows a priori classification of topics is lacking.
One approach that elides the need for topical evaluation altogether is offered by Rothman
and Updegraff (2011) who suggest that judgments of risk and certainty are a malleable
function of individuals.
Although the O’Keefe and Jensen (2006, 2007) meta-analyses did not consider person-cen
tered moderators, there is some evidence that attention to individual differences will be
needed to account for framing processes. Research that leverages the approach–avoid
ance distinction seems especially likely to bear fruit (Mann, Sherman, & Updegraff, 2004;
Sherman, Mann, & Updegraff, 2006). Although results vary somewhat across studies, the
typical finding is that of a frame–system matching effect (Gerend & Maner, 2011; Shen &
Dillard, 2007). That is, individuals are persuaded by gain-framed appeals to the extent
that they are high in approach motivation. Conversely, loss-framed messages are most ef
fective with those who are high in avoidance motivation. This pattern has been observed
using individual difference measures that assess chronic motivational levels (Mann et al.,
Page 9 of 19

2004), as well as in experimental studies that induce differing levels of approach or avoid
ance motivation (Yan, Dillard, & Shen, 2012). With regard to outcome variables, the
crossover interaction has been observed in attitude, intention, and recall data, as well as
in information-seeking behavior (Hong, 2012). Hence, the literature may be conceived as
X → Z with two moderators; that is, approach motivation (M1) and avoidance motivation
(M2).
Powerful and Powerless Language
Powerless speech is language that conveys a speaker’s uncertainty or ambivalence about

his or her position. It manifests as the frequent use of hedges (e.g., “sort of,” “I guess”),
hesitations (e.g., “uh,” “well,” “you know”), tag questions (e.g., “…don’t you think?”
“Don’t you agree?”), intensifiers (e.g., “really,” “so”), disclaimers (e.g., “I’m not sure,
but…,” “others may see it differently, but…”), and politeness (e.g., “Please,” “If you don’t
mind”). O’Barr’s (1982) sociolinguistic analyses of courtroom communication are usually
credited with originating the study of powerless language. (p. 183) Broadly, he found that
these features of language tended to co-occur and that they did so in the speech of low-
power/status witnesses. The reverse was true of persons with relatively high social pow
er, such as parole officers and physicians. In other words, powerful speech was marked
by the absence of forms that signal a lack of commitment or confidence.
Although the database is small, a quantitative summary indicates that powerful speech is
favorably associated with source credibility and with persuasion (Burrell & Koper, 1998).
Efforts to explain powerful(less) language effects can be summarized as:
• Powerful language conveys certainty, and judgments of source certainty cause belief
change in message recipients (Erickson, Lind, Johnson, & O’Barr, 1978).
• The effect of powerless forms is moderated by the medium of presentation, such that
they are most potent in fast modalities such as audio or video and least potent in slow
er modes such as print (Sparks, Areni, & Cox, 1998).
• Some components of powerless language influence depth of message processing,
whereas others function as peripheral cues (Blankenship & Holtgraves, 2005).
Differences in theoretical account aside, there are several potentially important qualifica
tions to the possibility of a general powerful(less) style effect. For one, there is remark
ably little evidence that the linguistic markers of power are part of an overall style.
O’Barr’s (1982) assertion remains the gold standard on this point, and, as noted, it was
based on highly contextualized courtroom interactions. Communication in legal settings
may be coherent and consequential in ways that remain distinct from behavior in other
contexts. And the degree to which a powerful(less) style is distinguishable from other lan
guage elements is uncertain. In this vein, one potential problem of classification was al
luded to earlier: should the phrase “Don’t you agree?” be viewed as powerlessness or as a
rhetorical question?
Page 10 of 19

There is also the question of the regularity with which powerless forms naturally co-oc
cur. Is it true that persons who use hedges are also likely to use hesitations and tag ques
tions? Or to what extent is there a single X construct? This issue is rendered complex by
evidence that different forms produce differences in perceived power in the context of a
job interview (Bradac & Mulac, 1984). To the extent that such findings hold in other con
texts as well, it means that components of powerless language cannot be substituted on a
one-to-one basis. Perhaps six hesitations are required to equal the judgmental impact of
one tag question. This alone does not invalidate the notion of a powerless style, but it
does underscore the challenges of approaching its study from a quantitative standpoint.
Further complications arise from the possibility that different forms influence different
aspects of the persuasion process (Holtgraves & Lasky, 1999). To this point, tag questions
can enhance or depress argument processing depending on source credibility (Blanken
ship & Craig, 2007) and the relevance of the topic to the individual (Blankenship & Holt
graves, 2005; Hosman & Siltanen, 2011). In contrast, hesitations seem to operate as pe
ripheral cues, exerting their influence primarily when involvement with the message is
low (Blankenship & Holtgraves, 2005). However, hedges produced negative evaluations
of the message irrespective of involvement (Blankenship & Holtgraves, 2005). This com
plicated pattern of nonparallel effects indicates a lack of homogeneity among the compo
nents alleged to comprise a powerful(less) style of speaking. The implication is that per
suasion research will be more likely to generate durable knowledge by focusing on the
powerful(less) language at the granular level rather than assuming the element consti
tutes a thematic style.
Domineering Language
Experimental studies frequently operationalize domineeringness as commands or orders

(e.g., “You must” or “You will”) versus formally polite (e.g., “Please”) or qualified terms
(e.g., “Perhaps” or “You may wish to”). As the examples suggest, domineering messages
are often explicit about what action is wanted from the target. But they are more than
just explicit (Miller, Lane, Deatrick, Young, & Potts, 2007). They convey something of the
speaker’s intensity about the issue and they imply a relationship in which the hearer is
subordinate. Both themes—speaker intensity and speaker–hearer relationship—are mani
fest in the theories that have been brought to bear on domineering language:
• Reactance theory emphasizes the control aspect of domineering language (Brehm &
Brehm, 1981). It supposes that individuals have the freedom to choose their own posi
tions and that domineering language causes a perceived threat to freedom.
• Politeness theory highlights the relational aspect of language (Brown & Levinson,
1978). All social actors are presumed to possess a desire for autonomy. What is termed
domineeringness (p. 184) here would be labeled a threat to negative face by Brown and
Levinson, that is, language that threatens the hearer’s desire for autonomy.
Page 11 of 19

Because these two theories orient toward different outcomes, they cannot be viewed as
competing positions and, thus, cannot be discriminated on the basis of data. But the ef
fect of domineeringness is not controversial: it typically produces a counterpersuasive ef
fect (Quick, Shen, & Dillard, 2013; Rains, 2013). This sometimes manifests as no persua
sion relative to some nondomineering control group or as a boomerang. But both cases
can be represented as X → Y → Z, where Y is an amalgam of anger and negative cogni
tions (Quick et al., 2013). One colorful application of domineering language can be seen
in Pennebaker and Sanders’s (1976) research in which they posted two differently worded
signs on college bathroom walls. One read “Do not write on these walls under any circum
stances” whereas the other read “Please don’t write on these walls.” Two weeks later, the
walls with the domineering message had substantially more graffiti on them. Rains (2013)
meta-analysis offers empirical support for the general nature of the effect, as well as for
the cognitive and emotional processes that underlie it.
Recent research on domineering language in the context of environmental messages sug

gests a potentially important side condition. In a series of studies that tapped the intensi
ty of participants’ pro-environmental position, Kronrod, Grinstein, and Wathieu (2012)
showed the typical counterpersuasive effect of strong language among persons whose en
vironmental stance was middle-of-the-road. However, domineering language (“You
must…”) proved effective at spurring pro-environment action among the subset of audi
ence members whose preexisting attitude toward the environmentalism was favorable.
Some earlier projects can be seen having as reached precisely the opposite conclusion.
For instance, when Bensley and Wu (1991) exposed light and heavy drinkers to a domi
neering antialcohol message, it was the heavy drinkers who reacted against the advocacy
by subsequently consuming more alcohol. Jointly, these two projects suggest more ex
treme and possibly opposed responses to domineering language as a function of message
recipients’ initial position.
Summary of Language Features and Persuasion

As with many scientific literatures, the research on language and persuasion is uneven.
For some areas, it is possible to generate relatively strong generalizations. Gain/loss
framing and domineering language are two such instances. For both, plausible theory
coupled with sizeable bodies of data has yielded a solid although by no means final under
standing of how these linguistic variations function to shape persuasion.
Other areas of inquiry offer less certainty. Although the findings for conclusion explicit
ness and self-referencing seem relatively clear, the (small) size of their respective data
bases means that the confidence intervals around our knowledge are large. For
powerful(less) language, there remain many unanswered questions about the coherence
of the elements as a style. Studies of the impact of specific elements such as hedges, hesi
tations, and tag questions are sufficient to stimulate additional theorizing and inquiry.
The picture is gloomier with regard to vividness. Although intuitively compelling, the con
Page 12 of 19

cept fails to meet the basic standards of construct explication, and the resulting literature
appears impervious to synthesis.
Future Directions in the Study of Language

and Persuasion
Developing research on language and persuasion may benefit from attention to several is
sues that emerged in the course of this review.
Messages Features Versus Message Effects
It is an error in scientific inquiry to conflate a cause with its effect. Illustration of the
message effects paradigm as X → Y → Z makes is graphically clear that treating X as Y is
unlikely to produce informative results. Yet this definitional misstep is altogether too com
mon. It was seen most clearly in the corpus of research on vividness, in which message
features were commingled with relational concepts that were further blended with mes
sage effects. The outcome is a literature that embodies creativity but is strongly resistant
to scientific synthesis. Future research should be heedful of the fundamental distinction
between message features and message effects (O’Keefe, 2003).
Style Is a Layered Concept
This chapter distinguished between granular and thematic features of language. Granular
features are more microscopic than thematic features, and their implementation is more
mechanical. Thematic features may be seen as regnant concepts that organize a coherent
set of granular elements. They might be considered styles, whereas granular features are
(p. 185) elements of style. Although the distinction is rudimentary, it underlies a broader
point, which is that style itself is a layered concept. Research at the lower levels of style
is more readily implemented because granular features are governed by simpler rules
and because there is less distance between concept and operationalization. Future re
search may delve into the questions of how to identify layers of style and how to illumi
nate the hierarchical relationships among them.
All Language Features Are Relational
If one takes as a premise that all communication implies a speaker–hearer relationship,

then language is merely a special case of that general principle. Buttressed by explicit as
sumptions in politeness and reactance theories, the notion of domineering language is
that one person is elevated relative to another. Similarly, powerful(less) style is predicat
ed on differences in status or dominance. Although less obvious, it can be said that all of
the language elements reviewed in this chapter have some implication for the speaker–
hearer bond. Even self-referencing calls for the message recipient to adopt a point of
Page 13 of 19

view: to look away from the message source and toward one’s self. Gain-loss framing
bends hearers toward one perspective on an issue and away from the alternatives.
Components of the vividness definition point toward another side of the claim that all lan
guage features are relational. Although relational implications may not be equipotent
across all elements of language, it is not implausible that all research would be better for
having considered the question.
Conclusion
As readers trained in social science will appreciate, our desiderata embody a particular
point of view, one colored by assumptions about the nature of cause, by a desire to identi
fy aspects of language that can be used to create informative experiments, and by the be
lief that close and effortful analysis of concepts may yield insights capable of moving the
ory forward. Indeed, the contents of the entire chapter are predicated on a premise devel
oped by ancient rhetoric: that it is meaningful to conceive of style more or less indepen
dently of other aspects of persuasive messages. Of course, there is an alternative to that
point of view. British author Martin Amis writes that style “is not something grappled on
to regular prose; it is intrinsic to perception.” In other words, content, structure, and
style are indivisible. Although this is not a position likely to aid the scientific study of lan
guage and persuasion, the Amis dictum provides a sounding board against which develop
ing ideas can be tested. To the extent that we can determine which elements of language
can be examined individually and which cannot, the study of language and persuasion
may grow into a literature that is better able to explain the sweet talk that is designed to
bring people over to the truth.
References
Bensley, L. S., & Wu, R. (1991). The role of psychological reactance in drinking following
alcohol prevention messages. Journal of Applied Psychology, 21(13), 1111–1124. doi:
10.1111/j.1559-1816.1991.tb00461.x
Blankenship, K. L., & Craig, T. Y. (2007). Tag questions as powerless speech or as inter
preted context. Journal of Experimental Social Psychology, 43(1), 112–118. doi: 10.1016/
j.jesp.2005.12.012
Blankenship, K. L., & Holtgraves, T. (2005). The role of different markers of linguistic
powerless in persuasion. Journal of Language and Social Psychology, 24(1), 3–24. doi:
10.1177/0261927X04273034
Blum-Kulka, S. (1987). Indirectness and politeness in requests: Same or different? Journal

of Pragmatics, 11, 131–146.
Bradac, J. J., & Mulac, A. (1984). A molecular view of powerful and powerless speech
styles. Communication Monographs, 51(4), 307–319.
Page 14 of 19

Brown, P., & Levinson, S. C. (1978). Universals in language use: Politeness phenomena. In
E. N. Goody (Ed.), Questions and politeness (pp. 56–289). London: Cambridge University
Press.
Brehm, S. S., & Brehm, J. W. (1981). Psychological reactance: A theory of freedom and
control. New York: Academic Press.
Burnkrant, R. E., & Unnava, H. R. (1989). Self-referencing: A strategy for increasing pro
cessing of message content. Personality and Social Psychology Bulletin, 38, 628–638. doi:
10.1177/0146167289154015
Burnkrant, R. E., & Unnava, H. R. (1995). Effects of self-referencing on persuasion. Jour

nal of Consumer Research, 22, 17–26.
Burrell, N. A., & Koper, R. J. (1998). The efficacy of powerful/powerless language on atti
tude and source credibility. In M. Allen & R. W. Preiss (Eds.), Persuasion: Advances
through meta-analysis (pp. 203–215). Cresskill, NJ: Hampton Press.
Chaiken, S., Liberman, A., & Eagly, A. H. (1989). Heuristic and systematic processing
within and beyond the persuasion context. In J. S. Uleman & J. A. Bargh (Eds.), Unintend
ed thought (pp. 212–252). New York: Guilford.
Chapman, H. H. (1957). Persuasion and sweet talk. Today’s Speech, 5(3), 5–5.
Cooper, L. (Ed.). (1932). The rhetoric of Aristotle. New York: Appleton-Century-Crofts.
Cruz, M. (1998). Explicit and implicit conclusions in persuasive messages. In M. Allen &
R. W. Preiss (Eds.), Persuasion: Advances through meta-analysis. Cresskill, NJ: Hampton
Press.
Dragojevic, M., Bell, R. A., & McGlone, M. (2013). Giving radon gas life through lan
guage: Effects of linguistic assignment in health messages about inanimate threats. Jour
nal of Language and Social Psychology. doi: 10.1177/0261927X13495738
(p. 186) Erickson, B., Lind, E. A., Johnson, B. C., & O’Barr, W. M. (1978). Speech style and
impression formation in a court setting: The effects of “powerful” and “powerless”
speech. Journal of Experimental Social Psychology, 14, 266–279.
Frank, J. (1990). You call that a rhetorical question? Forms and functions of rhetorical
questions in conversation. Discourse Processes, 14(5), 723–738. doi:
10.1016/0378-2166(90)90003-V
Frey, K., & Eagly, A. H. (1993). Vividness can undermine the persuasiveness of messages.
Journal of Personality and Social Psychology, 65(1), 32–44.
Gayle, B., Preiss, R., & Allen, M. (1994). The persuasive effect of rhetorical questions. In
M. Allen & R. W. Preiss (Eds.), Prospects and precautions in the use of meta-analysis (pp.
315–337). Dubuque, IA: Brown & Benchmark.
Page 15 of 19

Gerend, M. A., & Maner, J. K. (2011). Fear, anger, fruits, and veggies: Interactive effects
of emotion and message framing on health behavior. Health Psychology, 30, 420–423. doi:
10.1037/a0021981
Guadagno, R. E., Rhoads, K. V. L., & Sagarin, B. J. (2011). Figural vividness and persua
sion: Capturing the “elusive” vividness effect. Personality and Social Psychology Bulletin,
37(5), 626–638. doi: 10.1177/0146167211399585
Hamilton, M. A., Hunter, J. E., & Burgoon, M. (1990). An empirical investigation of an ax
iomatic model of the relationship between language intensity and attitude change. Jour
nal of Language and Social Psychology, 9, 171–195.
Hoffner, C. A., & Ye, J. (2009). Young adults’ responses to news about sunscreen and skin
cancer: The role of framing and social comparison. Health Communication, 24, 189–198.
doi: 10.1080/10410230902804067
Holtgraves, T., & Lasky, B. (1999). Linguistic power and persuasion. Journal of Language
and Social Psychology, 18(2), 196–205. doi: 10.1177/0261927X99018002004
Hong, T. (2012), Internet health search: When process complements goals. Journal of the
American Society for Information Science, 63, 2283–2293. doi: 10.1002/asi.22728
Hosman, L. A., & Siltanen, S. A. (2011). Hedges, tag questions, message processing, and
persuasion. Journal of Language and Social Psychology, 30(3), 341–349. doi:
10.1177/0261927X11407169
Hovland, C., & Mandell, V. (1952). An experimental comparison of conclusion-drawing by

the communicator and the audience. Journal of Abnormal and Social Psychology, 47, 581–
588.
Ilie, C. (1994). What e can I tel you?: A pragmatic study of English rhetorical questions as
discursive and argumentative acts. Stockholm: Almquist & Wiksell International.
Keller, P., & Block, L. (1995). Increasing the persuasiveness of fear appeals: The effect of
arousal and elaboration. Journal of Consumer Research, 22, 448–459.
Kisielius, J., & Sternthal, B. (1986). Examining the vividness controversy: An availability-
valence interpretation. Journal of Consumer Research, 12, 418–431.
Kronrod, A., Grinstein, A., & Wathieu, L. (2012). Go green! Should environmental mes
sages be so assertive? Journal of Marketing, 76, 95–102. doi. 10.1509/jm.10.0416
Mann, T. L., Sherman, D. K., & Updegraff, J. A. (2004). Dispositional motivations and mes
sage framing: A test of the congruency hypothesis in college students. Health Psychology,
23, 330–334. doi: 10.1037/0278-6133.23.3.330
Morrison, M. M., & Shaffer, D. R. (2003). Gender role congruence and self-referencing as
determinants of advertising effectiveness. Sex Roles, 49(5–6), 265–275.
Page 16 of 19

Meyerowitz, B. E., & Chaiken, S. (1987). The effect of message framing on breast self-ex
amination attitudes, intentions, and behavior. Journal of Personality and Social
Meyers-Levy, J., & Peracchio, L. A. (1996). Moderators of the impact of self-reference on

persuasion. Journal of Consumer Research, 22, 408–423.
Miller, C. H., Lane, L. T., Deatrick, L. M., Young, A. M., & Potts, K. A. (2007). Psychologi
cal reactance and promotional health messages: The effects of controlling language, lexi
cal concreteness, and the restoration of freedom. Human Communication Research,
33(2), 219–240. doi: 10.1111/j.1468-2958.2007.00297.x
Munch, J. M., & Swasy, J. L. (1985). Examining the target of receiver elaborations: Rhetor
ical question effects on source processing and persuasion. Journal of Consumer Research,
11(4), 877–886.
Nisbett, R. E., & Ross, L. (1980). Human inference: Strategies and shortcomings of social
judgment. Englewood Cliffs, NJ: Prentice Hall.
O’Barr, W. M. (1982). Linguistic evidence: Language, power, and strategy in the court
room. New York: Academic Press.
O’Keefe, D. J. (1997). Standpoint explicitness and persuasive effect: A meta-analytic re

view of the effects of varying conclusion articulation in persuasive messages. Argumenta
tion and Advocacy, 34(1), 1–12.
O’Keefe, D. J. (2003). Message properties, mediating states, and manipulation checks:

Claims, evidence, and data analysis in experimental persuasive message effects research.
Communication Theory, 13, 251–274. doi: 10.1111/j.1468-2885.2003.tb00292.x
O’Keefe, D. J., & Jensen, J. D. (2006). The advantages of compliance or the disadvantages
of noncompliance? A meta-analytic review of the relative persuasive effectiveness of gain-
framed and loss-framed messages. Communication Yearbook, 30, 1–43.
O’Keefe, D. J., & Jensen, J. D. (2007). The relative persuasiveness of gain-framed and loss-
framed messages for encouraging disease prevention behaviors: A meta-analytic review.
Journal of Health Communication, 12(7), 623–644. doi: 10.1080/10810730701615198
O’Keefe, D. J., & Jensen, J. D. (2009). The relative persuasiveness of gain-framed and loss-
framed messages for encouraging disease detection behaviors: A meta-analytic review.
Journal of Communication, 59(2), 296–313. doi: 10.1111/j.1460-2466.2009.01417.x
O’Keefe, D. J., & Wu, D. (2012). Gain-framed messages do not motivate sun protection: A
meta-analytic review of randomized trials comparing gain-framed and loss-framed ap
peals for promoting skin cancer prevention. International Journal of Environmental Re
search and Public Health, 9, 2121–2133. doi: 10.3390/ijerph9062121
Page 17 of 19

Pennebaker, J. W., & Sanders, D. Y. (1976). American graffiti: Effects of authority and re
actance arousal. Personality and Social Psychology Bulletin, 2, 264–267.
Petty, R. E., & Cacioppo, J. T. (1981). Attitudes and persuasion: Classic and contemporary
approaches. Dubuque, IA: Brown.
Petty, R. E., Cacioppo, J. T., & Heesacker, M. (1981). Effects of rhetorical questions on
persuasion: A cognitive response analysis. Journal of Personality and Social Psychology,
40(3), 432–440.
Quick, B. L., Shen, L., & Dillard, J. P. (2013). Reactance theory and persuasion. In J. P. Dil
lard & L. Shen (Eds.), The SAGE (p. 187) handbook of persuasion (2nd ed., pp. 167–183).
Thousand Oaks, CA: Sage.
Rains, S. (2013). The nature of psychological reactance revisited: A meta-analytic review.

Human Communication Research, 39, 47–73. doi: 10.1111/j.1468-2958.2012.01443.x
Reinhart, A. M., Marshall, H. M., Feeley, T. H., & Tutzauer, F. (2007). The persuasive ef
fects of message framing on organ donation: The mediating role of psychological reac
tance. Communication Monographs, 74, 229–255.
Reyes, R., Thompson, W., & Bower, G. (1980). Judgmental biases resulting from differing
availabilities of arguments. Journal of Personality and Social Psychology, 39, 2–12.
Rothman, A. J., & Updegraff, J. A. (2011). Specifying when and how gain- and loss-framed
messages motivate healthy behavior: An integrated approach. In G. Keren (Ed.), Perspec
tives on framing (pp. 257–278). London, UK: Taylor & Francis.
Roskos-Ewoldsen, D. R. (2003). What is the role of rhetorical questions in persuasion? In

J. Bryant, D. Roskos-Ewoldsen, & J. Cantor (Eds.), Communication and emotion (pp. 297–
321). Mahwah, NJ: Erlbaum.
Scherer, K. R., Schorr, A., & Johnstone, T. (Eds.). (2001). Appraisal processes in emotion:
Theory, methods, research. New York: Oxford University Press.
Shedler, J., & Manis, M. (1986). Can the availability of heuristic explain vividness effects?
Shen, L., & Dillard, J. P. (2007). The influence of behavior inhibition/approach systems
and message framing on the processing of persuasive health messages. Communication
Research, 34, 433–467. doi: 10.1177/0093650207302787
Sherman, D. K., Mann, T. L., & Updegraff, J. A. (2006). Approach/avoidance orientation,

message framing, and health behavior: Understanding the congruency effect. Motivation
and Emotion, 30, 164–168. doi: 10.1007/s11031-006-9001-5
Smith, S. M., & Petty, R. E. (1996). Message framing and persuasion: A message process
ing analysis. Personality and Social Psychology Bulletin, 22, 257–268.
Page 18 of 19

Smith, S. M., & Schaffer, D. R. (2000). Vividness can undermine or enhance message pro
cessing: The moderating role of vividness congruency. Personality and Social Psychology
Bulletin, 26, 769–779.
Sopory, P., & Dillard, J. P. (2002). The persuasive effects of metaphor: A meta-analysis. Hu
man Communication Research, 28, 382–419.
Sparks, J. R., Areni, C. S., & Cox, K. C. (1998). An investigation of the effects of language
style and communication modality on persuasion. Communication Monographs, 65, 108–
125.
Taylor, S. E., & Thompson, S. C. (1982). Stalking the elusive “vividness” effect. Psychologi
cal Review, 89, 155–181.
Tversky, A., & Kahneman, D. (1981 January). The framing of decisions and the psychology
of choice. Science, 211(30), 453–458.
Yan, C., Dillard, J. P., & Shen, F. (2012). Emotion, motivation, and the persuasive effects of
message framing. Journal of Communication, 62(4), 682–700. doi: 10.1111/j.
1460-2466.2012.01655.x
Zillman, D. (1972). Rhetorical elicitation of agreement in persuasion. Journal of Personali

ty and Social Psychology, 21(2), 159–165.
James Price Dillard
James Price Dillard, The Pennsylvania State University
Page 19 of 19

Language and Interpersonal Relationships

Steven W. Duck and Daniel Usera

Subject: Psychology, Social Psychology, Counseling Psychology, Clinical Psychology
This chapter discusses how language is used to constitute relationships, particularly

through the things we say and the things we show by saying them. Additionally, the chap
ter surveys the research in communication studies on six practices of everyday talk that
arise in various kinds of relationships: self-disclosure, metaphors, narrative, taken-for-
granted (TFGs), personal idioms, and intimate play. Four future directions of research are
then suggested.
Keywords: Everyday talk, relational history, speech acts, language use
Pleasant words are the food of love.
– Ovid
*
As scholars of communication and in the spirit of the Roman poet Ovid, we want to claim
that “pleasant words are the food of love,” as he wrote in his book on seduction The Art of
Loving. But we seek to affirm a more general claim upheld by our field, which is that com
munication is essential in all forms of relating, or more eloquently stated, “communica
tion is the lifeblood of relationships” (Knapp & Vangelisti, 2005). More specifically, we
wish to show that words are an expression of a person’s soul, character, personality, value
system, or “rhetorical vision” (Duck, 2011). Of course, we are aware that this claim is not
quite as drastic or perhaps “newsworthy” as one would expect. Researchers have already
established that in many cases of relational dissolution, former partners will cease talking
to each other completely, cutting off the lifeblood of relating (Battaglia, Richard, Datteri,
& Lord, 1998). Culture also restricts the notion of intimacy or personal relationships most
often to people we have personally met and know well through talk, although parasocial
relationships might be an exception (Rubin & McHugh, 1987).
Thus, what is intellectually insightful and what we find personally fascinating about the
general claim that “communication is the lifeblood of relationships” is not so much the
claim itself as much as the reasons and research supporting it. Therefore, in this chapter,
Page 1 of 23

we expound on the reasons that make communication the lifeblood of relationships and,
indeed, that communication “essentializes” relationships (Duck, 1995; Duck & Pond,
1989). We first show the role of language in establishing the social facts of relationships
and then show how it is used rhetorically to say things about the world and ourselves
(Duck & Pond, 1989). Once we have done that, we focus on the practice of everyday talk
to illustrate various linguistic practices that people (p. 189) in relationships of various
kinds use to make language and communication “the lifeblood” of those relationships.
We must agree with social psychologists Krauss and Fussell (1996, p. 65) who noted that
“communication is a complex and multidisciplinary concept and the several disciplines
that use the term have no consensus on exactly how to define it…[D]iscussions typically
pay little attention to the specific mechanisms by which the process works.” In their
analysis of communication, Krauss and Fussell note that many social psychological ap
proaches are essentially cognitive, whereas the processes of communication are neces
sarily social and involve at least the transmission of “something” from one person to an
other. Any use of a symbolic medium presumes shared knowledge (at least the sharing of
the meaning of the symbols). One particularly interesting consequence of this observation
is not only that words do things (Austin, 1962), but so, too, do the relationships upon
which any “sharing” is necessarily founded. Communication and relationships are thus
necessarily intertwined in subtle ways.
In addition to the “sharing” between two individuals that is implied in any relationship, all
relationships are socially contextualized by rules of politeness, such as those articulated
by Brown and Levinson (1987) who stressed the importance of “face” in the interpreta
tion of messages. Although Krauss and Fussell (1996) judged this model to be inadequate
and in need of revision, the social and relational dimensions of conversational context are
clear influences on the ways in which communication is interpreted. However, not all
communication consists of the requests, assertions, and status management that Brown
and Levinson focused on, and our central argument is that everyday mundane communi
cation includes many more elements and facets than research typically accommodates.
Language in Relating
Language and Social Facts
According to Searle (1995), language is what enables societies to establish institutional

and social facts. We are able to make meaningful claims like “Today is Sunday, August 8”
or “President Obama will veto the Senate bill” because we have a system of symbols that
enables us to socially establish these kinds of claims that would not normally be found in
the physical, prelinguistic world. In other words, although a prelinguistic animal may
have a sense of time, no prelinguistic animal would be able to literally think “Today is
Sunday, August 8” because that animal has no conception of language and would not be
able to articulate our 12-month calendar cycle and so forth. Similarly, no prelinguistic
species or societies would be able to have complex designations like “President” or “Se
Page 2 of 23

nate” and their respective balance of powers, including the President’s ability to “veto.”
All of these require symbols and structures that are represented and accepted in shared
language use. Use of such symbols requires coordination and recognition of statuses that
are established socially because they are not found naturally in the physical world—al
though if one ape clubs another and is subsequently treated by others with caution, then
some symbolic representation of power may be said to exist.
Similarly, when we say that Corey and Jaime are “in a relationship,” “hooking up,”
“friends with benefits,” or “just friends,” we are using language to attribute different
kinds of social claims about the state of their interactions and feelings about each other.
Each of these statuses has different expectations from the partner and society, which
makes language or “the label” important, since “labels” imply concordance of under
standing of the meaning of symbols. Thus, when we say that communication is key to re
lationships, we should note that much of “communication” is the use of language, a set of
a socially recognized system of symbols and syntaxes to express ourselves in a way that is
understandable to others. Furthermore, these symbols and syntaxes are ones that are
constitutive of institutional and social facts or claims, including relationships, and depend
on the concordant social acceptance of meanings.
Language as Rhetorical
Moreover, the usage of these labels has rhetorical implications in everyday life as well.
People can use language rhetorically to make arguments, build visions, and describe the
world to others as they see it. In regards to everyday talk, Carl and Duck (2004) state that
“talk is inherently rhetorical in that it implicitly offers a persuasive account of a view and
hence is an effort to attempt to persuade others to support said view of the world or
self” (p. 3). Whenever we describe the world, our experiences, or even make observa
tions, we are implicitly sharing our personal orders, our ways of making sense of the
world (Duck, 2011). As we are making our statements, claims, or rhetorical visions, we
are inviting others to share the viewpoint that we espouse (Duck, 2011). Additionally, the
notable rhetorician (p. 190) Kenneth Burke (1935) said that language often functions as a
“terministic screen.” As he puts it, “Every way of seeing is also a way of not seeing” (p.
70). Whenever we say something is X, we imply that it is not Y or Z. When we say some
thing is a square, we imply that it is not circular. When we say that Mike and John are
friends, we are also saying that they are not enemies or strangers. Thus, whenever we en
gage in talk, whatever utterances we make are bidirectional; we say something about the
world (and, implicitly, about what it is not), which can also say something about our per
sonal orders.
Language as Performative
So far, we have looked at how people use words to refer to things to establish social facts
and to construct realities of the world. However, language can also be used to perform ac
tions, which is what Austin (1962) refers to as “speech acts.” Speech acts are utterances
that have illocutionary force or, in other words, by virtue of being spoken, they are being
Page 3 of 23

done (e.g., “I hereby name this ship The Enterprise,” “This meeting is now adjourned,” or
“I promise to call you tomorrow”). Speech acts are considered to be what Austin (1962)
calls “performative” uses of language.
Performatives are an important part of relating. According to Duck (2011), “the nature of
a relationship requires you to perform it in particular ways that are persuasive and ac
cepted in a given society and play into a social order” (p. 44). One way of “performing a
relationship into being” linguistically is through the use of speech acts (Tracy, 2002).
Many relational events are marked by speech acts, including proposing, negotiating, re
questing, and promising. Many romantic relationships do not begin until a person has
“asked his or her partner out,” or they often do not end until one of the partners in the
dyad “breaks up with his or her partner,” which both normally involve some kind of
speech act of declaring or proposing. All of these speech acts involve language, but more
importantly they establish something, often a status, about the relationship. Moreover,
the speech acts are intelligible only because they take place in a society that gives them
status and function (Searle, 1995).
Moreover, along the lines of speech acts, when we say something, we also implicitly show
something about ourselves as well (Hauser, 2002). Imagine if we wanted to convince you
that we are funny, so we said to you, “We’re funny.” By this quote alone, you may not be
compelled to believe us because anyone can say that they are funny, but not everyone can
actually be funny. So, imagine us instead of saying “We’re funny,” we told you a really
good joke. By telling the joke, not only are we saying something humorous (in content),
but we are also showing that we are funny—or at least have a sense of humor (performa
tively). Similarly with interpersonal communication, when we ask a dear colleague how
his or her day is going, we are not just asking for information about his or her day, we are
also showing (or attempting to show) that we care about him or her. Thus, we do not just
say things with words; we do things with words.
Moreover, with our words, we can use material embodiments to establish our relation
ships. Duck (2011) discussed how people embody their relationships through physical and
material resources. To show that they are married, people usually wear wedding rings,
live together, and go through some sort of ceremonial declaration to legally or socially
recognize their marriage. To show they are friends, people become Facebook friends, buy
each other gifts, or wear friendship bracelets. All of these ceremonies and embodiments
rely on language along with a knowledge of social norms and customs (which also re
quires language) to render themselves intelligible (Duck, 2011).
Recapitulation
So far, we have argued the following: first, we need a verbal, syntactic language that is
meaningful to others in order to recognize and differentiate the kinds of relationships
that our culture finds acceptable or perhaps condemnable. Second, people use language
rhetorically to describe the world and their personal orders of belief. Third, people use
language to perform speech acts that are used in everyday relating and communication.
Page 4 of 23

Finally, people use physical embodiments to represent their relationships that ultimately
depend on the existence of language to render the use of those embodiments intelligible.
We have now given a general framework for the relation between language and interper
sonal relationships. There is huge potential for researchers to extend and clarify this con
ceptual relationship. Part of the enterprise of communication research is looking for pat
terns, commonalities, theories about how people talk to each other. For the remainder of
this chapter, we discuss studies in interpersonal communication that clarify the connec
tion of communication to relationships and resist a view that relationships are simply re
liant on “feelings.” Our argument is that, as people get to know each other, (p. 191) their
communication changes and takes on different features according to the type of relation
ship they establish. In particular, we choose to focus on the neglected role of everyday
talk because it is fundamental to all kinds of interpersonal communication and relating,
and yet, somewhat like the elephant in the room, it has not received enough direct atten
tion (Duck, 2011).
Three Functions of Everyday Talk

All appearances in research studies to the contrary, not all talk is special, significant, or
dramatic. Although self-disclosure has been studied inside and out, it in fact occurs only
2% of the time in everyday life (Dindia, 2000). In the search to find the extravagant re
sults that come from studying extravagant behavior, researchers have too readily over
looked the significance of the mundane.
One of the most fundamental practices of relating is everyday talk: the pedestrian, daily
interactions that we have with others. These interactions are not ones that are typically
picked out to be special relational events like “breaking up,” requesting a date, or offer
ing “social support.” It is the latter sorts of events that are overwhelmingly used by schol
ars to clarify the operation of “typical” relationships, usually with the caveat that the re
sults have “limitations” in their applications. Although such events may be important,
Carl and Duck (2004) argue that “It is in short, everyday communication within a relation
ship, which is seemingly unimportant, that gives the relationship its form and its life and
so provides the recognizable continuity to the experience of individuals as they continue
to believe day to day that they are ‘in’ a relationship” (p. 10).
Talking with one another in our daily lives is one form of symbolism about the status of a
relationship and what it means to be “in a relationship” (Duck, 2011). Three symbolic
functions of such everyday talk have been identified (Duck & Pond, 1989): instrumental,
indexical, and essentializing functions.
Instrumental Functions of Talk in Relationships
Instrumental functions obtain certain ends, whether they are material, physical, or rela
tional. In terms of material ends, a person may text his or her roommate to pick up some
groceries on the way home from work. People in “friends-with-benefits” relationships, al
Page 5 of 23

though talking also as friends, will instrumentally plan their next sexual encounter (Hugh
es, Morrison, & Asada, 2005). Relational ends are uses of everyday talk to develop the re
lationship directly, such as requesting a date.
These instrumental uses are symbolic indications of how one views the relationship. In
other words, the fact that a person does ask for material goods, requests the next sexual
encounter, or increases frequency of contact implicitly says something about that
person’s view of the relationship, as does the partner’s response. In Colombia, if a friend
refuses to do a favor for another (or act as a palanca—leverage/patron), then that refusal
can jeopardize the entire relationship because the amount of confianza (bondedness) be
tween the relational partners is undercut by the refusal (Fitch, 1998).
In the case of physical acts, maintaining daily contact and/or asking for sex are symbolic
of differing types of relationship. According to many researchers, the continued contact
between the partners is what distinguishes “friends with benefits” relationship from “one-
night stands” (Bradshaw, Kahn, & Saville, 2010; Grello, Welsh, & Harper, 2006; Lambert,
Kahn, & Apple, 2003; Manning, Giordano, & Longmore, 2006).
Finally, everyday talk is used to change or maintain the relationship. Sprecher and Duck
(1993) showed that the transition from a first date to a second date was based on an ex
pressed preference for a second meeting and increased frequency of conversational con
tact. Contrariwise, those who want to end a relationship initially symbolically “distance”
themselves by decreasing the frequency of talk (Baxter & Philpott, 1982; Cody, 1982).
Indexical Functions
The next symbolic use of everyday talk, its indexical function, indicates the state and na
ture of the relationship. Goldsmith and Baxter (1996) created a taxonomy of speech
events in everyday talk, and the five most commonly reported nonpurposive speech
events across different relationship types by participants were gossip, joking around,
catching up, small talk, and recapping the day’s events. All five of these events are ones
that are indexical in that they indicate the relationship’s current status. The topology re
veals that it is common for close relational partners to tease each other, role play, and
come up with inside jokes or personal idioms. The fact that partners feel that they can do
these things in a playful manner is indexical of their closeness and relational type.
Essentializing
The last symbolic function of everyday talk is called “essentializing” (Duck, 2011). Essen
tializing (p. 192) is how a person psychologically makes sense of all of the interactions
with a particular person to classify or embody the form of relationship and communicates
that embodiment. Thus, a person uses general, everyday talk with a partner as a source
of evidence that they are in a certain kind of relationship and whether it is one that is
healthy or not.
Page 6 of 23

This sort of function can be illustrated in the process of relational dissolution. Hess
(2002) found that people use both behavioral and cognitive tactics to create distance in a
relationship. The cognitive tactics he found include derogating the other person (thinking
the person as less capable, less important), disregarding messages (dismissing the
person’s communication as unimportant), and cognitive and emotional detachment (re
ducing emotions for the person and seeing each other as different people). All of these
tactics require reframing everyday talk to essentialize a different form of relationship.
Overall, the three symbolic functions of everyday talk are used to constitute and repre
sent a relationship in a specific form through simple, mundane, daily interactions.
Some simple and familiar examples of the basic representation of distance or “immedia
cy” in relationships were long ago pointed out by Brown (1965) in his discussion of the so
cial distances indicated in many languages (e.g., French, Spanish, Italian, German,
Dutch) by the use of the formal, plural, polite, “you” (dubbed the V-form from vous in
French) instead of the informal, singular, close (or patronizing) “you” (dubbed the T form
from tu in French, used to address family, friends, social inferiors, and children). Both
French (Tutoyer) and German (Dusagen) have verbs that grant permission to someone
else to use the more intimate form of language—namely, the singular you, rather than the
plural you; in English, this would be something akin to a person addressed as “Professor
Armstrong,” saying, “No, please call me Marc.”
Informality and inclusiveness are also indicated by pronoun use. To use the inclusive “we”
rather than “you and I” is to indicate acceptance and intimacy (Duck, 2008; Fitsimons &
Kay, 2004). Moreover, essentialization may be related to having a sense of cognitive inter
dependence, in which a partner centralizes the relationship in his or her life and uses
more plural pronouns as a representation of his or her identity; these have all been found
to be associated with romantic commitment (Agnew, Van Lange, Rusbult, & Langston,
1998). Such small everyday essentializing of relationships is also often accompanied by
social and language accommodation of the kind first discussed by Giles (1978). In these
cases, a higher status person can indicate acceptance of a lower status one by “converg
ing” in language and adopting the linguistic styles and tones of the lower status person.
Likewise, hostility to others can be indicated by “divergence” or deliberately adopting a
different style and tone from the one used by the other person.
Six Practices of Everyday Talk

The content, or subject matter, of everyday conversation changes in various relational
contexts. Relational partners tend to develop their own form of idiosyncratic communica
tion, which is often referred to as “relational culture” or “relational codes” (Fitch, 1998).
Although there are many avenues that we can explore, here, we specifically address six
particular discursive practices that arise in relating in everyday life and that are associat
ed with these private dialects or idioms. These six practices are self-disclosure,
metaphors, narrative, taken-for-granteds (TFG)s, idioms, and intimate play. We discuss
each in the following subsections.
Page 7 of 23

Self-Disclosure
As relational partners begin to get to know each other, they develop a sense of relational
history (Duck, 2002). Relational history involves the inferences a person makes about his
or her partner based on previous events between the two of them and future aspirations
for their relationship. As Duck (2002) puts it:
Yet many actions and comments are interpreted—that is, gain their personal
meaning—from both references to present or past experience with a speaker and
expectation (inferences) about future interaction with that person in a relation
ship. Talk does not dangle from the previous biography of the locutors because it
is guided also by expectation. (p. 50)
Thus, we have to consider how much knowledge each partner has about the other to truly
understand their interactions because we show later in this chapter that linguistic phe
nomena such as TFGs, idioms, and intimate play requires some sense of shared history if
it is to take place successfully.
A major way that partners learn about each other and develop a sense of sharedness com
municatively is through self-disclosure, which has been seen as “any information about
himself [or herself] which Person A communicates verbally to Person B” (Cozby, 1973, p.
73). There have been many (p. 193) studies throughout the years on self-disclosure con
cerning a variety of contexts such as social phobias (Sparrevohn & Rapee, 2009), interra
cial friendships (Shelton, Trail, West, & Bergsieker, 2010), adolescent friendships
(Bauminger, Finzi-Dottan, Chason, & Har-Even, 2008), and other contexts. Additionally,
there are strands of self-disclosure research that have investigated self-disclosure in on
line contexts such as Facebook (Ledbetter et al., 2011; Mazer, Murphy, & Simonds, 2007;
Park, Jin, & Jin, 2011; Walther, Van Der Heide, Hamel, & Shulman, 2009; Zhao, Gras
much, & Martin, 2008) and online dating (Ellison, Heino, & Gibbs, 2006; Gibbs, Ellison, &
Heino, 2006; Gibbs, Ellison, & Lai, 2011; Yurchisin, Watchravesringkan, & McCabe,
2005). Self-disclosure has been such a widely studied phenomenon that Altman and Tay
lor (1973) have proposed a theory—social penetration theory—that argues that relation
ships advance as a whole through stages based on changes in verbal communication sur
rounding self-disclosure.
Additionally, according to social penetration theory (Altman & Taylor, 1973), not only does
the content of self-disclosure change throughout the life course of a relationship, but how
people disclose changes as well. As people get to know each other, their interactions be
come more synchronized, which means that they are able to predict each others’ behav
iors and keep a conversation going with much more ease. Also, people are said to be
more careful in their choice of words and modes of self-expression around strangers than
around close relational partners (Altman & Taylor, 1973). In an experiment conducted by
Planalp and Benson (1992), participants were asked to listen to 36 prerecorded conversa
tions and determine whether those conversations had taken place between friends or
strangers. Planalp and Benson found that the participants were 80% accurate in deter
mining the relationship between the two people in the conversation. When asked how
Page 8 of 23

they judged, participants referred to a broad range of 21 different kinds of cues such as
mutual knowledge, relaxation, telling of stories, and laughter. Additionally, Hornstein
(1985) did a study of recorded phone conversations in three kinds of dyads: friends, ac
quaintances, and strangers. She found that friends asked each other more questions,
used more minimal response techniques to keep the conversation going (“mmhmm,”
“yeah,” etc.), and had more complicated closings (which involved some negotiation and
implicit strategies to indicate that the conversation was ending) compared to both the
stranger and acquaintance groups. Thus, self-disclosure not only involves people sharing
information about themselves with each other, it also involves creating a possibility for a
special style of communication that results from a level of familiarity with each other.
Nevertheless, although there have been many studies concerning the practice of verbal
self-disclosure, Dindia (2000) found that self-disclosure (as previously defined) occurs in
only about 2% of everyday interactions. However, Duck (2002) argues that self-disclosure
may not always occur through verbally and directly stating things about one self, but
from a partner observing and inferring things from a person’s behavior. As Goffman
(1959) said about impression management, there are signals that we give and there are
signals we “give off.” If we see self-disclosure not only in the sense of person A stating
things, but also of people “leaking” things about themselves and partners drawing infer
ences, then we can see that communication is not just about what we say, but about what
we do as well. Communication is not just about an intended message, but about the infer
ences that others draw from any behavior or its absence.
Metaphors
As relational partners get to know each other, they often begin using connotative linguis
tic devices. Hauser (2002) argues that we use language not just to point to things, but to
also create visions or represent reality. One linguistic device that people use for such
rhetorical ends is metaphors. Metaphors are words or phrases with their own meanings
that are used to connect one term with another sphere of discussion (as when we talk
about “steering the ship of state” and hence connect government to nautical maneuver
ing, although the classically literate will know that the term “government” itself derives
from the Latin for a naval steersman—gubernator). Thus, Kovesces (2003) argued that
language does not just help us categorize feelings, but it also enables us to constitute
emotional experiences by reference to other familiar concepts that are more easily under
stood (e.g., “I was burning with love,” “I was head over heels with adoration,” “our rela
tionship was cold and distant,” etc.). Similarly, metaphors can also help people to repre
sent complex relational experiences, suggesting stories and archetypical frameworks for
making sense of events. Indeed, metaphors provide habitual frameworks for couples’ ac
counts of their history (“I was stifled in the relationship”), their (p. 194) present state (“it’s
a fun ride”), and their future (“we will always stand by one another”). We focus here on
the cases of developing and dissolving relationships. (See also chapters in this volume by
Kaschak and by Giora.)
Page 9 of 23

Baxter (1992b) did a study of basic root metaphors used by couples in developing roman
tic relationships and found that the three most common usages are referred to work (“you
have to be willing to make the effort for the relationship to work”), a journey of discovery
(“a pathway toward marriage”), and an uncontrollable force that partners had to respond
to (“we just fell in love”). Other less often used metaphors include relationship develop
ment as danger (“fiery relationships”), as an organism (“we became one”), as an econom
ic exchange (“we are equals”), and as a game (“it was fun”). Additionally, Heino, Ellison,
and Gibbs (2010) investigated how the market/shopping metaphor is salient with online
dating. Many online daters saw online dating profiles as a “resume” to “market one’s
self” and that it was like “shopping out of a catalog.” One implication drawn from this
study was that the use of the marketing metaphor to characterize the online dating expe
rience led participants to justify their behaviors such as “increasing their own value” and
being aware of “exaggerated resume-like profiles,” which also affected how and who they
communicated with online. Thus, if relationship development in general is thought of as
“a game” (for example) in some cases, as Baxter’s (1992b) study found, such a characteri
zation may provide justification or influence as to how people communicate in courtship.
On the other hand, the view of online dating as somehow dishonorable, desperate, or
mercantile might increase people’s usage of trading metaphors to describe their behavior
in that environment.
In the case of dissolution, Owen (1993) found that the three main metaphorical themes
used were up/down (“the relationship was bumpy”), organic growth/deterioration (“the
relationship was dying, but we’re trying to save it.”), and presence/absence (“I felt like
there was something missing”). Each of these themes had subcategories as well. More
importantly, each metaphor has consequences and a rhetorical framework for its use. For
example, if a woman tells her partner that they are “falling apart,” she is implying that
she and her partner had “built something together” and that their relationship needed
“repair.” This way of characterizing a certain rhetorical situation frames the fitting re
sponse that she is calling for, which in this case is some kind of action on the partner’s
behalf to “fix” their relationship.
In fact, the metaphors we use may be influenced by the audience we are speaking to, as
Duck (2011) points out, because we may have differing rhetorical expectations for each
audience with which we discuss relationships. A public figure who wants to avoid disclos
ing the possibility of going through divorce might characterize his or her difficult mar
riage to the public as “a private matter involving a rough patch right now,” but tell close
friends that the relationship is “going down the tubes.”
The power and significance of metaphors have been explored in cross-cultural contexts
(Aksan & Kantar, 2008; Bayles, 2008; Dunn, 2004), education (Campbell, Parr, & Richard
son, 2009; Jensen, 2006; Peter, Ng, & Thomas, 2011), health communication (Angus &
Mio, 2011; Casarett et al., 2010), and social cognition (Haslam, Loughman, & Sun, 2011;
Landau, Meier, & Keefer, 2010; Ritchie, 2010). In all of these cases, metaphors give peo
Page 10 of 23

ple a tool to make sense of something, whether it is a relationship or how they construct
their identities as professionals. They are nontrivial aspects of communication.
Narrative
The use of narrative is another important but overlooked element of relationships. Narra
tive is typically conceived of as the telling of stories, although they can take many forms
(Riessman, 2008). Fisher (1985) argues that “a significant feature of compelling stories is
that they provide a rationale for decision and action. As such, they not only constrain be
havior, they may also determine it” (p. 364). Narratives have been studied in all sorts of
contexts including education (Stuhlman & Pianta, 2001), health (Carmack, 2010; Thorn
ton & Novak, 2010), and media studies (Hoebeke, Deprez, & Raeymaeckers, 2011; Sch
neider, Lang, Shin, & Bradley, 2004; Wise, Bolls, Myers, & Sterndori, 2009).
In terms of relationships, Custer, Holmberg, Blair, and Orbuch (2008) argue that how peo
ple tell stories of their relationship origination (including performance and content as
pects of stories) suggests how people use narratives to make meaning of their relation
ships to themselves and others. An important study of married couples attesting to the
importance of narrative in relating was conducted by Holmberg, Orbuch, and Veroff
(2004), which found that how married couples told their stories about their relationship
was associated with their overall relational well-being. In other words, there (p. 195) was
an association between how happy husbands and wives were and the kinds of stories they
told about their marriage.
Not only do narratives have impact on current relationships, they are also important in
accounts to others about past dissolved relationships. Boals and Klein (2005) found differ
ences in the use of emotion words when people talked about the dissolution of their rela
tionship versus the times when they were dating. Additionally, Felmlee (1995) studied “fa
tal attractions” by asking participants to list qualities that attracted them to a partner
whom they had recently broken up with and the qualities that they “least liked” about the
person. She found in many cases that a quality could be listed as attractive but also as a
reason for ending the relationship. Thus, people may reinterpret the “attractive” and “un
attractive” qualities of a partner as they go through time so that a characteristic does not
have its own properties independent of the observer and the time of observation.
Narratives are not only useful for sense-making, however. Tracy (2002) argues that the
act of telling narratives can also do relational work. If a person tells a story to another,
the storyteller implies that the listener is someone who would find this story worth hear
ing and that the story is relationally appropriate. Thus, if a spouse tells his or her beloved
the story of his or her abuse as a child, by the mere act of telling the story, the spouse is
suggesting that the beloved would find the story interesting and that the beloved has es
tablished a kind of relationship with the spouse to be able to hear such a story. Additional
ly, Koenig-Kellas and Trees (2006) discuss the practice of joint storytelling, in which two
or more people narrate a story together for an audience. Having the ability to jointly nar
rate a story together seems to imply a relational history and perhaps a kind of relational
Page 11 of 23

closeness, depending on the context of the story. Thus, narratives have both psychological
and interactional features that make them a prominent practice of everyday talk. (See
chapter by Pasupathi in this volume.)
Taken-for-Granteds
Earlier, we discussed indexing as being a function of everyday talk. As relational partners

continue to talk for the sake of talking and the relationship, they naturally develop con
versational shortcuts. In Hornstein’s (1985) telephone study, friends were found to use
nonindicative openings (“Hi, it’s me”), compared to both the stranger and acquaintance
groups. The use of implicit openings is an exemplary result of indexing; one could not call
a stranger, say “Hi, it’s me,” and expect a stranger to know who “me” is. However, in a
close relationship, this is possible because the partners in the dyad have a shared back
ground and recognize each other’s voice and identities. In fact, it may even be awkward if
you were to call a close friend and introduce yourself using your first and last name, al
though it may not be awkward with a stranger.
A result of indexing over time is demonstrated by TFGs. Hopper (1981a; 1981b) argues
that TFGs are particular utterances or experiences, within a given context or relation
ship, that do not require further explanation. They are often seen as taken-for-granted
presuppositions, enthymemes (Carl & Duck, 2004), or pragmatic implications (Bates,
1976). Although Hopper (1981a) argues that it is difficult to isolate a concrete list of TFGs
in a given interaction, people in relationships build interaction histories together to the
point at which a person can make utterances about certain topics without having to ex
plain any basic premises or descriptions with the partner.
An exemplary study of the use of TFGs is from Maynard and Zimmerman (1984), where
they analyzed the conversations between “acquainted” and “unacquainted” dyads who
were told to “get to know each other” before doing a debate in a laboratory setting. They
found that acquainted dyads were able to initiate “topicalized” conversation (in-depth dis
cussion about a certain topic) through the pertinent use of TFGs. Acquainted dyads would
initiate topicalized conversations through opening statements like “Joe came by the other
night.” In each of these cases, a certain topic is understood by both parties, even though
an outsider may have no idea who Joe was or where he went by last night. In both of
these cases, the topics and their background information are taken-for-granted by the
dyads.
Unacquainted dyads, on the other hand, used a conversational process called pretopical
sequencing. This process involves the use questions and answers for interlocutors to gen
erate some biographical information about each other in hopes of promoting more topi
calized conversation. For unacquainted dyads to initiate topicalized conversation, they of
ten first used questions such as, “What’s your major?,” “Where do you live?,” and others.
Depending on the response, interlocutors used different methods to attempt to reach top
icalized conversation, which Maynard and Zimmerman (1984) spell out in their article.
Our main point here is that because unacquainted dyads lacked much opportunity to
Page 12 of 23

have TFGs, they had to gather information about each other directly, which is why
(p. 196)
such direct questions had to be asked. Because the acquainted pairs already had enough
biographical information, they did not have to ask “Where are you from?” and instead
could focus on making statements like “I went to Jason’s Deli when I was in your area.”
Personal Idioms
Another feature of everyday talk that arises in various types of relationships is the use of
personal idioms, which are utterances that carry unique meanings to a particular rela
tionship. Hopper, Knapp, and Scott (1981) found that personal idioms developed through
partners’ shared history. Some of these sources were through linguistic distortions, per
sonal characteristics, memorable events, and external sources like songs or movies. When
one of the participants refers to his wife as “Futtbutt,” there is a story or history behind
the use of that word, which is the husband claiming that his wife tends to complain about
the size of her buttocks. Many of these types of expressions are akin to what Duck (2002)
calls hypertexts. Hypertexts in everyday talk can be likened to hypertexts on web pages:
they are utterances that have an apparent meaning, but if you “click on them” (or ask
what a person means by them), you see that they have more layers of meaning and histo
ry behind them.
In any case, Hopper et al. (1981) identified eight categories of idioms used by married
couples. Those categories are partner nicknames, expressions of affection (e.g., saying
“so much” instead of “I love you”), labels or nicknames for others outside of the relation
ship, confrontations (saying “jelly beans” when speaking over one’s head), requests and
routines (saying “What’s new?” when out with friends to indicate “I’m ready to leave”),
sex references and euphemisms (labels for different sexual organs), sexual invitations
(“Let’s go home and ‘watch TV’”), and teasing insults. An interesting finding by Hopper et
al. (1981) was that the couples reported that these idioms had an overall positive effect
on the relationship, even when the idioms were used to tease or bring displeasure. Bell,
Buerkel-Rothfuss, and Gore (1987) found that the number of idioms used was positively
related to the amount of commitment, love, and closeness in premarital relationships. Bell
and Healy (1992) also found the use of idioms to be related to relational solidarity. Dun
leavy and Booth-Butterfield (2009) found that couples who are de-escalating (declining in
intimacy) report using fewer idioms with less frequency, and the idioms used tend to be
more hurtful ones.
Moreover, Bruess and Pearson (1993) found that married couples in later stages of life
(child-bearing, mid-life, and retired couples) recalled using a lower number of idioms than
newlyweds. Although it could be an explanation of poorer memories, it does seem to sug
gest that idiom use is at least more salient in beginning marriages. Moreover, Hopper
(1981a) found that certain idioms “die out” (stop being used) for various reasons such as
situation changes, pleaded displeasure with the idiom by the partner, and the need for va
riety. This suggests that as couples go through time, they develop a conversational histo
Page 13 of 23

ry that enables them to develop new idioms, have stories of current language practices,
or declare some terms as “dead.”
Along with personal idioms, some researchers have identified “baby talk,” which is the
“relatively high-pitched and melodious way of speaking” that tends to be found in con
texts of care taking with small children and infants, also known as “motherese,” “par
entese,” and “infant-directed speak” (Bombar & Littig, 1996, p. 137). They found that ba
by talk is associated with intimacy, which indicates clearly that what and how people talk
to each other changes as they interact over time and therefore should be considered in
studies of relationships by psychologists and others.
Intimate Play
Along a similar vein, Baxter (1992a) did a study on “intimate play,” which is the less seri
ous, everyday interactions that relational partners have. In the study, Baxter (1992a)
found eight types of intimate play among friends and romantic partners including private
verbal codes (idioms), role playing (mocking a person or character outside of the relation
ship for fun), verbal teasing, prosocial physical play (e.g., pretending to waltz when see
ing each other), antisocial physical play (e.g., mock physical fighting), games (activities
with rules), gossiping, and public performances (e.g., pretending to fight to observe “the
reaction” of observers). Like Hopper et al. (1981), Baxter’s study found the use of idioms
in everyday talk, but went a step further in finding nonverbal forms of expression that are
unique to relating. Indeed, intimate play has been found to be a predictor of couple satis
faction and stability (Vanderbleek, Robinson, Casado-Kehoe, & Young, 2011) and positive
ly associated with positive emotion and relational satisfaction (Aune & Wong, 2002). Col
lege men and women report that they experience teasing (a form of intimate (p. 197) play)
on average once or twice a week and that teasing was done to show liking, fun, bonding,
and flirting (Becker et al., 2007). The key point for psychologists is that not all “communi
cation” is serious, yet these playful forms define the character of relationships and should
be taken into consideration in future studies.
It is not normal in American culture to go up to a stranger and pretend that you are about
to fight him or her, but in certain relationships (often male friendships) this is a form of
play. Glenn and Knapp (1987) observed couples engaged in intimate play and inquired as
to how they know when they are at play. Their study found that although couples often
use nonverbal signals like smiles or grins to indicate that they are playing, there were
some instances when none of these was displayed, and TFGs about the person’s behavior,
the setting, or past interactions were what signaled it. As Goffman (1974) argued, part
ners develop ways to “key” a particular moment of play; that is say, they learn to signal
that they are engaging in a kind of interaction not to be taken literally. Thus, it is possible
to see the roles of TFGs and indexing in the development of these types of linguistic prac
tices within relationship cultures.
Given these studies, we should frame the use of idioms—and all other forms of intimate
play—as a way that communication evolves, from simply the usage of words into symbolic
Page 14 of 23

action or performance of friendship, romance, or whatever relationship type the parties

perceive that they have. Communication is key to the performance of friendships, which
are not simply fixed forms of behavior derived from the internal psychological character
istics or feelings of the two partners.
Conclusion
Putting this all together, we argue that language is key to interpersonal relationships be
cause language is necessary to establish the social facts of relationships through its refer
ential and performative functions. Moreover, research on everyday talk suggests that at
least six discursive practices arise in relationships (some of which are tied to relational
well-being), which suggests that language and interpersonal communication are constitu
tive of relationships. We show how self-disclosure is used by partners to share their per
sonal orders explicitly and implicitly. Over time, partners begin to make sense of their pri
or interactions relationally by drawing on common metaphors and developing narratives
or stories of their relationship’s past, present, and future. As they develop a sense of
shared history, TFGs, personal idioms, and intimate play arise, and the use of these dis
cursive practices becomes a possible indicator of the state of the relationship. Hence, the
various types of relationships that exist in our culture today are possible only because of
language, which makes communication “the lifeblood of relationships” (Knapp & Vange
listi, 2005). Therefore, future relational researchers could benefit by seeking answers in
the directions suggested in the next section.
Future Directions
The first future direction for understanding the role of language and communication in in
terpersonal relationships is to begin to challenge the idea of “relationships as containers”
and to attempt a larger understanding of how culture and discourse help us make sense
of our relationships to ourselves and others. Some scholars already do this, but the path
is lightly trod. We should begin to understand in depth how the cultural use of idioms,
narratives, and media sources influence individuals in their relating to others (Duck &
McMahan, 2012). All of these cultural devices are based in language. Understanding how
these devices are incorporated into everyday relating will give us desperately needed
clarification of the role of everyday talk and the use of language in general in the continu
ation and development of relationships.
Second, researchers must attempt to understand the role of the mundane in everyday
talk. We believe that small talk (or what we really see as not-so-small talk) is not only
what keeps personal relationships afloat (Tracy, 2002; VanderVoort & Duck, 2000), but
that such contexts are also valuable in understanding everyday language use, whether it
is the use of yet-to-be-classified flirtation styles (Hall, Carter, Cody, & Albright, 2010), the
act of social support in a more mundane fashion (Badr, Acitelli, Duck, & Walter, 2001), or
the telling of mundane deceptions (DePaulo, Kashy, Kirkendol, Wyer, & Epstein, 1996).
We need more studies that investigate the texts of everyday talk, not just from an ethno
Page 15 of 23

graphic perspective (although they have been insightful) but also using all kinds of meth
ods and paradigms.
The third direction for understanding the relationship of language to interpersonal com
munication is to study how social media are changing language use and how people com
municate to each other. Although some scholars have studied the linguistics of texting
language (Choudhurry, Saraf, Jain, Mukherjee, Sarkar, & Basu, 2007), (p. 198) technologi
cal trends in social media are endlessly variable (see chapters in this volume by Kramer
and by Hancock) and will always be used for relational purposes (Duck & McMahan,
2012).
Last, scholars must study other contexts of relationships where intimacy and a sense of
closeness are not present. In this chapter, we focused much of our discussion on friend
ships and romantic relationships. However, future researchers must investigate “enemy
ships” (Wiseman & Duck, 1995), in which two people are in rivalry with or hate each oth
er, especially when they are co-workers, siblings, or share similar friends. How do people
talk to each other in situations in which they are forced to interact? What sort of relation
al culture do people in these sorts of antagonistic or boring contexts develop?
All of these future directions for research will bring us closer to understanding just how
communication is part of our everyday lives and is the lifeblood of our relationships.
References
Agnew, C. R., Van Lange, P. A. M., Rusbult, C. E., & Langston, C. A. (1998). Cognitive in
terdependence: Commitment and the mental representation of close relationships. Jour
nal of Personality and Social Psychology, 74(4), 939–954.
Aksan, Y., & Kantar, D. (2008). No wellness feels better than this sickness: Love
metaphors from a cross-cultural perspective. Metaphor and Symbol, 23, 262–291.
Altman, I., & Taylor, D. A. (1973). Social penetration: The development of interpersonal
relationships. New York: Holt, Rinehart, and Winston.
Angus, L., & Mio, J. (2011). At the “heart of the matter”: Understanding the importance of
emotion-focused metaphors in patient illness narratives. Genre, 44(3), 349–362.
Aune, K., & Wong, N. (2002). Antecedents and consequences of adult play in romantic re
lationships. Personal Relationships, 13, 279–286.
Austin, J. L. (1962). How to do things with words. Cambridge, MA: Harvard University
Press.
Badr, H., Acitelli, L., Duck, S. W., & Walter, C. (2001). Weaving social support and relation
ships together. In B. Sarason, & S. W. Duck (Eds.), Personal relationships: Implications for
clinical and community psychology (pp. 1–14). New York: John Wiley & Sons.
Bates, E. (1976). Language and context. New York: Academic Press.

Page 16 of 23

Battaglia, D., Richard, F., Datteri, D., & Lord, C. (1998). Breaking up is (relatively) easy to
do: A script for the dissolution of close relationships. Journal of Social and Personal Rela
tionships, 18(6), 829–845.
Bauminger, N., Finzi-Dottan, R., Chason, S., & Har-Even, D. (2008). Intimacy in adoles
cent friendship: The roles of attachment, coherence, and self-disclosure. Journal of Social
and Personal Relationships, 25(3), 409-428.
Baxter, L. (1992a). Forms and functions of intimate play in personal relationships. Human
Communication Research, 18(3), 336–363.
Baxter, L. (1992b). Root metaphors in accounts of developing romantic relationships.

Journal of Social and Personal Relationships, 9, 253–275.
Baxter, L., & Philpott, J. (1982). Attribution-based strategies for initiating and terminating
friendships. Communication Quarterly, 30(3), 217–224.
Bayles, B. (2008). Metaphors to cure by: Tojolab’al Maya midwifery and cognition. Anthro
pology and Medicine, 15(3), 227–238.
Becker, S., Clabaugh, S. E., Clark, R. A., Kosovski, M. C., Daar, R., Hefner, V., Kmetz, T.,
McDaniel, S., Miller, L., Moriarty, C., Qian, Z., Raja, S., Ramey, M., & Suri, R. (2007).
Teasing among college men and women. Communication Studies, 58(2), 157–172.
Bell, R., & Healey, J. (1992). Idiomatic communication and intepersonal solidarity among
friends in relational cultures. Human Communication Research, 18(3), 307–335.
Bell, R., Buerkel-Rothfuss, N., & Gore, K. (1987). “Did you bring the yarmulke for the cab
bage patch kid?”: The idiomatic communication of young lovers. Human Communication
Research, 14(1), 47–67.
Boals, A., & Klein, K. (2005). Word use in emotional narratives about failed romantic rela
tionships and subsequent mental health. Journal of Language and Social Psychology, 24,
252–268.
Bombar, M., & Littig, L. (1996). Babytalk as a communication of intimate attachment: An

initial study in adult romances and friendships. Personal Relationships, 3, 137–158.
Bradshaw, C., Kahn, A., & Saville, B. (2010). To hook up or date: Which gender benefits?
Sex Roles, 62, 661–699.
Brown, P. (1965). Social psychology. New York: Free Press.
bridge, UK: Cambridge University Press.
Bruess, C., & Pearson, J. (1993). “Sweet pea” and “pussy cat:” An examination of idiom
use and marital satisfaction over the life cycle. Journal of Social and Personal Relation
ships, 10, 609–615.
Page 17 of 23

Burke, K. (1935). Permanence and change: An anatomy of purpose. Berkeley: University

of California Press.
Campbell, T., Parr, M., & Richardson, C. (2009). From implicit intention to exquisite ex
pression: Finding metaphors for who we are and what we do. Journal of Transformative
Education, 17, 209–229.
Carl, W. J., & Duck, S. W. (2004). How to do things with relationships…and how relation
ships do things with us. Communication Yearbook, 1–35.
Carmack, H. (2010). Bearing witness to the ethics of practice: Storying physicians’ med
ical mistakes narratives. Health Communication, 25(5), 449–458.
Casarett, D., Pickard, A., Fishman, J., Alexander, S., Arnold, R., Pollack, K., et al. (2010).
Can metaphors and analogies improve communication with seriously ill patients? Journal
of Palliative Medicine, 39, 255–260.
Choudhurry, M., Saraf, R., Jain, V., Mukherjee, A., Sarkar, S., & Basu, A. (2007). Investiga
tion and modeling of the structure of texting language. International Journal on Docu
ment Analysis and Recognition, 10, 157–174.
Cody, M. (1982). A typology of disengagement strategies and an examination of the role

intimacy, reactions to inequity, and relational problems play in strategy selection. Commu
nication Monographs, 49, 148–170.
Cozby, P. (1973). Self-disclosure: A literature review. Psychological Bulletin, 79(2), 73–91.
Custer, L., Holmberg, D., Blair, K., & Orbuch, T. L. (2008). “So how did you two meet?:”
Narratives of relationship initiation. (p. 199) In S. Sprecher, A. Wenzel, & J. Harvey (Eds.),
Handbook of relationship initiation (pp. 453–470). New York: Psychology Press.
DePaulo, B., Kashy, D., Kirkendol, S., Wyer, M., & Epstein, J. (1996). Lying in everyday
life. Journal of Personality and Social Psychology, 70(5), 979–995.
Dindia, K. (2000). Self-disclosure, identity, and relationship development: A dialectical

perspective. In K. Dindia, & S. W. Duck (Eds.), Communication and personal relationships
(pp. 147–162). New York: John Wiley & Sons.
Duck, S. W. (1995). “Talking relationships into being.” Journal of Social and Personal Rela
tionships, 12, 535–540.
Duck, S. W. (2002). Hypertext in the key of G: Three types of “history” as influences on

conversational structure and flow. Communication Theory, 12(1), 41–62.
Duck, S. W. (2008). Silver anniversary essay: A past and a future for relationship re
search. Journal of Social and Personal Relationships, 25(1), 189–200.
Duck, S. W. (2011). Rethinking relationships. Thousand Oaks, CA: Sage Publications.
Page 18 of 23

Duck, S. W., & McMahan, D. T. (2012). Communication in everyday life. Thousand Oaks,
CA: Sage Publications.
Duck, S. W. & Pond K. (1989). Friends, Romans, Countrymen; lend me your retrospective
data: Rhetoric and reality in personal relationships. In C. Hendrick (Ed.), Close relation
ships (pp. 17–38). Newbury Park, Sage Publications.
Dunleavy, K., & Booth-Butterfield, M. (2009). Idiomatic communication in the stages of

coming together and coming apart. Communication Quarterly, 57(4), 416–432.
Dunn, C. (2004). Cultural models and metaphors for marriage: An analysis of discourse at
Japanese wedding receptions. Ethos, 32(3), 348–373.
Ellison, N., Heino, R., & Gibbs, J. (2006). Managing impressions online: Self-presentation
processes in the online dating environment. Journal of Computer-Mediated Communica
tion, 11, 415–441.
Felmlee, D. (1995). Fatal attractions: Affection and disaffection in intimate relationships.

Journal of Social and Personal Relationships, 12(2), 295–311.
Fisher, W. (1985). The narrative paradigm: An elaboration. Communication Monographs,

52, 347–367.
Fitch, K. (1998). Speaking relationally: Culture, communication, and interpersonal con

nection. New York: Guilford Press.
Fitzsimons, G. M., & Kay, A. C. (2004). Language and interpersonal cognition: Causal ef
fects of variations in pronoun usage on perceptions of closeness. Personality and Social
Psychology Bulletin, 30, 547–557.
Gibbs, J., Ellison, N., & Heino, R. (2006). Self-presentation in online personals. Communi
cation Research, 33(2), 152–177.
Gibbs, J., Ellison, N., & Lai, C. (2011). First comes love then comes Google: An investiga
tion of uncertainty reduction strategies and self-disclosure in online dating. Communica
tion Research, 38(1), 70–100.
Giles, H. (1978). Linguistic differentiation in ethnic groups. In H. Tajfel (Ed.), Differentia

tion between social groups (pp.361–393). London: Academic Press.
Glenn, P., & Knapp, M. L. (1987). The interactive framing of play in adult conversations.
Communication Quarterly, 35(1), 48–66.
Goffman, E. (1959). The presentation of self in everyday life. New York: Doubleday.
Goffman, E. (1974). Frame analysis: An essay on the organization of experience. New

York: Harper and Row.
Page 19 of 23

Goldsmith, D., & Baxter, L. (1996). Constituting relationships in talk: A taxonomy of

speech events in social and personal relationships. Human Communication Research, 44,
87–114.
Grello, C., Welsh, D., & Harper, M. (2006). No strings attached: The nature of casual sex
in college students. Journal of Sex Research, 43(3), 255–267.
Hall, J., Carter, S., Cody, M., & Albright, J. (2010). Individual differences in the communi
cation of romantic interest: Development of the flirting styles inventory. Communication
Quarterly, 58(4), 365–393.
Haslam, N., Loughnan, S., & Sun, P. (2011). Beastly: What makes animal metaphors offen
sive? Journal of Language and Social Psychology, 13, 311–325.
Hauser, G. (2002). Introduction to rhetorical theory (2nd ed.). Prospect Heights: Waveland
Press.
Heino, R., Ellison, N., & Gibbs, J. (2010). Relation shopping: Investigating the market
metaphor in online dating. Journal of Social and Personal Relationships, 27(4), 427–447.
Hess, J. (2002). Distance regulation in personal relationships: The development of a con

ceptual model and a test of representational validity. Journal of Social and Personal Rela
tionships, 17, 663–683.
Hoebeke, T., Deprez, A., & Raeymaekers, K. (2011). Heroes in the sport pages. Journalism
Studies, 12(5), 658–672.
Holmberg, D., Orbuch, T. L., & Veroff, J. (2004). Thrice told tales: Married couples tell
their stories. Mahwah, NJ: Lawrence Erlbaum Associates.
Hopper, R. (1981a). How to do things without words: The taken- for-granted as speech-ac
tion. Communication Quarterly, 5, 228–236.
Hopper, R. (1981b). The taken-for-granted. Human Communication Research, 7(3), 195–

211.
Hopper, R., Knapp, M. L., & Scott, L. (1981). Couples’ personal idioms: Exploring inti
mate talk. Journal of Communication, 31(1), 23–33.
Hornstein, G. (1985). Intimacy in conversational style as a function of the degree of close

ness between members of a dyad. Journal of Personality and Social Psychology, 49(3),
671–681.
Hughes, M., Morrison, K., & Asada, K. (2005). What’s love got to do with it?: Exploring
the impact of maintenance rules, love attitudes, and network support on friends with ben
efits relationships. Western Journal of Communication, 69(1), 49–66.
Jensen, D. (2006). Metaphors as a bridge to understanding educational and social con

texts. International Journal of Qualitative Methods, 5(1), 36–54.
Page 20 of 23

Knapp, M. L., & Vangelisti, A. (2005). Interpersonal communication and human relation
ships, 5th ed. Boston: Pearson Education.
Koenig Kellas, J. and A. R. Trees (2006). Finding meaning in difficult family experiences:
Sense making and interaction processes during joint family storytelling. Journal of Family
Communication, 6(1), 49–76.
Kovesces, Z. (2003). Metaphor and emotion: Language, culture, and body in human feel
ing. Cambridge, UK: Cambridge University Press.
Krauss, R. M., & Fussell, S. R. (1996). Social psychological models of interpersonal com
munication. In E. T. Higgins & A. Kruglanski (Eds.), Social psychology: Handbook of basic
principles (pp. 655–701). New York: Guilford.
Lambert, T., Kahn, A., & Apple, K. (2003). Pluralistic ignorance and hooking up. Journal of
Sex Research, 40(2), 129–133.
Landau, M. J., Meier, B. P., & Keefer, L. (2010). A metaphor-enriched social cognition. Psy
chological Bulletin, 136(6), 1045–1067.
(p. 200) Ledbetter, A., Mazer, J., DeGroot, J., Meyer, K., Mao, Y., & Swafford, B. (2011). At
titudes toward online social connection and self-disclosure as predictors of Facebook com
munication and relational closeness. Communication Research, 38(27), 27–53.
Manning, W., Giordano, P., & Longmore, M. (2006). Hooking up: The relationship contexts
of “nonrelationship” sex. Journal of Adolescent Research, 21, 459–483.
Maynard, D. W., & Zimmerman, D. H. (1984). Topical talk, ritual and the social organiza
tion of relationships. Social Psychology Quarterly, 47(4), 301–316.
Mazer, J., Murphy, R., & Simonds, C. (2007). I’ll see you on “Facebook”: The effects of
computer-mediated teacher self-disclosure on student motivation affective learning, and
classroom climate. Communication Education, 56(1), 1–17.
Ovid, & May, J. L. (1930). The love books of Ovid being the Amores, Ars Amatoria, Reme
dia Amoris, and Medicamina Faciei Femineae of Publius Ovidius Naso. J. L. May (Trans.)
Whtiefish, MT: Kessinger Publishing.
Owen, W. (1993). Metaphors in accounts of romantic relationship terminations. In P.

Kalbfleisch (Ed.), Interpersonal communication: Evolving interpersonal relationships (pp.
261–278). Hillsdale, NJ: Lawrence Erlbaum Associates.
Park, N., Jin, B., & Jin, S. (2011). Effects of self-disclosure on relational intimacy in Face
book. Computers in Human Behavior, 27, 1974–1983.
Peter, L., Ng, J., & Thomas, K. (2011). How career changers make sense of teaching
through professional metaphors. Action in Teacher Education, 17, 232–245.
Page 21 of 23

Planalp, S., & Benson, A. (1992). Friends’ and acquaintances’ conversations I: Perceived
differences. Journal of Social and Personal Relationships, 9, 483–506.
Riessman, C. (2008). Narrative methods for the human sciences. Thousand Oaks: Sage
Publications.
Ritchie, L. (2010). “Everybody goes down”: Metaphors, stories, and simulations in conver
sations. Metaphor and Symbol, 25, 123–143.
Rubin, R., & McHugh, M. (1987). Development of parasocial interaction relationships.

Journal of Broadcasting & Electronic Media, 31(3), 279–292.
Schneider, E., Lang, A., Shin, M., & Bradley, S. (2004). Death with a story: How story im
pacts emotional, motivational, and physiological responses to first-person shooter video
games. Human Communication Research, 30(3), 361–375.
Searle, J. R. (1995). The construction of social reality. Cambridge, UK: Cambridge Univer
sity Press.
Shelton, J., Trail, T., West, T., & Bergsieker, H. (2010). From strangers to friends: The in
terpersonal process model of intimacy in developing interracial friendships. Journal of So
cial and Personal Relationships, 27(1), 71–90.
Sparrevohn, R., & Rapee, R. (2009). Self-disclosure, emotional expression and intimacy
within romantic relationships of people with social phobia. Behavior Research and Thera
py, 47, 1074–1078.
Sprecher, S., & Duck, S. W. (1993). Sweet talk: The role of communication in consolidat
ing relationship. Personality and Social Psychology Bulletin, 20, 391–400.
Stuhlman, M., & Pianta, R. (2001). Teachers’ narratives about their relationships with
children: Associations with behavior in classrooms. School Psychology Review, 31(2),
148–163.
Thornton, L., & Novak, D. (2010). Storying the temporal nature of emotion work among
volunteers: Bearing witness to the lived traumas of others. Health Communication, 25(5),
437–448.
Tracy, K. (2002). Everyday talk: Building and reflecting identities. New York: Guilford
Press.
Vanderbleek, L., Robinson III, E., Casado-Kehoe, M., & Young, M. (2011). The relationship
between play and couple satisfaction and stability. Family Journal, 19, 132–139.
VanderVoort, L., & Duck, S. W. (2000). Talking about “relationships”: Variations on a

theme. In K. Dindia & S. W. Duck (Eds.), Communication and personal relationships (pp.
1–12). Chichester: John Wiley & Sons.
Page 22 of 23

Walther, J. B., Van Der Heide, B., Hamel, L. M., & Shulman, H. (2009). Self-generated ver
sus other-generated statements and impressions in computer-mediated communication: A
test of warranting theory using Facebook. Communication Research, 36(2), 229–253.
Wise, K., Bolls, P., Myers, J., & Sternadori, M. (2009). When words collide online: How
writing style and video intensity affect cognitive processing of online news. Journal of
Broadcasting & Electronic Media, 53(4), 532–546.
Wiseman, J. P., & Duck, S. W. (1995). Having and managing enemies: A very challenging
relationship. In S. W. Duck, & J. T. Wood (Eds.), Confronting relationship challenges [Un
derstanding relationship processes 5; pp. 43–72). Thousand Oaks, CA: Sage Publications.
Yurchisin, J., Watchravesringkan, K., & McCabe, D. (2005). An exploration of identity re-
creation in the context of internet dating. Social Behavior and Personality, 33(8), 735–
750.
Zhao, S., Grasmuck, S., & Martin, J. (2008). Identity construction on Facebook: Digital
empowerment in anchored relationships. Computers in Human Behavior, 24, 1816–1836.
Notes:
(*) The order of authorship is alphabetical and required by OUP. It does not reflect the rel
ative contributions of the two authors as prescribed by APA guidelines.
Steven W. Duck
Steven W. Duck, Department of Communication Studies, The University of Iowa
Daniel Usera
Daniel A. Usera, Department of Communication Studies, University of Iowa.
Page 23 of 23

Natural Language Use as a Marker of Personality

Molly E. Ireland and Matthias R. Mehl

Natural language has historically been integral to the study of personality. Yet research
on how personality is revealed through language use has only recently gained momen
tum. This chapter reviews research on how different aspects of personality are manifest
ed in the way people use words. The chapter provides the conceptual foundation for re
search on linguistic markers of personality, discusses the psychometric properties of nat
ural word use, and summarizes findings on how the Big Five personality dimensions (ex
traversion, agreeableness, neuroticism, conscientiousness, and openness), trait emotion
ality (trait negativity, trait anger, and trait positivity), and psychopathology-related per
sonality traits (Type A, depression, narcissism, Machiavellianism) are linked to patterns of
word use. With progress in stationary and mobile computing technology and parallel ad
vances in computational linguistics, the field is bound to experience a strong growth over
the next years.
Keywords: Computational linguistics, Big Five, linguistic inquiry and word count, LIWC, word use, text analysis,
linguistic markers
Natural language has been integral to the study of personality since the field’s inception.
Lexical approaches to personality factors analyze the words people use to describe others
in order to zero in on a small number of mostly independent personality traits. Early on,
this approach generated the Big Five model that remains a major force in personality re
search today (Costa & McCrae, 1992; Goldberg, 1981). More holistic narrative approach
es code content themes in the stories that people tell about their lives in order to under
stand how individuals construe their own personality and represent it to others
(McAdams & Pals, 2006).
Despite personality researchers’ early recognition that language and stories might accu
rately describe personality, the idea that natural language use also reflects personality
and might provide knowledge about individual differences above and beyond self-reports
did not gain traction until the past decade. Before the computer revolution, Walter Wein
traub (1981) spent decades amassing data on natural language use—primarily by training
coders to count words and phrases by hand—with little recognition. Although some ac
Page 1 of 33

cepted his premise that verbal behavior, like other actions, reflects psychological process
es, the psychological community was slow to adopt the burdens that came with carrying
out linguistic research on sufficiently large samples by hand (Mehl, 2006a).
The recent surge in the popularity of studying language in the social and behavioral sci
ences stems in large part from technological advances in computational linguistics. The
internet and computer science more generally have made it possible to easily compile and
analyze natural language corpora with word counts climbing into the millions and bil
lions. As of 2010, Google Books had digitized 12 percent of all books ever published and
has (p. 202) made those data available to interested researchers (Michel et al., 2010). So
cial networking and publishing sites like Livejournal and, more recently, Twitter allow re
searchers to download their users’ language use for free or for relatively low costs (e.g.,
Golder & Macy, 2011; Ramírez-Esparza, Chung, Kacewicz, & Pennebaker, 2008; Yarkoni,
2010). Facebook similarly allows researchers to access users’ status updates, although
they tend to have stricter restrictions on usage than other sites (e.g., Kramer, 2010;
Schwartz et al., 2013). The near-universal use of the internet has helped psychology
widen its sampling nets for a wide range of purposes. The rise of the internet has been
particularly fruitful for language research: observing verbal communication or other lan
guage behavior online is simple. Within minutes, online social media users often produce
hundreds of units of objectively and easily quantifiable (i.e., text-based) behavioral data.
Perhaps more importantly, sophisticated text analysis tools allow researchers to analyze
the increasingly large and diverse datasets that technological advances have made possi
ble. One of the first text analysis tools to enter into common usage is a computationally
simple word count program called the Linguistic Inquiry and Word Count (LIWC; Pen
nebaker, Francis, & Booth, 2007). LIWC is a computerized text analysis program that out
puts the percentage of words used in a text or batch of texts that fall into one or more of
more than 80 grammatical (e.g., articles, first-person singular pronouns), psychological
(e.g., positive emotion, insight), and topical (e.g., social, sex) categories. It does so by
comparing each word in a text against a set of internal word lists or dictionaries. Al
though the psychometrically developed content categories have a (minor) subjective com
ponent, the grammatical and linguistic categories are, for the most part, based on objec
tive and factual information about the established lexical members of that category.
LIWC, along with its predecessors the General Inquirer (Stone, Dunphy, Smith, & Ogilvie,
1966) and DICTION (Hart, 1984), focus on word frequencies alone irrespective of con
text. For example, the sentences, “I’ve never been less happy” and “I’m the happiest per
son alive” would be coded as containing identical proportions of positive emotion words
despite the very different meaning of each sentence. Critics have pointed out that, in ex
amples like these, the programs’ context blindness can lead to noisy or difficult to inter
pret results.
The role of word count programs, however, is not to obviate self- and observer reports but
to complement them and to clarify the gaps and inconsistencies they leave behind. A cur
sory reading of the two statements about happiness above or a single-item measure of
positive emotion could tell you that the speakers range from least to most happy, respec
Page 2 of 33

tively. However, much of word count programs’ incremental utility comes from the infor
mation that questionnaires and content coding often miss: programs like LIWC are able
to tell us that the speakers in these examples, despite their differences, are focusing on
happiness to similar degrees. Studies have borne out the intuition that it’s often speak
ers’ focus rather than their conveyed meaning that matters. For example, early expres
sive writing studies showed that people who wrote about past traumas using positive
emotion words tended to benefit more than those who used exclusively negative emotion
words (Pennebaker, 1997). This finding was particularly striking because, due to the trau
matic nature of the expressive writing topics, uses of positive emotion words were fre
quently negated (e.g., “she’ll never forgive me,” “I’m not a good person”).
Measurement and Psychometrics of Natural

Language Use
Before serious and widespread work could begin on the links between natural language
use and personality, two questions needed to be addressed: First, how can we measure
(i.e., analyze) natural language use? And, second, does natural language use fulfill the ba
sic psychometric requirements for a personality or individual difference variable (i.e., is it
moderately stable over time and across context)?
Before computerized text analysis, language analysis in both psychology and linguistics
was by necessity primarily qualitative. Linguists would often subject single utterances or
exchanges to intense scrutiny (see Tanenhaus & Trueswell, 1995). Others would draw in
ferences about human behavior in general based on everyday observations of speech pat
terns of social relationships (e.g., Lakoff, 1975). Psychologists who sought psychometri
cally tractable linguistic data tended to use content coding methods in which trained
judges rated texts according to a set of criteria that typically focused on writers’ or
speakers’ use of content themes, such as anxiety, hope, and health or sickness
(Gottschalk & Gleser, 1969). However, even the most reliable and statistically sound of
these methods were slow, labor-intensive, and subject to human error.
A major benefit of word count approaches is that their reliability is never compromised by
(p. 203) subjective biases or experimenter error. Computer programs output the same re
sults regardless of the mental state of the experimenter using them, and they will always
find exactly how many times a certain word or group of words occurs in a given text re
gardless of how easily overlooked those words might be (e.g., the notoriously invisible
of1). Indeed, the context-blindness of word count programs is beneficial in this respect.
Whereas a human coder might be bogged down by shades of meaning or biased assump
tions about a speaker or writer, computerized text analysis programs focus single-mind
edly on what language alone can tell us.
Once programs were available that could rapidly and reliably identify language patterns
in large samples of texts, it was possible to establish the basic psychometrics of natural
language use. One of the first investigations into the reliability of language use was con
Page 3 of 33

ducted by Gleser and colleagues (1959). In that study, participants told a personal story in
monologue for about 5 minutes. Transcripts were divided into two equal halves and cod
ed for several linguistic and psychological language categories, such as adjectives and
feelings. The average correlation between the two halves of these stories across all cate
gories was moderate to high, approximately r = .50. In other words, as anyone who has
either written or read a story could tell you, there is some variation in stories over time.
Authors use different words when laying out the setting of a story than when describing
the climax (Malin et al., 2014). Yet, despite these obvious differences, the linguistic fin
gerprint of an author or speaker tends to remain visible over the course of a narrative.
Similarly, individuals’ spoken language use tends to remain consistent during hour-long
life history interviews, with consistency coefficients (Cronbach’s α) ranging from .41 to .
64 for several stylistic and content language categories, despite stark differences in inter
viewer questions between the first and second halves (Fast & Funder, 2008).
Several studies have now demonstrated that natural language use evidences substantial
consistency not only over the course of a narrative, but over time as well. Schnurr et al.
(1986) asked medical students to describe their experiences coming to medical school in
two unscripted monologues spaced one week apart. They found that language use, as an
alyzed with the General Inquirer (Stone et al., 1966), was highly reliable over time.
Across 83 content categories measured in that study, including references to people,
work, affect, and evaluations, the average correlation between the two monologues was .
78. Later, around the time that computerized text analysis approaches were beginning to
gain mainstream momentum, Pennebaker and King (1999) further tested the limits of lin
guistic stability by comparing individuals’ language use over longer stretches of time,
lasting up to several years, and across diverse contexts, ranging from scientific articles to
students’ stream-of-consciousness essays. They once again found, using word frequencies
calculated by LIWC, that people maintain good linguistic consistency in most language
categories—often despite predictable situational fluctuations in language use.
The same consistency across time and place has been found for naturalistic, everyday
spoken language as well. In one of the first studies using the Electronically Activated
Recorder (EAR; Mehl, Pennebaker, Crow, Dabbs, & Price, 2001) methodology, Mehl and
Pennebaker (2003) recorded college students in their natural environments over the
course of two 2-day periods spaced 4 weeks apart. The EAR sampled 30 seconds of ambi
ent sounds roughly every 12 minutes. All captured talking was transcribed and coded for
participants’ location, activity, and mode of conversation (i.e., telephone or face-to-face).
With few exceptions, both linguistic and psychological categories were substantially cor
related across time, activity, and interaction mode. Interestingly, the authors also found
that function word categories, including grammatical parts of speech such as pronouns
and articles, were more consistent than were content-based psychological categories (av
erage function word r = .41; average psychological processes r = .24).
Taken together, this past research establishes that people’s natural language use is char
acterized by a good degree of temporal stability and cross-situational consistency. There
fore, the ways in which people spontaneously use language—for example, idiosyncrasies
Page 4 of 33

in word choices or speaking styles—satisfy psychometric requirements for personality or

individual difference variables.
Language Style Versus Language Content

Early analyses of personality and language tended to base their conclusions on tran
scripts that had been coded for content words and phrases (e.g., Smith, 1992). Indeed, in
computational linguistics, the function words that make up language style continue to be
referred to as “stop words” because they are usually ignored during automated language
processing. However, individual differences in (p. 204) language style are often more psy
chologically telling and psychometrically parsimonious than are differences in language
content. Language style is defined by a person’s use of function words, including pro
nouns, articles, conjunctions, and several other categories that make up the grammatical
structure of utterances (Table 13.1). Function words tend to be short, frequently used,
and have little meaning outside the context of a conversation. In part because of these
characteristics, they are processed fluently and largely automatically during both lan
guage production and comprehension (Bell, Brenier, Gregory, Girand, & Jurafsky, 2009;
Levelt, 1989). Language content, on the other hand, is defined by a person’s use of nouns,
verbs, adjectives, and most adverbs. In short, content words determine what a person
says, and function words determine how they say it.
Page 5 of 33

Table 13.1 Function word categories and examples
Category LIWC Label Examples
First-person singular pronouns I I, me, my
Third-person singular pronouns shehe she, his, her
Second-person singular pro you you, y’all, yours

nouns
Third-person singular pronouns we we, our, us
Third-person plural pronouns they they, their, them
Impersonal pronouns ipron it, those, that
Articles article a, an, the
Conjunctions conj and, but, be

cause
Prepositions preps in, under, about
Auxiliary verbs auxverb shall, be, was
High-frequency adverbs adverb quite, highly,

very
Negations negate no, not, never
Quantifiers quant much, few, lots
Only basic-level and not superordinate function word categories were

included. Categories are from LIWC2007 (Pennebaker et al., 2007).
To a greater degree than function words, content words are practically constrained by the
topic or context of conversation. For example, group members assigned to solve math
problems together will uniformly use content words related to that task (e.g., calculate,
solution) regardless of whether they are each individually thinking about the problems in
different ways. Function words, on the other hand, are more loosely constrained by the
topic of a conversation, allowing people to discuss the same content in different styles.
The versatility of function words allows researchers to measure differences in language
style across contexts rapidly and objectively. Despite some degree of natural verbal and
Page 6 of 33

nonverbal convergence between individuals during social interaction (Chartrand & van
Baaren, 2009; Ireland & Pennebaker, 2010; Pickering & Garrod, 2004), function words
used during conversation reliably reflect differences in social status, honesty, and leader
ship styles (Hancock, Curry, Goorha, & Woodworth, 2008; Pennebaker, 2011; Slatcher,
Chung, Pennebaker, & Stone, 2007; see Tausczik & Pennebaker, 2010, for a review).
A second reason that researchers might focus on language style instead of or in addition
to content is that function words tend to more directly reflect social cognition during con
versation. The relationship between function word use and social cognition is primarily a
practical matter: because function words have little meaning outside of the context of a
sentence, they require common ground or shared social knowledge to be interpreted
(Pennebaker, Mehl, & Niederhoffer, 2003). For example, to understand the sentence He
shut the dog in there, the speaker must know that the listener shares his knowledge of
the man, the dog, and the location in question. This mutual understanding of a situation,
(p. 205) its potential referents, and each conversation partner’s knowledge of the situation
is known as common ground and theoretically forms the foundation of any successful con
versation (Clark & Brennan, 1991). Given that interest in and attention to others’
thoughts and feelings is an integral aspect of personality (e.g., particularly Big Five extra
version and agreeableness), the ability to automatically extract individuals’ social cogni
tive styles from their language use could be a valuable addition to personality re
searchers’ toolboxes.
A final reason for paying attention to style in addition to content is purely psychometric.
Individual differences in language observed in early language research focused primarily
on phrases that include both language content and style. To use women as an example,
female speakers tend to use more uncertainty phrases, such as I wonder if, and extra-po
lite phrases, such as would you mind, than male speakers do (Holmes, 1995; Lakoff, 1975;
Poole, 1979; Rubin & Green, 1992; see Brown & Levinson, 1987). However, many of these
phrases can be measured more simply by counting the function words that they common
ly include—specifically, in the previous examples, first-person singular pronouns (e.g., I)
and auxiliary verbs (e.g., would). Indeed, an analysis of formal writing in the British Na
tional Corpus found that function words offered the most efficient way to classify texts as
authored by men or women (Koppel, Argamon, & Shimoni, 2003). Similarly, a corpus
analysis of spoken and written language collected in 70 studies revealed that function
words more reliably discriminated between male and female participants than did con
tent words (Newman, Groom, Handelman, & Pennebaker, 2008). In personality research
specifically, direct comparisons are relatively sparse. However, style appears to provide
the best classification accuracy for neuroticism, providing gains over content alone and
even over content and style combined (Argamon, Koppel, Pennebaker, & Schler, 2009).
Whether similar effects will be found for other personality traits and individual differ
ences has yet to be conclusively determined.
In the end, whether language content or style is a more reliable indicator of personality
and individual differences may depend to a large degree on what personality measures
are used to establish criterion validity. Although individual differences such as age and
Page 7 of 33

sex appear to be more strongly linked to language style, personality traits as measured by
Big Five scales are often more consistently and strongly linked with language content—
including both language categories and individual words—than language style. This may
be because research exploring the link between language use and the Big Five has over
whelmingly used personality self-reports as the gold standard of true personality, where
as demographic variables can be more objectively observed.
The pattern of personality self-reports matching language content and demographic indi
vidual differences matching language style may be due to a match between the auto
maticity of the behavior and the measurement (Eastwick, Eagly, Finkel, & Johnson, 2011).
For example, a neurotic person who realizes that neuroticism is socially undesirable may
—deliberately or not—project a cheerful exterior by using positive emotion words and by
downplaying his anxiety in self-report questionnaires. However, less accessible compo
nents of language use, such as increased use of self-references like I and me, may corre
late with less accessible behavioral indicators of worry such as compulsively checking the
status of a loved one’s flight or spending extensive time on WebMD. Given the abundance
of online language use—including emails, blogs, and online chats, which are often
archived by default (see Baddeley, 2011)—and the fact that many nonverbal behaviors can
be accessed simply by downloading browser histories, future personality research may be
able to incorporate naturalistic measures of individuals’ online behavior to triangulate
when and where language style and content and behavioral and self-reported personality
converge.
The Big Five Personality Domain

The literature on language and the Big Five is the largest of the subareas within the study
of personality and natural language. To accommodate its size, the following sections first
summarize the samples commonly used in this research and next address major findings
for each Big Five dimension individually.
Language Samples
The kind of naturalistic language that has perhaps most frequently been subjected to
computerized text analysis and linked with the Big Five is online or computerized lan
guage use. In the roughly 20 years since the internet was made accessible to the general
public, language has become the most accessible naturalistic behavior available to behav
ioral scientists. As the next sections will explore as well, everyday verbal behavior is car
ried out online and often automatically saved in blogs, social networking sites, email ac
counts, online chats, and text (p. 206) messages. More formal texts abound as well, includ
ing a huge range of academic submissions, ranging from admissions essays to published
scholarly work, not to mention nearly a fifth of the fictional novels, poetry collections, and
nonfictional books published in recorded history (see Michel et al., 2010).
Page 8 of 33

Considering that this goldmine of information is often free and accessible to anyone with
the necessary web programming or copying-and-pasting skills, it is surprising that only a
few studies linking the Big Five and the kind of quasi-naturalistic language use that oc
curs in these formats have been conducted. The studies that have been conducted show
great promise, however, for both understanding naturalistic manifestations of personality
and for the longstanding goal of automatically building personality profiles based on be
havioral data (Dodds & Danforth, 2010; Mairesse & Walker, 2011; Mairesse, Walker,
Mehl, & Moore, 2007).
A few studies have gone to the effort of collecting spoken language as it occurs in real
life. These studies were made possible by the advent of the EAR about a decade ago
(Mehl et al., 2001). The EAR is a programmable audio recorder that periodically records
snippets of ambient sounds (e.g., 30 seconds every 12 minutes). When the EAR records, it
captures any surrounding noise—including language used by subjects in their daily inter
actions with their social networks. Later, trained transcribers listen to the recordings,
type the language they hear, and typically also code for basic features of subjects’ mo
mentary social environments (e.g., location, activity).
Within studies that have looked at laboratory writing or dialog tasks, language use large
ly falls into two categories: tasks with face-valid relevance to personality, such as asking
people to talk about events that were important in shaping their identity, and those that
attempt a more circumspect route, such as asking students to describe an object (e.g., a
water bottle; Pennebaker, 2011). Not surprisingly, considering that the criterion for per
sonality is nearly always responses to face-valid self-report scales, language used in the
former tasks tends to correlate more strongly with personality dimensions. For example,
although Pennebaker and King found only a small number of modestly significant correla
tions (rs <.20) between self-reported Big Five traits and language used in stream-of-con
sciousness writing and essays about coming to college, Hirsh and Peterson (2008) and
Fast and Funder (2008) found a large number of moderate correlations (rs = .20–.40) be
tween self-reported personality and language used in separate tasks that asked partici
pants to describe their life stories.
Extraversion
Among the Big Five, extraversion tends to leave some of the strongest and most pre
dictable traces in individuals’ online and physical environments (Gosling, 2008) and, to a
slightly lesser degree, language (Mairesse & Walker, 2006). People who are rated by oth
ers and themselves as higher in extraversion use less inhibited (e.g., careful, avoid), ten
tative (e.g., doubt, maybe), and self-focused language; more positive emotion words (e.g.,
adorable, nice); are more talkative; and talk more about social topics, such as friends,
people, communication, and leisure activities (Augustine, Mehl, & Larsen, 2011; Dewaele
& Furnham, 2000; Fast & Funder, 2008; Mairesse & Walker, 2006; Mehl et al., 2006;
Oberlander & Gill, 2006; Qiu, Lin, Ramsay, & Yang, 2012; Yarkoni, 2010). Word- and
phrase-level analyses of large corpora made up of Facebook status updates (Kosinski &
Stillwell, 2012) and blog entries (Yarkoni, 2010) have found that party, bar, can’t wait are
Page 9 of 33

among the best indicators of high extraversion, and internet, computer, and cats are good
indicators of low extraversion.
In terms of language style, extraverts use more immediate first-person plural and second-
person pronouns such as we and you (i.e., pronouns used primarily to talk with rather
than about a person; Dewaele & Furnham, 2000; Holtgraves, 2011; Yarkoni, 2010). Al
though extraverts do not appear to use first-person singular pronouns at different rates
than their introverted counterparts overall (Mehl et al., 2006; Pennebaker & King, 1999;
Yarkoni, 2010), some evidence suggests that the link between extraversion and I is mod
erated by the words that first-person singular co-occurs with. A study that counted pairs
of co-occurring words rather than individual words found that extraverts used some
I-phrases more frequently and others less frequently than introverts do, leading to null
first-person singular correlations overall. For example, Gill and Oberlander (2002) found
that people asked to write emails to a close friend used a greater variety of I-phrases and
more bigrams containing negations (I don’t, I’m not) to the degree that they reported be
ing relatively introverted. In the same study, those higher in extraversion limited them
selves to a small number of less negative and perhaps implicitly more social first-person
phrases such as I’ll be and I was.
(p. 207) Agreeableness
A clear and intuitive indicator of agreeableness is linguistic positivity. Agreeable people

use more positive language, including both verbs (e.g., laughing) and modifiers (e.g., love
ly) and fewer negative emotion words (e.g., damn, jerk) in everyday speech and writing
(Augustine et al., 2011; Holtgraves, 2011; Küfner, Back, Nestler, & Egloff, 2010; Yarkoni,
2010). They have also been found to use more social words (Küfner et al., 2010) and self-
references (Mehl et al., 2006), which, in the context of generally cheerful language, may
suggest polite hedge phrases such as I think rather than the ruminative self-references
that characterize negative affective traits (see the section on Depression). First-person
singular also signifies lower social status relative to other conversation partners
(Kacewicz, Pennebaker, Davis, Jeon, & Graesser, 2013), further suggesting that I-words
used by agreeable individuals reflect polite self-effacement rather than neurotic self-con
sciousness (see Holtgraves, 2010, for a review of polite language use). Consistent with
this interpretation, one of the best classifiers of high agreeableness in Facebook status
updates is the phrase thank you (Schwartz et al., 2013).
A facet-level analysis of agreeableness demonstrated that it may be one of the most cohe
sive five-factor dimensions, linguistically and otherwise: Across most or all of its five
facets, people who rank higher in agreeableness talk more about home, family, and com
munication and avoid sensitive topics such as death (e.g., coffin, killer; Yarkoni, 2010).
Another face-valid linguistic correlate of agreeableness is swearing. Unsurprisingly, peo

ple swear less to the degree that they report being more agreeable (Holtgraves, 2010;
Mehl et al., 2006; Yarkoni, 2010; see Jay, 2009). Indeed, the five words that best discrimi
nate between individuals ranking high and low in agreeableness in Facebook status up
Page 10 of 33

dates are all swear words (Schwartz et al., 2013). The negative correlation between
agreeableness and swearing fits with lay theories of personality as well and is correctly
interpreted by outside observers of students’ EAR recordings (Mehl et al., 2006). Given
the low overall incidence of swearing—making up about 1/3 of a percent of spoken con
versation and 1/10 of a percent of emotional writing—these results essentially mean that
highly agreeable people are unlikely to swear even once in a given sample, whereas a
highly disagreeable person might swear only a few times. Nevertheless, swearing, like
negative emotion words, another low-frequency category, is a potent and reliable indica
tor of agreeableness and other key psychological variables (e.g., Robbins et al., 2011).
Neuroticism
Low emotional stability, or neuroticism, is characterized by negative thoughts, anxiety,

and ruminative self-focus (Teasdale & Green, 2004). The language use of individuals who
rate themselves as high in neuroticism reflects these characteristics in higher rates of
I-words, negations, and references to sadness, anger, and anxiety in written life stories,
bereavement narratives, stream-of-consciousness essays (in both Korean and English),
blogs, and text messages (Argamon et al., 2009; Baddeley & Singer, 2008; Holtgraves,
2011; Hirsh & Peterson, 2009 Lee, Kim, Seo, & Chung, 2007; Mairesse et al., 2007; Pen
nebaker & King, 1999; Qiu et al., 2012; Yarkoni, 2010). In Facebook status updates, the
words basketball and success were the best classifiers of emotionally stable individuals,
whereas the single best classifier of high neuroticism was the swear word fucking, closely
followed by variants of depression and the phrases I hate and sick of (Schwartz et al.,
2013).
In contrast with the majority of research cited earlier, everyday spoken language shows
very few signs of neuroticism (Mehl et al., 2006). The reason behind the discrepancy be
tween these sets of findings may lie in the extent to which individuals feel pressure to be
have in socially desirable or appropriate ways (Mehl & Holleran, 2008). Like depression,
neuroticism is known to be a socially undesirable trait, and the expression of negative
emotion is typically considered a private event. As such, individuals who are aware that
they tend to be dysphoric and anxious may mask their negative emotions from many of
the individuals they interact with directly in their daily lives. Thus, like depression, disclo
sure of negative emotions may turn out to be moderated by closeness with one’s interac
tion partners (Baddeley, 2011; see the section on Depression). In anonymous experimen
tal writing, during which showing obvious signs of neuroticism could not possibly matter,
or text messages and Facebook statuses, which are likely to be read by friends rather
than casual acquaintances or passersby, neurotic individuals may feel free to express
their chronic negativity (Holtgraves, 2011; Schwartz et al., 2013). Consistent with this in
terpretation, Mehl and Holleran (2008) found that private but not public negative emotion
word usage reflects increased neuroticism. As a result of this moderation, when language
is aggregated across conversation contexts, (p. 208) ranging from public conversations
with professors to private conversations with relationship partners, the typical markers of
neuroticism may be critically attenuated, resulting in surprisingly modest effect sizes for
Page 11 of 33

intuitively valid markers of neuroticism, such as anxiety words (Pennebaker & King,
1999).
Openness
The trait of openness is an interesting case that strongly recommends the inclusion of
facet-level traits in research on language and personality. First, openness is relatively dif
ficult to capture linguistically. Analyses of naturalistic spoken and text message conversa
tions have produced a number of null findings and a handful of significant correlations
that fail to paint a cohesive or theory-consistent picture (Holtgraves, 2011; Mehl et al.,
2006).
When meaningful patterns do emerge, the linguistic indicators of openness seem to re
flect only its intellectual aspects, ignoring facets related to artistic expression and emo
tionality. Bloggers whose self-reports indicated a higher degree of openness used more
articles and prepositions and fewer personal pronouns and references to family and home
(Yarkoni, 2010). On Facebook, highly open individuals are distinguished by more frequent
uses of the words universe, writing, and music (Schwartz et al., 2013). In life history in
terviews, people who ranked higher in openness in general and intellectualism specifical
ly used more articles as well (Fast & Funder, 2008), and in a Korean stream-of-conscious
ness writing sample, people who were higher in openness produced more sentences and
talked about sleeping and resting less (Lee, Kim, Seo, & Chung, 2007). In other words, at
least when talking about themselves or their interests at length, people who are high in
openness seemed to adopt a formal rather than narrative writing style. This distinction is
captured by Biber’s (1995) involved-informative dimension of language use, which de
scribes a high rate of verbs and pronouns at the involved end of the spectrum and nouns
(implied by the use of articles) and prepositions at the informative end. Informative lan
guage tends to be more characteristic of male language use, scripted speech, formal writ
ing, and, in the last US presidential elections, liberal political candidates (Koppel et al.,
2003; Pennebaker, 2011).
Intellectualism and liberalism may be salient and central characteristics of openness, but
they are hardly necessary or sufficient for a person to rank relatively highly on a broad
five-factor measure of openness. The other facets of openness concern interest in art,
emotionality, adventurousness, and imagination. In the only study that examined the rela
tionship between language and the finer-grained facets of the Big Five, Yarkoni (2010)
found that several of the language categories that predicted emotionality and artistic in
terest were at odds with those that predicted the remaining facets. People who are emo
tionally and artistically open use more personal pronouns, references to physical states,
positive emotions, and words related to leisure or rest; liberal, intellectual, imaginative,
and adventurous bloggers used fewer of each category. In studies that reported language
correlates of only the primary Big Five dimensions, this moderation by facet is likely to
have led to null results for categories that show full crossover effects between facets and
weakened results for categories that only occur for one or a few facets. For example, in
the Yarkoni (2010) sample, the modest negative correlation between references to home
Page 12 of 33

and total openness scores was diluted by the finding that references to home are unrelat
ed to emotionality and artistic interests.
Conscientiousness
Like agreeableness, conscientiousness appears to manifest in everyday language use as

polite speech. In blogs, as well as in students’ EAR recordings and stream-of-conscious
ness essays, conscientious individuals swear less and use fewer negative emotion words
regardless of their sex (Lee et al., 2007; Mairesse & Walker, 2006; Mehl at al., 2006; Pen
nebaker & King, 1999; Yarkoni, 2010). At the level of individual words as well, conscien
tiousness is best defined by the words that people high on the trait do not use. For the
facets of achievement striving and self-discipline in particular, almost every one of the top
10 strongest single-word predictors were negatively correlated (e.g., protest, boring), ex
cepting only a few positively correlated words, including ready and HR (human resources)
(Yarkoni, 2010). Schwartz and colleagues’ (2013) word- and phrase-level Facebook analy
sis corroborates these findings: Although highly conscientious people use the phrases
ready for, to work, and great day more than less conscientious people do, the strongest
correlates of conscientiousness were negative, including several swear words as well as
the words YouTube and bored. The high number of negative correlations suggests that,
perhaps especially for those high in achievement striving and self-discipline facets of con
scientiousness, increased conscientiousness is associated with regulation or inhibition of
those words that serve as indicators of individuality for other traits. That behavior pat
tern (p. 209) would be consistent with findings that people who rank higher in conscien
tiousness tend to value conformity more than less conscientious individuals do (Roccas,
Sagiv, Schwartz, & Knafo, 2002).
Surprisingly, women used second-person pronouns, or you, much less to the extent that
they were conscientious, whereas the correlation was the opposite for men—a pattern
that observers listening to participants’ daily recordings accurately (although modestly)
observed and used in their personality ratings (Mehl et al., 2006). Like many pronouns,
you is a versatile word. However, it most notably predicts hostility and is associated with
both state and trait anger (Simmons, Chambless, & Gordon, 2005; Weintraub, 1981). It
may be the case that, in the context of daily social interactions with friends and
strangers, conscientious people strive to follow societal gender norms that define men as
assertive (more you) and women as nonconfrontational (less you).
Moderation by Sex
Sex2 appears to significantly moderate the strength of behavioral signals of personality in

some cases and reverse them in others. Moderation appears to be particularly likely in
cases in which word categories have relatively low base rates and are taboo or socially
sensitive (characteristics that are obviously not unrelated). In an interesting study that
was designed to capture correlations between Big Five traits and swearing, a taboo lin
guistic dimension that is especially elusive in more controlled written language use, Fast
and Funder (2008) had participants complete a life history interview as well as a sponta
Page 13 of 33

neous face-to-face conversation with two strangers also in the experiment. When acquain
tances’ ratings of participants’ personality (i.e., informant reports) were compared with
swearing across these two natural dialog tasks, men but not women who swore were rat
ed by informants as more extraverted. (In that study, self-report personality measures
were not collected.) Mehl et al. (2006) found a similar moderation by sex: untrained
judges who listened to EAR recordings of students’ daily lives rated men but not women
as more extraverted when they swore. In that study, self-reported extraversion was unre
lated to swearing for either sex. In other words, in contrast with individuals’ self-ratings,
judges tend to interpret swearing as a sign of aggression or assertiveness that, for men
but not women, might be used as a heuristic for extraversion. Indeed, in Mehl et al.
(2006), men who argued more during their EAR recordings were also seen by judges—but
not by themselves—as more extraverted.
As mentioned earlier, the display of content-based indicators of neuroticism, such as neg

ative emotion words, is likely moderated by whether language use is public and identifi
able or private and anonymous. Consistent with the trait’s characteristic negative emo
tionality, worry, and tendency to overreport health symptoms (Watson & Pennebaker,
1989), women who are more neurotic use more negative emotion words and references to
health in naturalistic text messages (Holtgraves, 2011). However, men who are more neu
rotic not only fail to use more negative emotion words, but they also fail to use any lan
guage category that was examined more or less than those who are more emotionally sta
ble. Mehl et al.’s (2006) EAR recordings showed a similar effect: whereas for women, neu
roticism was reflected in decreased verbosity and laughter and increased arguing, men’s
neuroticism was reflected in more socializing and time spent outdoors. These results may
be due to gender differences in emotional display rules. It may be that the traditional so
cial sanction against male expressions of strong emotion causes men to regulate their
emotional language (even in private text messages) and, as an alternative, cope with the
negative emotions of neuroticism through social and physical activity to a greater degree
than women do.
Trait Emotionality
Presumably due to the strong presence of the Big Five approach in personality research,
trait positivity and negativity has received relatively little attention in language research.
Yet there are a few studies that suggest both face-valid and nonobvious relationships be
tween trait emotionality and natural language. Despite the apparent importance of the
trait, however, those studies that have investigated language and trait affect have not ex
amined its relationship with and stylistic aspects of natural language such as pronoun us
age. Rather the aim tends to be on improving the ability to unobtrusively gauge individu
als’ affect through automated text analysis (e.g., Cohen, Minor, Najolia, & Hong, 2009;
Dodds & Danforth, 2010).
Page 14 of 33

Trait Negativity
Unlike more specific measures of depression or neuroticism, general psychological dis

tress appears to be strongly related to more references to negative emotions and fewer
references to positive emotions. This is what Cohen (2011) found in a sample of students
who were asked to speak in monologue about a recent disagreement. In this (p. 210)
study, the author used custom positive and negative emotion dictionaries that excluded
positive emotion words with more common non-emotional meanings such as like and pret
ty. Psychological distress in that sample was unrelated to positive or negative emotion
word usage for LIWC, although LIWC’s negative emotion category was positively correlat
ed with depression symptoms in the same sample (Cohen, 2011). Part of the success of
this study in accurately tracking negative affect in dialogue may be due to the nature of
the task, which likely afforded greater use of emotional language than more neutral
monologues. The study also suggests, however, that LIWC, the most widely used text
analysis program in personality research, has room to improve its emotion categories
(O’Carroll Bantum & Owen, 2009).
Trait Anger
Walter Weintraub (1981), a pioneer in the study of personality and language, found in a
case study of a young man with an explosively angry temperament that angry speech is
characterized by a high rate of second-person pronouns (e.g., you, you’re). A follow-up
analysis of simulated anger in spontaneous monologues produced by two trained actors
replicated the second-person singular finding and further found that angry speech con
tained fewer uses of we, more uses of me, more swear words, and more negations (e.g.,
no, not). (Coming before the computer revolution, Weintraub and collaborators counted
pronouns by hand.) The parallels between the chronically angry young man and the ac
tors asked to simulate anger suggest that these linguistic characteristics reflect both
state and trait anger. Despite the small sample sizes used in Weintraub’s original studies,
his results presaged later findings regarding the roles of decreased we and increased you
and me as indicators of negativity during marital disputes (Simmons et al., 2005) and hos
tility during interactions between family members (Simmons, Chambless, & Gordon,
2008).
Trait Positivity
Positive emotionality or trait positive affect is another broad and important aspect of per
sonality. Beyond the fact that being happy is valuable for its own sake, happier people live
longer, a fact that is reflected in language use. In a sample of 180 autobiographical
sketches written by Catholic nuns when they first entered their convent, women who
used more positive language had lower mortality risk between the ages of 75 and 95
(Danner, Snowdon, & Friesen, 2002). However, Danner et al. (2002) did not measure trait
positive affectivity, optimism, or other enduring personality dimensions outside of nuns’
positive emotion references. As such, it is unclear whether their results were, as the lon
gitudinal results strongly suggest, due to trait rather than state positivity. If we assume
Page 15 of 33

that positive language tapped some chronic underlying trait in that study, it is not clear
which trait it specifically reflects.
More recent research sheds some light on what trait or traits positive references in Dan
ner et al. (2002) may have captured. In unstructured experimental language use, trait
positive affect is, as expected, positively correlated with positive emotion word usage. For
example, a sample of students who were instructed to talk aloud about any topic that they
chose for 3 minutes used more positive emotion words to the degree that they scored
higher on measures of trait positive affect and behavioral approach system sensitivity and
used more negative words to the degree that they ranked higher in trait negative affect
and behavioral inhibition system sensitivity (Cohen, Minor, Baillie, & Dahir, 2008).
Psychopathology-Related Personality Traits

Type A
Some of the earliest research on language use and personality interestingly came out of
behavioral medicine. In a series of studies, Scherwitz and colleagues (for reviews see
Scherwitz & Canick, 1988; Scherwitz, Graham, & Ornish, 1985) linked self-involvement to
the type A behavior pattern and coronary heart disease (CHD). Self-involvement was op
erationalized as the frequency with which participants used first-person singular pro
nouns during the structured type A interview. Type A emerged as positively related to the
use of first-person singular. Furthermore (and going beyond personality), use of first-per
son singular pronouns was related to blood pressure, coronary atherosclerosis, and
prospectively to CHD incidence and mortality. Interestingly, the relationship between
first-person singular use and CHD outcome often remained significant even after statisti
cally controlling for the type A behavior pattern (Scherwitz et al., 1985; Scherwitz &
Canick, 1988). Despite these promising early findings, links between pronoun use, type A,
and heart disease have not pursued in recent years (see Rohrbaugh, Mehl, Shoham, Reil
ly, & Ewy, 2008).
Depression
One theory of depression suggests that emotional pain—like physical pain—forces people
to pay (p. 211) attention to themselves (Pyszczynski & Greenberg, 1987). High degrees of
self-focus can be seen in people’s inattention to others and preoccupation with the self.
Other work on self-focus suggests that states of self-awareness are linked to elevated use
of first-person singular pronouns—especially the use of I-words such as I, I’m, I’ll (e.g.,
Davis & Brock, 1975).
Depressive states have been linked to higher use of I-words across several genres. In
some of the first systematic word count studies, Walter Weintraub (1981) found that
stream-of-consciousness writing produced much higher use of I-words in depressed pa
tients than in patients dealing with other medical disorders. More recently, Rude, Gort
Page 16 of 33

ner, and Pennebaker (2004) found that college students diagnosed with depression used
more I-words than did non-depressed students when writing essays about their college
experiences. Also, in natural speech captured over several days of tape recordings, use of
first-person singular is more frequent among those with high depression scores than
those with low depression scores (Mehl, 2006b).
The link between depression and increased self-focus is found not only in everyday lan
guage use but in published writing as well. Being a published poet is a surprisingly dan
gerous job. According to Kay Jamison (1995), published poets are up to 18 times as likely
as the general population to commit suicide. Her research suggests that poets have an
extraordinarily high rate of bipolar disorder, which is known to be closely linked to suici
dal behaviors. Motivated by this observation, Stirman and Pennebaker (2001) analyzed
the published work of 18 poets—nine who committed suicide and nine yoked controls.
They found that those who eventually committed suicide used first-person singular pro
nouns at higher rates than those who did not (Stirman & Pennebaker, 2001). Ironically,
suicidal poets did not use more negative emotion words than other poets. Overall, suici
dal poets’ language use showed that they were more self-focused and less socially inte
grated than nonsuicidal poets.
Surprisingly, depression diagnosis and subclinical symptomatology are only positively cor
related with references to negative emotions in certain contexts. For example, in natural
istic emails, talking, and blog posts, clinically depressed people and those with higher
subclinical depression scores use as many or more references to positive feelings overall
as never-depressed and less depressed individuals do (Baddeley, 2011; Rodriguez, Holler
an, & Mehl, 2010). In a sample of college students with subclinical depression symptoma
tology, Mehl (2006b) found no correlation between depression symptom severity and use
of positive emotion words, sadness, and anger words in everyday spoken language. It is
clear, however, that people suffering from depression experience more negative emotions
than do unaffected people and that those who are vulnerable to or currently experiencing
depression preferentially focus on negative over positive stimuli (Bistricky, Ingram, &
Atchley, 2008; Teasdale, 1988). Null results for negative emotion word use and depres
sion symptoms therefore suggest that depressed individuals inhibit the most obvious
markers of depression—negativity, sadness, irritability—by regulating their emotional lan
guage, perhaps in an attempt to avoid the negative social consequences of depression.
Depressed individuals may mask negative emotions in order to maintain their social net
work. Jenna Baddeley and colleagues have conducted two studies of naturalistic language
use among individuals with major depressive disorder (MDD) that may shed light on
where and to whom negative emotions are expressed during depression. In an EAR study
that recorded the daily lives of individuals from the community with and without MDD
over 3–4 days, she found that depressed individuals laugh, socialize, and talk as much as
nondepressed people (Baddeley, Pennebaker, & Beevers, 2012). In contrast with other
EAR findings, those participants with MDD did use negative emotions more frequently
than never-depressed participants; however, their negative emotion word use was moder
Page 17 of 33

ated by interaction partner, such that they primarily expressed negativity only with close
friends and relationship partners.
A similar pattern emerged in real-life emails. In a recent study, Baddeley (2011)

downloaded emails sent and received by individuals with and without MDD over the
course of 1 year. For MDD participants, that year included a depressive episode and a pe
riod of remission that each lasted at least 1 month. She found that depressed individuals
overall used more positive language than did matched controls and that when depressed
individuals used negative emotion words they did so primarily with correspondents they
rated as being close to them.
Depressed individuals’ tendency to hide their depression from all but their closest friends
may be an effective method of coping with the negative social consequences of depres
sion, such as being avoided by nondepressed friends (Schaefer, Kornienko, & Fox, 2011).
Content words, such (p. 212) as the negative emotion terms awful and cry, and their mean
ings are more salient during language processing than are function words like I and me;
as a result, they are easier to inhibit and control (Schmauder, Morris, & Poynor, 2000;
Townsend & Saltz, 1972). Therefore, when depressed individuals attempt to appear nor
mal, content markers of depression decrease and function word markers of depression re
main the same, results suggest.
The Dark Triad
Another area of psychopathology-related individual differences that language researchers

are increasingly interested in is the “dark triad” (Paulhus & Williams, 2002). The dark tri
ad is comprised of a group of socially antagonistic personality traits: Machiavellianism (a
tendency to strategically deceive and manipulate others), narcissism (a tendency to hold
inflated, grandiose self-views), and psychopathy (a tendency to show disregard and a lack
of empathy and remorse toward others). Similar to research on subclinical depression,
the dark triad personalities are typically considered subclinical, nonpathological versions
of “full blown” clinical personality disorders. Machiavellianism, narcissism, and psychopa
thy share a “dark” socially malevolent core but also have unique psychological aspects
and correlates (Paulhus & Williams, 2002). Among the dark triad personalities, subclini
cal narcissism has by far received the most scientific attention (Campbell & Miller, 2011).
The idea that narcissism is characterized by a frequent use of first-person singular pro
nouns was straightforward, given narcissists’ pervasive self-focus in social interactions.
This hypothesis was therefore also the first to be subjected to empirical tests. Raskin and
Shaw (1988) found that subclinical narcissism correlated positively with the use of first-
person singular and negatively with the use of first-person plural in tape-recorded im
promptu monologues. Such a bias toward linguistic egocentrism (Weintraub, 1981) ap
peared theoretically meaningful and suggested that narcissism is distinctly manifested in
people’s natural use of self-references.
Page 18 of 33

Building on Raskin and Shaw’s (1988) landmark study, other researchers have taken first-
person singular use as a face-valid linguistic indicator of self-focus. DeWall, Buffardi,
Bonser, and Campbell (2011) found that narcissistic Facebook users who did not draw at
tention to themselves in their profile pictures (e.g., by posting a sexy, revealing photo)
used first-person singular self-references at a high rate in their profile self-descriptions.
Similarly, narcissistic participants who did not use antisocial language (i.e., swear or
anger words) in personality essays (explaining why they possess the traits of honesty,
trustworthiness, and kindness) used first-person singular self-references at a high rate in
their essays. In both studies, presumably, the use of first-person singular indexed a com
pensatory verbal attention seeking strategy—although, interestingly, it did not show a (re
ported) significant zero-order correlation with narcissism.
Most recently, DeWall, Pond, Campbell, and Twenge (2011) tested the notion of a genera
tional increase in narcissism by tracking the use of self-references in popular US song
lyrics. The authors reported a linear trend pointing to an increase in first-person singular
pronouns between 1980 and 2007. Again, presumably, the use of first-person singular in
song lyrics indexes (cultural) egocentrism, and the data are meant to speak to the debate
on whether US society has become more narcissistic over time (Trzesniewski & Donnel
lan, 2010).
Several other studies, though, question the simple logic of I-use equals narcissistic self-fo
cus. Sampling language representatively from all of participants’ daily conversations (us
ing the EAR method), Holtzman, Vazire, and Mehl (2011) found no systematic association
between use of first-person singular and narcissism (r = .13, p = .25). Also, Fast and Fun
der (2010) found no simple correlation between use of first-person singular and narcis
sism among participants who underwent a life history interview (r = .02 for women and r
= .11, n.s. for men). Among men but not women, authority—a facet of narcissism—was
significantly positively related to I-use. However, self-sufficiency, yet another facet,
showed a significant negative association with I-use among men. Among women, I-use
was generally more strongly related to reported and observed anxiety and depressive
symptoms than it was to any facet of narcissism—a finding that is consistent with prior
research on the link between first-person singular self-references and depression (Mehl,
2006b; Rude, Gortner, & Pennebaker, 2004; Weintraub, 1981). The scientific question of
whether or when first-person singular use indexes narcissism is thus still open. Conceptu
ally, it seems important to distinguish between self-focus and self-awareness, with narcis
sists presumably ranging high on the former but low on the latter. Future research should
aim at reconciling the discrepant findings by identifying language contexts that particu
larly afford the (p. 213) expression of narcissism and neuroticism (or negative affect or de
pression), respectively.
Going beyond the narcissism–I link, Holtzman and colleagues (2010) found that narcis
sism in general, and especially its “toxic” components superiority/arrogance and exploita
tiveness/entitlement, were correlated with a more frequent use of swear and anger
words. The study further found that narcissists make more sexual references in their dai
ly language use. Finally, in the only language study on the dark triad personality trait of
Page 19 of 33

Machiavellianism, Ickes, Reidhead, and Patterson (1986) used text analysis to compare
Machiavellianism and self-monitoring. Following the idea that Machiavellianism is a form
of self-oriented, assimilative impression management, it was positively related to the use
of first-person singular (and plural) pronouns and negatively related to the use of second-
and third-person pronouns. Self-monitoring, however, as a form of other-oriented, accom
modative impression management, was negatively related to first-person singular use and
positively related to second-person pronouns.
Future Directions
The aim of this chapter has been to provide a blueprint of research on language and per
sonality up to this point that depicts both its structural soundness and need for additions
and improvements. In closing, we provide an overview of the existing studies in form of a
summary (see Table 13.2) and outline a few recommendations for future progress.
Page 20 of 33

Table 13.2 Summary of linguistic indicators of personality
Correlates Qualifier
(e.g., Mod
erator)
Big Five
Extraver second-person pronouns (+), first- Sex

sion person plural pronouns (+), positive
emotion (+), social (+), leisure (+),
sex (+), inhibition (–), tentativeness
(–)
Agreeable positivity (+), first-person singular

ness pronouns (+), social (+), home (+),
family (+), communication (+), death
(–), money (–), swearing (–)
Openness articles (+), prepositions (+), person Contradicto

al pronouns (–), family (–), home (–), ry facets
rest (–)
Conscien swearing (–), negative emotion (–) Sex

tiousness
Neuroti first-person singular pronouns (+), Sex, both on

cism negative emotion (+) ly in private
Psy
chopatholo
gy
Type A first-person singular pronouns (+)
Narcissism first-person singular pronouns (+/0), Sex, Commu

negative emotion (+), anger (+), nication Con
swear (+), sex (+), first-person plural text
pronouns (–)
Page 21 of 33

Machiavel first-person singular pronouns (+), Sex

lianism first-person plural pronouns (+), sec
ond person pronouns (–), third-per
son pronouns (–)
Depression first-person singular pronouns (+), Public vs.

negative emotion (+) private, cor
respondent
closeness
Trait Emo
tionality
Anger second-person singular pronouns (+)
Negative negative emotion (+) Writing or

emotions speaking
topic
Positive positive emotion (+)

emotions
See text for references. For the Big Five, only the most common and
universal correlates are listed.
Finding consistent threads among studies is sometimes made difficult by differing

methodologies. Even among studies that used the same text analysis tool, some focused
only on linguistic content rather than all categories, and others used different versions of
a program that include several nonoverlapping categories. The literature on language
and personality would no doubt benefit from more comprehensive reporting of effects, in
papers or in online supplemental materials. The existing studies suggest that both con
tent and style categories are critical. Although content words are more susceptible to
self-regulation and thus tend to be lower fidelity indicators of internal states, the degree
to which a person’s language use fails to reflect his or her self- or informant-reported per
sonality is often a telling indicator of self-regulatory personality processes and person ×
situation interactions (Baddeley, 2011; Baddeley, Pennebaker, & Beevers, 2012; Mehl et
al., 2006; Mehl & Holleran, 2008). Style words are often more challenging to interpret
but are valuable as the mostly automatic, and therefore more psychologically representa
tive, indicators of attentional focus and thinking styles (see Tausczik & Pennebaker,
2010). Content and style are two sides of a data-rich coin, and personality psychology has
much to gain from increasingly considering both aspects of language use. To correctly in
terpret the nature and true magnitude of effects, studies of language and personality may
also need to increasingly measure and consider a range of potential moderators or modi
fiers, including facet-level trait measures (Yarkoni, 2010), individuals’ sex (Mehl et al.,
Page 22 of 33

2006), whether language use is public or private (Mehl & Holleran, 2008), the closeness
of conversation partners (Baddeley, 2011), and linguistic co-occurrences (Gill & Oberlan
der, 2002). Specifically for function words, which are by definition extraordinarily versa
tile, research has shown that moderators matter. For example, whether I or you is said by
a man or a woman and in the context of an angry or cheerful communication can dramati
cally influence which psychological processes those words reflect (Fast & Funder, 2010;
Mehl et al., 2006; Tausczik & Pennebaker, 2010).
Context effects, such as the types of communication that a situation affords or demands,
are important considerations in any area of behavioral research. Studies of language use
are no exception. Just as a highly extraverted person would not be expected to behave
dramatically differently than an introverted person in a situation lacking the potential for
social interaction, personality traits that are predominantly defined by differences in so
cial interaction are likely to leave fewer observable traces in solitary writing such as
stream-of-consciousness essays. Furthermore, writing or speaking tasks that resemble
criterion measures of personality (e.g., self-report personality questionnaires and essays
describing one’s personality) are bound to be more highly correlated than naturalistic
measures of language (e.g., Hirsh & Peterson, 2009). However, perhaps in part due to the
influence of corpus linguistics, in which language from a wide range of communication
media are frequently compiled into a single dataset comprising billions of words, studies
of linguistic indicators of personality have only recently come to seriously consider com
munication context. Given that so many personality dimensions hinge on how people re
act to and interact with others, it is particularly important—in studies of natural language
use and beyond—for personality research to increasingly study the links between natural
ly occurring dialog, self-reports, and observer (p. 214) reports. As naturalistic language re
search expands with ongoing advances in audio recording technology and computer sci
ence methods, it should become easier to understand how linguistic signals are attenuat
ed and warped by contextual influences such as experimental task, communication medi
um, and motivation.
The accomplishments of computerized text analysis in the past 15 years have been extra
ordinary. However, the software designers, programmers, and data analysts behind this
revolution readily admit that there is room for improvement. Cohen and col
leagues’ (2009) and S. Cohen’s (2011) research on the measurement of trait affect points
to a possible need to improve word-count measurements of common positive emotion
words, which are often used in ways that do not reflect positivity (e.g., I was pretty bored,
someone like you), by considering their linguistic contexts. New discoveries made in func
tion word categories that are new to the most recent version of LIWC (Pennebaker, Booth,
& Francis, 2007) suggest that finer grained analyses based on words’ grammatical roles
have the potential to clarify mixed results in past research and shed light on the cognitive
mechanisms underlying personality dimensions.
Measures of within-text context—and the usability of tools that consider linguistic context
—are (p. 215) bound to improve studies of language and personality as well. A word’s loca
tion in a text or sentence (Vine & Pennebaker, 2009) and its neighboring words (Gill &
Page 23 of 33

Oberlander, 2004) clearly matter but are rarely considered in psychological text analyses.
Programs such as Latent Semantic Analysis (Landauer & Dumais, 1997) and WordSmith
(Scott, 2008) handle such variables and, as they become more widely known and user-
friendly, stand to greatly enrich future research.
Conclusion
In his famous monograph on personality, Allport (1937) wrote “language is a codification
of common human experience, and by analyzing it much may be found that reflects the
nature of human personality” (p. 373). Interestingly, the field of personality and language
use only started getting serious momentum more than half a century later. As the re
search reviewed in this chapter reveals, though, the field is now rich, vibrant, and has al
ready produced many important discoveries. We expect that the immense progress in
(stationary and mobile) computing technology and parallel advances in computational lin
guistics will create a strong push for the field over the next years and lead to critical im
provements in the complexity with which naturalistic language can be analyzed (Ran
ganath, Jurafsky, & McFarland, 2013; Schwartz et al., 2013). It is our sense that the field
will thrive to the extent that it uses these technologically driven, “bottom-up,” analytic ad
vances and, at the same time, balances them with innovative theoretical developments
and clarifications from the “top down.” To achieve this, it will undoubtedly become neces
sary for researchers from different fields to “cross-talk.” Social psychologists, personality
psychologists, cognitive psychologists, linguists, communication scholars, computer sci
entists, and other researchers will need to engage in conversations and collaborations
and thereby transcend (and hopefully reduce) traditional discipline boundaries to more
fully understand how our words reflect our selves.
References
Allport, G. W. (1937). Personality: A psychological interpretation. Oxford, England: Holt.
Argamon, S., Koppel, M., Pennebaker, J. W., & Schler, J. (2009). Automatically profiling the
author of an anonymous text. Communications of the ACM, 52, 119.
Augustine, A. A., Mehl, M. R., & Larsen, R. J. (2011). A positivity bias in written and spo
ken English, and its moderation by personality and gender. Social Psychology and Person
ality Science, 2, 508–515.
Baddeley, J. L. (2011). Email communications among people with and without major de
pressive disorder. Unpublished doctoral dissertation. University of Texas at Austin,
Austin, TX.
Baddeley, J. L., Pennebaker, J. W., & Beevers, C. G. (2012). Everyday social behavior dur
ing a major depressive episode. Social Psychology and Personality Science, 4, 445–452.
Page 24 of 33

Baddeley, J. L., & Singer, J. A. (2008). Telling losses: Personality correlates and functions
of bereavement narratives. Journal of Research in Personality, 42, 421–438. doi: 10.1016/
j.jrp.2007.07.006
Bell, A., Brenier, J., Gregory, M., Girand, C., & Jurafsky, D. (2009). Predictability effects on
durations of content and function words in conversational English. Journal of Memory and
Language, 60, 92–111. doi: 10.1016/j.jml.2008.06.003
Biber, D. (1995). Dimensions of register variation: A cross-linguistic comparison. Cam

Bistricky, S. L., Ingram, R. E., & Atchley, R. A. (2011). Facial affect processing and depres
sion susceptibility: Cognitive biases and cognitive neuroscience. Psychological Bulletin,
137, 998–1028. doi: 10.1037/a0025348
bridge: Cambridge University Press.
Campbell, W. K., & Miller, J. D. (2011). The handbook of narcissism and narcissistic per
sonality disorder: Theoretical approaches, empirical findings, and treatments. Hoboken,
NJ: John Wiley & Sons.
Chartrand, T. L., & van Baaren, R. (2009). Human mimicry. Advances in Experimental So
cial Psychology, 41, 219-274.

Levine, & S. D. Teasley (Eds.), Perspectives on socially shared cognition (pp. 127–149).
Washington, DC: APA Books.
Cohen, A. S., Minor, K. S., Baillie, L. E., & Dahir, A. M. (2008). Clarifying the linguistic sig
nature: Measuring personality from natural speech. Journal of Personality Assessment,
90, 559–563.
Cohen, A. S., Minor, K. S., Najolia, G. M., & Hong, S. L. (2009). A laboratory-based proce
dure for measuring emotional expression from natural speech. Behavior Research Meth
ods, 41, 204–12. doi: 10.3758/BRM.41.1.204
Cohen,.S. J.(2011). Measurement ofnegativity bias in personal narratives using corpus-

based emotion dictionaries. Journal of Psycholinguistic Reseach, 40: 119–135
Costa, P. T., Jr., & McCrae, R. R. (1992). Normal personality assessment in clinical prac
tice: The NEO Personality Inventory. Psychological Assessment, 4, 5–13.
Danner, D. D., Snowdon, D. A., & Friesen, W. V. (2002). Positive emotions in early life and
longevity: Findings from the Nun Study. Journal of Personality and Social Psychology, 80,
804–813.
Page 25 of 33

Davis, D., & Brock, T. C. (1975). Use of first-person pronouns as a function of increased
objective self-awareness and performance feedback. Journal of Experimental Social Psy
chology, 11, 381–388.
Dewaele, J-M., & Furnham, A. (2000). Personality and speech production: A pilot
(p. 216)
study of second language learners. Personality and Individual Differences, 28, 355–365.
DeWall, C. N., Buffardi, L. E., Bonser, I., & Campbell, W. K. (2011). Narcissism and implic
it attention seeking: Evidence from linguistic analyses of social networking and online
presentation. Personality and Individual Differences, 51, 57–62.
DeWall, C. N., Pond, R. S., Campbell, W. K., & Twenge, J. M. (2011). Tuning in to psycho
logical change: Linguistic markers of self-focus, loneliness, anger, antisocial behavior, and
misery increase over time in popular U.S. song lyrics. Psychology of Aesthetics, Art, and
Creativity, 5, 200–207.
Dodds, P. S., & Danforth, C. M. (2010). Measuring the happiness of large-scale written ex
pression: Songs, blogs, and presidents. Journal of Happiness Studies, 11, 441–456.
Eagly, A. H. (1995). The science and politics of comparing women and men. American Psy
chologist, 50, 145–158.
Eastwick, P. W., Eagly, A. H., Finkel, E. J., & Johnson, S. E. (2011). Implicit and explicit
preferences for physical attractiveness in a romantic partner: A double dissociation in
predictive validity. Journal of Personality and Social Psychology, 101, 993–1011.
Fast, L. A, & Funder, D. C. (2008). Personality as manifest in word use: correlations with
self-report, acquaintance report, and behavior. Journal of Personality and Social Psycholo
gy, 94, 334–346. doi: 10.1037/0022-3514.94.2.334
Fast, L. A., & Funder, D. C. (2010). Gender differences in the correlates of self-referent
word use: Authority, entitlement, and depressive symptoms. Journal of Personality, 78,
313–338. doi: 10.1111/j.1467-6494.2009.00617.x
Gill, A. J., & Oberlander, J. (2002). Taking care of the linguistic features of extraversion.
Proceedings of the 24th Annual Conference of the Cognitive Science Society, 363–368.
Gleser, G. C., Gottschalk, L. A, & John, W. (1959). The relationship of sex and intelligence
to choice of words: a normative study of verbal behavior. Journal of Clinical Psychology,
15, 182–91.
Goldberg, L. R. (1981). Language and individual differences: The search for universals in
personality lexicons. In L. Wheeler (Ed.), Review of personality and social psychology (pp.
141–165). Beverly Hills, CA: Sage.
Golder, S. A., & Macy, M. W. (2011). Diurnal and seasonal mood vary with work, sleep and
daylength across diverse cultures. Science, 333, 1878–1881.
Page 26 of 33

Gosling, S. D. (2008). Snoop: What your stuff says about you. New York: Basic Books.
Gottschalk, L. A., & Gleser, G. C. (1969). Measurement of psychological states through the
content analysis of verbal behaviour. Berkeley: University of California Press.
Hancock, J. T., Curry, L., Goorha, S., & Woodworth, M. (2008). On lying and being lied to:
A linguistic analysis of deception. Discourse Processes, 45, 1–23.
Hart, R. P. (1984). Verbal style and the presidency: A computer-based analysis. New York:
Academic Press.
Hirsh, J. B., & Peterson, J. B. (2009). Personality and language use in self-narratives. Jour
nal of Research in Personality, 43, 524–527. doi: 10.1016/j.jrp.2009.01.006
Holmes, J. (1995). Women, men and politeness. Harlow: Longman.
Holtgraves, T. (2011). Text messaging, personality, and the social context. Journal of Re
search in Personality, 45(1), 92–99. doi: 10.1016/j.jrp.2010.11.015
Holtgraves, T. (2010). Social psychology and language: Words, utterances and conversa
tions. In S. Fiske, D. Gilbert, & G. Lindzey (Eds.), Handbook of social psychology, 5th edi
tion.
Holtzman, N. S., Vazire, S., & Mehl, M. R. (2010). Sounds like a narcissist: Behavioral
manifestations of narcissism in everyday life. Journal of Research in Personality, 44, 478–
484. doi: 10.1016/j.jrp.2010.06.001
Ickes, W., & Reidhead, S., & Patterson, M. (1986). Machiavellianism and self-monitoring:
As different as “me” and “you.” Social Cognition, 4, 58–74.
Ireland, M. E., & Pennebaker, J. W. (2010). Language style matching in writing: Syn
chrony in essays, correspondence, and poetry. Journal of Personality and Social Psycholo
gy, 99, 549.
Jamison, K. R. (1995). Manic-depressive illness and creativity. Scientific American, 272,

62–62.
Jay, T. (2009). The utility and ubiquity of taboo words. Perspectives on Psychological
Science, 4, 153–161.
Kacewicz, E., Pennebaker, J. W., Davis, M., Jeon, M., & Graesser, A. C. (2013). Pronoun
use reflects standings in social hierarchies. Journal of Language and Social Psychology.
Koppel, M., Argamon, S., & Shimoni, A. (2003), Automatically categorizing written texts
by author gender. Literary and Linguistic Computing, 17, 401–412.
Kosinski, M., & Stillwell, D. (2012). myPersonality research wiki: myPersonality project.
In http://www.mypersonality.org/wiki/.
Page 27 of 33

Kramer, A. D. I. (2010). An unobtrusive behavioral model of “gross national happiness”.

Proceedings of the 28th International Conference on Human Factors in Computing Sys
tems—CHI, pp. 287–290.
Küfner, A. C. P., Back, M. D., Nestler, S., & Egloff, B. (2010). Tell me a story and I will tell
you who you are! Lens model analyses of personality and creative writing. Journal of Re
search in Personality, 44, 427–435. doi: 10.1016/j.jrp.2010.05.003
Lakoff, R. T. (1975). Language and woman’s place. New York: Harper & Row.
Landauer, T. K., & Dumais, S. T. (1997). A solution to Plato’s problem: The Latent Seman
tic Analysis theory of the acquisition, induction, and representation of knowledge. Psycho
logical Review, 104, 211–240.
Lee, C. H., Kim, K., Seo, Y. S., & Chung, C. K. (2007). The relations between personality
and language use. Journal of General Psychology, 134, 405–413.
Levelt, W. J. (1989). Speaking: From intention to articulation. Cambridge, MA: The MIT
Press.
Mairesse, F., & Walker, M. A. (2011). Controlling user perceptions of linguistic style:
Trainable generation of personality traits. Computational Linguistics, 37, 445–488.
Mairesse, F., & Walker, M. A. (2006). Words mark the nerds : Computational models of
personality recognition through language. Proceedings of the 28th Annual Conference of
the Cognitive Science Society, 543–548.
Mairesse, F., Walker, M. A., Mehl, M. R., & Moore, R. K. (2007). Using linguistic cues for
the automatic recognition of personality in conversation and text. Journal of Artificial In
telligence Research, 30, 457–500.
McAdams, D. P., & Pals, J. L. (2006). A new Big Five: Fundamental principles for an inte
grative science of personality. American Psychologist, 61, 204–217. doi:
10.1037/0003-066X.61.3.204
Mehl, M. R. (2006a). Quantitative text analysis. In M. Eid & E. Diener (Eds.), Handbook of
multimethod measurement in psychology (pp. 141–156). Washington, DC: American Psy
chological Association.
Mehl, M. R. (2006b). The lay assessment of sub-clinical depression in daily life. Psycholog
ical Assessment, 18, 340–345.
Mehl, M. R., Gosling, S. D., & Pennebaker, J. W. (2006). Personality in its natural
(p. 217)
habitat: Manifestations and implicit folk theories of personality in daily life. Journal of
Mehl, M. R., & Holleran, S. E. (2008). How taking a word for a word can be problematic:
Context-dependent linguistic markers of extraversion and neuroticism. Paper presented
Page 28 of 33

at the 11th Conference of the International Association for Language and Social Psycholo
gy, Tucson, Arizona.
Mehl, M. R., & Pennebaker, J. W. (2003). The sounds of social life: A psychometric analysis
of students’ daily social environments and natural conversations. Journal of Personality
Mehl, M. R., Pennebaker, J. W., Crow, D. M., Dabbs, J., & Price, J. (2001). The Electronical
ly Activated Recorder (EAR): A device for sampling naturalistic daily activities and con
versations. Behavior Research Methods, Instruments, & Computers, 33, 517–523.
Michel, J.-B., Shen, Y. K., Aiden, A. P., Veres, A., Gray, M. K., Te Google Books Team, Pick
ett, J. P., Hoiberg, D., Clancy, D., Norvig, P., Orwant, J., Pinker, S., Nowak, M. A., & Aiden,
E. L. (2010). Quantitative analysis of culture using millions of digitized books. Science.
doi: 10.1126/science.1199644
Newman, M. L., Groom, C. J., Handelman, L. D., & Pennebaker, J. W. (2008). Gender dif
ferences in language use: An analysis of 14,000 text samples. Discourse Processes, 45(3),
211–236. doi: 10.1080/01638530802073712
Oberlander, J., & Gill, A. J. (2006). Language with character: A corpus-based study of indi
vidual differences in e-mail communication. Discourse Processes, 42, 239–270.
O’Carroll Bantum, E., & Owen, J. E. (2009). Evaluating the validity of computerized con
tent analysis programs for identification of emotional expression in cancer narratives.
Psychological Assessment, 21, 79–88.
Paulhus, D. L., & Williams, K. M. (2002). The Dark Triad of personality: Narcissism,
machiavellianism, and psychopathy. Journal of Research in Personality, 36, 556–563. doi:
10.1016/S0092-6566(02)00505-6
Pennebaker, J. W. (1997). Opening up: The healing power of expressing emotion. New
York: Guilford Press.
Pennebaker, J. W. (2011). The secret life of pronouns: What our words say about us. New
York: Bloomsbury Press.
Pennebaker, J. W., Booth, R. J., & Francis, M. E. (2007). Linguistic Inquiry and Word Count
(LIWC): LIWC 2007 [Computer program]. Austin, TX: LIWC.net.
Pennebaker, J. W., & Ireland, M. E. (2011). Using literature to understand authors: The
case for computerized text analysis. Scientific Study of Literature, 1, 34–48.
Pennebaker, J. W., & King, L. A. (1999). Linguistic styles: Language use as an individual
difference. Journal of Personality and Social Psychology, 77, 1296–1312.
Page 29 of 33

Pennebaker, J. W., & Lay, T. C. (2002). Language use and personality during crises: Analy
ses of mayor Rudolph Giuliani’s press conferences. Journal of Research in Personality, 36,
271–282.
Pennebaker, J. W., Mehl, M. R., & Niederhoffer, K. G. (2003). Psychological aspects of nat
ural language. use: Our words, our selves. Annual Review of Psychology, 54, 547–577.
Pickering, M. J., & Garrod, S. (2004). Toward a mechanistic psychology of dialogue. The
Behavioral and Brain Sciences, 27, 169–90; discussion 190–226.
Poole, M. E. (1979). Social class, sex, and linguistic coding. Language and Speech, 22, 49–
67.
Pyszczynski, T., & Greenberg, J. (1987). Self-regulatory perseveration and the depressive
self-focusing style: A self-awareness theory of reactive depression. Psychological Bulletin,
102, 122–138.
Qiu, L., Lin, H., Ramsay, J., & Yang, F. (2012). You are what you tweet: Personality expres
sion and perception on Twitter. Journal of Research In Personality, 46, 710–718.
Ranganath, R., Jurafsky, D., & McFarland, D. A. (2013). Detecting friendly, flirtatious,
awkward, and assertive speech in speed-dates. Computer Speech & Language, 27, 89–
115.
Ramírez-Esparza, N., Chung, C. K., Kacewicz, E., & Pennebaker, J. W. (2008). The psychol
ogy of word use in depression forums in English and in Spanish: Testing two text analytic
approaches. Proceedings of the International Conference on Weblogs and Social Media
(ICWSM 2008).
Raskin, R., & Shaw, R. (1988). Narcissism and the use of personal pronouns. Journal of
Personality, 56, 2, 393–404.
Robbins, M. L., Focella, E. S., Kasle, S., Weihs, K. L., Lopez, A. M., & Mehl, M. R. (2011).
Naturalistically observed swearing, emotional support and depressive symptoms in
women coping with illness. Health Psychology, 30, 789–792.
Roccas, S., Sagiv, L., Schwartz, S. H., & Knafo, A. (2002). The Big Five personality factors
and personal values. Personality and Social Psychology Bulletin, 28, 789–801. doi:
10.1177/0146167202289008
Rodriguez, A. J., Holleran, S. E., & Mehl, M. R. (2010). Reading between the lines: The lay
assessment of subclinical depression from written self-descriptions. Journal of
Personality, 78, 575–598. doi: 10.1111/j.1467-6494.2010.00627.x
Rohrbaugh, M. J., Mehl, M. R., Shoham, V., Reilly, E. S., & Ewy, G. A. (2008). Prognostic
significance of spouse we talk in couples coping with heart failure. Journal of Consulting
and Clinical Psychology, 76, 781–789.
Page 30 of 33

Rubin, D. L., & Greene, K. (1992). Gender-typical style in written language. Research in
the Teaching of English, 7-40.
Rude, S. S., Gortner, E. M., & Pennebaker, J. W. (2004). Language use of depressed and
depression-vulnerable college students. Cognition & Emotion, 18, 1121–1133.
Schaefer, D. R., Kornienko, O., & Fox, A. M. (2011). Misery does not love company: Net
work selection mechanisms and depression homophily. American Sociological Review, 76,
764–785.
Scherwitz, L., & Canick, J. (1988). Self-reference and coronary heart disease risk. In K.
Houston, & C. R. Snyder (Eds.), Type A behavior pattern: Research, theory, and interven
tion. New York: John Wiley & Sons.
Scherwitz, L., Graham, L. E., & Ornish, D. (1985). Self-involvement and the risk factors
for coronary heart disease. Advances, 2, 6–18.
Schmauder, A. R., Morris, R. K., & Poynor, D. V. (2000). Lexical processing and text inte
gration of function and content words: Evidence from priming and eye fixations. Memory
& Cognition, 7, 1098–1108.
Schnurr, P. P., Rosenberg, S. D., Oxman, T. E., & Tucker, G. J. (1986). A methodological
note on content analysis: Estimates of reliability. Journal of Personality Assessment, 50,
601–609.
Schwartz, H. A. Eichstaedt, J. C., Dziurzynski, L., Kern, M. L., Blanco, E., Kosinski, M.
Stillwell, D., Seligman, M., & Ungar, L. H. (2013). Toward personality insights from lan
guage exploration in social media. AAAI-2013 Spring Symposium: Analyzing Microtext.
Stanford, California.
Schwartz, H. A., Eichstaedt, J. C., Kern, M. L., Dziurzynski, L., Ramones, S. M., Agrawal,
M., Ungar, L. H. (2013). (p. 218) Personality, gender, and age in the language of social me
dia: The open-vocabulary approach. PLoS One, 8(9), e73791.
Scott, M. (2008). WordSmith Tools version 5. Liverpool: Lexical Analysis Software.
Simmons, R. A., Gordon, P. C., & Chambless, D. L. (2005). Pronouns in marital interaction.
Psychological Science, 16, 932–936.
Simmons, R. A., Chambless, D. L., & Gordon, P. C. (2008). How do hostile and emotionally
overinvolved relatives view relationships? What relatives’ pronoun use tells us. Family
Process, 47, 405–419.
Slatcher, R. B., Chung, C. K., Pennebaker, J. W., & Stone, L. D. (2007). Winning words: In
dividual differences in linguistic style among U.S. presidential and vice presidential candi
dates. Journal of Research in Personality, 41, 63–75.
Page 31 of 33

Smith, C. P. (Ed.). (1992). Motivation and personality: Handbook of thematic content

analysis. Cambridge, MA: Cambridge University Press.
Stirman, S. W., & Pennebaker, J. W. (2001). Word use in the poetry of suicidal and non-sui
cidal poets. Psychosomatic Medicine 63, 517–522.
Stone, P. J., Dunphy, D. C., Smith, M. S., & Ogilvie, D. M. (1966). The general inquirer: A
computer approach to content analysis. Cambridge: MIT Press.
Tanenhaus, M. K., & Trueswell, J. C. (1995). Sentence comprehension. In Eimas & Miller
(Eds.) Handbook in perception and cognition: Vol. 11: Speech Language and Communica
tion (pp. 217–262). New York: Academic Press.
Tausczik, Y. R., & Pennebaker, J. W. (2010). The psychological meaning of words: LIWC
and computerized text analysis methods. Journal of Language and Social Psychology, 29,
24–54. doi: 10.1177/0261927X09351676
Teasdale, J. D. (1988). Cognitive vulnerability to persistent depression. Cognition and

Emotion, 2, 247–274.
Teasdale, J. D., & Green, H. A. C. (2004). Ruminative self-focus and autobiographical

memory. Personality and Individual Differences, 36, 1933–1943.
Townsend, D. J., & Saltz, E. (1972). Phrases vs. meaning in the immediate recall of sen
tences. Psychonomic Science, 29, 381–384.
Trzesniewski, K. H., & Donnellan, M. B. (2010). Rethinking “Generation Me”: A study of

cohort effects from 1976–2006. Perspectives in Psychological Science, 5, 58–75.
Malin, J. J., Vine, V. J., Stanton, A., Cannava, K., Bodie, G., & Pennebaker, J. W. (2014, Feb
ruary). The arc of narrative: Using language markers to identify stories. Poster presented
at the annual meeting of the Society for Personality and Social Psychology, Austin, TX.
Watson, D., & Pennebaker, J. W. (1989). Health complaints, stress, and disease: Exploring
the central role of negative affectivity. Psychological Review, 96, 234–254.
Weintraub, W. (1981). Verbal behavior: Adaptation and psychopathology. New York:

Springer.
Yarkoni, T. (2010). Personality in 100,000 Words: A large-scale analysis of personality and

word use among bloggers. Journal of Research in Personality, 44, 363–373. doi: 10.1016/
j.jrp.2010.04.001
Notes:
(1) . At some point, you may have received the following test in an email: “How many Fs
does the following passage contain? ‘Finished files are the result of years of scientific
Page 32 of 33

study combined with the experience of years.’” Finding only three Fs tends to result from
readers skipping ofs.
(2) . The term sex is used by default to refer to all differences in personality–language
links between men and women. However, gender may be more appropriate in cases
where linguistic differences seem to be more strongly influenced by gender norms than
biology (see Eagly, 1995).
Molly E. Ireland
Molly E. Ireland, Annenberg School for Communication, University of Pennsylvania
Matthias R. Mehl
Department of Psychology, University of Arizona, Tucson, USA
Page 33 of 33

Using Computerized Text Analysis to Track Social Processes
Using Computerized Text Analysis to Track Social

Processes
Cindy K. Chung and James W. Pennebaker
Print Publication Date: Sep 2014 Subject: Psychology, Social Psychology, Affective Science
Online Publication Date: Jun 2014 DOI: 10.1093/oxfordhb/9780199838639.013.037
What do our everyday words reveal about the quality and types of social processes at play
in interactions? This chapter focuses on the links between function and emotion words
and social behaviors. Computerized text analyses allow researchers to track social psy
chological processes in dyads, groups, and larger social networks. The authors review
studies on experimental and naturalistic archives that reveal relationship roles, process
es, and outcomes. Computer-based language analysis methods can be used to assess and
track ongoing social processes on a large scale. The expansion and application of lan
guage technologies can provide real-time feedback to shape group processes as they un
fold.
Keywords: Groups, language, LIWC, relationships, social interactions
For the first time in the history of the world, social scientists have access to massive lan
guage- based datasets with tools to analyze them. The availability of natural interactions
via email, blogs, Facebook, and other social media provides new ways to track interper
sonal dynamics in the past or those that are occurring in real time. Traditional research
on social and group processes have focused on small numbers of dyads and groups, par
ticularly on their motivations and outcomes. By building on research that has linked nat
ural language to psychological states and social processes, studies have begun to uncover
how social relationships and group dynamics unfold. The same metrics to assess social
and group processes via language in small experimental studies can be scaled up and ap
plied to more interactions and to entire communities (Pennebaker & Chung, in press). As
we outline in this chapter, the study of language and social psychology can now provide
real-time assessments of group processes and track the thinking of large groups of peo
ple over time and space.
Page 1 of 21

Tracking Group Processes

The dominant paradigm that has emerged in the group literatures is the input-process-
outcome model (IPO; see Guzzo & Dickson, 1996; Hackman & Morris, 1975). Most empiri
cal studies have focused on the input portion of the model. Studies on group size, compo
sition, role structure, context of the group, and the like have yielded an impressive num
ber of findings that are often relevant for specific types of groups. By the same token,
many studies have been conducted on group outcomes—ranging from measures of group
performance, cohesiveness, longevity, and the like.
The most discussed but the least studied or understood feature of the IPO model has been
the P—process. There is no doubt that how members of a group work with one another is
critical to understanding a group. The problem has been how to measure group interac
tions. Three recurring issues have challenged group process researchers: (1) identifying
the basic coordinating mechanisms of a working group, (2) agreeing on good measure
ment (p. 220) strategies that reflect relevant group processes, and (3) effectively manipu
lating the active ingredients in ways that can improve group processes and performance.
Identifying the Coordinating Mechanisms
In their review of teamwork, Salas, Goodwin, and Burke (2009) argue that groups are
most effective if they have good leadership, mutual performance monitoring, backup be
havior, team orientation, and are adaptable. These dimensions assume that the groups
have effective coordinating mechanisms, including shared mental models, effective com
munication patterns, and mutual trust. All three of these coordinating mechanisms essen
tially argue that the group members are thinking along the same dimensions and commu
nicate effectively.
Several other possible coordinating mechanisms have occasionally been identified, in
cluding positivity (Losada & Heaphy, 2004; West, Patera, & Carsten, 2009), task focus
(Jonassen & Kwon, 2001), cohesive subgroups (Foltz & Martin, 2009), and equality of con
tributions (Beebe & Masterson, 1997; Hare, 1962; Leavitt, 1951). Although not a coordi
nating mechanism per se, there is also the awareness that groups coordinate differently
as a function of time. McGrath (e.g., 1991) and others have argued, for example, that ear
ly group exchanges can sometimes predict long-term group behavior better than later
group interactions (for a nice example, see Sexton & Helmreich, 2000).
The Measurement Problem
Measuring ongoing group processes has been a stumbling block for researchers for
decades. In a live interacting group, there may be as many as twenty or more conversa
tional turns per minute with simultaneous topic threads. Because the interactions are dy
namic, the nature of the conversations may change drastically depending on situational
demands. Group researchers have dealt with measuring group processes in several ways.
The most common and efficient has been to ask the group members or expert judges to
Page 2 of 21

make ratings of the group’s interactions. In both cases, the ratings are generally summa
ry evaluations that attempt to capture the overall group process dimensions.
A second general strategy has been to hand code transcripts of group interactions in an
attempt to isolate group processes in the interactions themselves (for a good review, see
Harris & Sherblom, 2002). Although hand coding group interactions for coordination
strategies and communication has shown promise in predicting group performance, man
ual coding has proven to be slow and expensive, suffering from cross-judge agreement or
a lack of consistency in coding strategies across studies.
Automated analyses of group processes using multivariate methods such as latent seman
tic analysis (LSA; Foltz & Martin, 2009) have captured task focus in working groups. Sim
ilar to LSA methods of exam grading, other methods approximate the degree to which a
group uses language that reflects desired or correct solutions (see also Graesser et al.,
2004). These methods can also be used to evaluate the degree to which group members
discuss task content in overlapping ways.
Another strategy, also reported by Foltz and Martin (2009), is to capture talking patterns
within a group using sophisticated network analyses. Indeed, network analysis methods
have become quite fashionable in the computer science/social media world to identify net
work centrality (Hossain, Wu, & Chung, 2006), relative status (Scholand, Tausczik, & Pen
nebaker, 2010), knowledge transmission between divisions in a company (Hansen, 2002),
and the stability of social groupings within an organization (Tichy, Tushman, & Fombrun,
1979).
Language Measurement Strategy

The past 50 years of group research has yielded an impressive body of knowledge about
the I and the O of the IPO model. The theorizing about the process link has also been vig
orous. The roadblock has been in developing efficient ways to track ongoing group inter
action. Furthermore, without capturing real-time group dynamics, realistic empirically
based real-time group interventions are simply not practical.
Researchers interested in social interactions are standing on the threshold of a new

world. For the first time, scientists in social psychology, artificial intelligence, and com
munications are developing tools that can capture real-world social interaction through
people’s use of natural language. We can now begin to think about ways of measuring in
formation transfer, language coordination, emotional tone, and other processes as they
occur and, at the same time, distinguish which features of language are most related to
outcomes of interest. Most exciting is that we can begin to harness these findings and
provide immediate feedback to group members in ways that can shape their interaction
styles. This presents new ways to shape group interactions in communication settings
across education, forensic (p. 221) interviews, clinical sessions, and everyday social media
interactions.
Page 3 of 21

There are currently a large number of approaches to study natural language with com
puter-based tools. Most were developed to study the content of speech. This content-
based strategy is inherent in internet search engines, as well as in powerful programs
such as LSA (Landauer, Foltz, & Laham, 1998) and other topic modeling methods
(Steyvers & Griffiths, 2007) that rely on singular value decomposition (SVD) metrics. In
terestingly, most such programs exclude the analysis of function words, which makes
good sense if one is interested in the content of speech (Chung & Pennebaker, 2008). Un
like more complex systems that rely on SVD, ours is a much simpler word count ap
proach.
Linguistic Inquiry and Word Count (LIWC; Pennebaker, Booth, & Francis, 2007) is a soft
ware program that is made up of a processor and a dictionary. The processor references
the dictionary to count words in the specified text. The output is a matrix of text files
(rows) and each of the LIWC dictionary word categories (columns). The percentage of to
tal words in a file belonging to a particular category appears in each of the cells. The out
put also includes basic information about each text, such as the word count and the per
centage of words that the dictionary captured within the text, as well as the percentage
of words greater than six letters and punctuation characters appearing in the text.
The LIWC dictionary consists of approximately eighty categories that are grammatical
(e.g., articles, pronouns, prepositions), psychological (e.g., cognitive mechanisms, emo
tions, social), referring to relativity (e.g., time, space, motion), or content-based (e.g.,
achievement, home, religion). In the first phases of development, candidate words were
generated by referring to thesauruses, dictionaries, questionnaires, and by human
judges. To ensure high inter-judge reliability, a second phase then relied on three or four
human judges who independently decided if a given word was appropriate to a given cat
egory (see Pennebaker, Chung, Ireland, Gonzales, & Booth, 2007). Because grammatical
categories exist in relatively fixed groupings (e.g., articles, first person singular pro
nouns, conjunctions, etc.), these were also added to the dictionary. The resultant default
dictionary cannot be modified to allow direct comparison of results from one user to an
other. However, the LIWC processor allows users to upload custom dictionaries to process
text.
Each word or word stem in the LIWC default dictionary defines one or more word cate
gories or subdictionaries. For example, the word “cried” is part of five word categories:
sadness, negative emotion, overall affect, verb, and past tense verb. Hence, if it is found
in the target text, each of these five subdictionary scale scores is incremented. As in this
example, many of the LIWC categories are arranged hierarchically. All sadness words, by
definition, will be categorized as negative emotion and overall emotion words.
Language Content Versus Style
Hundreds of studies have used LIWC (see Tausczik & Pennebaker, 2010a) to assess the
psychology of individuals over time or within and between groups. The word categories
that are reliably used by people of various groups can tell us how speakers or authors are
Page 4 of 21

attending to their topics and social worlds. A consistent finding is that many of the word
categories used to reliably classify psychological states can be considered to be a part of
language style as opposed to language content. That is, how people say things is often
more revealing that what they are saying. Language style can be detected through the
use of function words (also known as closed-class words), whereas language content is
largely made up of content words (also known as open-class words).
Function words include pronouns, prepositions, articles, conjunctions, and auxiliary

verbs, and account for more than half of the words we use in daily speech. Because rates
of function word use cannot be readily manipulated or consciously computed by listeners
of a conversation or readers of text, the computerized text analysis of conversation tran
scripts and correspondence can be considered a remote sensor of social dynamics. That
is, features of a relationship can be inferred based on the computation of spoken/written
words without having to directly ask the interactants, who may be inaccessible for self-re
port surveys, biased, or lacking awareness. So what can computerized text analyses tell
us about a relationship between interactants within transcripts of overheard or archived
communications?
Note that a growing body of research is showing that the rates of use of LIWC categories
have reliable relationships to demographics, personality, and health status (see Ireland &
Mehl, this volume). This chapter focuses on research studies that have shown what can
be learned about dyads, groups, and communities using computerized text analysis,
(p. 222) primarily through the use of LIWC. Much of the research using LIWC to analyze
communications has shown that function words are powerful markers of social processes.
For example, pronouns mark the degree to which an individual is paying attention to him
self and to his social world. Most natural language processing approaches have excluded
function words from analyses using a stoplist (i.e., a list of words, typically closed-class
words, that are filtered out from the processing of natural language data). Accordingly,
since the growth of LIWC, more studies have investigated function word use as a marker
of social processes than had previously been examined.
Language Style Matching
Another strategy that relies on the measurement of function words is language style
matching (LSM; Ireland & Pennebaker; 2010; Niederhoffer & Pennebaker, 2002). The
idea of LSM is that two or more people in an interaction should, in theory, match in their
levels of function words when interacting as they are establishing common ground. For
example, when two people are chatting about Jedidiah’s bouillabaisse, they can both use
“he” and “it” instead of each repeatedly having to say “Jedidiah” and “bouillabaisse.” The
more engaged they are with each other, the greater the synchrony. Although there are
several ways to measure LSM, our goal was to create a metric that people could calculate
without the use of a complex computer system. For example, if we have a telephone tran
script between two people, we could calculate the degree to which their personal pro
nouns matched with the following equation:
Page 5 of 21

The absolute difference between the personal pronouns assures that the number is al
ways positive, and the addition of .001 in the denominator guarantees no division by zero.
In addition, the fraction is subtracted from 1.0 so that LSM ranges from 0 to 1.0, with
higher numbers indicating greater matching. Note that the actual LSM coefficient is aver
aged over nine dimensions of function words (e.g., prepositions, articles, conjunctions,
negations, etc.).
The psychometrics of LSM are, in the view of the authors, elegant. Across multiple stud
ies, the LSM of one function word category is positively correlated with the LSM of the
other function word categories. Together, these yield internal reliabilities of the LSM sta
tistic ranging from .49 to .80 across multiple studies (Gonzales, Hancock, & Pennebaker,
2010; Ireland & Pennebaker, 2010; Ireland et al., 2011). Depending on the sample, we
find that LSM scores can range from as low as .30 for Wikipedia editor interactions
(Tausczik & Pennebaker, 2010b) to .93 for speed dating interactions (Ireland et al., 2011).
This range of LSM scores reflects, in part, the degree to which the group members are
actively engaged with one another in the here and now. Very low LSM scores among
Wikipedia editors reflect interactions that may take place across vast gaps in physical
space and time between questions and answers. Language style matching can be calcu
lated between any sets of text, including people in groups, written text such as emails, or
even between two articles by the same person.
It should be noted that LSM was compared with the more traditional latent semantic
analysis (LSA; Landauer et al., 1998). Latent semantic analysis compares the content sim
ilarity between text samples, whereas LSM focuses only on linguistic style. In one study
in which college students were asked to match the style of writing prompts to complete a
story, independent judges compared the language style between the prompts and the
writings. Interestingly, ratings of style similarity were correlated with content words as
measured by LSA coefficients and unrelated to style words as measured by LSM (Ireland
& Pennebaker, 2010). Thus, even judges are unable to detect writing style and naturally
confuse it with writing content.
Style matching appears to occur automatically, outside the control of the speaker or
writer, and is difficult to detect by independent judges. Nevertheless, it is a reliable, inter
nally consistent measure related to a small number of demographic factors but not reli
ably linked to traditional personality measures (Ireland & Pennebaker, 2010).
Page 6 of 21

Using Language Markers to Track Social and

Group Processes
The analysis of function words can be used to track and understand ongoing group dy
namics. Here, we review a diverse group of studies that explore language markers of sta
tus, longevity, cohesiveness, performance, and other features of group dynamics.
(p. 223) Status, Power, Leadership, and Social Hierarchies
Even in purely social online computer-mediated communication (CMC) pairs in which the
two interactants never see one another, the two people quickly come to agree who has
more status based on a few “innocent” questions: where are you from, what’s your major,
what year in school are you, and the like. More striking, however, is the degree to which
relative status is revealed through function words—especially pronouns. Within dyadic
and group interactions, the person with higher status talks more, uses fewer I-words,
more we-words and you-words, and fewer impersonal pronouns (Pennebaker, 2011). The
effect sizes are reasonably strong and consistent across studies (d’s range between .70
and 1.2; Kacewicz, Pennebaker, Burris, David, & Rodgers, 2011).
Comparable effects have been found for real-world, long-lasting groups. In an analysis of
an online collaboration of eighteen engineers, economists, technology experts, and oth
ers, Scholand and colleauges (2010) were able to track the language of the group over an
18-month period. During this time, the professional work group underwent several hard
ships, with the group eventually falling apart because of funding issues. Consistent with
the lab studies, relative use of pronouns in general and I-words in particular were strong
correlates of leadership in the group. Similar patterns were found in our analysis of docu
ments and memos between members of Saddam Hussein’s military and administration
(Hancock et al., 2010).
Why are I-words correlated with lower standing in the social hierarchy? One argument is
that the use of I-words reflects people’s focus of attention (Tausczik & Pennebaker,
2010a). People who are depressed or insecure have been found to be more self-focused in
general (e.g., Baddeley, Daniel, & Pennebaker, 2011; Rude, Gortner, & Pennebaker, 2004).
One can see similar patterns in people’s body language. When threatened or forced into a
subordinate position, humans and other social animals look away from the high-status in
dividual and often focus on their own body. As has been emphasized before, relative use
of I-words is likely reflecting the status hierarchy rather than influencing it.
Beyond counts of function words, Danescu- Niculescu-Mizil and colleagues (2012)

examined 240,000 Wikipedia discussions and found that lower status editors changed
their language more (i.e., showed higher LSM) to match their higher status counterparts.
Similar effects were reported in the same paper in an analysis of more than 50,000 con
versational exchanges in oral arguments before the U.S. Supreme Court, in which lawyers
matched their language more to the Chief Justice than to Associate Justices. Knowing
one’s standing in a social hierarchy is adaptive across social situations, and, indeed, many
Page 7 of 21

studies show that status seems to be reliably marked in language through the use of vari
ous function words.
Group Stability and Longevity
A commonly held belief is that the use of we-words is a reliable indicator of group solidar
ity and identity. Not so fast. Several studies have indeed demonstrated that, when inter
viewed by others, couples that use we-words when talking about their marriages are
more likely to remain married (Seider, Hirschberger, Nelson, & Levenson, 2009) or to evi
dence better physical health (Rohrbaugh, Mehl, Shoham, Reilly, & Ewy, 2008). Ironically,
this pattern does not always hold up in experimental groups (based on reanalyses of data
from Kacewicz et al., 2013; Gonzales et al., 2010) or even in young dating couples (Slatch
er & Pennebaker, 2006).
Across various contexts, from weekly therapy groups (Odom, 2008), to airline cockpit
crews (Sexton & Helmreich, 2000), to the Beatles 10-year collaboration (Petrie, Pen
nebaker, & Sivertsen, 2008), we-words are linked to group longevity. That is, the longer a
group has been together, the more the people in the group use we-words when talking
with each other (Chung & Pennebaker, 2012; Pennebaker, 2011). Note that this reflects a
more nuanced understanding of the use of we-words.
The use of we-words reflects a general feeling of group shared history and group longevi
ty. Although a sense of group identity may be important in bringing people together and
in giving them a sense of group solidarity, it is not clear that group identity is sufficient to
get a group to work well together. Ultimately, it may be more important that the group
members all think about their work in the same way—to have a shared mental model of
team- and task-based outcomes (Mathieu, Heffner, Goodwin, Salas, & Cannon-Bowers,
2000). To the degree that researchers seek to understand group processes, more dynamic
language measures must be employed.
Group Cohesiveness and Performance
The LSM approach is a potentially powerful method by which to tap ongoing group be
havior. In a series of studies in small-group behavior, (p. 224) one study examined whether
group-level LSM was related to group cohesiveness and performance. Approximately 300
students were assigned to same-sex groups of four to six people to work together on a
problem-solving task (Gonzales et al., 2010). Half of the groups worked together in the
same room (face-to-face, or FTF condition), and the remaining participants worked in in
dividual rooms using online chat technology (CMC condition).
Language analyses of the FTF and CMC groups revealed that LSM was positively related
to subsequent ratings of group cohesiveness. That is, the more the group members tend
ed to use function words at comparable levels, the more they felt a sense of support and
shared perspectives. Language style matching was positively correlated with group task-
based performance. Interestingly, this effect was stronger for the FTF than the CMC con
dition. Note that a similar pattern of effects was found in a reanalysis of data of forty-one
Page 8 of 21

four-person groups from a simulation study conducted at the University of Illinois Busi
ness School (Burris, Rodgers, Mannix, Hendron, & Oldroyd, 2009; also Kacewicz et al.,
2013). Similarly, LSM has been associated with more cooperation in an experimental so
cial dilemma game (Scissors, Gill, Geraghty, & Gergle, 2009).
Real-World Group Dynamics
Language style matching does not necessarily indicate cooperation, but engagement
more specifically. For example, it has been found that LSM was higher in the words ex
changed between hostage takers and hostage negotiators in real-world hostage negotia
tions with better outcomes (Taylor & Thomas, 2008). That is, the more that hostage tak
ers and negotiators were trying to gain common ground, the better the outcomes of the
hostage situations.
Another finding suggesting that LSM is a marker of engagement more broadly than mere
cooperation was in a study of emails by people who reported having kept a major life se
cret (any secret, as long as it was reported to be devastating to one’s self or to others if
the secret were to have gotten out; Tausczik, Chung, & Pennebaker, under review). We
examined email communication patterns and language between sixty-one secret keepers
with their secret targets (i.e., contacts from whom they were keeping the secret), confi
dants (i.e., contacts with whom they were confiding about the secret), and uninvolved
contacts (i.e., contacts who were not relevant to the secret). More than 59,000 messages
sent only between the participants and contact (i.e., no group emails) over the course of a
year (beginning 1 month before secret keeping) were analyzed to and from 705 corre
spondents. Each email was an average of seventy words long, and the data was processed
using LIWC.
Instead of disengaging from communication, there was a slight trend for secret keepers
to email slightly more and to show greater engagement with secret targets relative to be
fore secret keeping. That is, upon secret keeping, LSM increased to all contacts, but only
significantly to secret confidants and uninvolved contacts. This could be due to a goal to
“appear normal” or to engage more with secret-irrelevant topics for relief from the stress
of secret keeping.
Overall, analyzing language in emails proved to be a useful way of assessing how relation
ships in an individual’s network change, as opposed to previous studies on secrets that
have focused on self-reports by secret keepers or by secret targets to get at the social ef
fects of keeping a major life secret. The study enabled the analysis of actual communica
tions during a time when the secret was ongoing. We used an unobtrusive method in the
sense that there was no reactance by participants of being in the study and no behavior
or language changes due to having participated in the study.
Page 9 of 21

Real-World Close Relationships
Previous LSM projects focused on groups of varying size in which group-relevant out
comes were objectively measured. Language style matching has been used to predict out
comes in heterosexual speed dating. The study examined transcripts of forty men and
forty women who participated in twelve 4-minute interactions with members of the oppo
site sex (Ireland et al., 2011). The analyses showed that “matches” (in which both parties
indicated interest in dating) were far more likely if LSM during the speed dating interac
tions was above the median. Indeed, LSM measures predicted successful matches better
than the postinteraction ratings of the individuals. In other words, LSM could predict if
they would go out on a date better than the couples could themselves.
In another corpus of speed dates, Jurafsky, Ranganath, and McFarland (2009) analyzed
991 4-minute speed-dating sessions for rates of language use, prosody, and other features
of dialogue. In terms of language features, they found that judgments of speakers by
daters as both friendly and flirting were correlated with the use of “you” by males, and
“I” by females; men perceived by dates as awkward used (p. 225) significantly lower rates
of “you.” In another study on the same corpus, the authors (Ranganath, Jurafsky, & Mc
Farland, 2009) found that the pronoun cues were generally accurate: men who reported
flirting used more “you” and more “we,” among other features; women who reported flirt
ing used more “I” and less “we.” Similar to the function word findings in the previous
speed-dating studies, language analyses to detect self-reported intent to flirt were much
better than daters’ own perceptions of their speed-date’s flirting.
Language style matching has also been shown to be predictive of the long-term success of
people who were already dating. In a reanalysis of an older study (Slatcher & Pennebak
er, 2006), the instant messages (IMs) between eighty-six heterosexual romantic couples
were downloaded, and an LSM score was computed over 10 days of IMs. Language style
matching predicted the likelihood of a romantic couple being together 3 months later bet
ter than self-reported ratings of relationship stability (Ireland et al., 2011, study 2).
Finally, the LSM effects generalize to historical relationships based on archival records,
such as the correspondence between Sigmund Freud and Carl Jung, and the poetry of
Elizabeth Barrett and Robert Browning as well as Sylvia Plath and Ted Hughes (Ireland &
Pennebaker, 2010). For each of these pairs, LSM reliably covaried with relationship har
mony.
Summary
Across a wide variety of groups, tasks, goals, and outcome measures, the analysis of func
tion words has yielded theoretically and practically important findings. Relative use of
pronouns as well as measures of LSM have been shown to reflect feelings of group cohe
siveness, group task performance, and relationship longevity. Taken together, the findings
suggest that it is possible to capture the degree to which any two or more people are mo
tivated to be coordinated in their thinking styles.
Page 10 of 21

Note that the various findings reflect psychological state and social relationships rather
than cause them. That is, our few attempts to directly manipulate language have not suc
cessfully affected group dynamics. Nevertheless, the findings provide clues to how group
processes are unfolding and ways to devise more meaningful interventions.
Toward a Language-Based Group Intervention
Given the complexity of studying, measuring, and understanding group process, it is not
surprising that very few process-oriented studies have attempted to directly manipulate
coordinating mechanisms or other dynamic features of groups. Taking a broad view, there
is reason for optimism that group processes can be influenced by interventions of various
types. A meta-analysis of team training interventions showed that team training signifi
cantly improves cognitive, affective, process, and performance-based outcomes (Salas et
al., 2008). However, most team training interventions involve changes in instructions to
the group prior to interactions (e.g., Paulus & Yang, 2000), or feedback midway through
group interactions (Leshed, Hancock, Cosley, McLeod, & Gay, 2007). Only rarely have
studies attempted to unobtrusively provide continuous tailored feedback within ongoing
group interactions.
Several computer-mediated group communication tools using LIWC have been developed
to provide feedback to group members based on the language produced during a group
interaction. GroupMeter (Leshed et al., 2007) computes several dimensions of language
based on LIWC and presents feedback to group members midway through the group in
teraction using a graphical display. Leshed and colleagues found that the more assents
(e.g., “uhhuh, yeah, agree”) used in a group chat, the more likely group members were to
report feeling positive about the group. Indeed, another study found that business stu
dent groups were more likely to form a successful coalition in negotiation tasks (Huffaker,
Swaab, & Diermeir, 2011) if there were more assents during the tasks. Leshed and col
leagues also found that the more words typed, the greater the task focus of its members.
This finding was similar to that of Sexton and Helmreich (2000), who found that the more
that members spoke in flight simulation studies, the better the teams’ performance.
More recently, as part of large introductory psychology classes of 500 students each, an
online computer system, TOWER (Texas Online World of Educational Research; Pennebak
er, Ferrell, & Gosling, 2012), was developed for testing and for in-class discussions. The
program allows for people to meet in online virtual groups of two or more individuals
where they work together on assigned problems. At prespecified intervals (typically every
1–5 minutes), each person’s typed words are analyzed by LIWC. That is, the TOWER sys
tem computes real-time analyses about the ways group members use words. From those
analyses, TOWER can be programmed to then provide textual feedback to group mem
bers about their word use.
As an initial pilot of TOWER (Tausczik & Pennebaker, 2013), students were re
(p. 226)
quired to read a relatively short reading assignment that dealt with intergroup relations
and close relationships. They knew that they would be tested on this material before log
Page 11 of 21

ging onto the computer. At specified times, students entered the TOWER site using a se
cure id and password. Students were first directed to a 10-item multiple-choice test that
covered the assigned reading. No feedback was given. The students were then directed to
a TOWER “waiting room” page, where they were randomly assigned to groups that varied
in size ranging from two to four (mean group size = 2.87 people). Once in the virtual chat
room, students were expected to interact for 20 minutes and discuss each of the topics
that had been covered on the previous test. Overall, there were 186 groups (composed of
534 people, of which 61 percent were female) that completed the assignment without any
technical problems.
After interacting for 20 minutes, students were directed to the post-test where each indi
vidual completed a second ten-item exam that covered the same basic material as the
first exam, although the questions were different. Following the exam, students complet
ed a series of questions about the group dynamics. Changes in performance from the first
to the second exam served as the performance outcome. Ratings of group liking and co
hesiveness captured group climate.
Drawing on the group process literature together with previous research on function
words, four dimensions were targeted: positivity (as measured by the use of LIWC’s posi
tive emotion words and exclamation points minus negative emotion words), equality of
participation (as measured by the Gini coefficient of its group members’ word counts, as
in Kittur & Kraut, 2008), engagement (as measured by LSM), and information exchange
(as measured by a composite of task-relevant words, articles, and word count). A single
measure was constructed by taking the sum of z-scores for the component language vari
ables.
The results were encouraging. The self-reported group climate measures were signifi
cantly positively correlated with the language measures of positivity, engagement, and
equality of participation, whereas only the information exchange variable was related to
improvements in grades.
In a follow-up pilot study with feedback, the LIWC program was used in TOWER to ana
lyze individual and group data once every 3 minutes and provide general feedback to the
group concerning the group’s overall LSM. If the group’s LSM score was 0.86 or higher,
everyone received a message that said, “Your group is working well together. Everyone
seems to be paying attention to the topic.” If the group’s LSM was quite low (below 0.65),
all members read, “Your group is not working very well right now. Pay more attention to
each other and listen to what the others are saying.”
In addition to the group feedback, individual members would occasionally receive feed
back that no one else saw. For example, if there were four people in the group, in which it
was assumed that each person should contribute 25 percent of the words, anyone who
spoke more than 40 percent received a message that read, “You are talking more than
your share. Encourage others to say more and listen to what they say.” Those who said
Page 12 of 21

fewer than 10 percent of the words received the message, “You are talking much less
than you should. Get involved and tell people your views.”
The primary goal in these pilot studies was to determine if a real-time feedback program
was feasible, user friendly, and could yield meaningful feedback for participants. Students
reported liking the system and generally found it helpful. Unfortunately, with no control
group, it was impossible to know if the test performance gains were meaningful. Since
this original pilot, the TOWER system has been upgraded so that it is faster, more reli
able, and can provide ongoing feedback for any LIWC or other language dimensions as
needed.
Conclusions and Future Directions

For the past 50 years, remarkable progress has been made in our understanding of social
and group processes. Most work has helped to delineate the factors that affect groups
(i.e., the input level of understanding) as well as the best ways to measure the outputs of
groups. Due to the enormous problem of tracking ongoing communication among groups,
research on group process has been particularly difficult.
Much of the work on identifying the roles of function words in social processes has been
at the individual as opposed to the group level (see Ireland & Mehl, this volume). Only re
cently has research shown the powerful role that function word analysis can contribute to
understanding group processes. For example, language can tap relative status, thinking
styles, and other socioemotional and cognitive concepts that have direct relevance for
studying group climate and performance. What is less clear is identifying precisely what
language features may best reflect successful group processes.
(p. 227) Currently, pilot studies have measured several group processes that can be mea
sured in word use: positivity, engagement, equality of participation, and information ex
change. Along with considerations for a more measurement-driven approach, a more
grounded theoretical strategy also makes sense. Specifically, constructs within theories
that are posited to be linked to successful group interactions should be linked to lan
guage features. All of these features should be tested with large and diverse samples, and
targeted features should be manipulated to see if it is possible to bring about changes in
group behavior.
Beyond finding the active ingredients or words that reflect social dynamics, one of the
great challenges in providing feedback to interactants will be in balancing the cognitive
load of feedback without detracting from engagement with the interaction itself. For ex
ample, if men in online speed dates are asked to attend more to their interaction partner
to increase their use of “you” and be perceived as more flirtatious and less awkward (as
found in Jurafsky et al., 2009), they may appear to be more awkward when faced with the
challenge of dividing their attention between their dates and the criticism and pressure to
be more suave. Similar challenges of divided attention and acceptance of feedback will be
Page 13 of 21

present in other real-time feedback technologies, for example, in online therapies, foren
sic interviews, or other social interactions.
Much research remains to be done in implementing feedback technologies. One chal

lenge for future research will be to encourage researchers to examine open-ended natu
ralistic language use. There are archives of naturalistic communication, and many more
technologies are available for data mining and analyzing language. Researchers should
take up the challenge of thinking about how to scale their current methods to be applied
to big data and how their questions might be analyzed and their findings applied in real-
time interactions.
Uncovering the psychometrics and social meanings of function words promises new ways
of thinking in the fields of psychology and communications. The computerized analysis of
words enables researchers to examine social processes as they unfold in an ecologically
valid way, using retrospective (archival) and prospective designs. More exciting is that
with simple-to-use text analysis tools and so many sources of natural (and increasingly
digital) language to explore, we are at the frontier of a new way to understand and poten
tially improve the ways people communicate and interact.
One of the exciting features of this new language approach is that it opens the door to a
number of new and exciting questions. Three promising ones include:
How can social psychology broaden its worldview? We know the answer: begin analyzing
naturally produced text using some revolutionary methods being developed in many other
disciplines. Although most of the tools and methods described in this chapter are relative
ly easy for social scientists to use, there are more mathematically involved approaches to
natural language developed by computer scientists, computational linguists, and engi
neers in the research area of machine learning. Given that machine learning experts are
interested in developing tools for data mining, analyzing, and building platforms for inter
actions involving language, a social science perspective can greatly enhance their work
by offering a more meaningful perspective on how people communicate and interact. At
the same time, the data and statistical expertise of more computational researchers can
enhance the scale of a social science research study and provide insights into applica
tions that social scientists may not have considered.
In an era in which psychologists have become mesmerized by functional magnetic reso

nance imaging (fMRI), bioinformatics, and other computationally heavy approaches, so
cial psychologists are only beginning to appreciate the ways we can understand social
processes through computerized text analyses.
How can computerized text analysis methods generalize across languages and culture?
Function words exist in all major languages that we have studied, and it is likely that
function words in other languages reflect social processes. For example, Asian languages
such as Japanese and Korean have specific morphemes and pronouns that mark relative
status. Many European languages have a formal versus informal second person singular
Page 14 of 21

pronoun, “you.” The Southern “y’all” is similar to the Northeastern “youse” or the more
formal American English “all of you.”
Given that computerized text analysis tools are available in many languages, including
translated versions of LIWC, it is now possible for social scientists to study social process
es across cultures through interactants speaking their native language. Much work is
needed to first validate the language features that reflect social processes in other lan
guages and cultures. It will then be possible to begin to understand social processes in
groups speaking another (p. 228) language. Future research will likely yield important dis
coveries about cross-cultural thinking through the use of computerized text analysis
methods.
What is the link between language use and psychological states? Does language reflect or
dictate psychological states? The answer is probably both. However, much of the lan
guage research dealing with function words in psychology indicates that language re
flects rather than influences an individual’s psychological state. Note also that language is
used for interpersonal communication, not only to reflect to others your own psychologi
cal state, but also to affect others’ psychological states. Accordingly, language has the po
tential to influence the psychological states of others and of a group’s dynamics.
Although it is difficult to instruct people to change their rates of function word use (for
example, consider the next to impossible task of trying to use 3.5 percent more third-per
son pronouns tomorrow than you did today), it is possible to give people instructions on
how to change their perspectives or behaviors. By doing so, language markers change,
and that is an objectively measurable change. Finding systematic relationships among
language, psychological states, and behaviors are the first steps in determining how one
might be able to change perspectives and behaviors in online therapies, e-health, or in
new social and organizational communication technologies.
Getting in on the Next Big Thing: Some Realis

tic Advice
We hope you can appreciate why we are so enthusiastic about text analysis as a viable
method for the social sciences in general and social psychology in particular. The analysis
of words can reveal information about personality, basic social processes, culture, biology,
just about anything. But then, the same has been said for self-reports, the IAT, fMRI, psy
choneuroimmunology, Baron & Kenny, and sliced pizza (we are not certain about the piz
za claim). Despite the apparent promise of text analysis, you should realistically consider
it just another tool in our toolbox of methods to understand social behavior.
As more researchers jump onto the text analysis bandwagon, they will quickly learn that
language use is highly variable across people and contexts. The effect sizes are often very
small and only apparent with very large samples. Many of the findings are highly depen
dent on dictionary words that need frequent updates or that are not relevant to different
Page 15 of 21

media, languages, or statistical approaches. If you are a researcher who is considering a

computerized text analysis method, keep your expectations in check. You will not find the
secret to the universe or even the key to understanding what makes people tick. You may,
however, stumble across a finding that may change the way you think about how people
interact and influence one another.
Acknowledgments
Correspondence should be addressed to either of the authors at Department of Psycholo
gy A8000, University of Texas at Austin, Austin, Texas 78712 or, via email, to
[email protected] or [email protected]. Preparation of this manu
script was aided by funding from the Army Research Institute (W91WAW-07-C-0029) and
the National Science Foundation (NSCC-0904913; BCS-1228693). The authors wish to
thank Yla R. Tausczik, Yitai Seih, and Jason D. Ferrell for their helpful comments.
Financial and Disclosure Notes: The LIWC2007 program, co-owned by Pennebaker, is

commercially available for $89 USD (full package), $29 USD (student version), with dis
counts for bulk purchases on www.liwc.net. LIWC2007 demos, downloads, and products
can be found on www.liwc.net. Text data for research purposes will be analyzed by Pen
nebaker free of charge. All profits that go to Pennebaker from LIWC2007 sales are donat
ed to the University of Texas at Austin Psychology Department.
References
Baddeley, J. L., Daniel, G., & Pennebaker, J. W. (2011). How Henry Hellyer’s language use
foretold his suicide. Journal of Crisis Intervention and Suicide Prevention, 32, 288–292.
doi: 10.1027/0227-5910/a000092
Beebe, S. A., & Masterson, J. T. (1997). Communicating in small groups: Principles and
practices (5th ed.). New York: Longman.
Burris, E. R., Rodgers, M. S., Mannix, E. A., Hendron, M. G., & Oldroyd, J. B. (2009). Play
ing favorites: The influence of leaders’ inner-circle on groups processes and performance.
Personality and Social Psychology Bulletin, 35, 1244–1257. doi:
10.1177/0146167209338747
Chung, C. K., & Pennebaker, J. W. (2008). Revealing dimensions of thinking in open-ended

self-descriptions: An automated meaning extraction method for natural language. Journal
of Research in Personality, 42(1), 96–132. doi:10.1016/j.jrp.2007.04.006
Chung, C. K., & Pennebaker, J. W. (2012). Linguistic Inquiry and Word Count (LIWC): Pro
nounced “Luke”…and other useful facts. In P. M. McCarthy & C. Boonthum (Eds.), Ap
plied natural language processing and content analysis: Identification, investigation, and
resolution. Hershey, PA: IGI Global.
Page 16 of 21

Danescu-Niculescu-Mizil, C., Lee, L., Pang, B., & Kleinberg, J. (2012). Echoes of power:
Language effects and power (p. 229) differences in social interaction. In Proceedings of
the 21st International World Wide Web Conference. New York: ACM.
Foltz, P. W., & Martin, M. J. (2009). Automated communication analysis of teams. In E.

Salas, G. F. Goodwin, & C. S. Burke (Eds.), Team effectiveness in complex organizations:
Cross-disciplinary perspectives and approaches (pp. 411–431). New York: Routledge, Psy
chology Press.
Gonzales, A. L., Hancock, J. T., & Pennebaker, J. W. (2010). Language indicators of social
dynamics in small groups. Communications Research, 37, 3–19. doi:
10.1177/0093650209351468
Graesser, A. C., Lu, S., Jackson, G. T., Mitchell, H., Ventura, M., Olney, A. M., et al. (2004).
AutoTutor: A tutor with dialogue in natural language. Behavioral Research Methods, In
struments & Computers, 36, 180–193. doi: 10.3758/BF03195563
Guzzo, R. A., & Dickson, M. W. (1996). Teams in organizations: Recent research on perfor
mance and effectiveness. Annual Review of Psychology, 47, 307–388. doi: 10.1146/
annurev.psych.47.1.307
Hackman, J. R., & Morris, C. G. (1975). Group tasks, group interaction process, and group
performance effectiveness: A review and proposed integration. In L. Berkowitz (Ed.), Ad
vances in experimental social psychology (vol. 8, pp. 45–49). New York: Academic Press.
Hancock, J. T., Beaver, D. I., Chung, C. K., Frazee, J., Pennebaker, J. W., Graesser, A. C., &
Cai, Z. (2010). Social language processing: A framework for analyzing the communication
of terrorists and authoritarian regimes. Behavioral Sciences in Terrorism and Political Ag
gression, Special Issue: Memory and Terrorism, 2, 108–132. doi:
10.1080/19434471003597415
Hansen, M. T. (2002). Knowledge networks. Organization Science, 13, 232–248. doi:

10.1287/orsc.13.3.232.2771
Hare, A. P. (1962). Handbook of small group research. New York: Free Press.
Harris, T. E., & Sherblom, J. (2008). Small group and team communication. Boston: Allyn
& Bacon.
Hossain, L., Wu, A., & Chung, K. K. S. (2006). Actor centrality correlates to project based
coordination. In Proceedings of Computer Supported Cooperative Work 2006 (CSCW06)
(pp. 363–372). New York: ACM.
Huffaker, D., Swaab, R. I., & Diermeier, D. (2011). The language of coalition formation in
multiparty online negotiations. Journal of Language and Social Psychology, 30, 66–81.
Ireland, M. E., & Pennebaker, J. W. (2010). Language style matching in reading and writ
ing. Journal of Personality and Social Psychology, 99, 549–571. doi: 10.1037/a0020386
Page 17 of 21

Ireland, M. E., & Matthias, M. R. (this volume). Natural language use as a marker of per
sonality. In T. Holtgraves (Ed.), Oxford handbook of language and social psychology. New
York: Oxford University Press.
Ireland, M. E., Slatcher, R. B., Eastwick, P. W., Scissors, L. E., Finkel, E. J., & Pennebaker,
J. W. (2011). Language style matching predicts relationship formation and stability. Psy
chological Science, 22, 39–44. doi: 10.1177/0956797610392928
Jonassen, D. H., & Kwon, H. (2001). Communication patterns in computer mediated ver
sus face-to-face group problem solving. Educational Technology Research and Develop
ment, 49, 35–51. doi: 10.1007/BF02504505
Jurafsky, D., Ranganath, R., & McFarland, D. A. (2009). Extracting social meaning: Identi
fying interactional style in spoken conversation. In Proceedings of the 2009 Annual Con
ference of the North American Chapter of the Association for Computational Linguistics—
Human Language Technologies (NAACL-HLT2009) (pp. 638–646). Stroudsburg: Associa
tion for Computational Linguistics.
Kacewicz, E., Pennebaker, J. W., Davis, M., Jeon, M., & Graesser, A. C. (2013). Pronoun
use reflects standings in social hierarchies. Journal of Language and Social Psychology,
33, 125–143. doi: 10.11777/0261927X13502654.
Kittur, A., & Kraut, R. E. (2008). Harnessing the wisdom of crowds in Wikipedia: Quality
through coordination. In Proceedings of CSCW 2008 (pp. 37–46). San Diego, CA: ACM.
Landauer, T. K., Foltz, P. W., & Laham, D. (1998). An introduction to latent semantic analy
sis. Discourse Processes, 25, 259–284. doi: 10.1080/01638539809545028
Leavitt, H. J. (1951). Some effects of certain communication patterns on group perfor

mance. Journal of Abnormal Psychology, 46, 38–50. doi: 10.1037/h0057189
Leshed, G., Hancock, J. T., Cosley, D., McLeod, P. L., & Gay, G. (2007). Feedback for guid
ing reflection on teamwork practices. In Proceedings of the 2007 International ACM Con
ference on Supporting Group Work. New York: ACM.
Losada, M., & Heaphy, E. (2004). The role of positivity and connectivity in the perfor
mance of business teams: A nonlinear dynamics model. American Behavioral Scientist, 47,
740–765.
Mathieu, J. E., Heffner, T. S., Goodwin, G. F., Salas, E., & Cannon-Bowers, J. A. (2000). The
influence of shared mental models on team process and performance. Journal of Applied
Psychology, 85, 273–283. doi: 10.1037/0021-9010.85.2.273
McGrath, J. E. (1991). Time, interaction, and performance (TIP): A theory of groups. Small
Group Research, 22, 147–174. doi: 10.1177/1046496491222001
Page 18 of 21

Niederhoffer, K. G., & Pennebaker, J. W. (2002). Linguistic style matching in social interac
tion. Journal of Language and Social Psychology, 21, 337–360. doi:
10.1177/026192702237953
Odom, S. D. (2006). A qualitative and linguistic analysis of an authority issues training

group. Unpublished doctoral dissertation, University of Texas at Austin, TX.
Paulus, P. B., & Yang, H. C. (2000). Idea generation in groups: A basis for creativity in or
ganizations. Organizational Behavior and Human Decision Processes, 82, 76–87. doi:
10.1006/obhd.2000.2888
Pennebaker, J. W. (2011). The secret life of pronouns: What our words say about us. New
York: Bloomsbury Press.
Pennebaker, J. W., Booth, R. J., & Francis, M. E. (2007). Linguistic inquiry and word count
(LIWC2007): A text analysis program. Austin, TX: LIWC.net.
Pennebaker, J. W., & Chung, C. K. (in press). Counting little words in big data: The psy
chology of communities, culture, and history. In J. László, J. Forgas, & O. Vincze (Eds.), So
cial cognition and communication. New York: Psychology Press.
Pennebaker, J. W., Chung, C. K., Ireland, M. E., Gonzales, A. L., & Booth, R. J. (2007). The
development and psychometric properties of Linguistic Inquiry and Word Count
(LIWC2007). Austin, TX: www.liwc.net.
Pennebaker, J. W., Ferrell, J. D., & Gosling, S. D. (2012). Texas Online World of Educational
Research (TOWER). Unpublished manuscript.
Petrie, K. J., Pennebaker, J. W., & Sivertsen, B. (2008). Things we said today: A linguistic
analysis of the Beatles. Psychology of Aesthetics, Creativity, and the Arts, 2, 197–202. doi:
10.1037/a0013117jur
Ranganath, R., Jurafsky, D., & McFarland, D. A. (2009). It’s not you, it’s me: Detecting
flirting and its misperception (p. 230) in speed-dates. In Proceedings of the 2009 Confer
ence on Empirical Methods in Natural Language Processing (pp. 334–342). Stroudsburg:
Association for Computational Linguistics.
Rohrbaugh, M. J., Mehl, M. R., Shoham, V., Reilly, E. S., & Ewy, G. A. (2008). Prognostic
significance of spouse “we” talk in couples coping with heart failure. Journal of Consult
ing and Clinical Psychology, 76, 781–789. doi: 10.1037/a0013238
Rude, S. S., Gortner, E. M., & Pennebaker, J. W. (2004). Language use of depressed and
depression-vulnerable college students. Cognition and Emotion, 18, 1121–1133. doi:
10.1080/02699930441000030
Salas, E., DiazGrandados, D., Klein, C., Burke, C. S., Stagl, K. C., Goodwin, G. F., &
Halpin, S. M. (2008). Does team training improve team performance? A meta-analysis.
Human Factors, 50, 903–933. doi: 10.1518/001872008X375009
Page 19 of 21

Salas, E., Goodwin, G. F., & Burke, C. S. (2009). Team effectiveness in complex organiza
tions: Cross-disciplinary perspectives and approaches. New York: Routledge/Taylor &
Francis Group.
Scholand, A. J., Tausczik, Y. R., & Pennebaker, J. W. (2010). Linguistic analysis of work
place computer-mediated communication. In Proceedings of Computer Supported Cooper
ative Work 2010. New York: ACM.
Scissors, L. E., Gill, A. J., Geraghty, K., & Gergle, D. (2009). In CMC we trust: The role of
similarity. In Proceedings of CHI 2009 (pp. 527–536). New York: ACM.
Seider, B. H., Hirschberger, G., Nelson, K., & Levenson, R. W. (2009). We can work it out:
Age differences in relational pronouns, physiology, and behavior in marital conflict. Psy
chology and Aging, 24, 604–613. doi: 10.1037/a0016950
Sexton, J. B., & Helmreich, R. L. (2000). Analyzing cockpit communications: The links be
tween language, performance, error, and workload. Human Performance in Extreme Envi
ronments, 5, 63–68.
Slatcher, R. B., & Pennebaker, J. W. (2006). How do I love thee? Let me count the words:
The social effects of expressive writing. Psychological Science, 17, 660–664. doi: 10.1111/
j.1467-9280.2006.01762.x
Steyvers, M., & Griffiths, T. (2007). In T. Landauer, D. McNamara, S. Dennis, & W. Kintsch
(Eds.), Latent semantic analysis: A road to meaning (pp. 427–448). Mahwah, NJ:
Lawrence Erlbaum Associates.
Tausczik, Y. R., Chung, C. K., & Pennebaker, J. W. (under review). Tracking secret-keeping
in emails.
Tausczik, Y. R., & Pennebaker, J. W. (2010a). The psychological meaning of words: LIWC
and computerized text analysis methods. Journal of Language and Social Psychology, 29,
24–54. doi:10.1177/0261927X09351676
Tausczik, Y. R., & Pennebaker, J. W. (2010b). Wikipedia editors in collaborative working

groups. Unpublished manuscript.
Tausczik, Y. R., & Pennebaker, J. W. (2013). Improving teamwork using real-time language
feedback. In Proceedings of the ACM 2013 conference on Computer-Human Interaction.
New York: ACM.
Taylor, P. J., & Thomas, S. (2008). Linguistic style matching and negotiation outcome. Ne
gotiation and Conflict Management Research, 1, 263–281. doi: 10.1111/j.
1750-4716.2008.00016.x
Tichy, N. M., Tushman, M. L., & Fombrun, C. (1979). Social network analysis for organiza
tions. The Academy of Management Review, 4, 507–519.
Page 20 of 21

West, B. J., Patera, J. L., & Carsten, M. K. (2009). Team level positivity: Investigating posi
tive psychological capacities and team level outcomes. Journal of Organizational Behavior,
30, 249–267. doi: 10.1002/job.593.
Cindy K. Chung
Cindy K. Chung, Department of Psychology, The University of Texas at Austin.
James W. Pennebaker
James W. Pennebaker is a professor in the Department of Psychology at the Universi

ty of Texas at Austin in Austin, TX.
Page 21 of 21

Language and Social Comprehension

Robert S. Wyer Jr.

Online Publication Date: Apr 2014 DOI: 10.1093/oxfordhb/9780199838639.013.029
This chapter reviews the factors that influence the comprehension messages that are con
veyed in the course of social communication, focusing on both informal conversation and
the information conveyed in the media (newspapers and advertisements. It also considers
the implications of these comprehension processes for both memory of the information
and judgments based on it. The modality of the information, the manner in which it is
conveyed, and the social context in which it occurs are discussed.
Keywords: Comprehension, conversations, memory, language
In discussing a painting in The Portrait of the Artist as a Young Man, James Joyce noted
that only a fraction of what an artist intends to express in a painting is actually transmit
ted to the people who view it. That is, some of the meaning that the artist imparts to the
canvas remains there undiscovered. At the same time, viewers often extract meaning
from the painting that the artist did not intend to transmit. So it is with social communi
cations in general.
Language is obviously fundamental to social communication. However, an understanding

of its role is by no means straightforward. For one thing, information is often transmitted
through bodily movements, facial expressions, and tone of voice, and these nonverbal be
haviors can qualify the meaning of the verbal content they accompany. Even leaving these
considerations aside, however, the meaning extracted from a communication cannot be
inferred from its semantic implications alone. It can depend on the structure and form in
which the message is presented and the context in which the message is encountered (an
informal conversation, a newspaper or television commercial, etc.). Moreover, as Joyce’s
observation suggests, it is affected by characteristics of the recipients themselves, includ
ing their prior knowledge about the topic, the purpose for which they expect to use the
information contained in the communication, and the processing strategy they use to ex
tract the message’s implications (e.g., the use of visual imagery).
Page 1 of 28

Social communications include all messages that are constructed by a specific source
with the intention of conveying information to a specified audience. Thus, they include
not only communications exchanged in face-to-face conversations or through e-mail but
also messages encountered in the media (e.g., newspapers or advertisements) that are di
rected to a subset of the general population. (In contrast, encyclopedias, which are in
tended to preserve archival knowledge, or novels, which are written without any particu
lar audience in mind, would not normally fall into this category.)
A consideration of the factors that influence the meaning transmitted in a social commu
nication makes salient the distinction between a communication’s literal meaning and
both (a) the meaning that the communicator actually intends to convey (p. 234) and (b) the
meaning that the recipient perceives to be intended. The factors that influence percep
tions of a communicator’s intended meaning are the subject of extensive theory and re
search in both psycholinguistics and social psychology (e.g., Grice, 1975, 1989; Higgins,
1981; Holtgraves, 2002; Krauss & Fussell, 1996; Wyer, 2004). In this chapter, I restrict at
tention to a relatively small subset of issues that are not often addressed in other discus
sions of these factors. In the first section, I consider characteristics that influence percep
tions of a verbal communication’s literal meaning, focusing on the mental representations
constructed in the course of construing a message’s implications, the processing strategy
that recipients employ, and the effects of these construals on judgments and decisions. In
the second section, I review the factors that affect recipients’ perception of the intended
meaning of social communications that are conveyed not only in informal conversations
but also in the media and in psychological experiments.
Perceptions of Literal Meaning

The Role of Visual Imagery in Language Comprehension
Individuals who receive a verbal statement are likely to interpret it in terms of semantic
concepts that are accessible in memory at the time. Thus, they may interpret a person’s
statement that he plans to cross the Atlantic in a sailboat as either “adventurous” or
“foolhardy,” depending on which concepts come to mind first. Because the effects of con
struct accessibility on the interpretation of information are well-established and have
been reviewed in detail elsewhere (Förster & Liberman, 2007; Higgins, 1996; Wyer,
2008), they will not be reiterated.
However, the interpretation of verbal statements in terms of abstract concepts does not
normally occur until a second stage of comprehension, after the statements’ components
and their relationship have been interpreted in terms of concepts at a concrete, perceptu
al level. For example, comprehension of the statement “John set fire to the cat’s tail” as
“sadistic” requires a prior encoding of the statement’s referents in terms of action and
noun concepts as well as the relation between them. The application of more abstract
(e.g., attribute) concepts may not be applied unless the concepts are necessary to attain a
specific objective that the recipient has at the time (e.g., to form an impression of the per
Page 2 of 28

son or event being described or to make a behavioral decision). Otherwise, recipients

may not spontaneously engage in this higher order comprehension process. Rather, they
may construct a concrete mental simulation of the event or situation described (e.g., a sit
uation model; see Radvansky & Zacks, 1991; Wyer & Radvansky, 1999). This representa
tion may consist in part of a visual image.
Visual images are, of course, hypothetical constructs and cannot be directly observed.
However, their importance in language comprehension seems incontrovertible. Their role
was compellingly demonstrated many years ago by Bransford, Barclay, and Franks
(1972). They found, for example, that apparently anomalous sentences (e.g., “the
haystack was important because the cloth would rip”) could not be remembered when
they were presented out of context, presumably because a coherent mental representa
tion of the situation they depicted could not be spontaneously constructed. However, the
sentences could be easily understood and consequently recalled when they were preced
ed by a single word (e.g., “parachute”). The cue word apparently activated a complex
body of knowledge that elicited a mental image of the situation described and the circum
stances surrounding it (e.g., an image of a parachutist landing in the hay rather than in a
tree). As this example indicates, the comprehension of a verbal statement does not simply
involve the application of semantic concepts to the individual components of the state
ment. In addition, it involves the construction of a more complex mental image of the situ
ation to which the statement pertains.
Other evidence that visual images are formed spontaneously in the course of compre
hending verbal information is also compelling (e.g., Black, Turner, & Bower, 1979; Gar
nham, 1981; Glenberg, Meyer, & Lindem, 1987). As but one example, participants in a
study by Glenberg et al. (1987) read a story about a man’s activities while going jogging.
In one condition, the story began, “John put on his sweatshirt and went jogging.” In a sec
ond, it began, “John took off his sweatshirt and went jogging.” The sweatshirt was never
again mentioned. However, participants apparently formed a mental representation of
John engaging in the events described in the story that included a visual image of him
with his sweatshirt on in the first case but not in the second. This image was retained
throughout the story. Consequently, in a later recognition memory task, participants were
significantly quicker to identify “sweatshirt” as having been presented in the first case.
Radvansky and Zacks (1991; see also Johnson- Laird, 1983, 1989) assume that when the
situation (p. 235) or event described by a statement is situationally and temporally con
strained (i.e., it occurs at a specific, although not necessarily specified, time and place),
individuals form a mental simulation of it (a situation model) spontaneously. Moreover, if
the implications of two or more statements can be localized in the same time and place,
they form a single model that includes them all. (For example, people form a single model
of “The book is on the table. The cup is on the table,” but form separate models of “The
book is on the table. The book is on the chair”; see Radvansky & Zacks, 1991). On the oth
er hand, if the situation described in the statement is not temporally or spatially con
strained (e.g., “The man owns a car;” “honesty is the best policy”), the statement is se
mantically coded but no mental image is formed. Similar considerations surround the
Page 3 of 28

comprehension of statements about behavioral events (Radvansky, Wyer, Curiel, & Lutz,
1997).
A theory of social comprehension by Wyer and Radvansky (1999; Wyer, 2004) formalizes
these processes. According to this theory, individuals first encode information they re
ceive in the modality in which it is presented. That is, they encode observed or pictured
situations in visual terms and verbal descriptions semantically. Then, if the event or situa
tion described by the verbally coded information is temporally and situationally con
strained, they spontaneously construct a visual representation of it. However, the reverse
is not true. That is, if the visual coding of a picture or observation is sufficient to compre
hend it, they store this coding in memory without assigning verbal labels to it.
Effects of Visual Perspective on Comprehension and Judgments

Visual images of a situation, like pictures of it, are necessarily formed from a particular
visual perspective (e.g., the perspective of someone in the situation or an outside observ
er, or the perspective of someone close to the object or at a distance). These perspectives
can be elicited by the language used to describe the situation. Black, Turner, and Bower
(1979), for example, found that statements of the form “Mary was reading in her room.
John came in to talk to her” are comprehended more easily than the statements “Mary
was reading in her room. John went in to talk to her.” This is because the event described
in the first statement is imagined from the perspective of someone in the room with Mary,
and the second statement is easier to comprehend if the event it describes was imagined
from the same perspective than if a shift in perspective is required.
Individual differences in the disposition to form images were identified by Jiang and Wyer
(2009) in a series of studies that also took into account the familiarity of the perspective
from which images are constructed. Male and female participants read a series of single
statements with instructions to press a button on the computer keyboard to indicate
whether the statements were comprehensible. Four target statements were embedded in
the series: “A person went into the men’s room,” “A person went into the ladies’ room,” “A
person came into the men’s room,” and “A person came into the ladies’ room.” The se
mantic implications of the first two statements are the same as the implications of the last
two. However, the visual images elicited by the two sets of statements depend on the per
spective from which the images elicited by the first two statements are formed (i.e., that
of someone inside the room or someone outside). Moreover, the familiarity of the images
formed from the perspective of someone inside the room depends on participants’ sex.
In Jiang and Wyer’s (2009) study, participants were classified as chronic visual informa
tion processers (“visualizers”) or chronic verbal information processers (“verbalizers”) on
the basis of their responses to Childers, Houston, and Heckler’s (1985) style-of-processing
scale. Their ease of comprehending the target statements was inferred from the time they
took to respond to them. Male and female visualizers found it easy to comprehend state
ments that a person went into the men’s or ladies’ room. However, male visualizers took
more time to comprehend statements that a person came into the ladies’ room than state
ments that a person came into the men’s room (2.81 s vs. 2.00 s) whereas female visualiz
Page 4 of 28

ers took less time to comprehend the former statements than the latter (2.23 s vs. 2.62 s).
In contrast, individuals with a disposition to process information verbally without con
structing visual images did not differ in the time they required to comprehend the state
ments.
The impact of these dispositional influences on the comprehension of information is also

evident in the judgments that are based on it. In a second study, Jiang and Wyer (2009)
asked participants to read descriptions of positive and negative events and to indicate
their emotional reactions to them. Some statements were written from the perspective of
someone in the situation described (e.g., “A man came into the restaurant and kissed his
girlfriend,” (p. 236) “A drunk came into the kitchen and threw up on the floor,” etc.). In
other cases, they were written from the perspective of someone outside the situation (“A
man went…,” “A drunk went…”). Chronic visualizers had more extreme positive or nega
tive reactions to statements that were written from the perspective of someone in the sit
uation than when they were written from the perspective of someone outside, whereas
chronic verbalizers had similar reactions in both cases.
Situational Differences in Processing Strategy

Differences in disposition to form visual images can also be influenced by situational fac
tors that are unrelated to the information to be comprehended. Moreover, the effects of
these differences are similar to the effects of chronic differences. In a study by Jiang et al.
(2010; see Wyer, Hung, & Jiang, 2008), some participants’ chronic dispositions to engage
in visual or verbal processing was assessed using the Childers et al. (1985) scale. Other
participants’ processing strategy was induced by asking them to perform either a hidden
figures task or a hidden word task, respectively. Then, as part of a different experiment,
participants read information about the attributes of a computer mouse. The attributes
were the same in all conditions. However, the product was described as either a standard
mouse with which participants were familiar or a trackball mouse with which they were
not. Individuals who were disposed to form visual images found it difficult to accomplish
this when the product was unfamiliar and a previously formed visual representation of it
did not exist in memory. Consequently, they evaluated the unfamiliar mouse less favorably
than the standard one. However, participants with a disposition to process information
verbally evaluated the familiar and unfamiliar products similarly. These effects were virtu
ally identical regardless of whether these dispositions were chronic or situationally in
duced.
Interference Effects of Processing Strategies

When verbal and visual processing strategies are activated in close temporal proximity,
they may conflict, and processing effectiveness suffers. An experiment by Dodson, John
son, and Schooler (1997) provides an example. Participants were exposed to a series of
faces. Then, they were asked to describe one of the faces verbally. Finally, they were giv
en a recognition memory task in which they were asked to identify both the face they had
written about and the other faces they had seen Relative to control conditions, writing
about a face decreased the recognition of not only this face but also other faces that par
Page 5 of 28

ticipants had seen but had not written about. Apparently, the task of describing the face
in writing activated a verbal processing strategy that interfered with the visual process
ing strategy required for effective recognition, and this interference was evident in re
sponses to both the face they had written about and the others.
Linguistic Determinants of Mental Representation
As Jiang and Wyer’s (2009) study suggests, the mental representation that individuals
form when comprehending verbal descriptions can depend on subtle differences in the
way the descriptions are worded. Other research provides further evidence of these dif
ferences. A classic study by Loftus and Palmer (1974) is an example. Participants who had
been shown a picture of the consequences of an automobile accident were later asked to
estimate either how fast the car was going when it “hit the tree” or, alternatively, how fast
it was going when it “smashed into the tree.” Participants estimated a faster rate of speed
in the second case than in the first. Moreover, in a subsequent memory task, they were
more likely to report seeing broken glass at the scene of the accident. Thus, although
both “hit” and “smashed” were appropriate concepts for describing the event, recipients
drew different inferences from the description that affected not only their estimates of its
implications but their reconstruction of the situation they had observed.
A more general case in which the wording of a question can influence the representation
constructed of its referent is suggested by the use of “marked” or “unmarked” adjectives
(Lachman, Lachman, & Butterfield, 1979). That is, one pole of an attribute dimension
(e.g., “honest”) is unmarked, and asking a question pertaining to it (e.g., “how honest is
John?”) carries no implications for the value along the dimension. In contrast, the other,
“marked” pole of the dimension does have these implications. That is, “How dishonest is
John?” carries with it the implication that the target is at least somewhat dishonest. Thus,
all other things being equal, a respondent is likely to infer that John is less honest if he is
asked the second question than if he is asked the first one. It is interesting to speculate
that questions that refer to marked and unmarked adjectives, like the questions asked in
Loftus and Palmer’s study, stimulate the construction of (p. 237) different representations
of the questions’ referents. These representations, once formed, influence individuals’ lat
er responses to their referents in ways that are not implied by the question being asked.
An analysis of individuals’ behavior descriptions by Semin (2000, Semin & Fiedler, 1988,
1991; see also Brown & Fish, 1983) is also worth considering in this context. Suppose one
person, A, punches another, B, in the nose. An observer of the event could describe it to
another using any of four terms: (a) a descriptive action verb, which preserves its percep
tual features (e.g., “A hit B”); (b) an interpretative action verb, which reflects an inference
that is drawn from the action (“A hurt B”); (c) a state action verb, which refers to the af
fective component of the observed event (“A hates B”); and (d) an adjective, which de
scribes a general attribute of the actor (“A is aggressive”). Although these characteriza
tions could each describe the action, they differ in the generality of their implications and
thus in the representation that is likely to be formed in the course of comprehending
them (Fiedler & Krüger, this volume). As suggested by several theoretical formulations
Page 6 of 28

(e.g., Trope & Liberman, 2003, 2010; Vallacher & Wegner, 1987), objectively irrelevant
situational factors that fortuitously influence the description an observer uses to charac
terize an action and communicate about it to other persons can potentially influence the
inferences that these persons make about the nature of the event, its consequences, and
characteristics of the actor.
Effects of Metaphors in Comprehension and Judgment

Mental representations of a stimulus can be difficult to construct, particularly when indi
viduals do not have a previously formed mental representation of the stimulus to draw up
on (Jiang et al., 2010). In such cases, communicators often increase the vividness of the
representation that individuals construct through the use of metaphors (Graesser, Mio, &
Mills, 1989). In many instances, metaphors are used and comprehended automatically,
thus having the status of idioms (Gibbs, 1983). For example, the comments that an overly
dependent graduate student is a “pain in the neck” or that salary “weighed heavily” in
one’s decision to take a job, are understood by most native English speakers without ac
cessing their literal meaning. However, metaphors can also be employed intentionally to
increase the comprehension of abstract descriptions of a stimulus or to increase the vivid
ness of its features. Thus, a person may describe love as either a “red, red rose” or as a
“battlefield,” depending on his feelings about his relationship with his partner at the time.
A detailed analysis of the use of metaphors in communication has been provided by Lakoff
and Johnson (1980; see also Graesser et al., 1989; for a review of the use of metaphors in
persuasion, see Ottati & Renstrom, 2010) and is beyond the scope of this discussion.
However, the effect of metaphors on comprehension can be indirect. A particularly in
triguing example is provided by Ottati, Rhoads, and Graesser (1999). Participants who
had either positive or negative attitudes toward sports listened to a persuasive communi
cation advocating the institution of a senior thesis requirement for graduation. The mes
sage contained either strong or weak arguments in favor of the position advocated. In
some cases, however, the communication contained a number of sports-related metaphors
(e.g., “If students want to play ball with the best…”), whereas in other cases, the literal
meaning was conveyed (“If students want to work with the best,…”). The communication
had generally greater effect on participants’ acceptance of the position advocated when it
contained strong arguments. However, the use of metaphors affected the attention that
recipients paid to the communication content in different ways, depending on whether
they considered the domain of experience to which the metaphors’ literal meaning per
tained to be desirable. Thus, metaphors increased the effect of argument strength on
sports enthusiasts’ attitudes but decreased its effect on participants who disliked sports.
Effects of Communication Structure on Comprehension
The mental representation formed in the course of comprehending a communication can

also be influenced by the way the communication is structured. As an example, partici
pants in a study by Adaval, Isbell, and Wyer (2007) received verbal descriptions of a polit
ical candidate’s activities over the course of his career. The effects described were either
predominantly favorable or predominantly unfavorable, thus conveying either a positive
Page 7 of 28

or negative impression of the candidate. However, the structure of the communication

varied. In some cases, the information was presented in the form of a narrative that indi
cated the temporal sequence in which the activities occurred. In other cases, the same in
formation was conveyed in an ostensibly unordered list. Finally, the verbal descriptions
were in some cases accompanied by an evaluatively neutral picture of the candidate and
in other pictures (p. 238) were not presented. After receiving this information, partici
pants evaluated the candidate.
When information was conveyed in the form of a narrative, participants apparently con
structed a mental representation of the sequence of events as a whole and based their im
pression on this representation rather than considering its individual features (see Pasu
pathi, Mansfield, & Weeks, this volume; Pennington & Hastie, 1988, 1992, for other evi
dence of narrative-based processing). Furthermore, pictures of the candidate provided
cognitive “glue” that tied the events together and consequently increased the extremity
of the impression based on them. When information was conveyed in an unordered list,
however, participants construed the semantic implications of each event description sepa
rately and combined these implications mechanistically (e.g., Anderson, 1971, 1981). In
this case, however, a picture of the candidate, which required visual information process
ing, interfered with this semantic integration process and decreased the extremity of the
impressions that were formed.
Comprehending information about a situation from the perspective of the individuals in
volved in the situation can increase the intensity of individuals’ reactions to it. Although
these effects were evident when information consisted of single statements (Jiang &
Wyer, 2009), they may be even more evident when information is in the form of a narra
tive. Green and Brock (2000) note that individuals often become “transported” into a situ
ation they read about. That is, they construe the events that occur from the perspective of
the protagonists and experience reactions similar to those they imagine the protagonists
to be experiencing. Although this disposition can occur spontaneously, it can also be in
duced by instructing persons to imagine themselves being in the situation. These instruc
tions can backfire, however, decreasing the effect of the communication if the image of
the situation is difficult to construct (Petrova & Cialdini, 2005).
In a study by Hung and Wyer (2009), for example, participants read an appeal for dona
tions to help combat child trafficking. The verbal content of the appeal encouraged partic
ipants to imagine the situation from the perspective of the victim. In this condition,
adding a picture of the victim increased the vividness of the representation of the situa
tion that was formed and consequently increased the amount of money that participants
were willing to donate. In some conditions, however, participants were asked before read
ing the appeal if they would be willing to help, thus inducing them to think about the situ
ation described from the perspective of a potential donor. In this case, increasing the
vividness of the situation from the victims’ perspective increased perspective conflict and
decreased the amount of money that participants donated.
Page 8 of 28

The Effect of Self-Generated Labels on Judgments and Behavior
The concepts that individuals arbitrarily use to communicate to others about a social
event can influence recipients’ inferences about the implications of the communication
and the conclusions they draw from it. Furthermore, if the concepts that communicators
use are stored in memory, the concepts may later be recalled and used as a bases for
communicators’ own judgments and decisions independently of the reasons that led the
concepts to be applied.
Several studies provide examples. In a study by Carlston (1980), individuals received de
scriptions of a person’s behavior that had favorable implications along one trait dimen
sion (kindness) and negative implications along a second (honesty). They first judged the
person with respect to one of the traits and later judged the person with respect to the
other trait. In communicating their first judgment, participants applied a concept of the
trait they were judging, and the evaluative implications of this concept influenced their
second judgment independently of the information they had originally received. (For ex
ample, if the target’s behavior was both kind and dishonest, participants who had initially
judged him as kind later judged him as honest whereas those who had initially judged him
as dishonest later judged him as unkind.) Moreover, these effects increased over time.
A study with more direct implications for the effect of generating a communication on in
dividuals’ own judgments was reported by Higgins and Rholes (1978). Participants read a
passage about a target person describing behaviors that could be interpreted in terms of
either favorable attribute concepts (adventurous, self-confident, etc.) or unfavorable ones
(foolhardy, conceited, etc.). Then, some participants were asked to write a general de
scription of the person that would permit him to be identified by someone who knew the
target and either liked him or disliked him. Others anticipated writing this description but
did not actually do so. Then, all participants evaluated the target person themselves.
When participants had not generated a description of the target, their evaluation of him
did not depend on whether he was liked or disliked by another. (p. 239) When participants
had written a description of the target, however, they used concepts in their description
that were evaluatively consistent with attitudes of the person to whom they were commu
nicating. Then, having done so, they used these concepts as bases for their own evalua
tion, reporting greater liking for the target when the concepts had favorable implications
than when they did not.
The language that individuals use in communicating can also affect their overt behavior.
Participants in a study by Sherman, Alma, Berman, and Lynn (1978) were asked to rate
the importance of a number of social issues that were normatively either very important
(e.g., nuclear disarmament) or trivial (e.g., pay-on-entry buses). In some cases, one of the
items rated was recycling (a moderately important issue), whereas in other cases this
item was not included. Later, in the experimenter’s absence, a confederate (posing as an
other subject) asked them if they would be willing to help out with a recycling project.
When participants had not rated recycling, they volunteered more help when the issues
they had considered were important than when they were not. However, participants who
Page 9 of 28

judged recycling rated it as less important in the context of important issues than in the
context of trivial ones and then, having done so, volunteered less help in the former case
than in the latter. Thus, although the context issues apparently had no effect on partici
pants’ perception of the importance of recycling at the time they rated it, they affected
the language that participants used to report their judgments along the response scale
they were given. Consequently, participants later recalled this rating out of context and
used it as a basis for their behavioral decision.
Summary
The research reviewed in this section calls attention to several important considerations.
First, the comprehension of a message’s literal meaning involves the construction of a
mental representation whose features are not always captured by the semantic concepts
to which the message directly refers. Second, the nature of this representation is influ
enced by situational and individual difference factors that dispose individuals to process
information visually or verbally, by linguistic characteristics of the information itself, and
by the form in which the information is presented. These mental representations, once
formed, can have implications for the referent that go beyond the semantic meaning of
the messages that gave rise to them.
However, the preceding review ignores a factor that is critical in understanding the com
prehension of information transmitted in a social context. That is, social communications
are normally transmitted for a purpose (to inform, to persuade, to entertain, etc.). More
over, recipients of the communications typically make implicit assumptions about the
communicator’s objectives. When recipients’ construal of a communication’s literal mean
ing is inconsistent with these assumptions, they may either (a) reinterpret the message’s
meaning in a way that is consistent with expectations or, alternatively, (b) revise their
perception of the communicator’s objectives. The next section considers these possibili
ties in more detail.
Perceptions of Pragmatic Meaning

As Grice (1975) was among the first to recognize, the meaning attached to communica
tions that are transmitted in a social context is governed by certain implicit rules or prin
ciples that are applied both by the communicator in generating his or her message and by
the recipient in interpreting it. For example, communications should be informative; that
is, they should convey information that the recipient does not already have. Moreover,
they should be truthful or should convey the truth as the communicator sees it. Also, they
should be relevant to the topic at hand. Although these principles are likely to apply in
general, others are more situation-specific. For example, the communications exchanged
in informal conversations should be polite (Brown & Levinson, 1987; Holtgraves, 2002);
that is, they should not offend or upset the recipient. They should also be modest and not
unduly extol one’s own skills and accomplishments at the listener’s expense. Participants
in informal conversations are usually expected to apply these principles in constructing
Page 10 of 28

their messages. Moreover, recipients assume that this is in fact the case and interpret the
messages accordingly. The role of Gricean principles in informal conversation has been
elaborated in research on psycholinguistics (Green, 1989; Sperber & Wilson, 1986) and
social psychology (Higgins, 1981; Schwarz, 1994; Strack, 1994). Some additional consid
erations regarding the applicability of these principles are nonetheless worth reviewing.
The aforementioned principles of communication are unlikely to be universal. Cultural

differences, for example, may exist in the role of modesty and politeness. Furthermore,
the principles apply primarily under conditions in which (p. 240) communicators’ primary
objective is to conduct the conversation in an effective and enjoyable fashion. In many sit
uations, however, one or both parties may have a different agenda. In a job interview, for
example, the modesty principle is less applicable. In other situations, one party may wish
to ingratiate or to convey a favorable impression. Or, he or she may be motivated to per
suade the recipient to adopt a particular point of view or engage in certain behavior.
When recipients are aware of these objectives, they may apply other normative principles
than the ones originally proposed by Grice. Under conditions in which a given set of prin
ciples are operative, however, the processes that underlie their impact on the generation
and comprehension of communications are likely to be similar independently of their spe
cific content.
In this section, I consider communications in three types of situations: (a) informal con
versations, in which individuals are either active participants or observers; (b) the public
media, in which messages are transmitted not only to inform but also to persuade; and (c)
surveys and psychological experiments in which individuals are expected to make re
sponses that are not only informative and truthful but also that the investigators will in
terpret correctly.
Communication in Informal Conversations
When individuals receive information about someone’s behavior from a dispassionate

source, they are likely to take the information at face value. In this case, their impression
of the person may be based on the descriptive implications of his or her behavior for vari
ous personality traits and the evaluative implications of this behavior for his likeableness
(Srull & Wyer, 1989). When the same information about the person is conveyed in an in
formal situation by an acquaintance of the individual, other considerations arise. In this
case, recipients’ perception of the validity of the behavior descriptions and their repre
sentativeness may depend in part on their impressions of the communicator and his or
her motive for providing the description. In many instances, the communicator and recipi
ent are not well acquainted. In this case, the communicator’s statements about another
person may reflect as much on the speaker as on the person that he or she is describing.
As a consequence, recipients’ impression of the individual being described is likely to be
quite different from the impression they might form on the basis of the same information
when it is conveyed out of its conversational context.
Page 11 of 28

This difference was demonstrated by Wyer, Budesheim, and Lambert (1990). Participants
in these studies listened to a tape-recorded conversation between two college students
about a third. Before engaging in the conversation, the communicators ostensibly wrote
down two or three trait adjectives that they considered to be particularly descriptive of
the person they were about to discuss, and these adjectives were passed on to the listen
ers. The communicators then exchanged anecdotes about a number of behaviors that the
target had performed, some of which were favorable (e.g., “invited a charity collector
home for dinner”) and others of which were unfavorable (e.g., “shouts and honks at slow
drivers”). After hearing the conversation, participants evaluated both the target person
and the speakers and then recalled the behaviors they had heard.
To provide a basis for interpreting the results of this study, consider participants’ reac
tions to identical information when it is conveyed out of the context of a conversation. In
this case, when the target’s behaviors are more likely to be considered representative of
his general personality (Srull & Wyer, 1989; Wyer & Srull, 1989), participants’ impres
sions of the target are typically based on the initial trait descriptions of the target inde
pendently of the behavior descriptions that follow it. However, behaviors that are evalua
tively inconsistent with these initial impressions are typically better remembered, indicat
ing that individuals think more extensively about these behaviors in an attempt to recon
cile their occurrence.
In Wyer et al.’s (1990) study, in which the same information was conveyed in the context
of a conversation, quite different effects were obtained. For one thing, speakers’ trait de
scriptions of the target had little direct effect on participants’ evaluations of him. Instead,
they influenced participants’ impressions of the speakers themselves; participants evalu
ated the speakers more favorably if the speakers had described the target positively than
if they described him negatively. Furthermore, participants’ recall protocols indicated
that, rather than attempting to reconcile the speakers’ descriptions of the target’s behav
ior with more general impressions of his personality, they used the descriptions to con
firm the impressions they formed of the speakers. Thus, participants had better recall of
the behaviors mentioned by one speaker when they were evaluatively inconsistent with
the trait description provided by (p. 241) the other speaker, suggesting that the speaker’s
trait description of the target reflected a characteristic of the speaker rather than of the
person the speaker was describing.
Reactions to Individuals’ Self-Descriptions

The aforementioned study confirmed the conclusion that when information is conveyed in
a conversation, listeners do not assume that speakers’ descriptions of the target were
consistent with the informativeness principle. Rather, they considered why the speakers
were likely to generate these descriptions and, in doing so, formed impressions of the
speakers themselves.
A second study (Wyer, Budesheim, Lambert, & Swan, 1994) confirmed this conclusion
and, in doing so, identified the effect of two other norms that are likely to govern conver
sations outside the laboratory. As noted earlier, individuals in a conversation are expected
Page 12 of 28

to be both polite and modest. Individuals who apply these norms when they observe a
conversation are likely to think more extensively about comments that violate these
norms, and this increased processing may lead the behaviors to be better recalled.
Wyer et al. (1994) confirmed this possibility using behavioral descriptions identical to
those employed in the previous study. In this case, however, the conversation was ostensi
bly about one of the speakers. Thus, one speaker described both favorable and unfavor
able behaviors that the other had performed, whereas the second speaker described fa
vorable and unfavorable behaviors that he personally had performed. After listening to
the conversation, participants reported their impressions of the speakers and then re
called the behaviors that each speaker had mentioned. As expected, participants thought
more extensively about communications that violated the politeness and modesty princi
ples, as reflected in their recall of the communications later. That is, they had better re
call of behaviors that one speaker mentioned about the other if the behaviors were unfa
vorable (thereby violating the politeness principle). On the other hand, they had better re
call of behaviors a speaker had mentioned about himself if they were favorable (and con
sequently violated the modesty principle).
Teasing and Sarcasm

Although the communication principles postulated by Grice clearly come into play in
everyday conversations, their effects can be hard to identify without knowing the atti
tudes and beliefs of the communication partners and the past history of their interaction.
For example, a corollary of Grice’s (1975) informative principle, denoted as the “maxim of
quantity,” is that conversing parties should provide neither too much nor too little infor
mation in response to a question. Rather, the amount of information they transmit should
be tailored to the recipient’s prior knowledge. Too little information is obviously uninfor
mative. However, excessive detail that is incomprehensible given the recipient’s knowl
edge can also be uninformative. For example, suppose a resident of Urbana, Illinois, is
asked where he comes from. His response “America” might be informative to a Chinese
resident in Hong Kong but would be uninformative to another resident of Urbana. On the
other hand, the response “401 W. Delaware Avenue” would be informative to the second
individual but not the first. An observer of the conversation who had no knowledge of ei
ther party would therefore have trouble knowing whether a conversational norm was vio
lated or not.
These considerations are particularly important in identifying whether a communicator’s

statement reflects his true attitude or opinion or, alternatively, is sarcastic. Individuals
who hear a statement that they consider to be clearly untrue may infer that its literal
meaning is not the meaning the communicator intends, but is ironic. To make this infer
ence, however, the recipient must have knowledge of the speaker’s a priori knowledge
and attitudes. For example, suppose someone asserts that “George W. Bush was the
greatest U.S. President since Abraham Lincoln.” Most people are likely to infer that the
communicator is being sarcastic and that his or her intended meaning is quite the oppo
site. However, not everyone is likely to interpret the statement as ironic. For this interpre
tation to be made, two things must be true. First, recipients must personally believe that
Page 13 of 28

the statement is false. Second, they must perceive that the communicator also believes it
is false. If either of these conditions is not met, the recipients may believe that the literal
meaning is, in fact, the intended meaning and may take the statement at face value. In
short, the recognition that a statement’s literal meaning is not its intended meaning re
quires knowledge about not only the topic of the message but also the communicator.
These considerations come into play in teasing. Suppose, after publishing his first book,
an American author is confronted by a colleague who remarks, “Congratulations on your
book. However, (p. 242) when are they going to publish the English version?” If the author
has confidence in his writing skills and believes that the colleague has general respect for
his abilities, he may infer that the communicator is intending to tease him and that the
implications of the colleague’s statement are not meant to be taken literally. If these con
ditions do not obtain, however, the recipient may infer that the colleague is intentionally
disparaging his ability and is being hostile.
Humor Elicitation
The considerations that underlie the identification of sarcasm and teasing are related to
the elicitation of humor. What is it that makes something funny? Although many answers
to this question have been suggested (Berlyne, 1971; Freud, 1960/1905; Suls, 1972; Zill
man & Cantor, 1976), one of the most complete formulations was proposed by Apter
(1982) and extended by Wyer and Collins (1992; see also Wyer, 2004). Apter assumes that
people spontaneously comprehend a communication in terms of concepts and knowledge
about its referents and arrive at a tentative understanding of the situation in terms of
these concepts. In some cases, however, this interpretation is obviously inapplicable, and
so a reinterpretation is required. If this reinterpretation diminishes the importance of the
situation or the individuals involved in it, humor is elicited.
To give an example, consider the following:
A blind man enters a department store, picks up his dog by its tail and begins
swinging it over his head. A clerk hurries over and asks, “Can I help you, sir?”
“No, thanks,” the man replies. “I’m just looking around.”
In this joke, “looking around” is initially interpreted in terms of its common, idiomatic
meaning, but is then reinterpreted in terms of its literal meaning along with knowledge
that the blind man relies on his dog to do his “looking.” This meaning trivializes the situa
tion and consequently elicits amusement.
The diminishment assumption permits one to conceptualize the humor elicited by not on
ly jokes but also witticisms, sarcasm, puns, and everyday life events (Wyer & Collins,
1992). It also permits one to distinguish between humor-eliciting reinterpretations of
stimulus events and reinterpretations that increase the importance of the events (e.g.,
scientific discoveries). Other factors also play a role in the amount of humor a communi
cation elicits (Wyer, 2004; Wyer & Collins, 1992). For example, the amusement experi
enced as a result of reinterpreting a communication is a nonmonotonic function of the
cognitive effort required to make this interpretation. That is, a joke is more humorous if a
Page 14 of 28

moderate amount of effort is required than if it is either too easy to comprehend or too
difficult. Second, the amusement experienced increases with the amount of cognitive
elaboration of its humor-eliciting implications that occurs subsequent to its reinterpreta
tion, but decreases with the elaboration of its nonhumorous implications. Thus, a joke
that is disparaging to women may be equally amusing to both men and women immedi
ately after it is told. However, the amusement experienced by men may increase as a re
sult of thinking about it whereas the amusement experienced by women may decrease
(Wyer & Collins, 1992).
Indirect Speech Acts and Emotional Communication

Related considerations arise in the interpretation of indirect speech acts (Gordon &
Lakoff, 1975; Searle, 1975; for a review, see Holtgraves, 2002). These acts are common
place in everyday conversation and, in many instances, their intended meaning is obvi
ous. For example, a man who is asked, “Can you tell me the time?” typically assumes that
the questioner already knows his ability to read a watch and intends to ask what time it
is. In some instances, however, the identification of indirect speech acts is less clear. For
example, suppose a woman who is giving a dinner party comments, “it’s cold in here.”
Her guests may interpret this as a statement of opinion and perhaps as an indirect sug
gestion to her husband to turn up the heat. Her spouse, however, who is aware of her re
peated complaints about his obsession with saving energy and unwillingness to set the
thermostat higher than 65 degrees, may interpret the remark as an expression of irrita
tion and anger.
Indirect expressions of emotion may be inherently ambiguous, however. Empirical evi

dence of this was obtained in a study of emotional communication between marriage
partners (Gaelick, Bodenhausen, & Wyer, 1985). Marriage partners first engaged in a 10-
minute tape-recorded discussion of a problem they were having in their relationship.
Then, in a second session, each partner viewed the tape of the conversation and identi
fied three statements that were made that had an important effect on their feelings. Fi
nally, in a third session, they reviewed the segments of the tape that contained these
statements. Partners who made (p. 243) the statements rated the feelings they intended to
convey and how they thought their partner would respond. They rated the statements
their partner made in terms of the feelings their partner intended to convey, how the
partner thought they would react, and their actual reactions. Factor analyses of these rat
ings indicated that they fell along two independent dimensions pertaining to love and hos
tility.
In general, partners attempted to reciprocate the feelings they perceived their spouse to
convey to them. However, they were only accurate in perceiving their spouse’s hostility.
Consequently, feelings of hostility were actually reciprocated, whereas feelings of love
were not. Thus, feelings of hostility were likely to escalate over the course of the interac
tion but feelings of love were not. In addition, when partners communicated in a way they
intended to be affectively neutral, male partners interpreted these messages as hostile
whereas female partners interpreted them as an expression of love. These differences
could indicate that different conversational norms govern male and female partner’s ex
Page 15 of 28

pectations for one another’s communications. That is, men expect their spouses to be lov
ing and consequently interpret messages that deviate from expectations as hostile,
whereas women expect their spouses to be hostile and interpret messages that deviate
from expectations as loving.
Communication Norms in Get-Acquainted Conversations.

Somewhat different principles of communication are likely to govern informal conversa
tions. In these situations, conversation partners are expected to convey information that
makes it easy for the other to respond, thereby keeping the exchange of information flow
ing smoothly. This is normally done either (a) by providing information that can stimulate
a response, based on the recipient’s knowledge about the topic at hand or (b) by asking a
question in return. Thus, for example, suppose one party in a conversation asks the other
“Do you live in Chicago?” If the recipient’s response is “Yes, I just moved here from
Boston,” the first speaker is likely to find it easy to respond, either by asking another
question (e.g., “How do you like the Midwest?”), by mentioning his own recent visit to
Boston, or the like. If the recipient answers “Yes. What about you?” the first speaker can
also respond easily. However, suppose the response is simply “Yes,” with no elaboration
and without asking a question in return. Then, the burden is placed on the first speaker to
come up with another topic to discuss. This increases the speaker’s difficulty in continu
ing the conversation and consequently is likely to decrease the speaker’s liking for both
the conversation and the partner.
A study by Wyer, Swan, and Gruenfeld (1995) confirmed these speculations. Male and fe
male college students engaged in a 10-minute get-acquainted conversation with another
student of the same sex. Before they engaged in the conversation, however, one partner
was told that to help control for the content of the conversations, he or she should ask the
other five standard questions (“Where are you from?” “Do you like school?” etc.), working
them into the conversation at points that seemed natural. The other partner, however,
was given the five questions the other would ask in advance and told to respond to each
question either (a) by giving a one- or two-word response or elaborating or (b) by either
asking a question in return or not doing so. The effects of these manipulations on part
ners’ reactions to the conversation depended on the sex of the conversation partners.
Males’ difficulty in conducting the conversation was influenced primarily by whether
their partner asked them questions, whereas females’ difficulty depended on whether
their partner elaborated. In each case, however, participants’ difficulty in carrying on the
conversation had a negative impact on their enjoyment of the conversation and on their
consequent liking for their partner.
Responses to Communications in the Media
According to Wyer and Radvansky (1999; Wyer & Gruenfeld, 1995), individuals who re
ceive a communication may first construe its literal meaning, and, if this interpretation is
consistent with the normative principles that govern communications in the context in
which the message is encountered, they are likely to accept its implications at face value.
If, however, the communication appears to violate one or more of these principles, they
Page 16 of 28

are likely to try to understand the reason for the violation. In doing so, they may reinter
pret the communication in a way that is more consistent with the norm that is violated.
Their means of doing so may depend on the particular norm violated and the context in
which the violation occurs.
Violations of Informativeness
Newspapers are expected to report new information. Therefore, if a news headline con
tains a statement whose validity seems to go without (p. 244) saying (and thus is uninfor
mative) people attempt to understand why the statement was made. In doing so, they may
speculate that there might actually be some reason to suppose that the statement is actu
ally untrue, thus making an assertion of its validity informative. If this occurs, it might
plant a kernel of doubt in the recipient’s mind as to the statement’s validity that had not
existed before. As a result, the assertion could have a boomerang effect, decreasing be
liefs in its validity rather than increasing it.
Note, however, that the effects are only likely to occur if recipients perceive that the
source intends to be informative. If this is not the case, the informativeness principle may
not be applied. In a study by Gruenfeld and Wyer (1992), participants read a series of
statements that were ostensibly taken from either newspapers (a source whose objective
is to be informative) or an encyclopedia (a source whose objective is simply to preserve
archival knowledge). Some statements asserted the validity of propositions that partici
pants in the study were unlikely to believe on a priori grounds (e.g., “Lyndon Johnson was
responsible for the assassination of John F. Kennedy”). Others denied the validity of these
propositions and thus were uninformative (e.g., “Lyndon Johnson was not responsible…”).
After reading the statements, both these participants and control participants completed
a questionnaire in which their own beliefs in the propositions were assessed. When the
statements that participants had read had allegedly come from an encyclopedia, informa
tive assertions increased their beliefs in the propositions, whereas the denials slightly de
creased these beliefs. When the statements had ostensibly come from newspapers, how
ever, the uninformative denials increased participants’ beliefs in the propositions, and this
increase was nearly as great as the increase that resulted from exposure to the asser
tions. Thus, participants who read denials in a newspaper apparently inferred that there
was some reason to believe that the statements were true (thus making the denials infor
mative), and so they increased their beliefs in the statements’ validity.
Some assertions, of course, cannot possibly be untrue. The validity of the assertion that
“the United States has never had a woman president” is not open to question. In this
case, individuals who encounter the assertion and question the reason for making it may
interpret the statement as an indirect expression of attitude (e.g., “The U.S. has never
had a woman president and this is a good/bad thing”), and this attitude might influence
their own opinion. Supplementary data by Gruenfeld and Wyer found some evidence for
these reactions as well.
Page 17 of 28

Informativeness and Truthfulness in Advertising
The primary objective of a newspaper is to be informative, and the statements attributed

to them in Gruenfeld and Wyer’s (1992) study clearly violated this norm. In many media
communications, this is not the case. For example, advertisements purport to provide new
and important information about the products they are promoting, but their primary ob
jective is to persuade. Consequently, it is often unclear whether the material conveyed in
an advertisement is actually informative or not.
Gricean principles of informativeness and truthfulness come into play in the comprehen
sion of an advertisement in three ways. First, individuals who assume that advertise
ments are informative may make unwarranted inferences on the basis of assertions that
are objectively meaningless. For example, individuals who read that “Brand X is better”
may mentally complete the claim in a way that is informative and relevant (“…better than
other brands”) rather than in other, less relevant ways (“…better than it used to be,” “…
better than the most inferior brands on the market”). This is not always the case, howev
er (Johar, 1995; Sawyer & Howard, 1991). That is, the effect appears to occur only when
participants are highly motivated to evaluate the product at the time they receive the in
formation about it. That is, less involved participants do not engage in the cognitive work
required to apply the informativeness principle and consequently are less affected by the
ambiguous claims than are involved participants.
Second, some ads use puffery, that is, attribute descriptions that purport to be important
but are actually meaningless. If these descriptions are inferred to be informative, they
can have ironic implications. For example, individuals who read a claim that “Brand X
contains Shedigira extract” presumably have no idea what this attribute is. If they infer
that it provides new information, however, they may infer not only that “Shedigira ex
tract” is desirable but also that other products do not have it. However, suppose the claim
asserts that “Brand X contains no Shedigira extract.” Then, recipients are likely to infer
that the attribute is undesirable but that most competitors do have it. In short, attempts
to make objectively meaningless attribute descriptions in a manner that conforms to the
informativeness principle can have an impact on (p. 245) not only recipients’ attitudes to
ward the product being advertised but also on their attitudes toward its competitors.
In many instances, however, the effect of puffery can depend on how much recipients
think they know about the product being advertised. In a study by Xu and Wyer (2010),
participants first completed a test of their knowledge about fabrics and received feed
back that they were either more or less knowledgeable than others in the general popula
tion. Then, as part of an unrelated study, they read an ad for a down jacket. In some cas
es, the ad contained meaningless attribute descriptions (e.g., “Omni-bloc HP Cyber
Stretch”) and in other cases, it did not. When the ad had ostensibly appeared in a maga
zine that was read primarily by fashion experts, the puffery generally increased partici
pants’ evaluation of the product. When the ad had appeared in a popular magazine, how
ever, puffery had a positive influence only when participants perceived themselves to
know less than others in the general population. When they considered themselves to
Page 18 of 28

know more about fabrics than the average person, they inferred that attribute descrip
tions they didn’t understand would not be understood by anyone and were consequently
intended only to persuade and not to inform. In this case, the puffery had a negative,
boomerang effect on their evaluations.
Third, the messages encountered in advertisements can violate not only the informative
ness principle but also the truthfulness principle. These latter violations may nonetheless
be consistent with expectations. That is, individuals generally believe that advertising
claims are likely to be exaggerated (Campbell & Kirmani, 2008; Friestad & Wright, 1994).
Consequently, they may correct for the bias in literal meaning of the ad’s claims in assess
ing its implications for the product being advertised. However, this may occur only if re
cipients are motivated and able to expend the cognitive effort required to make a correc
tion. In a study by Hung and Wyer (2008), participants received information about an ad
for a hair-restoring ointment that described both the amount of hair loss before using the
product and the amount that was evident afterward. However, each piece of information
was either described verbally or conveyed in a picture. When one of the two components
was pictured and the other was described verbally, participants attempted to construe the
implications of the verbal information in a way that was consistent with the picture that
accompanied it. However, the cognitive effort required to do this decreased the effort
they later expended in adjusting for the possibility that the ad’s claims were exaggerated.
Consequently, the ad’s effectiveness increased in this condition relative to conditions in
which both components were described verbally. When both components were pictured,
however, the implications of the ad were easy to comprehend, and so individuals were
willing to expend the effort required to correct for its biasing implications. Consequently,
the effectiveness of the ad decreased relative to conditions in which only one component
was pictured.
Informativeness and Truthfulness in Psychological Experiments
As pointed out many years ago (Wyer, 1974), psychology experiments are social situations
in which one person (the subject) attempts to understand what sorts of responses the oth
er (the experiment) wants or expects and responds accordingly. Informativeness and
truthfulness are likely to play a role in these situations. When the question to be an
swered concerns a matter of fact, the response that the experimenter considers to be
truthful and informative is clear. In many cases, however, the response to be made is sub
jective (e.g., an attitude) and must be reported in a language with which the respondent
is unfamiliar (i.e., along a category response scale). Then, the respondent must determine
the meaning that the experimenter gives to the various response alternatives in order to
respond in a way that he or she will interpret correctly.
These matters come into play in informal conversations as well, of course. If a person is
asked to describe his political orientation, the label he uses (“conservative, “moderate,”
“liberal,” “radical,” etc.) might depend on not only his actual position but also on his or
her perception of how the recipient will interpret his response. Thus, he might be less
Page 19 of 28

likely to describe himself as “liberal” if he is asked by a Tea Party advocate (who is likely
to equate “liberal” with “communist”) than if he is asked by a member of “Move On.”
These possibilities were formalized by Ostrom and Upshaw (1968). They note that when
individuals are asked to describe a stimulus along a subjective attribute dimension, they
infer the range of stimulus values that the questioner considers to be relevant and assign
the stimulus a value that reflects its relative position within this range. For example, they
may use the labels “tiny, “small,” “big,” and “enormous” to describe different sizes. How
ever, they may assign these labels to a lower range of (p. 246) physical stimulus values if
they are being asked the size of a baby than if they are asked the size of an apartment.
Thus, they may judge a baby as “big” but an apartment as “small” despite the fact that
babies are smaller than apartments.
Although these phenomena occur in informal conversations, they become particularly im
portant in psychological research, in which participants are often asked to report their
judgments along a numerical scale with verbally labeled endpoints (e.g., “very unfavor
able, “very favorable,” etc.). The objective values of the stimuli that correspond to these
endpoints are rarely indicated. To respond in a way that reflects their underlying judg
ment, the participants must first infer the range of the stimulus values that the experi
menter considers to be relevant and assign the stimulus being judged a value that corre
sponds to its position relative to these endpoints. However, individual and situational dif
ferences occur in the nature of these assumptions. In some cases, communicators may in
fer the range from the type of stimuli being judged, as in the “baby” versus “apartment”
example given earlier.
When several different stimuli are judged, participants may use a range of values of these
stimuli as an indication of the range they are supposed to consider (Parducci, 1965), and
so the stimuli may have a negative, contrast effect. Thus, for example, participants may
judge recycling to be lower along a subjective scale of importance if they have previously
encountered a number of very important issues, leading them to assume that the experi
menter is considering a high range of values, than if they have previously rated a number
of unimportant issues (Sherman et al., 1978).
In many psychology experiments, however, no clear indication is given of the range of val
ues that the experimenter considers to be relevant. In this case, respondents’ assump
tions concerning this range maybe influenced by irrelevant situational factors of which
they are unaware. In a provocative demonstration of this possibility, Adaval and Monroe
(2002) subliminally exposed participants to either high or low numbers while they per
formed an ostensibly irrelevant task. Later, they considered a product whose price was
listed and estimated both its cost and its desirability. Participants who had been sublimi
nally exposed to high numbers apparently positioned the response scales they were given
to include a relatively higher range of values than did those who had been subliminally
exposed to low numbers, and this was true regardless of the attribute to which the scale
pertained. Consequently, they rated the product both as less expensive and as less desir
able in the former case than in the latter.
Page 20 of 28

The Role of Gricean Principles in Survey Responding

The tendency for individuals to communicate information they believe the recipient will
consider informative is particularly evident in survey research (Schwarz, 1994; Strack,
1994). Although the effects of this tendency are reviewed in detail by (Conrad, Schober, &
Schwarz, this volume), an example may be worth noting in the present context. When in
dividuals are asked to report the frequency of their behavior along a scale of ordered al
ternatives, the nature of these alternatives can provide an indication of what responses
are normative, and this can influence not only participants’ reports of their behavior and
their perceptions of its implications.
For example, suppose people who personally watch 3 hours of television each day are
asked to report the amount of television they watch along a scale from “less than ½ hour/
day” to “more than 2½ hours/day.” In this case, they are likely to perceive themselves to
watch television a greater amount of time than others do. If they are asked to report their
viewing habits along a scale from “less than 2½ hours/day” to “more than 4 hours/day”
however, they are likely to perceive themselves to watch less television than average. To
this extent, if people perceive watching television to be a waste of time, they may feel less
satisfied with the way they spend their leisure time after responding along the first scale
than after responding along the second. This is in fact the case (Schwarz, Hippler,
Deutsch, & Strack, 1985). This and many other examples (Conrad et al., this volume) indi
cate that the impact of Gricean principles on survey responses is quite pervasive.
Conclusion
The research reviewed in this chapter does not begin to cover all of the phenomena that
must be considered to provide a complete understanding of the role of language in social
communication. However, it hopefully calls attention to a number of issues that are not of
ten addressed in this area. In discussing the role of language in the comprehension of a
communication’s literal meaning, I focused largely on the role of visual imagery in
(p. 247) the mental representations that are constructed from verbal information and the
effects of this imagery on judgments. In conceptualizing the factors that underlie percep
tions of a communication’s intended meaning, I considered not only the factors that stim
ulate a consideration of this meaning but also the amount of cognitive effort required to
do so, as reflected in both the recall of the information and judgments based on it. In the
last section, I attempted to show that general principles of communication of the sort con
sidered by Grice (in particular, informativeness, truthfulness, and relevance) affect com
prehension of messages not only in informal conversations but in social communications
more generally, including those transmitted in the media and in psychological experi
ments.
Despite the breadth of the phenomena reviewed in this chapter, many important issues
have been ignored. Despite its limitations, however, I hope it provides a few insights and
Page 21 of 28

suggestions for future research in this central and important area of psychological in
quiry.
Acknowledgments
The preparation of this chapter was supported by grants GRF 640011, GRF 453710, GRF
453110, and GRF 451812 from the Research Grants Council, Hong Kong.
References
Adaval, R., Isbell, L. M., & Wyer, R. S. (2007). The impact of pictures on narrative-based
impression formation: A process interference model. Journal of Experimental Social Psy
chology, 43, 352–364.
Adaval, R., & Monroe, K. B. (2002). Automatic construction and use of contextual infor
mation for product and price evaluations. Journal of Consumer Research, 28, 572–588.
Apter, M. J. (1982). The experience of motivation The theory of psychological reversals.

Anderson, N. H. (1971). Integration theory and attitude change. Psychological Review, 78,
171–206.
Anderson, N. H. (1981). Foundations of information integration theory. New York: Acade

mic Press.
Berlyne, D. E. (1969). Laughter, humor and play. In G. Lindzey & E. Aronson (Eds.), Hand
book of social psychology (vol. 3, pp. 795–852). Reading, MA: Addison-Wesley.
Black, J. B., Turner, T., & Bower, G. H. (1979). Point of view in narrative comprehension,
memory, and production. Journal of Verbal Learning and Verbal Behavior, 11, 717–726.
Bransford, J. D., Barclay, J. R., & Franks, J. J. (1972). Sentence memory: A constructive
versus interpretative approach. Cognitive Psychology, 3, 193–209.
Brown, P., & Levinson, S. C. (1987). Universals in language usage: Politeness phenomena.
In E. Goody (Ed.), Questions and politeness (pp. 256–289). Cambridge, UK. Cambridge
University Press.
Brown, R., & Fish, D. (1983). The psychological causality implicit in language. Cognition,
14, 237–273.
Campbell, M. C., & Kirmani, A. (2008). I know what you’re doing and why you’re doing it:
The use of the persuasion knowledge model in consumer research. In. C. P. Haugtvedt, P.
M. Herr, & F. R. Kardes (Eds.), Handbook of consumer psychology (pp. 549–574). Mah
wah, NJ: Lawrence Erlbaum Associates.
Page 22 of 28

Carlston, D. E. (1980). Events, inferences and impression formation. In R. Hastie, T. Os

trom, E. Ebbesen, R. Wyer, D. Hamilton, & D. Carlston (Eds.), Person memory: The cogni
tive basis of social perception (pp. 89–119). Hillsdale, NJ: Erlbaum.
Childers, T. L., Houston, M. J., & Heckler, S. E. (1985). Measurement of individual differ
ences in visual versus verbal information processing. Journal of Consumer Research, 12,
125–134.
Dodson, C. S., Johnson, M. K., & Schooler, J. W. (1997). The verbal overshadowing effect:
Why descriptions impair face recognition. Memory & Cognition, I25, 129–139.
Förster, J., & Liberman, N. (2007). Knowledge activation. In A. Kruglanski & E. T. Higgins
(Eds.), Social psychology: Handbook of basic principles (2nd ed., pp. 201–231). New York:
Guilford.
Friestad, M., & Wright, P. (1994). The persuasion knowledge model: How people cope
with persuasion attempts. Journal of Consumer Research, 21, 1–31.
Friestad, M., & Wright, P. (1995). Persuasion knowledge: Lay people’s and researchers’
beliefs about the psychology of advertising. Journal of Consumer Research, 22, 62–74.
Freud, S. (1960/1905). Jokes and their relation to the unconscious. New York: Norton.
Gaelick, L., Bodenhausen, G. V., & Wyer, R. S. (1985). Emotional communication in close
relationships. Journal of Personality and Social Psychology, 49, 1246–1265.
Garnham, A. (1981). Mental models as representations of text. Memory & Cognition, 9,

560–565.
Gibbs, R. W. Jr. (1983). Do people always process the literal meaning of indirect requests?
Journal of Experimental Psychology: Learning, Memory and Cognition, 9, 524–533.
Glenberg, A. M., Meyer, M., & Lindem, K. (1987). Mental models contribute to fore
grounding during text comprehension. Journal of Memory and Language, 26, 69–83.
Gordon, D., & Lakoff, R. T. (1975). Conversational postulates. In P. Cole & J. Morgan
(Eds.), Syntax and semantics 3: Speech acts (pp. 83–106). New York: Academic Press.
Graesser, A. C., Mio, J., & Millis, K. (1989). Metaphors in persuasive communication. In D.
Meutsch (Ed.), Models of literary understanding (pp. 131–154). Amsterdam: Elsevier.
Green, G. M. (1989). Pragmatics and natural language understanding. Hillsdale, NJ: Erl
baum.
Green, M. C., & Brock, T. C. (2000). The role of transportation in the persuasiveness of
public narratives. Journal of Personality and Social Psychology, 79, 701–721.
Page 23 of 28

Grice, H. P. (1989). Studies in the way of words. Cambridge, MA: Harvard University
Press.
Gruenfeld, D. H., & Wyer, R. S. (1992). The semantics and pragmatics of social influence;
How affirmations and denials affect beliefs in referent propositions. Journal of Personality
Higgins, E. T. (1981). The “communication game:” Implications for social cogni

(p. 248)
tion and persuasion. In E. T. Higgins, C. P. Herman, & M. P. Zanna (Eds.), Social cognition:
The Ontario Symposium (vol. 1, pp. 342–392). Hillsdale, NJ: Erlbaum.
Higgins, E. T. (1996). Knowledge activation: Accessibility, applicability, and salience. In E.

T. Higgins & A. Kruglanski (Eds.), Social psychology: Handbook of basic principles (pp.
133–168). New York: Guilford.
Higgins, E. T., & Rholes, W. S. (1978). “Saying is believing:” Effects of message modifica
tion on memory and liking for the person described. Journal of Experimental Social Psy
chology, 14, 363–378.
Hung, I. W., & Wyer, R. S. (2008). The role of implicit theories in the impact of problem-
solving print advertisements. Journal of Consumer Psychology, 18, 223–235.
Hung, I. W., & Wyer, R. S. (2009). Differences in perspective and the influence of charita
ble appeals: When imagining oneself as the victim is not always beneficial. Journal of
Marketing Research, 46, 421–434.
Jiang, Y. S., Steinhart, Y., & Wyer, R. S. (2010). The role of visual and semantic processing
strategies in consumer information processing. Unpublished manuscript, Hong Kong Uni
versity of Science and Technology.
Jiang, Y. S., & Wyer, R. S. (2009). The role of visual perspective in information processing.
Journal of Experimental Psychology, 45, 486–495.
Johar, G. V. (1995). Consumer involvement and deception from implied advertising claims.
Journal of Marketing Research, 32, 267–279.
Johnson-Laird, P. N. (1983). Mental models: Towards a cognitive science of language, in

ference and consciousness. Cambridge, MA: Harvard University Press.
Johnson-Laird, P. N. (1989). Mental models. In M. I. Posner (Ed.), Foundations of cognitive

science (pp. 469–500). Cambridge, MA: MIT Press.
Krauss, R. M., & Fussell, S. R. (1996). Social psychological models of interpersonal com
munication. In E. T. Higgins & A. W. Kruglanski (Eds.), Social psychology: Handbook of
basic principles (pp. 655–701). New York: Guilford.
Page 24 of 28

Lachman, R., Lachman, J. L., & Butterfield, E. C. (1979). Cognitive psychology and infor
mation processing. Hillsdale, NJ: Erlbaum.
Lakoff, G., & Johnson, M. (1980). Metaphors we live by. Chicago: Chicago University
Press.
Loftus, E. F., & Palmer, J. C. (1974). Reconstruction of automobile destruction: An exam

ple of the interaction between language and memory. Journal of Verbal Learning and Ver
bal Behavior, 13, 585–589.
Ostrom, T. M., & Upshaw, H. S. (1968). Psychological perspective and attitude change. In
A. G. Greenwald, T. M. Ostrom, & T. C. Brock (Eds.), Psychological foundations of attitude
(pp. 217–242). New York: Academic Press.
Ottati, V. C., & Renstrom, R. A. (2010). Metaphor and persuasive communication: A multi
functional approach. Social and Personality Psychology Compass, 4(9), 783–94.
Ottati, V. C., Rhoads, S., & Graesser, A. C. (1999). The effect of metaphor on processing
style in a persuasion task: A motivational resonance model. Journal of Personality and So
Parducci, A. (1965). Category judgment: A range-frequency model. Psychological Review,

72, 407–418.
Pennington, N., & Hastie, R. (1988). Explanation-based decision making: Effects of memo
ry structure on judgment. Journal of Experimental Psychology: Learning, Memory and
Cognition, 14, 521–533.
Pennington, N., & Hastie, R. (1992). Explaining the evidence: Tests of the story model for
juror decision making. Journal of Personality and Social Psychology, 62, 189–206.
Petrova, P. K., & Cialdini, R. B. (2005). Fluency of consumption imagery and the backfire
effects of imagery appeals. Journal of Consumer Research, 32, 442–452.
Petrova, P. K., & Cialdini, R. B. (2008). Evoking the imagination as a strategy of influence.
In C. P. Haugtvedt, P. M. Herr, & F. R. Kardes (Eds.), Handbook of consumer psychology
(pp. 505–523). New York: LEA/Psychology Press.
Radvansky, G. A., Wyer, R. S., Curiel, J. M., & Lutz, M. F. (1997). Situational models and
abstract ownership relations. Journal of Experimental Psychology: Learning, Memory and
Cognition, 23, 1233–1246.
Radvansky, G. A., & Zacks, R. T. (1991). Mental models and he fan effect. Journal of Ex
perimental Psychology: Learning, Memory, and Cognition, 17, 940–953.
Sawyer, A. G., & Howard, D. J. (1991). Effects of omitting conclusions in advertisements

to involved and uninvolved audiences. Journal of Marketing Research, 28, 467–474.
Page 25 of 28

Schwarz, N. (1994). Judgment in a social context: Biases, shortcomings, and the logic of
conversation. In M. P. Zanna (Ed.), Advances in experimental social psychology (vol. 26,
pp. 123–162). San Diego: Academic Press.
Schwarz, N., Hippler, H. J., Deutsch, B., & Strack, F.(1985). Response scales: Effects on
behavioral reports and comparative judgments. Public Opinion Quarterly, 49, 388–395.
Searle, J. R. (1975). Indirect speech acts. In P. Cole & J. Morgan (Eds.), Syntax and seman
tics 3: Speech acts (pp. 59–82). New York: Academic Press.
Semin, G. R. (2000). Communication: Language as an implementational device for cogni

tion. European Journal of Social Psychology, 30, 585–612.
chology, 54, 558–568.
Semin, G. R., & Fiedler, K. (1991). The linguistic category model, is bases, applications
and range. In W. Stroebe & M. Hewstone (Eds.), European review of social psychology
(vol. 2, pp. 1–30). Chichester, UK: Wiley.
Sherman, S. J., Ahlm, K., Berman, L., & Lynn, S. (1978). Contrast effects and their rela
tionship to subsequent behavior. Journal of Experimental Social Psychology, 14, 340–350.
Sperber, D., & Wilson, D. (1986). Relevance. Cambridge, MA: Harvard University Press.
Srull, T. K., & Wyer, R. S. (1989). Person memory and judgment. Psychological Review, 96,
58–83.
Strack, F. (1994). Response processes in social judgment. In R. S. Wyer& T. K. Srull

(Eds.), Handbook of social cognition (2nd ed., vol. 1, pp. 287–322). Hillsdale, NJ: Erlbaum.
Suls, J. M. (1972). Two-stage model for the appreciation of jokes and cartoons: Informa
tion-processing analysis. In. J. H. Goldsein & P. E. McGhrr (Eds.), The psychology of hu
mor (pp. 81–100). San Diego, CA: Academic Press.
Trope, Y., & Liberman, N. (2003). Temporal construal, Psychological Review, 110, 403–
421.
Trope, Y., & Liberman, N. (2010). Construal-level theory of psychological distance. Psy
chological Review, 117, 440–443.
Vallacher, R. R., & Wegner, D. M. (1987). What do people think they are doing? Ac
(p. 249)
tion identification and human behavior. Psychological Review, 94, 3–15.
Wyer, R. S. (1974). Cognitive organization and change: An information processing ap

proach. Hillsdale, NJ: Erlbaum.
Page 26 of 28

Wyer, R. S. (2004). Social comprehension and judgment: The role of situation models, nar
ratives, and implicit theories. Mahwah, NJ: Lawrence Erlbaum Associates.
Wyer, R. S. (2008). The role of knowledge accessibility in cognition and behavior: Implica
tions for consumer information processing. In C. Haugtvedt, P. Herr & F. Kardes (Eds.),
Handbook of consumer psychology (pp. 31–76). Mahwah, NJ: Erlbaum.
Wyer, R. S., Adaval, R., & Colcombe, S. J. (2002). Narrative-based representations of so
cial knowledge: Their construction and use incomprehension, memory and judgment. In
M. P. Zanna (Ed.), Advances in experimental social psychology (vol. 34, pp. 133–197). San
Diego: Academic Press.
Wyer, R. S., Budesheim, T. L., & Lambert, A. J. (1990). The cognitive representation of
conversations about persons. Journal of Personality and Social Psychology, 58, 218–238.
Wyer, R. S., Budesheim, T. L., Lambert, A. J., & Swan, S. (1994). Person memory and judg
ment: Pragmatic and semantic influences on impressions formed in a social context. Jour
Wyer, R. S., & Collins, J. E. (1992). A theory of humor elicitation. Psychological Review,
99, 663–688.
Wyer, R. S., & Gruenfeld, D. H. (1995). Pragmatic information processing in social con
texts: Implications for social memory and judgment. In M. Zanna (Ed.), Advances in ex
perimental social psychology (vol. 27, pp. 49–92). San Diego, CA: Academic Press.
Wyer, R. S., Hung, I. W., & Jiang, Y. S. (2008). Visual and verbal processing strategies in
comprehension and judgment. Journal of Consumer Psychology, 18, 244–257.
Wyer, R. S., & Radvansky, G. A. (1999). The comprehension and validation of social infor
mation. Psychological Review, 106, 89–118.
Wyer, R. S., & Srull, T. K. (1989). Memory and cognition in its social context. Hillsdale, NJ:
Erlbaum.
Wyer, R. S., Swan, S., & Gruenfeld, D. H. (1995). Impression formation in informal conver
sations. Social Cognition, 13, 243–272.
Xu, A. J., & Wyer, R. S. (2010). Puffery in advertisements: The effects of media context,
communication norms and consumer knowledge. Journal of Consumer Research, 37, 329–
343.
Zillman, D., & Cantor, J. R. (1976). A disposition theory of humor and mirth. In A. J. Chap
man & H. C. Foot (Eds.), Humor and laughter: Theory, research and application (pp. 93–
116). New York: Wiley.
Robert S. Wyer Jr.
Page 27 of 28

Robert S. Wyer, Department of Marketing, HKUST Business School, The Hong Kong
University of Science and Technology
Page 28 of 28

Language and Attribution: Implicit Causal and Dispositional Information
Contained in Words
Language and Attribution: Implicit Causal and Disposi

tional Information Contained in Words
Klaus Fiedler and Tobias Krüger

This chapter is concerned with the role of language in mediating attributions, conceived
as speech acts that explain other people’s and one’s own behavior. Attributions entail
judgments about causes and reasons, motives and goals of action, about guilt, responsi
bility, authorship, and origins of emotions. As explained in the linguistic category model
and in construal-level theory, strategic shifts between abstract and concrete language use
can account for fundamental attribution error, actor-observer bias, egocentric bias, lin
guistic-intergroup bias, and behavioral phenomena associated with psychological dis
tance. Language can also trigger the attribution of truth and suspicion. The reviewed re
search highlights the fact that attributional inferences are wired into the semantic mean
ing of lexical stimuli, in particular verbs, adjectives, and nouns.
Keywords: abstractness, actor-observer bias, concreteness, consensus, consistency, construal-level theory, distinc
tiveness, fundamental attribution error, linguistic category model, linguistic intergroup bias, truth assessment
Verbal language constitutes an extremely rich and elaborate symbol system that renders
human communication fundamentally different from the communication systems of other
animals. Language represents an important cultural value in its own right, a target of
aesthetic evaluation and enjoyment, as evident in literature as a manifestation of linguis
tic art that can greatly contribute to human well-being and life fulfillment. Its uniqueness
and creativity gives rise to claims of authorship and renders language skills and dialects
major determinants of personal impressions and social categorization (Giles, Coupland, &
Coupland, 1991). Research in the behavioral sciences, though, is hardly interested in the
aesthetic and cultural value of language per se. It is rather interested in language as a
tool for social action and interaction (Glucksberg & Danks, 1975; Holtgraves, 1986, 2002;
Semin, 2000).
To highlight the behavioral functions of language, pragma-linguists (Searle, 1969) have

coined the terms “speech act” or “communicative act” to distinguish between different
pragmatic functions that language can fulfill: asking for explanations, making requests,
flirting and ingratiation, persuading, insulting, or expressing emotions. However, among
Page 1 of 27

Contained in Words
the multitude of functions of language, hardly any speech act is as ubiquitous, socially
meaningful, and consequential as making attributions—perceiving and judging the causes
of behavior (Jones et al., 1987).
Attributions are judgments about who is to blame or to praise, about causes of success
and failure, motives and goals of action, about guilt, responsibility, authorship, origins of
emotions, and sources of conflict. Such judgments determine reward and punishment, the
nature of resulting emotions, resource allocation and investment, (p. 251) approach and
avoidance behavior, and the potential to make rational decisions and engage in reason
able action. For example, the attribution of one’s physical health (to controllable vs. un
controllable conditions) determines future food and alcohol consumption and sports activ
ities (Steptoe & Wardle, 2001). Students’ attributions of their success and failure affect
their self-concept and achievement motivation (Weiner & Kukla, 2000). Jurors’ and
judges’ attributions of criminal behaviors determine sentencing decisions and punish
ment in the courtroom.
The aim of this chapter is to illuminate the manifold ways in which attributions depend on
the language used to describe social behavior. In this section, we first point out that the
semantic and pragmatic meaning of lexical terms can trigger specific attributions be
cause word semantics largely overlap with the kind of information that is required for at
tributional inferences. Indeed, specific attributions can be shown to be wired into the lexi
con, in accordance with the rules of prominent attribution theories. In a major review sec
tion, we then provide an overview of pertinent empirical evidence. Specifically, we illus
trate how language and in particular linguistic abstractness influences such phenomena
as fundamental attribution error, actor-observer bias, egocentric bias, linguistic-inter
group bias, and the relationship between abstractness and psychological distance. More
over, we illuminate the influence of implicit verb causality on internal versus external at
tributions, and further subsections will address the role of language in the attribution of
truth and pragmatic inference schemes. The concluding section addresses the claim that
linguistically driven attributions afford effective tools for social influence and social com
parison.
Dimensions of Attribution
Although distinct attribution theories focus on different subsets of dimensions, they over
lap in many respects. The “locus of attribution”—that is, whether an individual’s behavior
is explained in terms of internal traits and dispositions within the individual or in terms of
external constraints of the situation—depends on judgments on the following dimensions:
Does the individual’s behavior appear intentional and controlled (Jones & McGillis, 1976;
Weiner, 1985)? Is the behavior consensual (Kelley, 1967) or abnormal (Hilton & Slugosky,
1986) and at variance with cultural norms (Morris & Peng, 1994)? Does the behavior oc
cur consistently over time and occasions? Does the behavior appear to be directed at a
specific target or behavioral object? And, how salient is the individual relative to the situ
ational background (McArthur & Ginsberg, 1981)? For instance, an aggressive reaction is
Page 2 of 27

Contained in Words
more likely attributed to an internal cause within the individual subject if that subject is
the focus of attention, if the behavior is unusual and unlikely to be observed in other indi
viduals, and if the subject exhibits the same behavior consistently and under intentional
control.
In the most prominent attribution model, advocated by Kelley (1973), causal inferences
are supposed to follow an analysis of variance in a three-factorial design, involving the di
mensions consensus (How many other subjects exhibit the same behavior?), distinctive
ness (Does the behavior generalize to other objects?), and consistency (Does the behavior
generalize across time or opportunities?). The model predicts that a factor will be per
ceived as a cause if it bears the strongest covariation with the occurrence of an event or
behavior. Thus, a behavior is attributed to the subject when it covaries with the subject
factor (i.e., low consensus means that only this subject shows the behavior but no other
subjects) but does not covary with objects (low distinctiveness) and opportunities (high
consistency).
In a similar but notably distinct fashion, Jones and Davis’s (1965) correspondent-infer
ence model predicts that an observed behavior should be attributed to a corresponding
trait or person disposition if the person’s behavior is unusual (low consensus) and if the
assumption of intentionality is justified because the behavior is targeted toward a specific
goal and the subject is knowledgeable and capable enough to execute the behavior delib
erately (Jones & McGillis, 1976).
Mapping Word Semantics onto Attributional

Dimensions
Given this wide agreement between reasonable attribution theories as well as common
sense, it is intriguing to note that the semantic meaning of many words in the lexicon can
be described in terms of distinct coordinates on all major dimensions of attribution. One
might indeed say that implicit attribution models are built into the semantic meaning of
words. Based on a lexical analysis of the word classes that can appear as predicates in be
havior descriptions, the linguistic-category model (LCM; Semin & Fiedler, 1988, 1991)
distinguishes among four word classes with clearly different attributional parameters
(see Table 16.1).
Page 3 of 27

Contained in Words
Table 16.1 Four word classes that can be used as predicted in social
behavior descriptions, according to the linguistic category model
(Semin & Fiedler, 1988)
Word Features Exam

Class ples
Descrip Singular behav talk Low

tive Ac ioral episode of dial linguis
tion one particular touch tic ab
Verbs behavior with an push stract
(DAV) invariant physi ness
cal attribute
Reference to con
crete situation /
specific object
Context-depen
dent meaning
Usually no con
notation/valence
Interpre Singular behav cheat

tive Ac ioral episode flat
tion from a general ter
Verbs class of behav harm
(IAV) iors help
Reference to con
crete situation /
specific object
Context-indepen
dent meaning
Positive or nega
tive connotation/
valence
State Cognitive or af like

Verbs fective states envy
(SV) without clear be trust
ginning and end hate
Page 4 of 27

Contained in Words
Reference to a
social object, but
no specific situa
tion
Positive or nega
tive connotation/
valence
Adjec Enduring person reli High lin

tives al disposition able guistic
(ADJ) No reference to ag abstract
specific behavior, gres ness
object, or situa sive
tion intel
Highly interpre ligent
tive fair
Positive or nega
tive connotation/
valence
(p. 252) At the highest level of linguistic abstractness, the semantic meaning of most ad
jectives (ADJs; such as honest, clever, or hostile) entails low consensus (e.g., an honest
person is supposed to be unusually honest), low distinctiveness (honest behavior general
izes across objects), and high consistency (i.e., being honest is a stable trait over time).
Thus, the specific conditions for internal attributions to the sentence subject, according
to the Kelley model, are wired into the semantic meaning of ADJs, which afford a power
ful linguistic tool to suggest dispositional attribution.
At an intermediate level of linguistic abstractness, using interpretive action verbs (IAVs;

such as intimidate, insult, protect) as predicates also suggests a cause within the subject
but for different reasons. The meaning of these words does not emphasize temporal con
sistency or low consensus because all people sometimes show these behaviors. However,
they highlight the chief mediator of internal attribution in correspondent-inference theory
(Jones & McGillis, 1976), namely, intentional control. The semantic meaning of most IAVs
implies that the behavior described in a sentence was initiated intentionally and knowing
ly by the subject.
Whereas IAVs imply causal origins in the subject and emotional consequences in the sen
tence object (e.g., anger in Tom if “Paul insults Tom”), the reverse holds for emotional
(and cognitive) state verbs (SVs; like hate, loathe, admire). This class of verbs suggests a
causal origin in the object that stimulates an emotional reaction in the sentence subject
(e.g., respect in Paula for an asset in Mary if “Paula admires Mary”). Consequently, SVs
afford linguistic tools for external explanations and excuses of the subject’s behavior
Page 5 of 27

Contained in Words
(e.g., paraphrasing “Paul insults Tom” as “Paul abhors Tom” means to redirect the attri
bution to something aversive in Tom rather than something rude in Paul).
Finally, descriptive action verbs (DAVs such as write, touch, or kick) provide the most con
crete and attributionally most abstemious word category of the LCM. The attribution and
evaluation of behaviors expressed by these verbs depend heavily on the external situation
rather than the inherent meaning of the verb. Touching or writing can be positive or neg
ative, caused by the subject or the object, stable or unstable, normal or abnormal, de
pending on the behavioral context and the embedding situation. Although many DAV sen
tences are by default interpreted as (trivially) caused by the subject (i.e., the actor who
writes, touches, or kicks), this shallow internal attribution does not imply any stable or
unusual person attribute, nor does it greatly constrain the attributor’s future behavior to
ward that subject.
Note that the LCM suggests two possible strategies to influence attributions through lexi
cal word (p. 253) choice, abstractness, and locus of causality. As ADJs maximize the ab
stractness of behavior, they point to a stable and general disposition within the subject’s
personality as the most natural explanation, independently of the object person. In con
trast, DAVs minimize abstractness and therefore reveal little about stable dispositions
and person-specific traits. At an intermediate level of abstractness, both IAVs and SVs
represent behaviors as transitive relations peculiar to a specific dyad of subject and ob
ject. However, IAVs and SVs differ markedly in implicit verb causality (Holtgraves & Ray
mond, 1995; Rudolph & Försterling, 1997), that is, in the locus of causality implied by the
verb meaning.
Empirical Research on Lexical Influences on

Attribution
What empirical evidence is there to substantiate the notion developed so far, namely, that
the semantic meaning and the pragmatic use of language mediates and constrains the at
tribution process? This section is devoted to a review and synopsis of pertinent empirical
findings, along with a discussion of their theoretical and practical implications.
Abstractness and the Attribution of Behavior
As spelled out in Vallacher and Wegner’s (1987) action-identification theory, an action to
ward a certain goal can always be represented at different levels of linguistic abstract
ness. For instance, a student’s behavior observed in a situation that calls for resistance to
temptation might be expressed as saying “no” (DAV level), or as withstand (IAV level), or
as being morally strong (ADJ level). Raising the identification level from the lowest (DAV)
to the highest (ADJ) identification level is tantamount to attributing the behavior to an in
ternal disposition within the student (i.e., her moral strength). Using purely descriptive
language (DAV) amounts to refraining from attributions and evaluations. In any case, the
choice of a preferred abstractness level determines whether the details of contextualized
Page 6 of 27

Contained in Words
behavior (DAV), the actor’s intention (IAV), or a personality trait (ADJ) is the focus of com
munication.
Action-identification theory (Vallacher & Wegner, 1989) was one of the first frameworks
to conceptualize the causes and consequences of different abstractness levels used to
identify behaviors. Presupposing that the prevalent identification level has a systematic
impact on behavior, Vallacher and Wegner assumed that more abstract, higher level rep
resentations are generally preferred to low-level representations but that difficulties to
maintain an abstract level of representation cause a shift toward lower identification lev
els. This postulated preference for high identification levels suggests a linguistic basis for
a general preference for dispositional attributions, which provides the starting point for
the following empirical review.
Fundamental Attribution Bias

The so-called fundamental attribution bias (Nisbett & Ross, 1980) reflects the broad ten
dency to explain behavior in terms of internal characteristics within the person while ne
glecting the causal influence of situational factors. Linguistically, this bias can be inter
preted as a shift toward higher verbal abstractness (e.g., from behaviors expressed in
verbs to traits expressed in adjectives). Fiedler, Semin, and Bolten (1989) demonstrated
this general trend toward abstractness in an experimental retelling game (cf. serial repro
duction; Bartlett, 1932), in which participants were asked to pass on verbal descriptions
of different social roles (e.g., a male porter, a female teacher) to other participants, who
in turn had to reproduce the descriptions in their own words. Although all predicates of
the original texts were either only IAVs or only SVs, the reproductions were replete with
spontaneous dispositional attributions. The average ADJ rate rose from 0 percent to 37.6
percent in the first and 41.5 percent in the second round of reproduction.
In a similar but more refined study by Kashima (2000), a description of a man and a
woman who showed both gender-typical and gender-atypical behaviors was transmitted
through five-person communication chains. Although stereotype-atypical information was
more readily reproduced at the beginning, reproductions of typical behaviors that con
firmed the gender stereotype were predominant toward the end. This shift of verbal re
productions toward more and more expected information, together with the well-estab
lished tendency to use abstract language to communicate expected information (Wig
boldus, Semin, & Spears, 2000), testifies to the impact of verbal communication on the
formation of abstract language and consequent maintenance of stereotypes.
Additional studies by Lyons and Kashima (2003) corroborate that the increasing preva
lence of stereotypical attributions in serial event reproductions is indeed a genuine com
munication effect. That is, a strong shift toward stereotype-consistent (p. 254) information
occurred only when participants had the intention to communicate with somebody else,
but not when the reproduction task was introduced as a pure memory test. This finding
suggests that trying to tell a communicable story may be a crucial catalyst of the bias.
Consistent with this interpretation in terms of a communicative intention, the gradual
shift toward expected attributions proved to be particularly strong when communicators
Page 7 of 27

Contained in Words
believed that their audience and other community members shared and endorsed their
stereotypes (Lyons & Kashima, 2003).
Construal Level Theory and Psychological Distance

Having established the systematic influence of linguistic abstractness on attribution, the
next question pertains to the determinants of abstract language use. The evidence from
serial reproduction experiments indicates that a trend toward abstract language and ex
pectancy-congruent internal attributions is inherent to the communication game itself.
Another significant moderator of linguistic abstraction is psychological distance, as em
phasized in Trope and Liberman’s (2003, 2010) construal-level theory, which suggest a
general link between an object’s distance and its abstractness of mental representation.
In a typical study, Liberman and Trope (1998) asked participants to imagine engaging in
various activities “tomorrow” (short distance) or “next year” (long distance) and to de
scribe these activities (e.g., “reading a science fiction book,” “moving into a new apart
ment,” “taking an exam,” or “watching TV”). The resulting descriptions were then ana
lyzed in terms of their hierarchical linguistic structure (cf. Hampson, John, & Goldberg,
1986). Superordinate, high-level descriptions are assumed to fit the scheme “[descrip
tion] by [activity],” whereas subordinate low-level descriptions are assumed to match an
“[activity] by [description]” structure. For instance, you can broaden your horizon by
reading a book, but not vice versa. Conversely, you cannot flip pages by reading a book,
but it makes sense to say that you read a book by flipping pages. Accordingly, when de
scribing the activity “reading a book,” the former and latter expressions can be consid
ered high-level and low-level, respectively. As predicted by construal level theory, in
Liberman and Trope’s (1998) study, thinking about the far-away future solicited clearly
more abstract descriptions than thinking about the near-by future.
Nussbaum, Trope, and Liberman (2003) addressed the impact of temporal distance on the
fundamental attribution error in the original Jones and Harris (1967) paradigm. Partici
pants read an essay expressing a political position that was ostensibly imposed on the
writer through explicit instruction (situational constraint) or freely chosen by the essay
writer (disposition). Participants were asked to predict the author’s future attitude re
garding either the following day (concrete construal) or a year later (abstract construal).
As predicted, the fundamental attribution error—that is, the tendency to infer the writer’s
attitude from the essay regardless of external constraints—was more pronounced in the
distant as compared to the proximal future condition. Taken together these results cor
roborate the proposed association between abstractness and dispositional attribution.
Actor-Observer Bias
The actor-observer perspective, which can be conceived as a moderated version of the
fundamental attribution error, can also be interpreted within the framework of linguistic
abstractness. The actor-observer bias (Jones & Nisbett, 1972) is the finding that the fun
damental attribution error is more pronounced when observers make attributions for oth
ers’ behavior than when actors make attributions for their own behavior. As demonstrat
ed by Semin and Fiedler (1989), manipulations of the actor versus observer perspective
Page 8 of 27

Contained in Words
also elicit differential levels of linguistic abstractness and, consequently, different rates of
dispositional attribution. For instance, asking participants to describe why others like
their best friend (observer perspective) elicited more abstract answers than asking them
to describe why they like their own best friend (actor perspective).
Differential linguistic abstractness in actor- observer perspectives may be explained by

differences in the psychological distance between the self and others as the target. Be
cause the self is necessarily psychologically closer than others are, according to constru
al-level theory (Trope & Liberman, 2010), it should be construed in more concrete, con
textual terms. Conversely, others are more distant and, therefore, should be construed
rather abstractly; that is, in terms of dispositions. A study by Stephan, Liberman, and
Trope (2010) supports this idea: participants used more polite language when communi
cating to temporally or spatially distant others, and polite language, in turn, was associat
ed with the usage of rather abstract language.
(p. 255) Egocentric Bias

In apparent contradiction to the predictions of the actor-observer bias, people sometimes
tend to claim more responsibility for themselves than attributing responsibility to others.
This egocentric bias (Ross & Sicoly, 1979) was demonstrated, for instance, for self- and
partner attributions in close relationships. Typically, both partners rate themselves to be
more responsible than their partners for a number of actions like planning joint leisure
activities or causing arguments (Thompson & Kelley, 1981).
Using Thompson and Kelley’s set of items, Fiedler, Semin, and Koppetsch (1991) aimed to
disentangle the seeming contradiction between actor-observer and egocentric bias. As it
turned out, both an actor-observer bias toward inferring more dispositions from the part
ner than the self, as well as an egocentric bias toward attributing more contributions to
the self than the partner, emerged. However, these two attribution patterns were found
on different levels of linguistic abstractness. On the one hand, at the level of ADJs, the ac
tor-observer discrepancy was manifested in a higher number of trait ascriptions to the
partner in comparison to the self. On the other hand, at the level of IAVs, the egocentric
bias was observed in terms of more interpreted actions ascribed to the self in comparison
to the partner.
In retrospect, these findings highlight how linguistic analyses can illuminate theoretical
conflicts in attribution research: whereas many traditional actor-observer bias studies
were concerned with trait ascription at the more abstract ADJ level, egocentric-bias stud
ies typically dealt with actions and accomplishments identified at the more concrete IAV
level (cf. Jones & Nisbett, 1972; Ross & Sicoly, 1979). Such a fine-grained analysis of dif
ferential attribution patterns on different levels of linguistic abstraction provides an ex
cellent example of how the language approach fosters empirical integration and theoreti
cal refinement.
Page 9 of 27

Contained in Words
Self-Serving Bias and Ingroup-Serving Bias
Just as the actor-observer bias moderates the fundamental attribution bias, a self-serving
bias in attribution moderates the actor-observer bias. That is, the tendency toward more
dispositional attributions of others’ than one’s own behavior is most pronounced for nega
tive behaviors, when motivational forces work against self-attribution (Hastorf, Schneider,
& Polefka, 1970). The same tendency is reduced or even reversed for positive behaviors,
which are more readily ascribed to one’s own dispositions. By analogy, there is an in
group-serving bias by which people provide relatively more dispositional attributions for
positive ingroup and negative outgroup behaviors than for negative ingroup and positive
outgroup behaviors.
Linguistic Intergroup and Linguistic Expectancy Bias

The linguistically mediated version of such an ingroup-serving bias has been termed lin
guistic intergroup bias (LIB; Maass, this volume; Maass, 1999; Maass, Salvi, Arcuri, &
Semin, 1989). The LIB reflects a three-way interaction among abstractness, reference
groups, and valence. More precisely, the LIB is evident in more abstract and stable predi
cates used to describe positive ingroup and negative outgroup attributes but more con
crete and situation-specific predicates used to describe negative ingroup and positive out
group attributes. Such an ingroup-favoring attribution pattern has been found to charac
terize communications in contexts as diverse as local horse-racing teams (Maass et al.,
1989), political group membership (Anolli, Zunoli, & Riva, 2006), academic affiliation
(Wigboldus et al., 2000), and gender groups (Fiedler, Semin, & Finkenauer, 1993).
Originally, the LIB was explained in terms of self-serving, motivated reasoning (cf. Kunda,
1990; Maass et al., 1989). However, more recent evidence puts forward a different, essen
tially prosocial motive. The linguistic expectancy bias (LEB; Maass, Milesi, Zabbini, &
Stahlberg, 1995; Wigboldus et al., 2000) refers to people’s tendency to express expected
information on a higher level of abstractness than unexpected information. Expected in
formation is knowledge-consistent and typical and, therefore, can be communicated ab
stractly. Unexpected information, however, is knowledge-inconsistent and atypical, requir
ing a concrete communication style that allows for taking all the details into account.
Accordingly, the LEB is able to explain the attribution patterns proposed by the LIB: usu
ally, what is shared and expected is positive ingroup and negative outgroup information,
whereas negative ingroup and positive outgroup information is unshared and unexpected
because communication partners more likely stem from the same group than not. Hence,
the former kind of information is expressed more abstractly than the latter kind. A proso
cial, cooperative reason for the differentially (p. 256) abstract communication of expected
and unexpected information may be found in Grice’s (1975) maxim of quantity: shared
and expected information is communicated more efficiently in a more condensed, ab
stract format, just as required by the quantity maxim. In contrast, to communicate un
shared and unexpected information, it is most often necessary to go into more detail and
to provide a more elaborate picture of the situational background.
Page 10 of 27

Contained in Words
Fiedler, Bluemke, Friese, and Hofmann (2003) extended this line of research on the LIB
and LEB by taking strategic motives into account. They showed that when the communi
cation goal is to cooperatively share knowledge, then expected and unexpected informa
tion is communicated abstractly and concretely, respectively, consistent with the LEB.
Contrary to the LEB principle, however, when the communication goal is to persuade or
move the recipient toward an attitude or behavior, then unshared and unexpected infor
mation needs to be expressed in a more abstract fashion (e.g., mentioning more traits)
because abstract statements carry more meaningful interpretations, with more far-reach
ing implications. For instance, persuasive advertising has to be explicit about the assets
of a recommended product or political candidate and about the deficits of the alterna
tives. In an empirical investigation supporting this idea, Douglas and Sutton (2003)
showed that when participants were induced with the goal to describe cartoons in clearly
favorable or unfavorable terms, they resorted to abstract language descriptions, which
clearly conveyed the positive or negative valence, respectively.
Neither the debate about a LIB or a LEB underlying varying levels of linguistic abstrac
tion nor the latter findings on strategic language usage (Douglas & Sutton, 2003; Fiedler
et al., 2003) qualify the basic truth that abstract communication serves to induce inter
nal, dispositional attributions. Rather, the preceding discussion on moderating conditions
of the LIB and LEB was meant to emphasize the fact that a language approach to attribu
tion has led to considerable theoretical refinement. Growing research highlights the need
to take the type of communicative act into account. For instance, mechanisms that have
been demonstrated in cooperative communication can differ drastically from mechanisms
in settings that encourage noncooperative communication, such as deception, flattering,
or instrumental manipulation (Fiedler, 2008a).
Nomina Sunt Omina

Going beyond different classes of verbs and adjectives, a recent set of studies by Car
naghi, Maass, and colleagues (2008) investigated the inductive consequences of nouns on
attributions generated by listeners. Providing participants with expectancy-consistent
nouns (e.g., artist, athlete, genius) instead of trait adjectives (e.g., artistic, athletic, bril
liant) in descriptions of fictitious target persons led to stronger inferences of stereotypi
cal person attributes. In other words, noun phrases like “is a criminal” or “is a homosexu
al” caused even more far-reaching stereotype-consistent inferences about the target than
ADJ phrases like “is criminal” or “is homosexual.” However, expectancy- inconsistent
nouns were more effective in disconfirming stereotypic attributions than were ADJs. A
positive, prosocial impression of a target is more likely changed if “he is a criminal” than
if “he is criminal.” Thus, the authors conclude that nouns, in comparison to ADJs, do not
facilitate all kinds of inferences, but specifically those inferences that are based on ex
pected attributes.
In any case, many nouns may be even more informative, more enduring, more general,
and thus more predictive for future events than ADJs (Carnaghi, Maass, et al., 2008).
However, in spite of these abstract implications, nouns are also more strongly related to
easily imaginable and materially existing objects, which may render them more concrete.
Page 11 of 27

Contained in Words
Most likely, the word class of nouns—ranging from very concrete (e.g., kitchen table, cof
fee cup) to highly abstract (e.g., love, wisdom)—is too diverse to establish a general rule.
In other words, nouns that represent different levels of abstraction may generate differ
ent inference processes.
Specific Attributions Triggered by Abstractness

In addition to suggesting dispositional, stable, and general causes, abstract language has
been shown to facilitate several other, more specific attributional inferences. For exam
ple, strategic variation between abstract and concrete language can be used to regulate
psychological distance (Semin, 2007). Abstract terms induce and signal higher social dis
tance, and perhaps a stronger status asymmetry, than do concrete expressions. Also, ab
stractness and psychological distance are associated with polite language (Stephan et al.,
2010).
In a related vein, Hansen and Wänke (2011) conducted several experiments showing that
both consumers and advertisers describe luxury products (p. 257) in more abstract lan
guage than they do ordinary products. Conversely, abstractly described products are ex
perienced as more luxurious than are concretely described products. Because luxury
goods are usually perceived as distant and hard to obtain for people of low social status
and only accessible to high-status people, the authors interpret the link between abstract
ness and luxury as mediated by status distance.
The same interpretation applies to Magee, Milliken, and Lurie’s (2010) finding that highly
abstract language indicates high position power. Again, the attribution of power is proba
bly mediated by the increasing degree of psychological distance associated with increas
ing levels of linguistic abstractness (Smith & Trope, 2006).
Fluency
Fluency, or “the experienced ease of ongoing conceptual or perceptual cognitive process
es” (Unkelbach, 2006), has also been found to be an important antecedent of linguistic
abstractness (Alter & Oppenheimer, 2008). Asking participants to describe the city of
New York in a few sentences was significantly impacted by the fluency experience evoked
by the item’s presentation style. In particular, independent judges rated participants’ de
scriptions as more abstract when the experimental material was presented in an unusual,
low-fluency (10-point, 25 percent gray, italicized Arial) in contrast to an easy-to-read,
high-fluency font (12-point Times New Roman).
Implicit Verb Causality
Abstractness is by no means the only semantic aspect of language that impacts the attri
bution process. As already mentioned, linguistic stimuli also differ systematically in the
locus of causality they imply. The phenomenon of implicit verb causality (Kanouse & Abel
son, 1967; Brown & Fish, 1983; Fiedler & Semin, 1988; Garvey & Caramazza, 1974; Holt
graves & Raymond, 1995; Rudolph & Försterling, 1997), in particular, refers to the fre
quently demonstrated finding that almost all action verbs (IAVs) attribute the cause of be
Page 12 of 27

Contained in Words
havior to the sentence subject. In contrast, the vast majority of emotional or cognitive
subjective state verbs (SVs) suggest a causal origin in the sentence object. Thus, depend
ing on whether the same prosocial behavior is expressed in an IAV (Peter votes for Paul)
or in a SV (Peter admires Paul), the behavior is implicitly attributed to the subject (Peter)
or the object (Paul), respectively. Likewise, the verbs used in newspaper headlines about
politicians or celebrities may either facilitate internal attributions (using IAVs like insult,
cheat, support) or external attributions (using SVs like abhor, hate, admire).
Explaining Implicit Verb Causality

Various theoretical explanations have been suggested to account for the strength and reg
ularity of the implicit verb causality phenomenon. Brown and Fish (1983) considered and
discarded a lexical account, which states that, in the English lexicon, most ADJs derived
from IAV word stems point to the logical subject (e.g., helpful is an attribute of the sub
ject who helps), whereas most ADJs derived from SV word stems belong to the logical ob
ject (e.g., admirable is an attribute of the object who is admired). However, a closer analy
sis of the lexicon shows that there is a substantial number of verbs whose derived adjec
tives are inconsistent with this rule. For instance, the SV respect implies object causation
even though lexical knowledge includes the subject-referent ADJ respectful. Moreover,
implicit verb causality is less pronounced in other languages, like German, which offer
both subject- and object-referent ADJs for the majority of verbs. Morphological con
straints alone (i.e., word-stem constraints) are therefore unlikely to provide a sufficient
account of implicit verb causality.
Several researchers have tried to explain verb causality in terms of implicit knowledge
about the implicit consensus, distinctiveness, and consistency of the verbs, in accordance
with Kelley’s (1967) attribution model (cf. Kanouse & Abelson, 1967; Rudolph & Förster
ling, 1997). Indeed, given an IAV sentence like A attacks B, people are typically reluctant
to assume that many other subjects also attack others (low consensus), but they do as
sume that the same subject will attack other objects as well (low distinctiveness). Appar
ently, then, the IAV sentence expresses something peculiar to the subject but independent
of the object, and such a pattern should lead to subject attribution. In contrast, SV sen
tences like A abhors B give rise to relatively higher implicit consensus (other subjects
probably also abhor the object) and higher distinctiveness (little generalization to other
objects), thus suggesting an object attribution (cf. Rudolph & Försterling, 1997).
The question, however, is whether causal inferences are consequences or antecedents of

the implicit consensus, distinctiveness, and consistency of the verbs. That Martha encour
ages Jim implies a behavior caused by Martha may be evident (p. 258) immediately, well
before it generates implicit estimates of how many other subjects encourage Jim and how
many other objects Martha encourages. This would suggest that causal inferences pre
cede generalizing inferences, not the other way around. The priority and independence of
verb causality is also evident in the finding that IAVs and SVs conserve their causal impli
catures if temporal quantifiers obscure the generalizing inferences (Fiedler, 1978). Thus,
Page 13 of 27

Contained in Words
if Martha always encourages Jim, Martha still appears as the causal origin, although it is
quite unlikely that Martha also always encourages many other people.
Holtgraves and Raymond (1995) advanced a priming approach to explaining the locus of
causality associated with IAVs and SVs. They demonstrated that, in an unexpected memo
ry test, the cued recall performance was systematically enhanced for the subjects of IAV
sentences but not for the objects included in SV sentences. Note that, in this account, too,
the implicit causality of the verb is manipulated as an independent variable, whereas the
specificity of the verb for the subject (consensus) or object (distinctiveness) is assessed as
a dependent variable. Thus, rather than supporting the notion that verb causality reflects
the subject- and object-specificity of verbs, these findings may rather indicate that prima
ry causal meaning is an antecedent condition of secondary inferences about specificity
versus generality.
Even the syntactic or grammatical form of the verb can have a systematic influence on
the attribution of intentions to the sentence subject, as demonstrated in recent research
by Hart and Albarracín (2011). If participants read what a person was doing rather than
what a person did, the cognitive accessibility of intention-related concepts was enhanced.
The cognitive representation elicited by the imperfective aspect, which highlights the on
going and incomplete state of action, apparently facilitates intentional attributions. This
effect of the imperfective aspect was mediated by the imagination of more detailed ac
tions from which the person’s intention could be reasonably inferred.
Other demonstrations of syntactic influences on attribution refer to subject-object order

ing and the direction of comparison statements. Thus, the question “Did you dance with
him?,” in which “you” is the sentence subject, tends to solicit more internal self-attribu
tions than the question “Did he dance with you?,” in which the responder (addressed as
“you”) is in the object position. Accordingly, the responder would feel more responsible
for causing the behavior in the former than in the latter case (see also Wänke & Reutner,
2011, for a review of linguistic influences on comparison judgments).
Implicit Causality Affecting Explicit Attribution Tasks

Consistent with such a claim of linguistic relativity (Fiedler, 2008; Lucy, 1997), some evi
dence suggests that linguistic recoding or paraphrasing may strongly moderate the attri
bution of behavior presented on video rather than in text. In a series of experiments by
Fiedler, Armbruster, Nickel, Walther, and Asbeck (1996), participants watched a video-
taped adversarial discussion between agents of consumer organizations and managers of
the food industry. Although the discussants’ actual behavior was manifest and experimen
tally controlled in the video film, the final attribution of traits was strongly moderated by
the type of verb used for the intermediate question items. If participants had to verify the
observation of positive IAVs in a target discussant (Did he help, encourage, support oth
ers?), their subsequent attributions of positive traits corresponding to these IAVs in
creased, even when the IAV questions had been denied. Similarly, merely thinking about
negative IAVs (Did he insult, attack, devalue others?) induced relatively more negative
trait attributions. Interestingly, however, and consistent with the notion of implicit verb
Page 14 of 27

Contained in Words
causality, merely thinking about negative SVs led to less negative trait attributions, ap
parently because SVs suggest external causation or excuses for negative behaviors. By
the same token, positive trait attributions decreased when positive SV questions served to
suggest external causes for positive behaviors.
Attribution of Truth
One ubiquitous attribution task is truth assessment. Whenever we engage in everyday

communication, negotiations, job interviews, intimate discussions in close relationships,
or eyewitness questioning, or when we read a newspaper or science article, the question
is whether the communication is completely true or somehow biased toward our commu
nication partner’s interest, misunderstandings, or self-presentation concerns. In this con
text, an external attribution suggests attribution of truth (e.g., the used-car dealer’s en
thusiasm really reflects the quality of the car he is praising), whereas an internal attribu
tion raises suspicion concerning (p. 259) other causes or motives (e.g., he has vested inter
est or lacks sufficient expertise).
Concreteness and Apparent Truth

Consistent with the definition in the American Heritage Dictionary of the English Lan
guage that concrete means “actual” or “existing in reality,” whereas abstract means
“apart from concrete existence” (Houghton Mifflin, 2000), propositions expressed in con
crete language are more likely to be truthful and veridical than are abstract propositions.
According to Hansen and Wänke (2010), there are at least three reasons for the impact of
concreteness on apparent truth: concrete phrases are recognized more quickly (Kroll &
Merves, 1986), they are more imaginable and memorable (ter Doest & Semin, 2005), and
ease of imagination and mental simulation increases subjective likelihood (Garry, Man
ning, Loftus, & Sherman, 1996). Several findings reported by Hansen and Wänke (2010)
lend support to language concreteness as a determinant of truth attributions. First, when
equivalent statements were presented concretely (The Naab flows into the Danube)
rather than abstractly (The Naab is a confluent of the Danube), they were more often
judged to be true. Second, this truth advantage of concrete statements increased when
participants were primed with a concrete mindset (through answering “how” questions)
rather than an abstract mindset (answering “why” questions). And third, concrete state
ments appeared particularly truthful when they were presented in a spatially proximal
rather than distant location. These findings lend themselves to an interpretation in accor
dance with construal-level theory’s (Trope & Liberman, 2010) notion that low distance
implies high probability and realism.
An alternative account of the concreteness effect can be found in Johnson and Raye’s
(1981) source-monitoring and reality-monitoring framework, according to which the vivid
perceptual and affective cues that characterize concrete statements are associated with
actually experienced as opposed to mentally imagined or generated representations (Ja
coby & Whitehouse, 1989). Interestingly, within this theoretical context, the ultimate im
pact of concreteness may depend on the attributor’s hypothesis, as shown by Johnson,
Bush, and Mitchell (1998). Rich detail and embellishment increased the attribution of
Page 15 of 27

Contained in Words
truth if, in a nonsuspicious condition, the hypothesis was that the statements are true. In
contrast, in a suspicious condition, detail richness and embellishment increased the sub
jective validity of the hypothesis that statements are deceptive.
In criminal law, when expert witnesses have to decide on the credibility of witness re
ports (typically in trials dealing with rape and sexual offense), the standard procedure of
criteria-based content analysis (CBCA; Steller & Köhnken, 1989; Vrij, 2005) relies heavily
on the linguistic symptoms of truth. In addition to the detail richness criterion, 18 other
linguistic criteria (cf. Table 16.2) determine legal experts’ attributions of the witness’ ver
bal report either to real experience or to internal generation or fabrication. In the ab
sence of physical evidence, responsible and highly consequential legal decisions are thus
based on the belief that the causal origin of witness reports is wired into language. Al
though the validity of CBCA is restricted and subject to several problematic assumptions
(Vrij, 2005), the incremental validity gained by this methodology can improve court deci
sions, particularly when judgmental uncertainty is very high.
Page 16 of 27

Contained in Words
Table 16.2 Linguistic criteria used for credibility analysis in criminal

law relying on criteria-based content analysis (Steller & Köhnken,
1989)
General characteristics
1. Logical structure
2. Unstructured production
3. Quantity of details
Specific contents
4. Contextual embedding
5. Descriptions of interactions
6. Reproduction of conversation
7. Unexpected complications during the incident
8. Unusual details
9. Superfluous details
10. Accurately reported details misunderstood
11. Related external associations
12. Accounts of subjective mental state
13. Attribution of perpetrator’s mental state
Motivation-related contents
14. Spontaneous corrections
15. Admitting lack of memory
16. Raising doubts about one’s own testimony
17. Self-deprecation
18. Pardoning the perpetrator
Offence-specific elements
19. Details characteristic of the offence
Modern methods of computer-based text analysis (Tausczik & Pennebaker, 2010) have al
so been applied successfully for the detection of deception. (p. 260) In a review provided
by Newman, Pennebaker, Berry, and Richards (2003), liars compared to truth-tellers were
shown to be lower in linguistic-cognitive complexity, to use fewer self- and other-refer
ences, and to use more negative emotion words. By and large, these findings about lin
guistic markers in deceptive communication appear to validate the aforementioned no
tion (cf. Hansen & Wänke, 2010) that truth decreases with increasing psychological dis
tance (as manifested in reduced complexity, paucity of references, and increased negativi
ty).
Linguistic style and contents not only determine the outcomes of attributions to a consid
erable degree, but also when attributions will be made. Given that unexpected events are
more likely to solicit causal reasoning and attributions than expected events (Hastie,
1984), the linguistic framing of narrative episodes as surprising or expectancy-incongru
Page 17 of 27

Contained in Words
ent can be used to trigger attributions. Conversely, engaging in causal explanations of be
haviors presupposed to be real results in elaborate memory representations of these
events so that they are more likely to be recalled and considered to be true. Thus, the re
lationship between detailed representation and apparent truth seems to be bidirectional.
Pragmatic Inference Schemes
Considering truth and validity from a rational point of view, it is interesting to note that
linguistic terms greatly constrain the a priori likelihood that specific attributions are true.
From propositional logic, it is well known that an existence statement like “some people
can be hypnotized” is more likely true than a universal statement like “all people can be
hypnotized”; in other words, statements with an “all” quantifier are more easily falsified
than weaker statement with a “some” qualifier. Similar constraints exist in ordinary lan
guage because the words we use to interpret and evaluate behavior vary systematically in
their implicit quantifiers or scope, to use a term coined by Gidron, Koehler, and Tversky
(1993). Some attributes, like honest or reliable, have a high implicit quantifier; somebody
must always or almost always engage in honest and reliable behaviors to elicit the trait
attributions of honest and reliable. Other terms with intermediate implicit quantifiers,
such as clever or aggressive, already allow for and suggest trait attributions if corre
sponding behaviors are exhibited sometimes. Very few observations are required to call
somebody dishonest, and a single behavioral proof is sufficient for the attribution murder
er. Thus, semantic knowledge imposes different thresholds on the attribution of different
traits. Less evidence is required to verify, and more evidence is needed to falsify, many
negative attributes (like dishonest, murderer) compared to many positive attributes (hon
est, friendly).
Related to the scope of an attribute is the diagnosticity of behaviors with regard to attri
butional inferences (Skowronski & Carlston, 1987). Behaviors corresponding to attributes
with a low implicit quantifier (e.g., dishonest) are typically more diagnostic than are be
haviors corresponding to attributes with a high implicit quantifier (e.g., honest). As Reed
er and Brewer (1979) have shown in their seminal work on attributional inference
schemes, the “big two” in personality research, morality and ability, have radically differ
ent implications for the relative diagnosticity of positive and negative behaviors. In the
morality domain, negative behaviors tend to be more diagnostic than positive behaviors.
A single lie is sufficient to call somebody a liar or a dishonest person, whereas many con
sistent observations of truthful communications are necessary to justify an honesty attri
bution. To the contrary, in the ability domain, a single instance of artistic skiing perfor
mance is enough to attribute high ability, whereas even repeated failures do not provide
unequivocal evidence for lack of ability.
Semantic knowledge not only constrains the inferential relations between behaviors and
attributed traits, but also the hierarchical relation between superordinate and subordi
nate attributes. The meaning of talkative is included in the meaning of extravert, but ex
travert is not included in talkative. According to Hampson et al. (1986), the availability of
traits at different hierarchical levels of mental representation provides another degree of
Page 18 of 27

Contained in Words
freedom to personality psychologists and language users alike. Language offers a chance
to frame the domain of attributional judgments more or less restrictively, thus allowing ei
ther a broad variety of behaviors (e.g., friendly), or only very few specific behaviors (e.g.,
corrupt), to verify the attribute focused in language. These constraints not only apply to
the original formation of attributions, but also to their modification and falsification.
There are more ways of falsifying the attribution of friendliness than corruptness.
Conclusion
Starting from an analysis of the semantic properties of lexical stimuli, which strongly cor
respond to the dimensions that determine attributional inferences, we have reviewed var
ious ways in which the language used to describe behaviors or events can (p. 261) influ
ence their explanation and evaluation. It is no wonder that the semantic information con
veyed by words and phrases concerning the abnormality (consensus), specificity (distinc
tiveness), consistency, emotionality, and intentionality (controllability) of behavior have a
systematic impact on the outcome of attributions that are expected to be determined by
exactly these properties.
Although a general tendency toward linguistically abstract descriptions of social behav

iors relates to the fundamental attribution bias, the tendency to describe others in more
abstract terms than oneself produces an actor observer-bias. Likewise, the linguistic in
tergroup bias (i.e., the selective tendency to describe negative outgroup and positive in
group behaviors in more abstract terms than positive outgroup and negative ingroup be
havior) constitutes a language-driven ingroup-serving bias. In addition to such influences
mediated by linguistic abstractness, we have also dealt with the implicit causality of ac
tion verbs versus state verbs and with the implications of fluency and specific semantic
cues on the attribution of truth, and with markedly different inference schemes triggered
by the semantics of ability and morality.
A final comment is in order regarding the theoretical and practical implications of the evi
dence presented in this chapter. One crucial question that suggests itself concerns the
causal direction of the relationship between language and attribution. Does language on
ly reflect and convey the outcome of cognitive attribution processes that take place with
in the human mind, or does language, as a toolbox and as a semantic system, acquire its
own autonomy and causal power that can impact on cognition and behavior? This old
question of the cognitive primacy of linguistic relativity has intrigued and occupied stu
dents of language from Aristotle and Plato to modern experimental psychology (Fiedler,
2008). The answer proposed by the research presented in this chapter suggests that lan
guage can indeed function as a causal antecedent of judgments, decisions, and social ac
tions. The linguistic tools used by attorneys to persuade judges and jurors in the court
room, readers’ inference schemes triggered by journalists, and the verbal tools used by
interviewers to evaluate respondents’ answers are rarely determined by cognitive reason
ing processes and deliberate intentions. Very often, language users have no introspective
insight into and control over the linguistic tools they are using and their implications and
Page 19 of 27

Contained in Words
side effects (Maass, 1999). Teachers, tourist guides, or novel writers may resort to ab
stract language with overly strong implications and evaluative meaning in order to raise
their communication partners’ attention or provide them with a meaningful context and
thereby fail to notice that they are inadvertently distributing unwarranted stereotypes.
Survey researchers are typically not aware of the degree to which the response options
they are providing determine the responses they get (Schwarz & Sudman, 1996). As well,
experts and politicians hardly understand that the verbal framing of risky decisions, for
instance in terms of either mortality or survival rates (Tversky & Kahneman, 1981), has a
profound effect on risk taking in medical, economic, and political contexts.
To the extent that the interpersonal consequences of language and communication are
detached from the communicators’ intrapersonal cognition and intention, it seems justi
fied to consider language a causal origin in and of itself. Although the strength of the in
fluence that words and verbal phrases can exert on people’s cognition and behavior is
certainly limited, this influence can nevertheless be substantial and sometimes stronger
than alternative means of social influence. For instance, if there is no doubt that a defen
dant in a court trial has done great harm to a victim, neither hypnosis nor conditioning
nor any other influence technique can create a positive impression of an apparently mean
person. However, whereas the defendant’s negative behaviors cannot be denied, and por
traying his behavior as prosocial would be worthless, linguistic means can still be used to
downplay negative aspects (e.g., by using concrete language for the manifest facts) and
to augment positive aspects (e.g., by using abstract language for nice features in the
defendant’s mentality). Because linguistic effects on attribution are subtle and unobtru
sive and evade the communication partners’ awareness, they may often have a more real
istic chance to exert social influence than intrusive arguments or coercive forces that are
likely to be met with resistance and reactance. One prominent task for future research,
indeed, is to investigate the impact of language on attribution in the context of social in
fluence processes, in such applied domains as advertising, therapy, education, and negoti
ation.
Acknowledgments
The research underlying this chapter was supported by a Koselleck grant awarded to the
first author by the Deutsche Forschungsgemeinschaft. (p. 262) Helpful comments by Alex
Koch and Michaela Wänke on a draft of this paper are gratefully acknowledged. Corre
spondence should be addressed to [email protected]
References
Alter, A. L., & Oppenheimer, D. M. (2008). Effects of fluency on psychological distance
and mental construal (or why New York is a large city, but New York is a civilized jungle).
Psychological Science, 19(2), 161–167.
Anolli, L., Zurloni, V., & Riva, G. (2006). Linguistic intergroup bias in political communica
tion. Journal of General Psychology, 133, 237–255.
Page 20 of 27

Contained in Words
Bartlett, F. (1932). Remembering: a study in experimental and social psychology. Cam
bridge, England: Cambridge University Press.
Brown, R., & Fish, D. (1983). The psychological causality implicit in language. Cognition,
14(3), 237–273. doi:10.1016/0010-0277(83)90006-9
Carnaghi, A., Maass, A., Gresta, S., Bianchi, M., Cadinu, M., & Arcuri, L. (2008). Nomina
sunt omina: On the inductive potential of nouns and adjectives in person perception. Jour
nal of Personality and Social Psychology, 94(5), 839–859. doi:10.1037/0022–
3514.94.5.839
Douglas, K. M., & Sutton, R. M. (2003). Effects of communication goals and expectancies
on language abstraction. Journal of Personality and Social Psychology, 84(4), 682–696.
doi: 10.1037/0022–3514.84.4.682
Fiedler, K. (1978). Kausale und generalisierende Schluesse aufgrund einfacher Saetze.

Zeitschrift fuer Sozialpsychologie [Causal and generalizing inferences based on simple
sentences], Zeitschrift fuer Sozialpsychologie, 9, 37–49.
Fiedler, K. (2008a). Language—A toolbox for sharing and influencing social reality. Per
spectives on Psychological Science, 3, 38–47.
Fiedler, K., Armbruster, T., Nickel, S., Walther, E., & Asbeck, J. (1996). Constructive biases
in social judgment: Experiments on the self-verification of question contents. Journal of
Personality And Social Psychology, 71(5), 861–873. doi:10.1037/0022-3514.71.5.861
Fiedler, K., Bluemke, M., Friese, M., & Hofmann, W. (2003). On the different uses of lin
guistic abstractness: From LIB to LEB and beyond. European Journal of Social
Psychology, 33(4), 441–453. doi: 10.1002/ejsp.158
Fiedler, K., & Semin, G. R. (1988). On the causal information conveyed by different inter
personal verbs. Social Cognition, 6, 21–39.
Fiedler, K., Semin, G. R., & Bolten, S. (1989). Language use and reification of social infor
mation: Top-down and bottom-up processing in person cognition. European Journal of So
cial Psychology, 19(4), 271–295. doi: 10.1002/ejsp.2420190403
Fiedler, K., Semin, G. R., & Finkenauer, C. (1993). The battle of words between gender
groups: A language-based approach to intergroup processes. Human Communication Re
search, 19, 409–441.
Fiedler, K., Semin, G. R., & Koppetsch, C. (1991). Language use and attributional biases
in close personal relationships. Personality and Social Psychology Bulletin, 17, 147–156.
Garry, M., Manning, C. G., Loftus, E. F., & Sherman, S. J. (1996). Imagination inflation:
Imagining a childhood even inflates confidence that it occurred. Psychonomic Bulletin &
Review, 3, 208–214.
Page 21 of 27

Contained in Words
Garvey, C., & Caramazza, A. (1974). Implicit causality in verbs. Linguistic Inquiry 5(3),
459–464.
Gidron, D., Koehler, D. J., & Tversky, A. (1993). Implicit quantification of personality traits.
Personality and Social Psychology Bulletin, 19(5), 594–604. doi:
10.1177/0146167293195011
Giles, H., Coupland, J., & Coupland, N. (Eds.). (1991). Contexts of accommodation: devel
opments in applied sociolinguistics. Cambridge, UK: Cambridge University Press.
Glucksberg, S., & Danks, J. H. (1975). Experimental psycholinguistics: An introduction.

Oxford England: Lawrence Erlbaum.
Grice, H. P. (1975). Logic and conversation. In P. Cole & J. L. Morgan (Eds.), Syntax and
semantics 3: Speech acts (pp. 41–58). New York: Academic Press.
Hampson, S. E., John, O. P., & Goldberg, L. R. (1986). Category breadth and hierarchical
structure in personality: Studies of asymmetries in judgments of trait implications. Jour
nal of Personality and Social Psychology, 51(1), 37–54. doi:10.1037/0022–3514.51.1.37
Hansen, J., & Wänke, M. (2010). Truth from language and truth from fit: The impact of
linguistic concreteness and level of construal on subjective truth. Personality and Social
Psychology Bulletin, 36(11), 1576–1588. doi:10.1177/0146167210386238
Hansen, J., & Wänke, M. (2011). The abstractness of luxury. Journal of Economic Psycholo
gy, 32(5), 789–796. doi:10.1016/j.joep.2011.05.005
Hart, W., & Albarracín, D. (2011). Learning about what others were doing: Verb aspect
and attributions of mundane and criminal intent for past actions. Psychological Science,
22(2), 261–266. doi:10.1177/0956797610395393
Hastorf, A. H., Schneider, D. J., & Polefka, J. (1970). Person perception. Oxford England:
Addison-Wesley.
Hastie, R. (1984). Causes and effects of causal attribution. Journal of Personality and So
cial Psychology, 46(1), 44–56. doi:10.1037/0022–3514.46.1.44
Hilton, D. F., & Slugoski, B. (1986). Knowledge-based causal attribution: The abnormal
conditions focus model. Psychological Review, 93, 75–88.
Houghton Mifflin. (2000). American heritage dictionary of the English language (4th ed.).
Retrieved from www.thefreedictionary.com
Holtgraves, T. (1986). Language structure in social interaction: Perceptions of direct and

indirect speech acts and interactants who use them. Journal of Personality and Social Psy
chology, 51(2), 305–314. doi:10.1037/0022–3514.51.2.305
Holtgraves, T. (2002). Language as social action: Social psychology and language use.
Mahwah, NJ: Lawrence Erlbaum.
Page 22 of 27

Contained in Words
Holtgraves, T., & Raymond, S. (1995). Implicit causality and memory: Evidence for a
priming model. Personality and Social Psychology Bulletin, 21(1), 5–12. doi:
10.1177/0146167295211002
Jacoby, L. L., & Whitehouse, K. (1989). An illusion of memory: False recognition influ
enced by unconscious perception. Journal of Experimental Psychology: General, 118(2),
126–135. doi:10.1037/0096–3445.118.2.126
Jones, E. E., & Davis, K. E. (1965). From acts to dispositions: The attribution process in
person perception. In L. Berkowitz (Ed.), Advances in experimental social psychology, Vol
ume 2 (pp. 219–266). New York: Academic Press.
Jones, E. E., & Harris, V. A. (1967). The attribution of attitudes. Journal of Experimental
Social Psychology, 3(1), 1–24.
Jones, E., Kanouse, D. E., Kelley, H. H., Nisbett, R. E., Valins, S., & Weiner, B.
(p. 263)
(Eds.). (1987). Attribution: Perceiving the causes of behavior. Hillsdale, NJ: Lawrence Erl
baum.
Jones, E. E., & McGillis, D. (1976). Correspondent inferences and the attribution cube: A
comparative appraisal. In J. H. Harvey, W. J. Ickes & R. F. Kidd (Eds.), New directions in
attribution research (Vol 1, pp. 389–420). Hillsdale, NJ: Erlbaum.
Jones, E. E., & Nisbett, R. E. (1972). The actor and the observer: Divergent perceptions of
the causes of behavior. In E. E. Jones, D. E. Kanouse, H. H. Kelley, R. E. Nisbett, S. Valins,
& B. Weiner (Eds.), Attribution: perceiving the causes of behavior (pp. 79–94). Morris
town, NJ: General Learning Press.
Johnson, M. K., Bush, J. G., & Mitchell, K. J. (1998). Interpersonal reality monitoring:
Judging the sources of other people’s memories. Social Cognition, 16(2), 199–224.
Johnson, M. K., & Raye, C. L. (1981). Reality monitoring. Psychological Review, 88(1), 67–
85. doi:10.1037/0033–295X.88.1.67
Kanouse, D. E., & Abelson, R. P. (1967). Language variables affecting the persuasiveness
of simple communications. Journal of Personality And Social Psychology, 7(2, Pt.1), 158–
163. doi:10.1037/h0025003
Kashima, Y. (2000). Maintaining cultural stereotypes in the serial reproduction of narra

tives. Personality and Social Psychology Bulletin, 26(5), 594–604.
Kelley, H. H. (1967). Attribution theory in social psychology. In D. Levine (Ed.), Nebraska

symposium on motivation, Vol. 15 (pp. 192–238). Lincoln: University of Nebraska Press.
Kelley, H. H. (1973). The processes of causal attribution. American Psychologist, 28(2),

107–128. doi:10.1037/h0034225
Page 23 of 27

Contained in Words
Kroll, J. F., & Merves, J. S. (1986). Lexical access for concrete and abstract words. Journal
of Experimental Psychology: Learning, Memory, and Cognition, 12, 92–107.
Kunda, Z. (1990). The case for motivated reasoning. Psychological Bulletin, 108(3), 480–
498. doi: 10.1037/0033–2909.108.3.480
Liberman, N., & Trope, Y. (1998). The role of feasibility and desirability considerations in
near and distant future decisions: A test of temporal construal theory. Journal of Personal
ity and Social Psychology, 75(1), 5–18. doi:10.1037/0022–3514.75.1.5
Lyons, A., & Kashima, Y. (2003). How are stereotypes maintained through communica
tion? The influence of stereotype sharedness. Journal of Personality and Social
Psychology, 85(6), 989–1005. doi:10.1037/0022–3514.85.6.989
Lucy, J. A. (1997). Linguistic relativity. Annual Review of Anthropology, 26 291–312. doi:

10.1146/annurev.anthro.26.1.291
Maass, A. (1999). Linguistic intergroup bias: Stereotype perpetuation through language.

In M. P. Zanna (Ed.), Advances in experimental social psychology (vol. 31, pp. 79–121).
Maass, A., Milesi, A., Zabbini, S., & Stahlberg, D. (1995). Linguistic intergroup bias: Dif
ferential expectancies or in-group protection? Journal of Personality and Social Psycholo
gy, 68, 116–126.
Maass, A., Salvi, D., Arcuri, L., & Semin, G. R. (1989). Language use in intergroup con
texts: The linguistic intergroup bias. Journal of Personality and Social Psychology, 57,
981–993.
Magee, J. C., Milliken, F. J., & Lurie, A. R. (2010). Power differences in the construal of a
crisis: The immediate aftermath of September 11, 2001. Personality And Social Psycholo
gy Bulletin, 36(3), 354–370. doi:10.1177/0146167209360418
McArthur, L. Z., & Ginsberg, E. (1981). Causal attribution to salient stimuli: An investiga
tion of visual fixation mediators. Personality and Social Psychology Bulletin, 7(4), 547–
553. doi:10.1177/014616728174004
Morris, M. W., & Peng, K. (1994). Culture and cause: American and Chinese attributions
for social and physical events. Journal of Personality and Social Psychology, 67(6), 949–
971. doi:10.1037/0022–3514.67.6.949
Newman, M. L., Pennebaker, J. W., Berry, D. S., & Richards, J. M. (2003). Lying words:
Predicting deception from linguistic styles. Personality and Social Psychology Bulletin,
29(5), 665–675. doi:10.1177/0146167203029005010
Nisbett, R. E., & Ross, L. (1980). Human inference: strategies and shortcomings of social
judgment. Upper Saddle River, New Jersey: Prentice Hall.
Page 24 of 27

Contained in Words
Nussbaum, S., Trope, Y., & Liberman, N. (2003). Creeping dispositionism: The temporal
dynamics of behavior prediction. Journal of Personality and Social Psychology, 84(3), 485–
497. doi: 10.1037/0022–3514.84.3.485
Reeder, G. D., & Brewer, M. B. (1979). A schematic model of dispositional attribution in

interpersonal perception. Psychological Review, 86(1), 61–79. doi:10.1037/0033–295X.
86.1.61
Ross, M., & Sicoly, F. (1979). Egocentric biases in availability and attribution. Journal of
Personality and Social Psychology, 37(3), 322–336. doi: 10.1037/0022–3514.37.3.322
Rudolph, U., & Försterling, F. (1997). The psychological causality implicit in verbs: A re
view. Psychological Bulletin, 121, 192–218.
Schwarz, N., & Sudman, S. (1996). Answering questions: Methodology for determining
cognitive and communicative processes in survey research. San Francisco: Jossey-Bass.
Searle, J. R. (1969). Speech acts. Cambridge: Cambridge University Press.
Semin, G. R. (2000). Communication. Language as an implementational device for cogni

tion. European Journal of Social Psychology, 30, 595–612.
Semin, G. R. (2007). The regulation and distance. In K. Fiedler (Ed.), Social communica
tion (pp. 389–408). New York: Psychology Press.
chology, 54, 558–568.
Semin, G. R., & Fiedler, K. (1989). Relocating attributional phenomena within a language-
cognition interface: The case of actors’ and observers’ perspectives. European Journal of
Semin, G. R., & Fiedler, K. (1991). The linguistic category model, its bases, applications,
and range. In W. Stroebe & M. Hewstone (Eds.), European review of social psychology
(vol. 2, pp. 1–30). Chichester: Wiley.
Skowronski, J. J., & Carlston, D. E. (1987). Social judgment and social memory: The role of
cue diagnosticity in negativity, positivity, and extremity biases. Journal of Personality and
Social Psychology, 52(4), 689–699. doi:10.1037/0022–3514.52.4.689
Smith, P. K., & Trope, Y. (2006). You focus on the forest when you’re in charge of the
trees: Power priming and abstract information processing. Journal of Personality and So
cial Psychology, 90(4), 578–596. doi:10.1037/0022–3514.90.4.578
Steller, M., & Köhnken, G. (1989). Criteria-based content analysis. In D. C. Raskin (Ed.),
Psychological (p. 264) methods in criminal investigation and evidence (pp. 217–245). New
York: Springer-Verlag.
Page 25 of 27

Contained in Words
Stephan, E., Liberman, N., & Trope, Y. (2010). Politeness and psychological distance: A
construal level perspective. Journal of Personality and Social Psychology, 98(2), 268–280.
Steptoe, A., & Wardle, J. (2001). Locus of control and health behaviour revisited: A multi
variate analysis of young adults from 18 countries. British Journal of Psychology, 92(4),
659–672.
ter Doest, L., & Semin, G. R. (2005). Retrieval contexts and the concreteness effect: Dis
sociations in memory for concrete and abstract words. European Journal of Cognitive Psy
chology, 17, 859–881.
Tausczik, Y. R., & Pennebaker, J. W. (2010). The psychological meaning of words: LIWC
and computerized text analysis methods. Journal of Language and Social Psychology,
29(1), 24–54. doi:10.1177/0261927X09351676
Thompson, S. C., & Kelley, H. H. (1981). Judgments of responsibility for activities in close
relationships. Journal of Personality and Social Psychology, 41(3), 469–477. doi:
10.1037/0022–3514.41.3.469
Trope, Y., & Liberman, N. (2003). Temporal construal. Psychological Review, 110(3), 403–
421.
Trope, Y., & Liberman, N. (2010). Construal-level theory of psychological distance. Psy
chological Review, 117(2), 440–463.
Tversky, A., & Kahneman, D. (1981). The framing of decisions and the psychology of
choice. Science, 211(4481), 453–458. doi:10.1126/science.7455683
Unkelbach, C. (2006). The learned interpretation of cognitive fluency. Psychological

Science, 17(4), 339–345. doi:10.1111/j.1467-9280.2006.01708.x
Vallacher, R. R., & Wegner, D. M. (1987). What do people think they’re doing? Action iden
tification and human behavior. Psychological Review, 94, 3–15.
Vallacher, R. R., & Wegner, D. M. (1989). Levels of personal agency: Individual variation
in action identification. Journal of Personality and Social Psychology, 57, 660–671.
Vrij, A. (2005). Criteria-based content analysis: A qualitative review of the first 37 stud
ies. Psychology, Public Policy, and Law, 11(1), 3–41. doi:10.1037/1076–8971.11.1.3
Wänke, M., & Reutner, L. (2011). Direction-of-comparison effects: How and why compar
ing apples with oranges is different from comparing oranges with apples. In G. Keren
(Ed.), Perspectives on framing (pp. 177–193). New York: Psychology Press.
Weiner, B. (1985). An attributional theory of achievement motivation and emotion. Psycho

logical Review, 92(4), 548–573. doi:10.1037/0033–295X.92.4.548
Page 26 of 27

Contained in Words
Weiner, B., & Kukla, A. (2000). An attributional analysis of achievement motivation. In E.
Higgins, A. W. Kruglanski (Eds.), Motivational science: Social and personality perspec
tives (pp. 380–393). New York: Psychology Press.
Wigboldus, D., Semin, G. R., & Spears, R. (2000). How do we communicate stereotypes?
Linguistic bases and inferential consequences. Journal of Personality and Social Psycholo
gy, 78, 5–18.
Klaus Fiedler
Department of Psychology University of Heidelberg Heidelberg, Germany
Tobias Krüger
Tobias Krüger is a PhD candidate at the University of Heidelberg.
Page 27 of 27

Me and My Stories
Me and My Stories
Monisha Pasupathi , C.D. Mansfield , and T. L. Weeks

This article reviews the literature linking personal narratives to self and identity across
the lifespan. Evidence supports the idea that personal narratives both reflect existing as
pects of a narrator’s self and identity, broadly understood, and serve to construct, main
tain, and even alter the narrator’s self and identity over time. These claims are based, as
the article reveals, on cross-sectional, prospective longitudinal, and experimental evi
dence. Some of the strongest evidence within the field focuses on well-being and narra
tive. Further, narrative and self/identity relationships reflect the broader sociocultural
contexts within which they are created. Areas with significant remaining questions in the
area of narrative and self are also reviewed, including the role of other people’s stories in
self and identity development and the implications of social media for self and narrative.
Keywords: Narrative, self, identity, autobiographical memory, well-being, lifespan development
This was around 2 weeks ago at my in-laws house in [nearby city]. I told my hus
band he was being a moron. This occurred during a game of Yahtzee and, due to
my somewhat overly competitive nature, I got a little carried away with myself (I
was doing badly) and Derek, my husband, offered some advice. I was irritated and
fed up with the game and displaced my ill-will towards it on innocent Derek. With
an angry look and voice I said, “I don’t need advice from a moron like you!”
I felt horrible afterwards! I was somewhat shocked at what had just blurted from
my mouth and I looked at him with an equally shocked expression—I could tell he
was shocked too. Just a split-second later, however, he regained his normal compo
sure and then began to laugh. In a mock-offensive tone he said, “Well excuse me”
and put up his hands as though to ward off my anger. I was embarrassed and
guilty—we were playing with his Mom and brother as well, and my loss of temper
and control made me look and feel bad. They, however, shrugged it off with a
laugh and we moved on with the game. Thank goodness for a forgiving, under
standing husband!
Page 1 of 27

Me and My Stories
Our stories have everything to do with our sense of ourselves. With this relatively trivial,
everyday event narrative, we get a rich sense of the character of the narrator. Her narra
tive was drawn from a large sample of written narratives about experiences of self-contra
diction solicited for an ongoing research project (Mansfield & Pasupathi, in preparation).
In it, she expresses several ideas about herself in relation to her behavior and finds ways
to both see the event as consistent and continuous with an ongoing sense of self, as well
as ways to dismiss the implications of the event about other potential qualities she might
have. Her narrative, we would argue, serves to both express and maintain a sense of her
self, in this instance, in the face of some challenging evidence. It does so in ways that are
multifaceted, multilayered, and multidirectional.
(p. 266) The Social Nature of Storytelling and Self

Both self and narrative are socially shaped creations, and a review of the ways in which
self-development and self in adulthood are deeply social is well beyond our present scope
(e.g., Harter & Monsour, 1992; Leary, 1993; McNulty & Swann, 1994; Mead, 1934). Here,
we focus on narratives. Narratives as we have defined them are linguistic constructions—
and many contributions in this volume revolve around the deeply social nature of lan
guage. This social nature of language extends from the immediate ways in which people
construct and comprehend single utterances to the process of meaning making and story
telling about the self that we are concerned with in this chapter. At a proximal level, all
language interactions are shaped by both speakers and their listeners via the process of
grounding—establishing mutuality of understanding (Clark, 1996). For stories, at least
two points are worth making in particular. Just as contributions in conversations are
structured by turn-taking dynamics (Clark, 1996) and the establishing of shared under
standing (Clark & Schaefer, 1989), so are storytelling interactions. For storytelling, the
grounding process includes the establishment of self-views and story meanings as mutual
ly understood (Bavelas, Coates, & Johnson, 2000), and the turn-taking dynamics require
negotiation of the speaker’s right to insert a monologue (Clark, 1996). In our example,
this is not necessary because the study context demands the generation of a written
monologue, but we can speculate that the participant has in mind psychological re
searchers as an audience and that her account might therefore be more psychologically
minded—as evidenced by her attentiveness to traits, emotional expressions, and psycho
logical elements of her experience—than if she were narrating to a friend or parent.
In conversation, the narrator would have to also establish the right to tell this story. In
some of our earlier work looking at bids to narrate a past experience, some such bids are
simply ignored by conversation partners (e.g., Pasupathi, Lucas, & Coombs, 2002). If tak
en up, there are negotiations about who may take the role of primary narrator, leaving
others to take on other roles—mentors who facilitate the telling or monitors who address
issues of accuracy. When taken up, partners may agree or disagree about factual and
meaning-laden aspects of the events as narrated—including what the event means about
the narrator him- or herself (Manier, Pinner, & Hirst, 1996; McLean & Mansfield, 2012;
Pasupathi et al., 2002). Other work suggests that it is common for people to engage in
Page 2 of 27

Me and My Stories
chained storytelling in which one participant’s story receives, as a response, a related sto
ry from another participant—mutual reminiscing (Hyman & Faries, 1992; Pasupathi &
Carstensen, 2003)—and this chained storytelling makes the negotiation of monologues
easier. These chained stories are termed “second stories” in the discourse literature
(Page, 2012) and simultaneously narrate a new story while also serving as a commentary/
analysis on the prior story. This commentary/analysis can be supportive—constructing
connections between two participants—or disagreeable, conveying distance. In our exam
ple, imagine that the participant narrated the account to a friend, who responded with an
account of feeling competitive and hostile but exercising forbearance. Such an account
can be expressed as a rebuke. More recent work (Cuc, Koppel, & Hirst, 2007) suggests
that in such exchanges of stories, participants not only overtly express support or dis
agreement, but also subtly shape the structures of the stories that are told so that they
parallel one another in terms of which contextual details may be included and which are
not relevant.
In addition to these more immediate ways in which narrating is social, there are also
many findings revealing the ways in which narratives reflect broader, more distal aspects
of the social context. In many cases, these broader and more distant contextual factors
are conceptualized in terms of cultural differences (Han, Leichtman, & Wang, 1998;
McAdams, 2006; Wang, 2004) or master narratives (Thorne & McLean, 2003) and often
examined via work on social class differences (Miller & Moore, 1989), ethnic differences
(Brice-Heath, 1983; Pasupathi, Wainryb, & Twali, 2012), or gender differences (Bohanek,
Marin, & Fivush, 2008; Fivush, 1998; Thorne & McLean, 2003). These approaches to sto
ries differ in some of their details, but most posit that cultures and subcultures share ap
proaches to storytelling that are distinctive. For example, Wang has argued that Chinese
storytelling emphasizes relational connections and obligations and moral themes more so
than American storytelling, whereas Americans emphasize individual distinctiveness and
preferences (e.g., Wang, 2004). These storytelling differences are then linked, conceptual
ly, to differences in the nature of self-conceptions between Asian and Western cultures
(Markus, Mullally, & Kitayama, 1997). The example narrative, told by a young European-
American woman, is quite focused on the distinctive, individual qualities (p. 267) of the
self—although it also emphasizes relational obligations, those are less about roles than
about mutual reciprocity and affectional ties.
Defining Some Basic Terms: Narrative and Self
Thus far, we’ve focused primarily on our example while ignoring foundational questions of
definitions. We define narratives as temporally ordered, linguistically based accounts of
events that involve descriptions of actions, but also of the beliefs, purposes, and reactions
of characters—the evaluations—that make those actions cohere (see also Bruner, 1990;
Labov & Waletsky, 1967). This definition is still quite broad, and in this chapter, we pri
marily limit our scope to autobiographical narratives ranging from single-event accounts
to accounts of an individual’s entire life. This choice is motivated by the established sig
nificance of such narratives for an individual’s sense of self and psychological health and
by the availability of data on the development, maintenance, and changes such narratives
Page 3 of 27

Me and My Stories
undergo across the lifespan. Later, we return to questions of how the conclusions from
our review may be affirmed, expanded, or altered by a consideration of other people’s
narratives (including fiction) and emerging social media.
Just as narrative is a potentially broad term, researchers also use the term “self” to refer
to an enormous array of concepts (e.g., Tesser, Crepaz, Beach, Cornell, & Collins, 2000)
that refer to many different elements of the person. In this case, rather than restrict our
examination, we attempted to be relatively broad in exploring relations between self and
narratives, in keeping with the models we have outlined in previous work linking self and
narrative (McLean, Pasupathi, & Pals, 2007). The self has contents—it possesses traits,
abilities, and preferences that are enduring across the life course (e.g., Harter, 1998;
McAdams & Cox, 2010). The self also can be described in terms of roles—employee,
mother, brother, husband, wife, child (Linville, 1987; Steele, 1988). It can be evaluated
positively or negatively, and, in that process, those traits, abilities, and preferences are
not all equal (Crocker, Brook, Niiya, & Villacorta, 2006; Harter, 1998). These evaluations
may be examined in terms of self-esteem—global or specific (Harter, 1998). The self can
also be examined in structural ways—as having more or less clarity, more or less coher
ence or consistency, more or less complexity, more or less compartmentalized structure
(Campbell, Assanand, & Di Paula, 2003; Linville, 1987; Showers, 1992). The self is also an
agent—a being whose actions and reactions are intimately connected to goals and beliefs,
albeit sometimes in complex ways (McAdams & Cox, 2010; Pasupathi & Wainryb, 2010a)
—and that agent has moral character (Frimer, Walker, Dunlop, Lee, & Riches, 2011; Pasu
pathi & Wainryb, 2010a). In our review, we examine virtually all of these facets of self as
they have been linked to narratives.
Identity is another way that self is conceptualized—the commitments and beliefs to which
individuals subscribe and that give them a sense of continuity over time (Chandler,
Lalonde, Sokol, & Hallett, 2003; Erikson & Erikson, 1997). Sometimes identity develop
ment is assessed in terms of whether someone has explored potential commitments and
beliefs and whether the person has subsequently committed to some set of ideologies
(Erikson & Erikson, 1997; Marcia, 1966), and such exploration and commitment can be
linked to features of narratives (McLean & Pratt, 2006; Pasupathi et al., 2012). However,
one prominent perspective on self and narrative, in fact, is that the individual’s identity is
equivalent to a particular multiepisode narrative—the individual’s life story (McAdams,
1996). The life story is a selective autobiography including important and personality-
shaping episodes and events—the story of “how I came to be me” (Bluck & Habermas,
2000; Habermas & Bluck, 2000; McAdams, 1996). Among the proposed episodes are turn
ing points, low points, high points, and illustrative events that demonstrate typical quali
ties of the person. Many researchers working within the narrative and self arena study
selected episodes from the life story (e.g., Frimer, et al., 2011; Matsuba & Walker, 2005;
McLean & Pratt, 2006). Notably, our opening example represents an everyday episode
unlikely to be part of the narrator’s life story. The story could be narrated somewhat dif
ferently and presented as part of the life story. For example, it could be told as a turning
point—a moment when the narrator came to the realization that she must learn to damp
Page 4 of 27

Me and My Stories
en her competitive spirit lest it interfere with other important dimensions of her interper
sonal life.
Our example suggested that self and narrative could be related in many ways, and the de
finitions just outlined provide different possibilities for such relationships. Our overarch
ing framework, however, is that relations between narrative and self are more than sim
ply the reflection of self in narrative. Rather, we proceed with the assumption that rela
tionships between narrative and self are reciprocally influential across developmental
time (e.g., McLean (p. 268) et al., 2007; Pasupathi, 2001). What this means is that, even as
the narrator in our example tells a story that reflects her current views about herself, she
is simultaneously strengthening some self-views, altering others, and thus prospectively
affecting her sense of self over time. Of course, it is unlikely that any single telling of a
story has the causal force to change self-views. Rather, it is in the incremental, repeated
telling of one story and others like it that stories may eventually serve to shape the self
prospectively. This set of assumptions is grounded in work on memory, self, and narrative
(e.g., Conway, 2005; Conway & Pleydell-Pearce, 2000; McLean et al., 2007; Pasupathi,
2001; Schacter & Addis, 2007), in which narratives are constructed from memory in ways
that reflect not only stored associations, but also the present context. In turn, by recon
structing a memory within narrative, those stored associations may be expanded or al
tered (e.g., Adler, 2012; Bauer & McAdams, 2010; McLean & Pasupathi, 2011; Pasupathi,
Stallworth, & Murdoch, 1998; Weeks & Pasupathi, 2011).
Developmental Considerations About Narrative and Self: Childhood

Antecedents
The set of relationships among memory, narrative, and self is also one that emerges over
developmental time, and child developmental antecedents of both narrative and self are
deeply social in nature. Both self (e.g., Damon & Hart, 1988; Harter, 1998) and narrative
abilities (e.g., Fivush, Haden, & Adam, 1995; Pasupathi & Wainryb, 2010b) are developing
across childhood and adolescence, and, in both cases, there are clear relationships be
tween the social context and children’s evolving selves and stories. Children’s narratives
are more or less elaborated in ways that reflect their early experiences creating narra
tives with their caregivers (Fivush & Nelson, 2004). Children whose mothers engage in
elaborative joint narration of shared events with them, prompting for details, providing
pieces of information, and helping them to recount their experiences, have better memo
ry for those events weeks later than do children whose mothers use a less elaborative
style (Fivush & Nelson, 2004). Longitudinal studies of mother–child joint reminiscing sug
gest that children with elaborative mothers tend to adopt an elaborative style themselves
over time and that, as children become more elaborative in their storytelling, mothers re
spond in kind (Reese, Haden, & Fivush, 1993). Finally, intervention studies have suggest
ed that enhancing maternal elaborative style results in changes in children’s storytelling
(Peterson, Jesso, & McCabe, 1999).
Page 5 of 27

Me and My Stories
Cross-sectional studies with preschool and kindergarten-aged children relate children’s

self- esteem and self-concept clarity and consistency to the joint construction of memory
narratives—reminiscing—by parents and children (Bird & Reese, 2006; Marin, Bohanek,
& Fivush, 2008; Reese, Bird, & Tripp, 2007). Generally, the more elaboration and complex
discussion of emotions that parents and children do while narrating, the clearer and more
consistent the child’s sense of self and the higher the child’s self-esteem. Parents’ inter
ested, engaged response to the selves their children’s stories portray helps children un
derstand those selves, both in the past and in the present, as they begin to connect the
causes and outcomes of their emotions and behaviors.
The stories children tell also reflect their sense of agency in various ways. The increasing
elaboration of psychological contents in narratives that occurs as children move into ado
lescence (Pasupathi & Wainryb, 2010b) can be interpreted as an increasing focus on
agency—relations between what the narrator (and others) wanted and what the narrator
did or did not do to seek the desired outcome. By the time they are teenagers, children
become capable of expressing their agency through stories in multiple ways—by express
ing their goals and desires and how those relate to actions (Pasupathi & Wainryb, 2010b)
—and also as authors, by highlighting or editing details, withholding or sharing informa
tion, or exploring or ignoring listener’s offerings (Weeks & Pasupathi, 2010). Finally, dur
ing the teen years, identity in the form of a life story emerges (Habermas & Bluck, 2000;
Habermas & de Silveira, 2008; Habermas, Ehlert-Lerche, & de Silveira, 2009). Although
different facets of the life story emerge earlier rather than later, a full-blown life story is
the developmental achievement of late adolescence and early adulthood.
Self and Story in Adulthood

Both narrative abilities and self are considered “established” by early adulthood. Self and
identity researchers have suggested that adults are actively engaged in ways of making
sense of their experiences that serve to support their existing self views via attention,
memory, and social partner preferences (Swann & Ely, 1984; Swann, 2012). Narrating ex
periences involves meaning making, draws attention to some features of events and not
others, and typically is done with social partners. Consequently, (p. 269) narrative repre
sents a potent process by which people might maintain an existing self-concept across
adulthood. Furthermore, narrative might represent a route to change, which, although
less prevalent, does occur as people encounter new relationships and challenging events.
Throughout this section, we consider (1) relations between features of self and narrative
in adult samples and (2) prospective or experimental evidence that narrative can influ
ence self-stability or change. Our review is organized around the idea that relations be
tween self and narrative are multifaceted, seen in various aspects of both self and narra
tives; that such relations are multilayered, involving both conscious and nonconscious as
pects of self and narration; and that such relations are multidirectional, in that self can
influence story, and story can influence self.
Page 6 of 27

Me and My Stories
Self and Story: Self-Perceptions and Goals Are Reflected in Narra

tives
The narratives people tell are clearly related to the way people conceive of themselves in
trait and identity terms (McAdams et al., 2004). Unsurprisingly, people whose narratives
were complex rated themselves higher on openness to experience, people whose narra
tives emphasized communion rated themselves higher on agreeableness, and those whose
narratives were negatively toned reported higher levels of neuroticism (McAdams et al.,
2004; see Pennebaker & King, 1999, for similar findings using linguistic markers). The
trait of neuroticism is notable in its array of associations with narrative features. More
neurotic individuals tell narratives with lower frequencies of personal growth themes or
relationship development themes (Bauer & McAdams, 2005), increased presence of se
quences in which an initially affectively positive scene is turned affectively negative
(Adler, Kissel, & McAdams, 2006), and an increased likelihood of interpreting experiences
as having negatively influenced the self (Lilgendahl & McAdams, 2011). Conversely, Lil
gendahl and colleagues (Lilgendahl & McAdams, 2011) showed that building a positive
story about change in the wake of trauma was more likely among those who were both
low in neuroticism and also viewed the self as open to change.
Relatedly, self-conceptions of expertise also influence narration, with those who feel their
expertise is greater providing more contributions to joint narration (Pasupathi, Alderman,
& Shaw, 2007; see also Manier et al., 1996). Ethnic identity has been linked to narratives
(Pasupathi et al., 2012; Syed & Azmitia, 2008; 2009). People with a stronger sense of eth
nic identity are more likely to generate narratives about belonging and discrimination
(Syed & Azmitia, 2008), and, for minority individuals, stronger ethnic identity was associ
ated with greater elaboration and meaning making in discrimination experiences (Pasu
pathi et al., 2012).
Narratives are typically told to communicate some perspective to an audience, and that
communicative goal matters for the story that gets told. The goal of supporting one side
of a conflict, for example, shapes narratives people construct about other people’s experi
ences (McGregor & Holmes, 1999; Tversky & Marsh, 2000), as well as about their own
experiences (McGregor & Holmes, 1999; McLean, 2005; Pasupathi, 2007a; Weinstein, De
ci, & Ryan, 2011). People include more information about insights and lessons in their
narratives when telling those narratives for self-explanation rather than to entertain
(McLean, 2005), and their narratives tend to be richer in content that conveys insight
about the self (Pasupathi, 2007a). The correlational and experimental basis for this work
provides strong evidence for the causal influence of goals.
Beyond communicative goals, other motivational constructs are also reflected in narra
tives. Individual differences in motives for autonomy affect the extent to which people in
tegrate past identity-relevant events—as indexed by greater use of personal pronouns in
their memories (Weinstein et al., 2011). Personal strivings have been associated with sto
ries of self-defining experiences (Blagov & Singer, 2004; Moffitt & Singer, 1994).
McAdams (1982) has found that people who express power motivation are likely to en
Page 7 of 27

Me and My Stories
dorse memories of power experiences as personally meaningful, whereas those who ex
press intimacy motivation are likely to endorse memories of experiences of closeness with
others as personally meaningful (see also McAdams, Hoffman, Mansfield, & Day, 1996).
Finally, experimental findings show that when people are motivated to display a particu
lar self-concept, they more easily access relevant memories for potential narration (Sani
tioso, Kunda, & Fong, 1990). Goals clearly exert causal effects on the stories people tell.
Narrative Features and the Adult Developmental Self
Narratives are also connected to features of self that are more developmentally based. In
these developmental perspectives, researchers either emphasize how the self-concept—as
traditionally conceptualized in social psychology—becomes more clear, complex, and sta
ble from early adulthood into (p. 270) midlife (Labouvie-Vief, Chiodo, Goguen, Diehl, & Or
woll, 1995) or track the developmental emergence of specific capacities of the self, such
as wisdom or generativity (Ardelt, 2000; Baltes, Staudinger, Maercker, & Smith, 1995;
Erikson & Erikson, 1997).
Age differences in narratives are consistent with proposed age differences in self, al
though the few available studies typically did not directly measure self-views and link
them to stories. For example, McLean (2008) showed that older adults’ self-defining mem
ory narratives were more thematically coherent and focused on a stable sense of self than
were the narratives of younger adults. Relatedly, Rice and Pasupathi (2010) asked older
and younger adults to narrate self-discrepant and self-confirming experiences. Their find
ings suggested that older adults were less preoccupied with self-construction in narra
tives about self-relevant events. These findings can be interpreted as consistent with the
idea that a clearly held self-view is less subject to change and threat than a less mature,
less clearly held view.
Two elements of a developmentally mature self in adulthood are generativity and wisdom,
both of which have been linked to narratives. Generativity is defined in terms of concerns
with giving back to future generations or with one’s legacy (Erikson & Erikson, 1997).
McAdams and colleagues (McAdams, Diamond, de St. Aubin, & Mansfield, 1997) found
that highly generative community members, compared with those who were similar in ed
ucation and background but lower on generativity, created distinctive life stories. Specifi
cally, generative individuals’ life stories communicated an early advantage that differenti
ated the generative adult from people less fortunate, and this helped establish early on
the notion that people need to be stewards for others in life. In turn, the narrator com
mits to this idea, and stewardship and care for others becomes a life-long value that is
closely held and reinforced over time. In addition to the early advantage and desire to
give back themes, generative adults organized negative and positive scenes in the life sto
ry in a critically different way than did the less generative. Generative adults’ low points
were consistently followed by high points, which McAdams terms redemption sequenc
ing. Those who prize moral values such as benevolence, justice, and caring show similar
narrative characteristics (Arnold, Pratt, & Hicks, 2004; Colby & Damon, 1992; Matsuba &
Walker, 2005; Pratt, Skoe, & Arnold, 2004) and are additionally more likely than others to
Page 8 of 27

Me and My Stories
integrate agentic and communal themes in their life stories and, in particular, to employ
agency in the service of communal goals (Frimer et al., 2011).
Two other developmental aspects of self have also been linked to narratives: ego-develop
ment and wisdom. Ego development is understood as a structure through which one
makes sense of life events, with higher levels of ego development suggesting the capacity
to think about self and society in complex but integrated ways (Adler, Wagner, &
McAdams, 2007), and wisdom is defined as broad and insightful knowledge about human
dilemmas, oriented toward promoting the collective good (Ardelt, 2000; Baltes et al.,
1995; Staudinger & Baltes, 1996). Narrative researchers interested in how these con
structs relate to how people tell stories find that greater psychosocial maturity is often
associated with nuanced and complex narration. For example, Bauer and McAdams
(2004b) found that people who had higher standings on ego development narrated life
transitions in a way that integrated multiple aspects of self. King and her colleagues
(King, Scollon, Ramsey, & Williams, 2000) found that parents at the highest levels of ego
development constructed narratives about learning that their child had Down’s syndrome
that were more likely to express ideas about change in the self. Even when people low on
ego development do create complex narratives about self-growth, they simultaneously ex
press a sense of greater cognitive effort (McLean & Fournier, 2008). Those high on wis
dom are more likely to narrate growth from transgressions (Mansfield, McLean, & Lilgen
dahl, 2010). Broadly, more complex and growth-oriented narratives are linked to more
mature and wise selves.
A common theme among these findings is consistent relationships between features of

narratives that entail exploring an experience, acknowledging adversity, and finding in it
a sense of growth and possibility, personal characteristics that are generally positive in
terms of their implications for well-being and adult development. Thus, it is not surpris
ing that psychological well-being is linked to narratives in many studies.
Psychological Well-Being and Narrative
There is a great deal of evidence linking narratives to well-being. Narratives that are
elaborated, meaning-laden, and convey a sense of acknowledging difficulty while main
taining a sense of growth and possibility are consistently related to higher levels of well-
being. Lilgendahl and McAdams (p. 271) (2011) coded life story interviews from adults
who ranged in age from 34 to 68 years and found that narratives with complex process
ing were linked to a greater sense of personal meaning and pleasure. McAdams and col
leagues (McAdams, Reynolds, Lewis, Patten, & Bowman, 2001) found that people who
narrated challenge and adversity in their lives with redemptive sequences, indicating pos
itive outcomes from negative experiences, had higher levels of well-being and self-esteem
compared to those who did not. Interestingly, redemption sequencing is important not
merely because it is positive but because of the movement from negative to positive (see
also Adler et al., 2006; McLean & Lilgendahl, 2008).
Page 9 of 27

Me and My Stories
Goals and themes that people articulate within their narratives also matter for well-being.
For example, Bauer and McAdams (2004a; 2004b) showed that using intrinsic themes in
life transition narratives—themes that are linked to social connection, autonomy, and
competence—was associated with higher levels of well-being. In fact, the thematic con
tent of memory narratives is linked to specific facets of self-reported well-being, such as
positive relations with others or purpose in life (Demiray & Bluck, 2011). People whose
narratives are laden with themes of personal growth, for example, also endorse a
stronger sense of personal growth on self-report measures of well-being. This relationship
holds for both young and middle-aged adults and is strongest between recent memories
and current well-being. Just as these findings suggest the importance of narrating an
elaborated sense of growth in the face of adversity and challenge, these findings also sug
gest that narrating a sense of autonomy, purpose, competence, and connection is also re
lated to better well-being. Beyond psychological well-being, the literature on the writing
cure has also demonstrated effects of narration on physical well-being (see Smyth & Pen
nebaker, 2008, for a review).
Self as Expressed Within the Story

Although the just mentioned work relates elements of narratives to aspects of self or well-
being assessed separately, another approach to narratives and self is to examine elements
of the story that are prima facie evidence of self-construction. These have been opera
tionalized in various ways. For example, self–event connections are instances in narra
tives that construct a relationship between the experience and the sense of self (Pasu
pathi, Mansour, & Brubaker, 2007). Such connections can involve stability of self-views,
as well as changes in self-views that are catalyzed by an experience (Pasupathi et al.,
2007b; see also Habermas & Paha, 2001). The opening example includes a stability-en
hancing self-event connection, “Due to my competitive nature” that may stabilize self-
views by linking them to new events. Self-event connections can also involve dismissing
experiences as atypical or irrelevant to the self and thus provide for the stability of self-
views by rejecting events that press for changes. For example, one of our participants de
scribed having stolen money from her mother with the statement “it was not at all like me
and I would never do it again.” Other narratives convey change connections (also termed
causal relations or connections, see Habermas & de Silveira, 2008; Habermas & Paha,
2001; Lilgendahl & McAdams, 2011). These connections express how a past experience
caused a change in one’s self. Like stability connections (explain and dismiss), change
connections are important narrative mechanisms that help people to establish continuity
of self across time (Pasupathi et al., 2007).
All self–event connections amount to a snapshot of the self as it is expressed in narrative

and allow researchers to examine the process of self-construction over time. McLean and
Pasupathi examined self–event connections in narratives about important experiences as
told to new romantic partners (McLean & Pasupathi, 2011). Across two studies, the re
sults showed that the majority of connections were stability-enhancing and positively va
lenced connections, consistent with the self-enhancing tendencies and high self-esteem in
Page 10 of 27

Me and My Stories
American collegiate samples; likewise, when change connections were made, they were
more likely to be positively than negatively valenced.
Self and Narrative Links Are Multilayered: Consciousness and Time
Not all aspects of self that relate to information processing are likely to be consciously
available to the individual, but more tacit aspects of self may also be evident in narra
tives. Woike and colleagues (Woike & Bender, 2009) have shown that implicit motives au
tomatically direct attention and, consequently, influence both what people narrate and
how they narrate it. For example, people with implicit achievement motives record more
stories about achievement as memorable than do those people with an implicit need for
intimacy (Woike, 1995; Woike & Polo, 2001). (p. 272) Being agentically versus communally
motivated (seeking achievement/self-expansion vs. intimacy motivations, respectively)
promotes different ways of structuring narrative (Woike, Gershkovich, Piorkowski, & Po
lo, 1999; Woike & Polo, 2001). Agentic individuals tend to differentiate contents of per
sonality-consistent memories whereas communal individuals tend to integrate across ele
ments in personality-consistent narratives (Woike et al. 1999; Woike & Polo, 2001).
Woike’s work suggests that there is room to link narratives to a broader array of implicit
aspects of the self and to the interplay of explicit and implicit selves. For example, Peters
and Gawronski (2011) conducted studies examining the extent to which implicit and ex
plicit self-concepts mutually influence self-construal. They found that when people were
asked to recall autobiographical memories of behaving in extroverted ways (or introvert
ed depending on randomly assigned condition), they were later more likely to report be
ing extroverted (or introverted) but that this relationship was mediated by implicit self-
conceptions. That is, it was those who implicitly viewed themselves as extroverts who
were more influenced by the recall task. Importantly, past research has also shown that
implicit and explicit self-conceptions can also be discordant (Asendorpf, Banse, & Mücke,
2002). Although largely unexplored, results such as these raise interesting new questions
for narrative researchers. One concerns the potential for different ways of narrating to
contribute to coherence or incoherence between implicit and explicit self-views, and an
other is whether and to what extent might storytelling help resolve or exacerbate discrep
ancies between implicit and explicit self-concepts.
Narratives are also linked to past, present, and future self-views. Although most of the
work reviewed earlier focuses on current or present views and their links to narrative fea
tures, there are some exceptions. Findings on constructing a sense of change in narra
tives can be interpreted as reflecting past selves juxtaposed with present and future
selves. Although not focused on narrative, per se, Ross and Wilson have conducted exper
iments on why people tend to see the current self as superior to past selves (Ross & Wil
son, 2003; Wilson & Ross, 2003). They show that people perceive flawed past selves as
farther from, and positive past selves as subjectively closer to, the current self. Impor
tantly, narrative processes offer various ways to make events feel more distant or more
present, depending on the extent to which people are immersed in the past experience
they narrate (see, e.g., Pillemer, Desrochers, & Ebanks, 1998; Rice & Pasupathi, 2010).
Page 11 of 27

Me and My Stories
So, subjective temporal distance can arise from variations in how people narrate the past,
and moreover, narrative features can be linked to past and future self-views, as well as to
present ones.
Self and Story Relations Are Multidirectional: Stories Shaping Selves

or Selves Shaping Stories?
One of the central assumptions in our approach is that stories and selves mutually shape
one another over time in reciprocal ways. However, longitudinal and experimental work is
important to demonstrate that both directions of influence can be demonstrated, and
much of the work just described is correlational and cross-sectional. Furthermore, the
idea that narratives are highly contextual constructions whereas self-conceptions are
characterized, at least in adulthood, by greater stability means that the burden of proof
lies primarily in showing that stories can prospectively influence self. It is clear that sto
rytelling on one occasion can influence memory on subsequent occasions (Pasupathi &
Hoyt, 2010; Pasupathi et al., 1998; Tversky & Marsh, 2000). It is also clear that self-con
nections expressed in the context of narrating may be retained over short periods of time
(McLean & Pasupathi, 2011). Fortunately, an emerging body of findings lends support to
the idea that narration can prospectively influence the person, especially in terms of well-
being, and indicators of the maturity of the person’s self, such as ego development. Much
of this work emphasizes aspects of narrative that indicate growth.
Pals (2006) examined midlife women’s (age 52) narratives about their most difficult expe
rience since being in college and looked at links between those narratives and responses
to individual difference questionnaires at ages 21 and 61. Narratives were coded for dif
ferences in interpretations of difficult experiences along two dimensions: exploratory nar
rative processing (richness of exploratory narrative processing and open exploratory vs.
closed/minimizing processing) and coherent positive resolution. Participants with higher
levels of coping openness at age 21 showed higher levels of ego maturity at 61, and this
relationship was explained in part by exploratory narrative processing in the age 52 nar
ratives. Further, coherent positive resolution in narratives of difficult experiences was re
lated to increased life satisfaction in midlife.
Over 3.5 years, Bauer and McAdams (2010) showed that initial levels of well-being and
maturity (p. 273) were linked to goals in narratives, and those goals in turn were related
to changes in well-being and ego development. Specifically, those with high initial levels
of subjective well-being and high conscientiousness narrated events with an emphasis on
communal socioemotional growth, and this in turn promoted increased well-being over
the next 3 years. Ego-development over 3 years was significantly predicted by intellectual
growth goals. In our lab, we have conducted a series of short-duration longitudinal stud
ies examining how people understand the self and how this relates to their narration of
personal transgressions, We are finding that people who score high on self-acceptance
are more likely to narrate a personal transgression in a manner that promotes growth
and that narrating in a growth-promoting fashion predicts enhanced self-compassion over
a 1-month period (Mansfield, Pasupathi, & McLean, under review). Other work exploring
Page 12 of 27

Me and My Stories
prospective relationships between aspects of self and features of narratives has begun to
link narratives prospectively to the development of ethnic identity (Syed & Azmitia, 2009)
and to link people’s goal-related strivings to the emotional qualities of the narratives they
construct (Sutin & Robins, 2008).
Taken together, these findings are showing that self shapes story and story shapes self
but that the specific patterns of such relationships depends on the aspect of self (e.g.,
well-being vs. ego-development vs. goal strivings) and narrative (exploration, growth, and
resolutions) at stake.
Proximal and Distal Contexts: The Roles of Lis

teners and Cultures in Adulthood
In childhood, parents play a crucial role as co-narrators of children’s experiences and co-
authors of their developing selves. Responsive, engaged listeners continue to be impor
tant in helping adult narrators to create elaborated, richly meaning-laden narratives
about both everyday experiences (Pasupathi & Hoyt, 2009; Pasupathi & Rich, 2005;
Weeks & Pasupathi, 2011) and more significant, self-defining events (Jennings, McLean,
& Pasupathi, in press). Emerging findings suggest that listeners in adulthood, primarily
friends and romantic partners, are particularly invested in promoting the stability of their
partners’ self in narration and in promoting a positive view of their partner in narrative
(McLean & Pasupathi, 2011). When listeners are distracted and unable to provide respon
sive, engaged listening, speakers become less likely to endorse the self-related meaning
of an experience after narrating that event (Pasupathi & Rich, 2005). Speakers are more
likely to retain self-related meanings that they create with listeners when they are suc
cessful at getting the listener to understand that meaning (McLean & Pasupathi, 2011),
underscoring the importance of social support for narrative meanings in ensuring that
those meanings are lasting. Finally, among adults, the line between listeners and narra
tors is not always very sharply drawn. Norrick (1997) and Hirst (Manier & Hirst, 1996)
have written from diverse disciplinary backgrounds about the way that groups of people
negotiate the joint narrating of shared experiences. Culture also matters for narrative
and self among adults. European-American adults tell stories that focus on an individual
as the central character, whereas Asian-American adults focus on general memories of
routine events and social interactions (Wang & Ross, 2005). European-Americans are less
likely to recall social and historical events than are Chinese adults, less likely to invoke
moral elements of memories, and less likely to focus on social interactions (Wang & Con
way, 2004).
Proximal and distal contextual factors collide when researchers attempt to experimentally
vary the strength of contextual factors. For example, priming specific self-perspectives,
such as the collective versus the private self, alters which aspects of a story Asian and
Caucasian American adults remember and shapes memories to conform with whichever
self-perspective was primed (Wang & Ross, 2005). These findings suggest that one route
Page 13 of 27

Me and My Stories
by which broad cultural variations in narrative and self emerge is via the more microlevel
processes of narration within social and cultural context.
Plot Turns the Field Could Take: Fiction and Social Media
We focused earlier on the ways in which stories we tell about our own lives serve to shape
our selves over time and the way that process is also embedded within our social and cul
tural contexts. The emerging findings make a compelling case for the way that stories
and selves are entwined in ways that are multifaceted, multilayered, and multidirectional.
They also reveal a number of gaps in our understanding, as well as point to the fact that
those relationships will be complex and vary depending on the elements of self and story
at stake. In this last section, we want to highlight two arenas for narrative researchers
that are largely uncharted territory: the stories of others and emerging social media
forms.
(p. 274) Other People’s Stories, Real or Not
Are other people’s stories, including fiction, relevant for our sense of ourselves? Research
on fiction has often focused on ways in which fiction is not like other types of text (Abra
ham, von Cramon, & Schubotz, 2008; Gordon, Gerrig, & Franklin, 2009). Gerrig (1993)
emphasized that, to understand narratives, readers construct a simulation of the world—a
simulation in which they hold preferences for certain outcomes over others. What this
means is that when readers attempt to understand text, they involve their self intimately
in that process, a claim supported by the recruitment of areas of the brain involving self-
referential thinking for reading fiction (Abraham et al., 2008). Building on these ideas,
Green and her colleagues have examined individual and situational differences in narra
tive transportation—the capacity to immerse oneself fully in a story (Green, 2005). Green
argued that transportation will determine the extent to which people are influenced in
their beliefs and attitudes by a story, and the idea was initially tested extensively in terms
of persuasion and marketing-relevant concerns. The more transported people are by a
story, the more influenced they are by its content.
Transportation is influenced heavily by the ability to identify with a character—which sug

gests that a person’s existing sense of self is likely to affect the likelihood of transporta
tion, in combination with the features of a given story (Cohen, 2006; Green, 2005). How
ever, more recently, Sestir and Green (2010) showed that transportation and identifica
tion with fictional characters temporarily shifts self-concept toward similarity with the
characters. This work also raises the possibility that people might be more or less trans
ported or immersed in their own narratives. Imagine, for example, narrating the account
of one’s dissatisfaction with a product to ever higher ranked individuals within a
company’s customer service department. Linguistic markers of this type of transportation
might include increased present tense narration, dramatic re-enactments, and greater
use of self-focused pronouns, although relatively few studies have addressed this issue di
rectly (Pennebaker & King, 1999; Pillemer et al., 1998; Rice & Pasupathi, 2010). The ex
Page 14 of 27

Me and My Stories
tent to which personal narratives reflect and can influence the self may be meaningfully
related to being transported by one’s own story.
Beyond fiction, many stories that people tell concern other people rather than them
selves, and these stories may also affect the narrator’s sense of self, based on some initial
findings. Fivush and colleagues have begun to examine intergenerational stories—stories
told to children about the past experiences of mothers, fathers, and other family mem
bers, and have been able to demonstrate relationships between these intergenerational
stories and children’s sense of self, narrative styles, and well-being (Zaman & Fivush,
2011). Hirst and colleagues (Cuc et al., 2007) have demonstrated that hearing another
person’s story about an event similar to our own serves to influence the details that we
include when telling our own story. The collective result of this social shaping of one
another’s stories may be what has been termed a shared reality—socially shared under
standings of stories, selves, and worldviews—with all the attendant costs and benefits
that implies (Hardin & Conley, 2001; Thorne & McLean, 2003).
Social Media: Blogs, Vlogs, and Tweets, and Facebook
Over the past decade, new electronic and social media have provided new platforms for
narration, including blogs, vlogs, Twitter, and Facebook. Blogs and vlogs are perhaps the
most narrative-like of these platforms, providing space for longer monologues over a
range of topics. Social media sites like Facebook and Twitter offer the opportunity for dif
ferent kinds of personal narratives—briefer and with opportunities to avoid or engage in
integrative connections. Much of the psychological attention to these outlets has focused
on the quality of relationships established in such forums (Vergeer & Pelzer, 2009) and re
lationships between involvement in these types of media and addiction (Weinstein &
Lejoyeux, 2010), whereas issues of identity and social media are more prominent in com
munications research (e.g., Page, 2012).
These forums are likely to allow people to reflect, create, and maintain identity in the
ways that traditional narration appears to do and to have similar relations to well-being
and adaptation (Hoyt & Pasupathi, 2008) and to individual differences like gender (Page,
2012); they can be studied using many of the same approaches outlined earlier. However,
these new platforms also differ from traditional face-to-face narration in several ways that
may be important for the way that they function vis-à-vis identity (see Page, 2012, for a
more comprehensive treatment). These forums focus more on here-and-now narration
and may be read and displayed from present to past. Space limitations in Facebook and
Twitter constrain elaboration and meaning making in narratives or require the (p. 275)
distribution of narration across multiple communications in ways open to incoherence. Al
though blogs and vlogs offer the capacity to produce and disseminate longer, more com
prehensive accounts of events, they allow people to narrate excerpts without creating a
coherent life story, and they permit the narration of events that would be too trivial to tell
in face-to-face contexts. Thus, social media raise the issue of how narratives differ from
any communicative behavior.
Page 15 of 27

Me and My Stories
The nature of social media means that blogs, status updates, and tweets are delivered to
a wide audience—in some cases, an audience unknown to the narrator. In everyday narra
tion, moment-to-moment feedback from listeners helps narrators to craft understandable
and meaningful stories and to establish the plausibility of the meanings they are creating.
In social media, communication is asynchronous. With respect to coherence, it may be dif
ficult to achieve coherence and clarity when constructing narratives for a somewhat un
known audience—but at the same time, Hoyt and colleagues have argued that familiar au
diences can fill in so much from contextual knowledge that they may demand less coher
ence (Hoyt et al., 2010).
The audience characteristics of social media also open the narrator to influence from peo
ple who are not in the individual’s real-world social network. In social media, people may
acquire difficult-to-find support for stigmatized or minority identities in ways that pro
mote healthy, adaptive self-development. However, people may also find validation for
identities and selves that are less healthy—consider, for example, the pro-anorexia web
sites reviewed by Page (2012), which have ambiguous implications for adaptive develop
ment at best. Page also notes that, for readers, celebrity tweets offer a sense of private
access, but that sense is patently false—created by the celebrity and often a publicist.
Given that, does the sense of intimacy that usually would accompany sharing of self via
narrative for the narrator also get compromised by the nature of the audience?
Many social media can be created with partial or complete anonymity, and authorship can
be falsely presented as well. These aspects of the social media world raise important
questions about the relationship between social media narration and authenticity of self.
They also may open the possibility of examining the negotiation of multiple identities—a
problem that remains unsolved and in many respects unaddressed in narrative identity
work (e.g., Fivush & Marin, 2007; Pasupathi, 2007b; Thorne & Nam, 2007). Finally, the
lingering nature of narratives and identities on the internet means that flexibility for
changing one’s self could be constrained in the future.
Conclusion
The stories we tell are a means to reflect on, create, and maintain or change our self
views across the lifespan. In our review, some relations among stories and selves are well-
established. For example, it is clear that traits and goals affect narratives, and it is in
creasingly clear that narratives may exert prospective influences on aspects of self-con
ceptions and well-being across the lifespan. It is also clear that storytelling is a vehicle by
which social and cultural influences on the self may be examined. However, there are are
nas in which narrative approaches to self and identity have been less explored. Among
the most important unaddressed questions in this area are those relating to less con
scious aspects of self and story, as well as questions about others stories and emerging
technologies. The perils and promise of storytelling are changing with emerging technolo
gies, but even in this era of vast technological changes, face-to-face storytelling is likely
Page 16 of 27

Me and My Stories
to remain an important way that we express and create self and provide what might be
considered a “narrative foundation” for most of us.
References
Abraham, A., von Cramon, D. Y., & Schubotz, R. I. (2008). Meeting George Bush versus
meeting Cinderella: The neural response when telling apart what is real from what is fic
tional in the context of our reality. Journal of Cognitive Neuroscience, 20(6), 965–976.
Adler, J. M. (2012). Living into the story: Agency and coherence in a longitudinal study of
narrative identity development and mental health over the course of psychotherapy. Jour
nal of Personality and Social Psychology, 102(2), 367–389.
Adler, J. M., Kissel, E. C., & McAdams, D. P. (2006). Emerging from the CAVE: Attribution
al Style and the Narrative Study of Identity in Midlife Adults. Cognitive Therapy and Re
search, 30(1), 39–51.
Adler, J. M., Wagner, J. W., & McAdams, D. P. (2007). Personality and the coherence of psy
chotherapy narratives. Journal of Research in Personality, 41(6), 1179–1198.
Ardelt, M. (2000). Antecedents and effects of wisdom in old age: A longitudinal perspec
tive on aging well. Research on Aging, 22(4), 360–394.
Arnold, M. L., Pratt, M. W., & Hicks, C. (2004). Adolescents’ representations of parents’
voices in family stories: Value lessons, personal adjustment, and identity development. In
M. W. Pratt & B. H. Fiese (Eds.), Family stories and the life course: Across time and gen
erations (pp. 163–186). Mahwah, NJ: Erlbaum.
Asendorpf, J. B., Banse, R., & Mücke, D. (2002). Double dissociation between implicit and
explicit personality (p. 276) self-concept: The case of shy behavior. Journal of Personality
and Social Psychology, 83(2), 380–393.
Baltes, P. B., Staudinger, U. M., Maercker, A., & Smith, J. (1995). People nominated as
wise: A comparative study of wisdom-related knowledge. Psychology and Aging, 10, 155–
166.
Bauer, J. J., & McAdams, D. P. (2004a). Growth goals, maturity, and well-being. Develop
mental Psychology, 40, 114–127.
Bauer, J. J., & McAdams, D. P. (2004b). Personal growth in adults’ stories of life transi
tions. Journal of Personality, 72, 573–602.
Bauer, J. J., & McAdams, D. P. (2005). Interpreting the good life: Growth memories in the
lives of mature, happy people. Journal of Personality and Social Psychology, 88, 203–216.
Bauer, J. J., & McAdams, D. P. (2010). Eudaimonic growth: Narrative growth goals predict
increases in ego development and subjective well-being 3 years later. Developmental Psy
chology, 46(4), 761–772.
Page 17 of 27

Me and My Stories
Bird, A., & Reese, E. (2006). Emotional reminiscing and the development of an autobio
graphical self. Developmental Psychology, 42, 613–626.
Blagov, P., & Singer, J. A. (2004). Four dimensions of self-defining memories (specificity,
meaning, content, and affect) and their relationships to self-restraint, distress, and re
pressive defensiveness. Journal of Personality, 72, 481–511.
Bluck, S., & Habermas, T. (2000). The life story schema. Motivation and Emotion, 24,
121–147.
Bohanek, J. G., Marin, K. A., & Fivush, R. (2008). Family narratives, self, and gender in
early adolescence. The Journal of Early Adolescence, 28(1), 153–176.
Brice-Heath, S. (1983). Ways with words: Language, life and work in communities and
classrooms. Cambridge, mass.: Cambridge University Press.
Bruner, J. (1990). Acts of meaning. Cambridge, MA: Harvard University Press.
Campbell, J. D., Assanand, S., & Di Paula, A. (2003). The structure of the self-concept and
its relation to psychological adjustment. Journal of Personality, 71(1), 115–140.
Chandler, M. J., Lalonde, C. E., Sokol, B. W., & Hallett, D. (2003). Personal persistence,
identity development, and suicide. Oxford, England: Blackwell.
Clark, H. H. (1996). Using language. Cambridge, MA: Cambridge University Press.
259–294.
Cohen, J. (2006). Audience Identification with Media Characters. In J. Bryant & P. Vorder
er (Eds.), Psychology of entertainment. (pp. 183–197). Mahwah, NJ: Lawrence Erlbaum
Associates Publishers.
Colby, A., & Damon, W. (1992). Some do care: Contemporary lives of moral commitment.
New York: Free Press.
Conway, M. A. (2005). Memory and the self. Journal of Memory and Language, 53(4),
594–628.
Conway, M. A., & Pleydell-Pearce, C. W. (2000). The construction of autobiographical

memories in the self-memory system. Psychological Review, 107(2), 261–288.
Crocker, J., Brook, A. T., Niiya, Y., & Villacorta, M. (2006). The Pursuit of Self-Esteem:
Contingencies of Self-Worth and Self-Regulation. Journal of Personality, 74(6), 1749–1771.
Page 18 of 27

Me and My Stories
Cuc, A., Koppel, J., & Hirst, W. (2007). Silence is not golden: A case for socially shared re
trieval-induced forgetting. Psychological Science, 18(8), 727–733.
Damon, W., & Hart, D. (1988). Self-understanding in childhood and adolescence. Cam
bridge, England: Cambridge University Press.
Demiray, B., & Bluck, S. (2011). The relation of the conceptual self to recent and distant
autobiographical memories. Memory, 19(8), 975–992.
Erikson, E. H., & Erikson, J. M. (1997). The life cycle completed: Extended version. New
York: W. W. Norton & Company.
Fivush, R. (1998). Gendered narratives: Elaboration, structure, and emotion in parent-

child reminiscing across the preschool years. In C. P. Thompson, D. J. Herrmann, D.
Bruce, J. D. Read, D. G. Payne, & M. P. Toglia (Eds.), Autobiographical memory: Theoreti
cal and applied perspectives (pp. 79–104). Mahwah, NJ: Erlbaum.
Fivush, R., Haden, C. A., & Adam, S. (1995). Structure and coherence of preschoolers’
personal narratives over time: Implications for childhood amnesia. Journal of Experimen
tal Child Psychology, 60, 32–56.
Fivush, R., & Marin, K. A. (2007). Place and power: A feminist perspective on self-event
connections. Human Development, 50(2–3), 111–118.
Fivush, R., & Nelson, K. (2004). Culture and language in the emergence of autobiographi
cal memory. Psychological Science, 15, 586–590.
Frimer, J. A., Walker, L. J., Dunlop, W. L., Lee, B. H., & Riches, A. (2011). The integration
of agency and communion in moral personality: Evidence of enlightened self-interest.
Journal of Personality and Social Psychology, 101(1), 149–163.
Gerrig, R. J. (1993). Experiencing Narrative Worlds: On the Psychological Activities of

Reading. New Haven: Westview Press.
Gordon, R., Gerrig, R. J., & Franklin, N. (2009). Qualitative characteristics of memories
for real, imagined, and media-based events. Discourse Processes, 46(1), 70–91.
Green, M. C. (2005). Transportation into narrative worlds: Implications for the self. In A.
Tesser, J. V. Wood, & D. A. Stapel (Eds.), On building, defending and regulating the self: A
psychological perspective. (pp. 53–75). New York: Psychology Press.
Habermas, T., & Bluck, S. (2000). Getting a life: The development of the life story in ado
lescence. Psychological Bulletin, 126, 748–769.
Habermas, T., & de Silveira, C. l. (2008). The development of global coherence in life nar
ratives across adolescence: Temporal, causal, and thematic aspects. Developmental Psy
chology, 44(3), 707–721.
Page 19 of 27

Me and My Stories
Habermas, T., Ehlert-Lerche, S., & de Silveira, C. (2009). The development of the tempo
ral macrostructure of life narratives across adolescence: Beginnings, linear narrative
form, and endings. Journal of Personality, 77(2), 527–560.
Habermas, T., & Paha, C. (2001). The development of coherence in adolescent’s life narra
tives. Narrative Inquiry, 11(1), 35–54.
Han, J. J., Leichtman, M. D., & Wang, Q. (1998). Autobiographical memory in Korean, Chi
nese, and American children. Developmental Psychology, 34, 701–713.
Hardin, C. D., & Conley, T. D. (2001). A relational approach to cognition: Shared experi
ence and relationship affirmation in social cognition. In G. B. Moskowitz (Ed.), Cognitive
social (p. 277) psychology: The Princeton symposium on the legacy and future of social
cognition (pp. 3–17). Mahwah, NJ: Lawrence Erlbaum.
Harter, S. (1998). The development of self-representations. In N. Eisenberg & W. Damon

(Eds.), Handbook of child psychology, 5th edition (Vol. 3, pp. 553–617). New York: John
Wiley and Sons.
Harter, S., & Monsour, A. (1992). Developmental analysis of conflict caused by opposing
attributes in the adolescent self-portrait. Developmental Psychology, 28, 251–260.
Hoyt, T., & Pasupathi, M. (2008). Blogging about trauma: Linguistic markers of apparent
recovery. E-Journal of Applied Psychology, 4(2), 56–62.
Hoyt, T., Pasupathi, M., Smith, B. W., Yeater, E. A., Kay, V. S., & Tooley, E. (2010). Disclo
sure of emotional events in groups at risk for posttraumatic stress disorder. International
Journal of Stress Management, 17(1), 78–95.
Hyman, I. E., & Faries, J. M. (1992). The functions of autobiographical memory. In M. A.

Conway, D. C. Rubin, H. Spinnler, & W. A. Wagenaar (Eds.), Theoretical perspectives on
autobiographical memory (pp. 207–221). Dordrecht, the Netherlands: Kluwer Academic
Publishers.
Jennings, L., McLean, K. C., & Pasupathi, M. (in press). Intricate lettings out and lettings
in: Listener Scaffolding of Narrative Identity in Newly-Dating Romantic Partners. Self and
Identity.
King, L. A., Scollon, C. K., Ramsey, C., & Williams, T. (2000). Stories of life transition: Sub
jective well-being and ego development in parents of children with Down Syndrome. Jour
nal of Research in Personality, 34(4), 509–536. doi: 10.1006/jrpe.2000.2285
Labouvie-Vief, G., Chiodo, L. M., Goguen, L. A., Diehl, M., & Orwoll, L. (1995). Represen
tations of self across the life span. Psychology and Aging, 10, 404–415.
Labov, W., & Waletsky, J. (1967). Narrative analysis: Oral versions of personal experience.
In J. Helm (Ed.), Essays in the verbal and visual arts (pp. 12–44). Seattle: University of
Washington Press.
Page 20 of 27

Me and My Stories
Leary, M. R. (1993). The interplay of private self-processes and interpersonal factors in

self-presentation. Hillsdale, NJ: Lawrence Erlbaum.
Lilgendahl, J. P., & McAdams, D. P. (2011). Constructing stories of self growth: How indi
vidual differences in patterns of autobiographical reasoning relate to well being in
midlife. Journal of Personality, 79(2), 391–428.
Linville, P. W. (1987). Self-complexity as a cognitive buffer against stress-related illness

and depression. Journal of Personality and Social Psychology, 52, 663–676.
Manier, D., & Hirst, W. (1996). Remembering as communication: A family recounts its
past. In D. C. Rubin (Ed.), Remembering our past (pp. 271–290). New York: Cambridge
University Press.
Manier, D., Pinner, E., & Hirst, W. (1996). Conversational remembering. In D. Hermann,
C. McEvoy, C. Hertzog, A. Hertel, & M. K. Johnson (Eds.), Basic and applied memory re
search (vol. 2, pp. 269–286). Mahwah, NJ: Erlbaum.
Mansfield, C. D., McLean, K. C., & Lilgendahl, J. P. (2010). Narrating traumas and trans
gressions: Links between narrative processing, wisdom, and well-being. Narrative In
quiry, 20(2), 246–273.
Mansfield, C. D., & Pasupathi, M. (in preparation). Goals and modes of talking about self-
discrepant experiences predict differences in narrative identity and well-being. University
of Utah.
Mansfield, C. D., Pasupathi, M., & McLean, K. C. (under review). Developing through Dif
ficulty: Does narrating growth of self in stories of transgressions promote self-acceptance
and self-compassion? University of Utah.
Marcia, J. E. (1966). Development and validation of ego identity status. Journal of Person
ality and Social Psychology, 3, 551–558.
Marin, K. A., Bohanek, J. G., & Fivush, R. (2008). Positive effects of talking about the neg
ative: Family narratives of negative experiences and preadolescents’ perceived compe
tence. Journal of Research on Adolescence, 18(3), 573–593.
Markus, H., Mullally, P. R., & Kitayama, S. (1997). Selfways: Diversity in modes of cultural
participation. In U. Neisser & D. A. Jopling (Eds.), The conceptual self in context: Culture,
experience, self-understanding. (pp. 13–61). New York: Cambridge University Press.
Matsuba, M. K., & Walker, L. J. (2005). Young adult moral exemplars: The making of self
through stories. Journal of Research on Adolescence, 15, 275–297.
McAdams, D. P. (1982). Experiences of intimacy and power: Relationships between social

motives and autobiographical memory. Journal of Personality and Social Psychology,
42(2), 292–302.
Page 21 of 27

Me and My Stories
McAdams, D. P. (1996). Personality, modernity, and the storied self: A contemporary

framework for studying persons. Psychological Inquiry, 7, 295–321.
McAdams, D. P. (2006). The redemptive self: Stories Americans live by. USA: Oxford Uni
versity Press.
McAdams, D. P., Anyidoho, N. A., Brown, C., Huang, Y. T., Kaplan, B., & Machado, M. A.
(2004). Traits and stories: Links between dispositional and narrative features of personal
ity. Journal of Personality, 72, 761–784.
McAdams, D. P., & Cox, K. S. (2010). Self and identity across the lifespan. In R. Lerner, A.
M. Freund & M. Lamb (Eds.), Handbook of Lifespan Development: Social and emotional
development (Vol. 2, pp. 158–207). Wiley.
McAdams, D. P., Diamond, A., de St. Aubin, E., & Mansfield, E. D. (1997). Stories of com
mitment: The psychosocial construction of generative lives. Journal of Personality and So
McAdams, D. P., Hoffman, B. J., Mansfield, E. D., & Day, R. (1996). Themes of agency and
communion in significant autobiographical scenes. Journal of Personality, 64, 339–377.
McAdams, D. P., Reynolds, J., Lewis, M., Patten, A. H., & Bowman, P. J. (2001). When bad
things turn good and good things turn bad: Sequences of redemption and contamination
in life narrative and their relation to psychosocial adaptation in midlife adults and in stu
dents. Personality & Social Psychology Bulletin, 27, 474–485.
McGregor, I., & Holmes, J. G. (1999). How storytelling shapes memory and impressions of
relationship events over time. Journal of personality and social psychology, 76, 403–419.
McLean, K. C. (2005). Late adolescent identity development: Narrative meaning-making

and memory telling. Developmental Psychology, 41, 683–691.
McLean, K. C. (2008). Stories of the young and the old: Personal continuity and narrative
identity. Developmental Psychology, 44(1), 254–264.
McLean, K. C., & Fournier, M. A. (2008). The content and processes of autobiographical
reasoning in narrative identity. Journal of Research in Personality, 42(3), 527–545. doi:
10.1016/j.jrp.2007.08.003
McLean, K. C., & Lilgendahl, J. P. (2008). Why recall our highs and lows: Relations
(p. 278)
between memory functions, age, and well-being. Memory, 16(7), 751–762. doi:
10.1080/09658210802215385
McLean, K. C., & Mansfield, C. D. (2012). The co-construction of adolescent narrative

identity: Narrative processing as a function of adolescent age, gender, and maternal scaf
folding. Developmental Psychology, 48(2), 436–447.
Page 22 of 27

Me and My Stories
McLean, K. C., & Pasupathi, M. (2011). Old, new, borrowed, Blue? The emergence and re
tention of personal meaning in autobiographical storytelling. Journal of Personality, 79(1),
135–164.
McLean, K. C., Pasupathi, M., & Pals, J. L. (2007). Selves creating stories creating selves:
A process model of self-development. Personality and Social Psychology Review, 11(3),
262–278.
McLean, K. C., & Pratt, M. W. (2006). Life’s Little (and Big) Lessons: Identity Statuses and
Meaning-Making in the Turning Point Narratives of Emerging Adults. Developmental Psy
chology, 42(4), 714–722.
McNulty, S. E., & Swann, W. B. Jr. (1994). Identity negotiation in roommate relationships:
The self as architect and consequence of social reality. Journal of Personality and Social
Psychology, 67, 1012–1023.
Mead, G. H. (1934). Mind, self and society: From the standpoint of a social behaviorist
(with intro by C. W. Morris.). In C. W. Morris (Ed.), Mind, self & society from the stand
point of a social behaviorist, edited, with introduction. Chicago: University of Chicago
Press.
Miller, P. J., & Moore, B. B. (1989). Narrative conjunctions of caregiver and child: A com
parative perspective on socialization through stories. Ethos, 17, 428–449.
Moffitt, K. H., & Singer, J. A. (1994). Continuity in the life story: Self-defining memories,
affect, and approach/avoidance personal strivings. Journal of Personality, 62, 21–43.
Norrick, N. R. (1997). Twice-told tales: Collaborative narration of familiar stories. Lan

guage in Society, 26, 199–220.
Page, R. E. (2012). Stories and Social Media: Identities and Interaction.
Pals, J. L. (2006). Narrative identity processing of difficult life experiences: Pathways of

personality development and positive self-transformation in adulthood. Journal of Person
ality, 74, 1079–1110.
Pasupathi, M. (2001). The social construction of the personal past and its implications for
adult development. Psychological Bulletin, 127(5), 651–672.
Pasupathi, M. (2007a). Telling and the remembered self: Linguistic differences in memo
ries for previously disclosed and previously undisclosed events. Memory, 15(3), 258–270.
Pasupathi, M. (2007b). Whither unity, and at what cost? Fragmentation in the life story.
Human Development, 50, 124–126.
Pasupathi, M., Alderman, K., & Shaw, D. (2007). Talking the talk: Collaborative remem
bering and self-perceived expertise. Discourse Processes, 43, 55–77.
Page 23 of 27

Me and My Stories
Pasupathi, M., & Carstensen, L. L. (2003). Age and emotional experience during mutual
reminiscing. Psychology and Aging, 18(3), 430–442.
Pasupathi, M., & Hoyt, T. (2009). The development of narrative identity in late adoles
cence and emergent adulthood: The continued importance of listeners. Developmental
Psychology, 45(558–574).
Pasupathi, M., & Hoyt, T. (2010). Silence and the shaping of memory: How distracted lis
teners affect speakers’ subsequent recall of a computer game experience. Memory, 18(2),
159–169.
Pasupathi, M., Lucas, S., & Coombs, A. (2002). Functions of autobiographical memory in
discourse: Long-married couples talk about conflicts and pleasant topics. Discourse
Processes, 34, 163–192.
Pasupathi, M., Mansour, E., & Brubaker, J. (2007). Developing a Life Story: Constructing
relations between self and experience in autobiographical narratives. Human Develop
ment, 50(2/3), 85–110.
Pasupathi, M., & Rich, B. (2005). Inattentive Listening Undermines Self-Verification in

Personal Storytelling. Journal of Personality, 73(4), 1051–1086.
Pasupathi, M., Stallworth, L. M., & Murdoch, K. (1998). How what we tell becomes what
we know: Listener effects on speakers’ long-term memory for events. Discourse Process
es, 26(1), 1–25.
Pasupathi, M., & Wainryb, C. (2010a). Developing moral agency through narrative. Hu
man Development, 53, 55–80.
Pasupathi, M., & Wainryb, C. (2010b). On telling the whole story: Facts and interpreta
tions in memory narratives from childhood through adolescence. Developmental Psycholo
gy, 46, 735–746.
Pasupathi, M., Wainryb, C., & Twali, M. (2012). Narrative construction of ethnicity-based
discrimination and its relationships to ethnic identity exploration and commitment in eth
nic majority and minority members. Identity: An International Journal of Theory and Re
search, 12(1), 53–73.
Pennebaker, J. W., & King, L. A. (1999). Linguistic styles: Language use as an individual
difference. Journal of Personality and Social Psychology, 77, 1296–1312.
Peters, K. R., & Gawronski, B. (2011). Mutual influences between the implicit and explicit
self-concepts: The role of memory activation and motivated reasoning. Journal of Experi
mental Social Psychology, 47(2), 436–442.
Peterson, C., Jesso, B., & McCabe, A. (1999). Encouraging narratives in preschoolers: An
intervention study. Journal of Child Language, 26, 49–67.
Page 24 of 27

Me and My Stories
Pillemer, D. B., Desrochers, A. B., & Ebanks, C. M. (1998). Remembering the past in the
present: Verb tense shifts in autobiographical memory narratives. In C. P. Thompson, D. J.
Herrmann, D. Bruce, J. D. Read, D. G. Payne, & M. P. Toglia (Eds.), Autobiographical mem
ory: Theoretical and applied perspectives (pp. 145–162). Mahwah: Lawrence Erlbaum As
sociates.
Pratt, M. W., Skoe, E. E., & Arnold, M. L. (2004). Care reasoning development and family
socialisation patterns in later adolescence: A longitudinal analysis. International Journal
of Behavioral Development, 28, 139–147.
Reese, E., Bird, A., & Tripp, G. (2007). Children’s self-esteem and moral self: Links to par
ent-child conversations. Social Development, 16(3), 460–478.
Reese, E., Haden, C. A., & Fivush, R. (1993). Mother-child conversations about the past:
Relationships of style and memory over time. Cognitive Development, 8(4), 403–430.
Rice, C. I., & Pasupathi, M. (2010). Reflecting on self-relevant experiences: Adult age dif
ferences. Developmental Psychology, 46, 479–490.
Ross, M., & Wilson, A. (2003). Autobiographical memory and conceptions of self: Getting
better all the time. Current Directions in Psychological Science, 12(2), 66–69.
Sanitioso, R., Kunda, Z., & Fong, G. T. (1990). Motivated recruitment of autobiographical
memories. Journal of Personality and Social Psychology, 59, 229–241.
(p. 279) Schacter, D. L., & Addis, D. R. (2007). On the constructive episodic simulation of
past and future events. Behavioral and Brain Sciences, 30(3), 331–332.
Sestir, M., & Green, M. C. (2010). You are who you watch: Identification and transporta
tion effects on temporary self-concept. Social Influence, 5(4), 272–288.
Showers, C. (1992). Compartmentalization of positive and negative self-knowledge: Keep

ing bad apples out of the bunch. Journal of Personality and Social Psychology, 62(6),
1036–1049.
Smyth, Joshua M., & Pennebaker, James W. (2008). Exploring the boundary conditions of
expressive writing: In search of the right recipe. British Journal of Health Psychology,
13(1), 1-7. doi: 10.1348/135910707X260117
Staudinger, U. M., & Baltes, P. B. (1996). Interactive minds: A facilitative setting for wis
dom-related performance. Journal of Personality and Social Psychology, 71, 746–762.
Steele, C. M. (1988). The psychology of self-affirmation: Sustaining the integrity of the

self. Advances in experimental social psychology, 21, 261–302.
Sutin, A. R., & Robins, R. W. (2008). Going forward by drawing from the past: Personal
strivings, personally meaningful memories, and personality traits. Journal of Personality,
76(3), 631–663.
Page 25 of 27

Me and My Stories
Swann, W. B., Jr. (2012). Self-verification theory. In P. A. M. Van Lange, A. W. Kruglanski &
E. T. Higgins (Eds.), Handbook of theories of social psychology (vol 2, pp. 23–42). Thou
Swann, W. B., Jr., & Ely, R. J. (1984). A battle of wills: Self-verification versus behavioral
confirmation. Journal of Personality and Social Psychology, 46, 1287–1302.
Syed, M., & Azmitia, M. (2008). A narrative approach to ethnic identity in emerging adult
hood: Bringing life to the identity status model. Developmental Psychology, 44(4), 1012–
1027.
Syed, M., & Azmitia, M. (2009). Longitudinal trajectories of ethnic identity during the col
lege years. Journal of Research on Adolescence, 19(4), 601–624.
Tesser, A., Crepaz, N., Beach, S. R. H., Cornell, D., & Collins, J. C. (2000). Confluence of
self-esteem regulation mechanisms: On integrating the self-zoo. Personality and Social
Psychology Bulletin, 26(12), 1476–1489.
Thorne, A., & McLean, K. C. (2003). Telling traumatic events in adolescence: A study of
master narrative positioning. In R. Fivush & C. Haden (Eds.), Connecting culture and
memory: The development of an autobiographical self (pp. 169–185). Mahwah, NJ:
Lawrence Erlbaum.
Thorne, A., & Nam, V. (2007). The life story as a community project. Human Development,
50(2–3), 119–123.
Tversky, B., & Marsh, E. J. (2000). Biased retellings of events yield biased memories. Cog
nitive Psychology, 40(1), 1–38.
Vergeer, M., & Pelzer, B. (2009). Consequences of media and Internet use for offline and
online network capital and well-being. A causal model approach. Journal of Computer-Me
diated Communication, 15(1), 189–210.
Wang, Q. (2004). The emergence of cultural self-constructs: Autobiographical memory

and self-description in European- American and Chinese Children. Developmental Psy
chology, 40, 3–15.
Wang, Q., & Conway, M. A. (2004). The stories we keep: Autobiographical memory in
American and Chinese middle-aged adults. Journal of Personality, 72, 911-938.
Wang, Q., & Ross, M. (2005). What we remember and what we tell: The effects of culture
and self-priming on memory representations and narratives. Memory, 13(6), 594–606.
Weeks, T. L., & Pasupathi, M. (2010). Autonomy, identity, and narrative construction with
parents and friends. In McLean & Pasupathi (Eds.), Narrative development in adoles
cence: Creating the storied self. (pp. 65–91). New York: Springer Science + Business Me
dia.
Page 26 of 27

Me and My Stories
Weeks, T. L., & Pasupathi, M. (2011). Stability and change self-integration for negative
events: The role of listener responsiveness and elaboration. Journal of Personality, 79,
469–498.
Weinstein, A., & Lejoyeux, M. (2010). Internet addiction or excessive Internet use. The
American Journal of Drug and Alcohol Abuse, 36(5), 277–283.
Weinstein, N., Deci, E. L., & Ryan, R. (2011). Motivational determinants of integrating
positive and negative past identities. Journal of Personality and Social Psychology, 100(3),
527–544.
Wilson, A., & Ross, M. (2003). The identity function of autobiographical memory: Time is
on our side. Memory, 11(2), 137–149.
Woike, B. A., Gershkovich, I., Piorkowski, R., & Polo, M. (1999). The role of motives in the
content and structure of autobiographical memory. Journal of Personality and Social Psy
chology, 76, 600–612.
Woike, B. A., & Polo, M. (2001). Motive-related memories: Content, structure, and affect.
Journal of Personality, 69(3), 391–415.
Woike, B. A. (1995). Most memorable experiences: Evidence for a link between implicit
and explicit motives and social cognitive processes in everyday life. Journal of Personality
and Social Psychology, 68(6), 1081–1091.
Woike, B. A., & Bender, M. (2009). Implicit motives as a way to understand cognitive
processes. Social and Personality Psychology Compass, 3(5), 702–710.
Zaman, W., & Fivush, R. (2011). When my mom was a little girl…Gender differences in
adolescents’ intergenerational and personal stories. Journal of Research on Adolescence,
21(3), 703–716.
Monisha Pasupathi
Monisha Pasupathi, Department of Psychology, University of Utah
C.D. Mansfield
C.D. Mansfield, Department of Psychology, University of Utah
T. L. Weeks
T.L. Weeks, Department of Psychology, University of Utah
Page 27 of 27

The Role of Language on the Perception and Experience of Emotion
The Role of Language on the Perception and Experi

ence of Emotion
Jennifer M.B. Fugate and Lisa Feldman Barrett
Print Publication Date: Sep 2014 Subject: Psychology, Social Psychology, Affective Science
This essay first reviewsthe role of language in several prominent theories of emotion.
Next, empirical evidence for the role of language in emotion perception is reviewed and
the neural basis of language and emotion is considered. In doing so, implications for the
field of emotion perception are discussed and a possible mechanism through which words
might exert their influence on perception is considered. The essay ends with a discussion
of how language is embodied in emotion experience and considers the role of language in
emotion perception and experience from several special populations.
Keywords: language, emotion, facial depictions, levels of processing, psychological constructionism
The Role of Language in Theories of Emotion

Humans experience emotion. For many, experience serves as an emotion’s central and
defining aspect (c.f. Barrett, 2006a). We assume that experiencing an emotion gives us an
understanding of an emotion itself. That is, we often assume that people experience emo
tion because people have “emotions”—internal mechanisms for a small set of reactions
(typically, happiness, anger, sadness, fear, disgust, etc.) that, once triggered, can be mea
sured in a clear and objective fashion.
Some scientific models reflect this belief. That is, they assume that different emotions
(e.g., happiness, anger, sadness, etc.) are associated with individual mental states and ob
jective indicators that reliably and consistently discriminate individual emotion cate
gories. As a result, one is able to measure behavior, facial muscle activity, vocal acoustics,
autonomic physiology, or brain activation and know what emotion it is that you are feeling
(i.e., emotion experience) or what emotion another person is feeling (i.e., emotion percep
tion). The guiding hypothesis among these models is that emotions are natural kinds that
exist independent of a perceiver, in which each emotion corresponds to an inherited, ar
chitecturally distinct circuit in the brain (see Barrett, 2006a, 2006b for a review). In other
models, emotions are not natural kinds, but constructed when a perceiver conceptualizes
Page 1 of 28

individual sensations from the world or body as belonging to a particular emotion catego
ry. This conceptualization may occur through a variety of means, with one powerful
source being language (i.e., either when someone explicitly labels an instance as “anger”
or “fear” or implicitly by anchoring emotion categories).
In the following paragraphs, we briefly describe four models of emotion and describe how
each model views the role of language. We follow this section with a review of studies
that show how emotion words affect emotion perception at various levels of processing.
We then review the neuroimaging and neurophysiological evidence showing how lan
guage plays a role in emotion perception and experience before suggesting a mechanism
by which it may do so. Next, we discuss how the experience of emotion (p. 281) is embod
ied and how emotion words help to situate individual instances of emotion. Finally, we
discuss evidence for the role of language in disordered emotion and in bilingual popula
tions.
The Basic View of Emotion

According to one natural kinds model of emotion, the Basic View (e.g., Ekman, 1992;
Izard, 1971; Tomkins, 1962), language is viewed as independent of emotion. In this view,
the experience of emotion (or the perception of emotion) occurs when a stimulus triggers
an ancillary set of coordinated changes in behavior, bodily physiology, and experience
(e.g., Allport, 1924; Ekman, 1992; Izard, 1971; McDougall, 1921/1908; Tomkins, 1962).
These changes result in a discrete emotion being triggered. According to this view, all in
stances of an emotion category are consistently activated and distinct from other emo
tions (e.g., anger from fear from sadness, etc.). As a result, there is little or no ambiguity
as to what you feel or what emotion you perceive when you look at another person. Emo
tion perception, in this case, is purely a “reading” of these changes in another individual.
Conceptual knowledge (e.g., information that a person brings to bear on incoming sensa
tions), including language, is not needed to know what emotion one is feeling or perceiv
ing. The fact that different categories of emotion are labeled differently simply reinforces
the structural specificity for emotional indicators among categories. That is, children ac
quire different emotion words because there is a natural, biological reality to anger (or
fear, sadness, etc.) (see Barrett, 2006a).
The Appraisal View of Emotion
According to appraisal views of emotion (Arnold, 1960a, 1960b), emotions are intensional
states (i.e., they are caused by and therefore refer to an object or situation in the world).
A person performs a meaning analysis of the situation to make a specific instance belong
to a discrete emotion category (e.g., Arnold, 1960a, 1960b; Ellsworth & Scherer, 2003;
Frijda, 1986; Lazarus, 1991; Roseman, 1991). Emotions are not merely triggered, as in
the Basic View, but arise from a meaningful interpretation of an object by an individual.
In some models of appraisal theory, emotions are the appraisal (e.g., Clore & Ortony,
2008), whereas in others, appraisals produce emotions (e.g., Scherer, 1984). In both cas
Page 2 of 28

es, however, language is viewed as a byproduct of the communication process. That is,
emotion words constrain the categories of emotion in which appraisals are placed, but
language itself does not hold any inherent influence on the process, and, therefore has no
explanatory power on the analysis.
The Social Constructionist View of Emotion
According to social constructionist views of emotion, emotions are defined by their social
function (e.g., Averill, 1980; Harré, 1986; Hochschild, 1983; Lutz, 1988; Ratner, 1989; for
a recent review, see Averill, 2012). In one view, emotions are afforded and constrained by
the history of the relationship, as well as by future projections of where the relationship
may go (see Boiger & Mesquita, 2012). “Construction,” in this view, can play out in at
least three different contexts—moment-to-moment interactions, developing and ongoing
relationships, and sociocultural contexts. For example, at the moment-to-moment level,
the development of anger in the context of a marital conflict may depend on whether the
spouse reciprocates with anger, sadness, or indifference. At the level of ongoing relation
ships, anger during marital conflict might be different depending on both the relationship
quality and the expected future of the relationship. And, at the level of sociocultural con
text, anger during marital conflict might be seen as functional for the development of
long-term relationships (Boiger & Mesquita, 2012).
To the extent that a culture’s language reflects such components, language will be in
volved in emotion experience and perception. For example, whether or not a behavior,
bodily, or physiological response is considered to belong to an emotion category will be
determined, at least in part, on what role emotions are meant to play in the culture (e.g.,
dominance or independence vs. submission or interdependence). For example, in some
cultures that value relational interdependence and harmony, the expression of anger may
be discouraged and perceived as immature and childish (Azuma, 1984; White & LeVine,
1986). The perception of emotion is also subject to social and cultural norms. For exam
ple, Westerners rely primarily on the target individual when perceiving how he or she
feels, whereas Japanese make more strategic use of the information available in the peo
ple around the target individual (Masuda et al., 2008), consistent with the idea that West
erners conceptualize emotions as located within the individual whereas Japanese concep
tualize emotion as reflecting the relationships between people.
(p. 282) The Psychological Constructionist View of Emotion
According to psychological constructionist views of emotion, an emotion emerges when a

person’s internal state is understood in some way as related to or caused by the situation.
That is, a person makes meaning of internal sensory information involved in affective
states. In some models, this analysis may be instinctual (e.g., James, 1884, 1890) or result
from processes like categorization (e.g., Barrett, 2006a, 2006b) or attribution (Russell,
2003). All forms of psychological constructionism posit more basic psychological ingredi
ents that come together (in different ways) to produce mental states like emotion. Fur
thermore, in psychological construction, the ingredients that constitute the psychological
Page 3 of 28

states that people colloquially refer to as “emotion” also constitute other mental states
that people refer to as “thoughts,” “beliefs,” and “memories.” As in social construction
ism, emotion categories are assumed to be culturally relative, although the ingredients
that go into making individual instances of emotion are assumed to correspond to biologi
cal systems (in one form or another) within the brain of a human being.
In our psychological constructionist model, called the conceptual act model, emotions
emerge when people make meaning out of sensory input from the body and from the
world using knowledge of prior experiences (Barrett, 2006a, 2006b, 2009, 2012). Emo
tions are “situated conceptualizations” (Barsalou, 2003) because the emerging meaning is
tailored to the immediate environment and prepares the person to respond to sensory in
put in a way that is adaptive to the situation (Barsalou, 2003). One way in which meaning
is created is through language. Emotion words, like “anger,” “sadness,” and “fear” are
used to name folk categories that help divide up the continuous and highly variable mea
surable outcomes (e.g., bodily physiology, behavior, etc.) (Barrett, 2006a, 2006b, 2009). In
fact, our model of emotion is the only model that postulates an explicit role for language:
emotion words are constitutive of emotion generation and perception.
Specifically, we formulated the language-as-context hypothesis (Barrett, Lindquist, & Gen

dron, 2007), which suggests that emotion words provide an “internal context” that helps
to constrain this continuous flow of information. Emotion words are essential to our mod
el because we assume that the instances of any emotion category (e.g., anger) do not
have strong statistical regularities in the real world or firm natural category boundaries
(for a discussion of the empirical evidence, see Barrett, 2006a, 2009; Barrett et al., 2007).
As is the case with most categories that lack strong statistical regularities, humans use
words as the glue that binds individual instances of a category together. In fact, even in
fants routinely use the phonological form of words to make conceptual inferences about
novel objects that share little perceptual similarity (e.g., Dewar & Xu, 2009; Xu, 2002). In
this view, language plays a strong causal role in the conceptual development (Gentner &
Goldin-Medows, 2003) of emotion knowledge. Children acquire emotion categories that
conform to their culture, not because there is some natural, biological reality to anger (or
fear or sadness, etc.), but because this level of categorization is socially functional (c.f.
Barrett, 2006a).
Language’s Role in Emotion Perception

Accumulating evidence indicates that words are important, if not necessary, for emotion
perception. In this section, we review evidence from tasks that tap different levels of pro
cessing to see whether and how emotion words affect emotion perception. Specifically,
we look at evidence from emotion perception studies in which words are required or used
as part of the task (i.e., conceptual judgments), studies in which words are not required
but could be used to solve the task (i.e., perceptual judgments), and, finally, evidence
from studies that are purely visual and therefore thought to be impermeable to the effect
of words.
Page 4 of 28

Conceptual Tasks
Many studies of emotion perception require a perceiver to look at a picture of a face and
pick an emotion word that describes how the person in the picture feels. Such a task is
called forced choice. When participants are shown emotional faces and asked to make a
forced choice, they show relatively high agreement (i.e., “accuracy,”) about the emotion
depicted especially when the faces used are highly posed and exaggerated. Yet, when per
ceivers are asked to spontaneously generate a label for such faces, agreement among
participants is significantly reduced (c.f. Russell, 1994; e.g. Boucher & Carlson, 1980).
Moreover, when more naturalistic or less intense (e.g., morphed) emotional faces are
used, agreement is significantly lowered in both forced choice and free labeling tasks (see
Russell, 1994). Participants are also quicker and more accurate to (p. 283) categorize a
face as belonging to an emotion category when the face is primed by a related category
word compared to when the face is primed with an unrelated word (Carroll & Young,
2005). In addition, participants are more likely to falsely recognize an emotion in a non-
prototypical face when asked with a semantic label (e.g., to detect “happy”) than when
asked with a visual descriptor (e.g., to detect a smiling face) (Fernandez-Dols, Carrera,
Barchard, & Gacitua, 2008). Finally, children are better able to match a facial depiction of
emotion to an emotion word than they are at matching two faces posing similar emotions
(e.g., two scowling faces; Widen & Russell, 2003). Thus, emotion words, when used as
part of the task affect conceptual judgments of emotion perception.
Perceptual Judgments
Even in studies that do not use emotion words in the experiment or in which words are
not necessary to solve the task (e.g., a perceptual judgment), emotion words seem to af
fect emotion perception. For example, when a participant’s accessibility to emotion words
is temporarily reduced using a standard laboratory task (via semantic satiation), his or
her ability to judge whether two faces match in emotional content is quite low (Lindquist,
Barrett, Bliss-Moreau, & Russell, 2006). When a participant is primed with an emotion
word (compared to a control word), he or she is more likely to mistakenly identify a less
intense version of the emotional face as the target face, suggesting that emotion words
“fill in” information to the stimulus (Fugate & Barrett, under review). Participants also re
member morphed faces as more “angry” when they are paired with the word “anger”
than when not paired with a word (Halberstadt & Niedenthal, 2001).
Patients who have lost the ability to use words (due to stroke or in semantic aphasia) do
not perceive discrete emotions in another’s face but rather see more generalized affec
tive content (Lindquist, Gendron, Dickerson, & Barrett, 2014; Roberson, Davidoff, &
Braisby, 1999). Thus, emotion words might be necessary to perceive discrete emotion cat
egories, a phenomenon often called categorical perception (see Fugate, 2013 for a re
view).
Specifically, categorical perception occurs when stimuli that vary along a continuum are
perceived as belonging to distinct categories marked by a sharp perceptual boundary
Page 5 of 28

(Harnad, 1987). As a result, perceivers are better at discriminating between stimuli that
belong to different categories than stimuli that belong to the same category, even when
the perceptual distance between stimuli is held constant (called the “between-category
advantage,” Goldstone, 1994). Many studies show that facial depictions of emotion are
perceived in a categorical fashion (Etcoff & Magee, 1992; Calder, Young, Perrett, Etcoff,
& Rowland, 1996; Young et al., 2005). The common conclusion drawn from these studies
is that the structural information in the face is sufficient for the phenomenon to occur.
Yet, in all of these studies participants must also sort the emotional faces into categories
anchored by emotion words. The result is that the structural information in the face alone
might not be sufficient for discrete emotion categorization; rather, a perceiver might use
the provided emotion words to help disambiguate more general affective information in a
face.
Several studies are consistent with this possibility. For example, categorical perception
for human facial depictions of emotion is eliminated when perceivers are first placed un
der verbal load, which makes activating emotion words nearly impossible (Roberson &
Davidoff, 2000; Roberson, Damjanovic, & Pilling, 2007). Deaf signers do not show cate
gorical perception for affective faces (e.g., posed emotional faces), although they do for
linguistic faces (e.g., signs for emotional faces) (McCullough & Emmorey, 2009). In anoth
er study (Fugate, Gouzoules, & Barrett, 2010, study 2), naïve perceivers were first
trained to learn categories of chimpanzee facial expressions (similar to those produced by
humans, but for which people do not readily evoke emotion categories) with either la
beled or unlabeled pictures. Participants showed categorical perception for similar faces
only if they learned the categories of chimpanzee faces with labels (even though the la
bels were not repeated in the task). Those participants who learned the same categories
without a label did not show categorical perception (despite being equally good at detect
ing small structural differences in the faces). In a follow-up study, some participants re
ceived the learned labels as part of the task and some did not. Those participants who
had explicit access to the labels showed different patterns of sorting the morphs than did
those participants who did not have the labels in the task (Fugate, Lynn, & Barrett, un
published). That is, just having the learned labels present in the task changed how faces
were categorized.
Visual Judgments
We now have evidence that emotion words affect not only how faces are judged, but also
how they are seen and encoded in the first place. For example, (p. 284) when the accessi
bility of the meaning of an emotion word is reduced (again, using semantic satiation), par
ticipants do not show decreased reaction times to recognize a previously seen face (a
phenomenon called repetition priming), suggesting that faces seen without access to the
meaning of an emotion word are encoded differently than when words are available (Gen
dron, Lindquist, Barsalou, & Barrett, 2012). Words also influence the contents of visual
consciousness (i.e., what information is seen in the first place). Using a binocular rivalry
paradigm, in which a facial depiction of one emotion was presented to one eye and anoth
er facial depiction to the other eye, participants primed with an emotion word (compared
Page 6 of 28

to a control word) experienced the relevant emotion face for a longer period in visual con
sciousness (Fugate & Barrett, under review). Thus, emotion words affect not only how a
face is encoded but also whether a face is seen in the first place.
Neuroimaging Evidence
Results of several recent meta-analyses of emotion neuroimaging studies show that the
circuitry underlying emotion perception and experience is quite distributed in the brain.
For purposes of this essay, we limit our summary of brain activation found in emotion per
ception and experience tasks to describe those areas known to be integral to category
knowledge (i.e., meaning-making) and semantic processing. For additional brain activa
tion results from meta-analyses of emotion, we refer the reader to works by Kober and
Lindquist (Kober et al., 2008; Lindquist, Wager, Kober, Bliss-Moreau, & Barrett, 2012).1
Meaning-Making Network
Meta-analyses of emotion neuroimaging studies show consistent activation in a network

of regions that include portions of the ventral and dorsal medial prefrontal cortex
(vmPFC, dmPFC) and regions of the medial temporal lobe, including the hippocampus,
entorhinal cortex, parahippocampal cortex, and the posterior cingulate cortex/retrosple
nial area. These areas together are known for their role in conceptualization and are im
portant for making meaning out of incoming sensory and internal bodily cues. Together,
these regions have been referred to as the “default” network (Raichle et al., 2001), the
“context” network (Bar, 2007), and simply the “core” network (Andrews-Hanna, Reidler,
Sepulcre, Poulin, & Buckner, 2010). This network has been implicated in remembering
the past and envisioning the future, theory of mind tasks, default mental activity (sponta
neous, highly associative mental activity), first impressions, fictitious imaginings, emotion
regulation, context framing, and moral decision making (c.f. Buckner & Carroll, 2007; for
reviews, see Adolphs, 2001; Bar, 2007; Blakemore, Winston, & Frith, 2004; Lane &
McRae, 2004; Ochsner, et al., 2004; Yeo, et al., 2011). Edelman (1989) refers to this gen
eral network as constructing “the remembered present.” We hypothesize that these brain
areas use stored representations of prior experiences to make meaning out of incoming
sensory or bodily changes from the self, and additionally when we observe such changes
in others (Barrett, 2006a, 2006b; Lindquist et al., 2012). Our findings indicate that the
conceptualization network is integral in the experience and perception of discrete emo
tions and are consistent with the hypothesis that prior episodic experience helps “mold”
experienced or perceived affect into meaningful instances of emotion (c.f. Lindquist et al.,
2012). Interestingly, some aspects of dmPFC might be functionally selective for conceptu
alization during emotion perception, whereas other aspects of dmPFC support conceptu
alization during emotion experience.2
Page 7 of 28

Language Network
Meta-analyses of emotion perception and experience also show consistent activation in

regions of the inferior gyrus (IFG) (pars orbitalis), ventrolateral PFC (vlPFC), and anterior
temporal lobe (ATL). These areas are broadly implicated in a host of cognitive processes
including those related to language (Binder, Desai, Graves, & Conant, 2009; Gitelman,
Nobre, Sonty, Parrish, & Mesulam, 2005) and the goal-related retrieval of conceptual
knowledge (Badre, Poldrack, Pare-Blagoev, Insler, & Wagner, 2005). For example, labeling
an emotional face increases neural activity in the right IFG, but decreases neural activity
in the amygdala (Lieberman, Eisenberger, Crockett, Tom, Pfeifer, & Way, 2007). This re
duction in amygdala activity was positively correlated with conceptualizing the structural
information of the face. Inferior frontal activation during emotion perception has also
been confirmed with magnetoencephalography (MEG) (Streit et al., 2003). In addition,
patients with brain lesions to the IFG show impairment in emotion perception (Hornak,
Rolls, & Wade, 1996) and specific deficits in phonological, articulatory, and syntactic pro
cessing (see Binder et al., 2009).
The vlPFC has been implicated in the retrieval, maintenance, and manipulation of
(p. 285)
conceptual knowledge stored elsewhere in the brain (Gabrieli, Poldrack, & Desmond,
1998; Martin & Chao, 2001; Poldrack et al., 1999; Wagner, Maril, Bjork, & Schacter,
2001), as well as in semantic processing tasks (e.g., Gitelman et al., 2005), categorization
of objects (e.g., Freedman, Riesenhuber, Poggio, & Miller, 2001), representation of fea
ture-based information for abstract categories (e.g., Freedman, Riesenhuber, Poggio, &
Miller, 2002; Miller, Freedman, & Wallis, 2002), selection among competing response rep
resentations (e.g., Badre & Wagner, 2007; Schnur et al., 2009), and inhibition of respons
es (Aron, Robbins, & Poldrack, 2004). As is the case with the IFG, it is not clear whether
the vlPFC’s role is functionally specific to language: rather, it may be involved in direct
ing attention to salient stimuli in the environment (Corbetta & Shulman, 2002; Corbetta,
Patel, & Shulman, 2008) (c.f. Lindquist et al., 2012).
Last, the ATL (namely the right ATL) has been implicated in the representation of ab
stract social concepts (e.g., Ross & Olson, 2010; Zahn et al., 2007; Zahn et al., 2009). Pa
tients with semantic dementia have focal atrophy to the ATL, difficulty utilizing semantic
knowledge, and exhibit deficits in emotion perception (Rosen et al., 2004) and empathy
(Rankin et al., 2006).
In fact, in a recent meta-analysis of semantic processing neuroimaging studies, the “se

mantic system” proposed a striking similarity to the “default network” (c.f. Binder et al.,
2009). One explanation is that semantic processing may constitute a large component of
cognitive activity during resting states (Binder et al., 1999). Another explanation is that
default activity may actually be analogous to offline processing, in which conceptual
knowledge is operating on internal stores of knowledge built from past experiences
(Binder et al., 1999). Both of these explanations are consistent with the idea that lan
guage is part of meaning making and involved in the use of conceptual knowledge associ
ated with emotional experience and perception.
Page 8 of 28

Evoked Response Potential Evidence
Both emotional faces and emotion words affect similar evoked response potential (ERP)
components across a variety of shared temporal components, although recent direct com
parisons of the time course of both emotional faces and emotion words have produced
varied results (see Früholtz, Jellinghaus, & Herrman, 2011; Rellecke, Palazova, Sommer,
& Schacht, 2011). For example, whether or not emotional words can affect the earliest vi
sual ERP components (as do emotional faces) is unclear (see Fischler & Bradley, 2006;
Kissler, Assadollahi, & Herbert, 2006). Some studies show that emotion words have very
early effects (50–100ms) in P1 (lateral occipital regions) and C1 (parieto-occipital re
gions) (e.g., Bernat, Bunce, & Shevrin, 2001; Scott, O’Donnell, Leuthold, & Sereno, 2009;
Skrandies, 1998; van Hooff, Dietz, Sharma, & Bowman, 2008), whereas other studies only
show later effects. Any semantic information that does get through prior to 150 ms may
be the result of a perceptual association between the word and the affective connotation
through continuous exposure, rather than of being semantically processed (see Rellecke
et al., 2011).3 In this scenario, perceptual input might be sufficient to activate emotional
aspects via magnocellular pathways without or prior to activating in-depth semantic asso
ciations (Hofmann, Kuchinke, Tamm, Võ, & Jacobs, 2009; Kissler et al., 2006; Ortigue et
al., 2004) (c.f. Rellecke et al., 2011).
Emotional words seem to have more consistent effects on later levels of processing, in
cluding the N170 component (130–200 ms) in occipitotemporal regions (Scott et al.,
2009). In one study (Landau, Aziz-Zadeh, & Ivry, 2010), the magnitude of the N170 was
larger in the left hemisphere when face stimuli were preceded by semantic descriptions
of faces compared to semantic descriptions of places. Face stimuli elicited a smaller
N170, however, when they were preceded by visual descriptions of faces compared to vi
sual descriptions of places. The results suggest that semantic information produces con
tent-specific interference on subsequent visual processing. In another study (Baggott,
Palermo, & Fox, 2011), participants saw an emotional word (e.g. “furious”) superimposed
on top of emotional faces and were asked to categorize the word as “sad,” “angry,” “fear
ful,” or neutral. When the words and faces were incongruent, the N170 was more nega
tive compared to when the two were congruent (although both congruent and incongru
ent information had significant effects on the mean amplitude of the N170 compared to
the neutral condition) (see also Zhu, Zhang, Wu, Luo, & Luo, 2010).
Finally, emotion words also affect the “early posterior negativity” (EPN) component (200–
350 ms) over occipitotemporal regions (Herbert, Kissler, Junghöfer, Peyk, & Rockstroh,
2006; Kissler, Herbert, Peyk, & Junghöfer, 2007; Kissler, Herbert, (p. 286) Winkler, &
Junghöfer, 2009; Schacht & Sommer, 2009; Scott et al., 2009) and the “late parietal com
plex/positivity” (LPC/LPP) component (400ms) (Fischler & Bradley, 2006; Herbert et al.,
2006; Herbert, Junghöfer, & Kissler, 2008; Kissler et al., 2009; Schacht & Sommer, 2009).
In one study (Kissler et al., 2007), emotional words (pleasant and unpleasant) were asso
ciated with an enhanced EPN signal compared to neutral words. One possibility put forth
by Kissler and colleagues (2007) is that bidirectional connections between the extrastri
ate ventral stream and the amygdala may amplify the ERP response to emotional words.
Page 9 of 28

Such connections in monkeys have been extensively described (Amaral, Behniea, & Kelly,
2003). Accordingly, left amygdala lesions selectively impair detection of emotional words
as they appear on a computer screen (Anderson & Phelps, 2001).
A Possible Mechanism for Language’s Effect on Emotion Perception
Given both anatomical and neuroimaging (along with neuropsychological) evidence, one
way in which language might aid emotion perception is in helping contribute to identify
ing an affective stimulus. The orbitofrontal cortex (OFC) is one region in the brain that is
important for object recognition and specifically influenced by top-down knowledge. In
one model of object recognition, the OFC is essential for combining perceptual informa
tion from visual regions with top-down information (Kveraga, Ghuman, & Bar, 2007; Bar,
2003). According to this model, the OFC uses early incoming visual information from the
dorsal stream to make a prediction about what the object might be. The OFC then sends
this information to areas like the inferior temporal cortex, which updates the prediction
with incoming bottom-up information until eventually the object is recognized and its af
fective significance registered (Barrett & Bar 2009). In a recent study, activity in the OFC
was correlated with subjectively perceived face categories (e.g., gender), but not with ob
jective (stimulus-based) judgments (Freeman, Rule, Adams, & Ambady, 2010). This find
ing is consistent with the idea that neurons in the OFC reflect the learned properties of a
visual stimulus more than the actual physical properties (Freeman et al., 2010). The OFC
is also connected with the amygdala, as well as with parts of the vlPFC, including the IFG,
which is important for semantic processing (for a recent review, see Binder et al., 2009).
Through such connections, emotion words may contribute to the creation and identifica
tion of an emotional percept.
Another region of the PFC that may be important in identifying an emotional percept is
the vmPFC. The vmPFC serves as a hub that connects conceptual information from multi
ple systems representing episodic memories, signals from the body, and adaptive respons
es (Roy, Shohamy, & Wager, 2012). The idea that the vmPFC is a “meaning making cen
ter” is also supported by anatomical connectivity. The vmPFC is unique among cortical re
gions in that it projects directly to other parts of the PFC (including OFC, dmPFC, and
dlPFC), as well as to subcortical nuclei, the amygdala, hypothalamus, and periaqueductal
gray (PAG) (Roy et al., 2012).
Language’s Role in the Experience of Emotion:

Embodiment
Historically, the activation of conceptual knowledge (including language) was considered
to be coded amodally and in separate storage locations from motor and sensory informa
tion in the brain (e.g., as in semantic memory, E.E. Smith, 1978; E.E. Smith & Medin,
1981). In an embodied view of cognition, however, conceptual knowledge about a stimu
lus is not recoded, but rather is stored modally in sensory and motor systems (see Barsa
lou, 1999, 2005). For example, conceptual knowledge, such as that about emotion, is es
Page 10 of 28

tablished by a distributed collection of context-specific instances captured in the body

and brain across all instances of a category (Niedenthal, 2007; Niedenthal, Barsalou,
Winkielman, Krauth-Gruber, & Ric, 2005).
In our psychological constructionist view of emotion, a person experiences an emotion

when he or she categorizes current information from the body using conceptual knowl
edge and contextual conditions. A specific emotion experience is generated via a multi
modal simulation that reinstates how these conditions have been experienced in the past,
and this conceptual representation interacts with the current information from the body
(see Barrett, 2006a, 2006b; Wilson-Mendenhall, Barrett, Simmons, & Barsalou, 2011).
The embodiment of emotion has a rich history (see Niedenthal et al., 2005 for a review).
In a now classic study, participants who were asked to hold a pen in their mouth (activat
ing the smiling muscle, zygomatic major) reported finding cartoons funnier than partici
pants in a control condition (who held the pen in their teeth) or who held the pen using
only the lips and not the teeth (activating the frowning muscle, corrugator supercilii)
(Strack, Martin, & Stepper, 1988). In addition, ERP measures show that participants who
observe a smile (p. 287) (or frown) engage in making similar facial movements (e.g., Dim
berg & Petterson, 2000; Dimberg, Thunberg, & Elmehed, 2000).
One way in which these varied embodied instances become conceptualized as emotion (or
a specific emotion) might be through a person applying the same emotion word to these
instances (Barrett, 2006a). For example, knowledge about the category anger is estab
lished when sensations from the body and individual behaviors are noted and labeled as
“anger”. As a result, the precise nature of anger knowledge activated at any particular
time will be heterogeneous and reflect the context in which it was first acquired or active
ly applied (Barrett, 2006a).
A wealth of other studies now shows us that embodiment leads to increased comprehen
sion of emotional material (e.g., Barsalou, 2008; Glenberg, 2008; Semin & Smith, 2008;
see Kaschak, this volume). For example, participants are quicker to judge the valence of a
sentence when their facial movements match (compared to mismatched) the sentence
meaning (Havas, Glenberg, & Rinck, 2007). Participants who read verbs related to either
positive or negative emotions (e.g., smile) show increased congruent facial activity (e.g.,
positive verbs engaged the zygomatic major muscle and negative verbs engaged the cor
rugator supercilii muscle) (Foroni & Semin, 2009). In another study, participants in a sad
mood were quicker to make lexical decisions about sad words than they were to make de
cisions about control words. These participants were also quicker to name sad words
compared to control words. Participants in a happy mood made lexical decisions about
sad words more slowly than about control words, and happy participants (compared to
neutral or sad participants) were quicker to name happy words (Niedenthal, Halberstadt,
& Setterlund, 1997).
According to an indexical hypothesis (Glenberg & Robertson, 1999, 2000), language com
prehension (e.g., understanding the verb to smile) leads to physical simulation of the
events to be comprehended. According to this hypothesis, simulation is necessary for lan
Page 11 of 28

guage comprehension. Many imaging studies are also consistent with embodied lan
guage. In a key example, participants who read or listen to phrases or words about specif
ic bodily actions show activation in corresponding regions of the premotor cortex used to
generate these actions (e.g., Aziz-Zadeh & Damasio, 2008; Aziz-Zadeh, Koski, Zaidel,
Mazziotta, & Iacobonni, 2006; Hauk, Johnsrude, & Pulvermuller, 2004; Hauk & Pulver
muller, 2004; Tettamanti et al., 2005). Thus, there is a shared neural representation for
the conceptual knowledge of action.
Language’s Role in Emotion in Special Popula

tions
Semantic Dementia and Frontotemporal Dementia
Emotion reports from various special populations can also shed light on the role of lan
guage in emotion experience and perception. Frontotemporal lobe degeneration (FTLD)
is a neurodegenerative disorder predominantly affecting the frontal and temporal lobes.
Patients typically experience behavioral and personality changes, as well as cognitive im
pairment, including most commonly loss of empathy (Rankin et al., 2006) and emotion
recognition disturbances (Gorno-Tempini et al., 2004b; Kipps & Hodges, 2006; Lough et
al., 2006; Viskontas, Possin, & Miller, 2007). Some studies suggest that negative emotions
are more difficult for FTLD patients to recognize than positive emotions (Diehl-Schmid et
al., 2007; Fernandez-Duque & Black, 2005; Keane, Calder, Hodges, & Young, 2002). Se
mantic dementia (SD), a related neurodegenerative disorder with similar emotion impair
ments, is characterized by ATL atrophy (Gorno-Tempini, et al., 2004a; Patterson, Nestor,
& Rogers, 2007). Both SD and FTLD patients have been shown to have difficulty match
ing, naming, and discriminating negative facial depictions of emotion compared to age-
matched control participants (Rosen et al., 2004). In one report, an SD patient was able
to match pictures of different facial depictions but was unable to match emotion words to
these pictures (Calabria, Cotelli, Adenzato, Zanetti, & Miniussi, 2009). Consistent with
these results, three SD patients were able to detect structural differences among facial
depictions of emotion as well as control participants, but were not able to categorize the
faces into discrete emotion categories. Rather, the SD patients sorted faces into three af
fective piles (i.e., a “neutral” pile, a “positive” pile, which contained faces posing happi
ness, and a “negative” pile, which contained faces posing anger, sadness, disgust, and
fear) (Lindquist et al. 2014). The results suggest that the linguistic deficits from which
these people suffer affect their ability to use emotion words to help make meaning of
more generalized affective information.
Bilingual Populations
Additional evidence for the role of language in emotion experience and perception comes
from (p. 288) bilingual populations. For example, bilinguals (Spanish and English speak
ers) are better able to recall emotion and emotion-laden words in their native language
Page 12 of 28

(Anooshian & Hertel, 1994, but see Ayçiçegˇi & Harris, 2004). In addition, several studies
show that bilinguals’ own emotional experiences are more accessible in their native lan
guage than in a second language (e.g., Altarriba & Santiago-Rivera, 1994; Santiago-
Rivera & Altarriba, 2002), and participants give more intense ratings to emotional scenar
ios when presented in their own language (Matsumoto, Anguas-Wong, & Martinez, 2008).
Conclusion
In the mid 1900s, Benjamin Whorf proposed the revolutionary idea that language shapes
thought (Whorf, 1956). The idea that words can contribute and even determine how a
person thinks has been argued over the years, especially for “perceptually defined” cate
gories or those considered natural kinds. For a long time, emotion, like color, was argued
to be impervious to language. This idea is reflected in many prominent theories of emo
tion. In many of these views, language is orthogonal to emotion (experience or percep
tion) or merely communicates predetermined categories of emotions that are shared by
all cultures, biologically innate, and evolutionarily conserved.
In this essay, we have shown that, contrary to many views of emotion, language affects
emotion experience and perception at various levels of processing, as shown with a vari
ety of behavioral tasks. Evidence from functional imaging and neural activity is consistent
with behavioral evidence and suggests that the networks which are important for detect
ing and interpreting incoming sensory information from the self or from others work in
tandem with those networks that are important for language. Research with bilingual in
dividuals and patient populations also show the role of language in emotion perception
and experience. In fact, language helps us feel and see diverse instances of changes in
bodily sensations and behaviors as emotional. To the extent that a culture has a word for
an emotion and another does not, we might be able to explain cross-cultural differences
in emotion and why perceiving emotion in individuals from cultures different than our
own is more challenging (e.g., Elfenbein & Ambady, 2002; see Wierzbicka, 1992, for a dis
cussion).
These ideas are consistent with the idea that language is for more than just communica
tion (see A. Clark, 1998, 2008; Columbetti, 2009). Language allows us to extend our
minds into the world, providing a “scaffolding” that enhances our cognitive abilities (A.
Clark, 1998, 2008). Accordingly, language not only allows a person to report on how he or
she feels, but also to make feelings explicit and more precise (Columbetti, 2009). Said an
other way, putting one’s feelings into words can clarify and change them. This idea is con
sistent with the idea that, without words, there are only background feelings that are not
conscious to the individual (see Damasio, 1999). Being able to label instances of bodily
changes or behaviors as emotional makes them real. In this sense, emotion words have
causal force (Barrett, 2011) and allow us to engage in socially constructed categories
with those who share the same language and conceptual systems.
Page 13 of 28

Acknowledgments
We would like to thank members of the IAS lab for insightful discussion that led to many
of the ideas discussed in this essay. This essay was written under a National Institutes of
Health Director’s Pioneer Award (DP1OD003312) and NSF grant (BCS 0721260) to Lisa
Feldman Barrett. The content is solely the responsibility of the authors and does not nec
essarily represent the official views of the National Institutes of Health.
References
Abrams, R. L., & Greenwald, A. G. (2000). Parts outweigh the whole (word) in uncon
scious analysis of meaning. Psychological Science, 11(2), 118–124. doi:
10.1111/1467-9280.00226
Adolphs, R. (2001). The neurobiology of social cognition. Current Opinion in

Neurobiology, 11, 231–239. http://dx.doi.org/10.1016/S0959-4388(00)00202-6
Allport, F. H. (1924). Social psychology. New York: Houghton Mifflin.
Altarriba, J., & Santiago-Rivera, A. L. (1994). Current perspectives on using linguistic and
cultural factors in counseling the bilingual Spanish-speaking client. Professional Psycholo
gy: Research and Practice, 25, 388–397. doi: 10.1037/0735-7028.25.4.388
Amaral, D. G., Behniea, H., & Kelly, J. L. (2003). Topographic organization of projections
from the amygdala to the visual cortex in macaque monkey brain. Neuroscience, 118(4),
1099–1120. http://dx.doi.org/10.1016/S0306-4522(02)01001-1
Anderson, A. K., & Phelps, E. A. (2001). Lesions of the human amygdala impair enhanced
perception of emotionally salient events. Nature, 411(6835), 305–309. doi:
10.1038/35077083
Andrews-Hanna, J. R., Reidler, J. S., Sepulcre, J., Poulin, R., & Buckner, R. L. (2010). Func
tional-anatomic fractionation of the brain’s default network. Neuron, 65(4), 550–562. doi:
10.1016/j.neuron.2010.02.005
Anooshian, L. J., & Hertel, P. T. (1994). Emotionality in free recall: Language specificity in
bilingual memory. Cognition and Emotion, 8(6), 503–514. doi:
10.1080/02699939408408956
Arnold, M. B. (1960a). Emotion and personality: Vol. 1. Psychological aspects. New York:
Columbia University Press.
Arnold, M. B. (1960b). Emotion and personality: Vol. 2. Physiological aspects. New York:
Columbia University Press.
Aron, A. R., Robbins, T. W., & Poldrack, R. A. (2004). Inhibition and the right inferior
frontal cortex. Trends Cognitive Science, 8(4), 170–177. doi:10.1016/j.tics.2004.02.010
Page 14 of 28

Averill, J. R. (1980). A constructivist view of emotion. In R. Plutchik & H. Kellerman

(Eds.), Emotion: Theory, research and experience: Vol. I. Theories of emotion (pp. 305–
339). New York: Academic Press.
Averill, J. R. (2012). The future of social constructionism: Introduction to a special section

of Emotion Review. Emotion Review, 4(3), 215–220. doi: 10.1177/1754073912439811
Ayçiçegˇi, A., & Harris, C. L. (2004). Bilinguals’ recall and recognition of emotion words.
Cognition and Emotion, 18(7), 977–987. doi:10.1080/02699930341000301
Aziz-Zadeh, L., & Damasio, A. R. (2008). Embodied semantics for actions: Findings from
functional brain imaging. Journal of Physiology-Paris, 102(1–3), 35–39. http://dx.doi.org/
10.1016/j.jphysparis.2008.03.012
Aziz-Zadeh, L., Koski, L., Zaidel, E., Mazziotta, J., & Iacobonni, M. (2006). Lateralization
of the human mirror neuron system. Journal of Neuroscience, 26(11): 2964–2970. doi:
10.1523/JNEUROSCI.2921-05.2006
Azuma, H. (1984). Secondary control as a heterogeneous category. American

Psychologist, 39I(9), 970–971. doi: 10.1037/0003-066X.39.9.970
Badre, D., Poldrack, R. A., Pare-Blagoev, E. J., Insler, R. Z., & Wagner, A. D. (2005). Disso
ciable controlled retrieval and generalized selection mechanisms in ventrolateral pre
frontal cortex. Neuron, 47, 907–918. http://dx.doi.org/10.1016/j.neuron.2005.07.023
Badre, D., & Wagner, A. D. (2007). Left ventromedial prefrontal cortex and the cognition
control of memory. Neuropsychologia, 45(13), 2883–2901. http://dx.doi.org/10.1016/
j.neuropsychologia.2007.06.015
Baggott, S., Palermo, R., & Fox, A. M. (2011). Processing emotional category congruency
between emotional facial expressions and emotional words. Cognition and Emotion, 25,
369–379. doi: 10.1080/02699931.2010.488945
Bar, M. (2003). A cortical mechanism for triggering top-down facilitation in visual object
recognition. Journal of Cognitive Neuroscience, 15, 600–609. doi:
10.1162/089892903321662976
Bar, M. (2007). The Proactive Brain: Using analogies and associations to generate predic
tions. Trends in Cognitive Sciences, 11(7), 280–289. http://dx.doi.org/10.1016/j.tics.
2007.05.005
Barrett, L. F. (2006a). Are emotions natural kinds? Perspectives on Psychological Science,

1(1), 28–58. doi: 10.1111/j.1745-6916.2006.00003.x
Barrett, L. F. (2006b). Solving the emotion paradox: Categorization and the experience of
emotion. Personality and Social Psychology Review, 10(1), 20–46. doi: 10.1207/
s15327957pspr1001_2
Page 15 of 28

Barrett, L. F. (2009). The future of psychology: Connecting mind to brain. Perspectives in

Psychological Science, 4(4), 326–339. doi: 10.1111/j.1745-6924.2009.01134.x
Barrett, L. F. (2011). Was Darwin wrong about emotional expressions? Current Directions
in Psychological Science, 20(6), 400–406. doi: 10.1177/0963721411429125
Barrett, L. F. (2012). Emotions are real. Emotion, 12(3), 413–429. doi: 10.1037/a0027555
Barrett, L. F., & Bar, M. (2009). See it with feeling: Affective predictions during object
perception. Philosophical Transactions of the Royal Society B., 12(364), 1325–133. doi:
10.1098/rstb.2008.0312
Barrett, L., Lindquist, K. A., & Gendron, M. (2007). Language as context for the percep
tion of emotion. Trends in Cognitive Sciences, 11(8), 327–332. doi: 10.1016/j.tics.
2007.06.003
Barrett, L. F., & Wager, T. D. (2006). The structure of emotion: Evidence from neuroimag
ing Studies. Current Directions in Psychological Science, 15(2), 79–83. doi: 10.1111/j.
0963-7214.2006.00411.x
577–660.
Barsalou, L. W. (2003). Abstraction in perceptual symbol systems. Philosophical Transac

tions of the Royal Society of London B., 358(1435), 1177–1187. doi: 10.1098/rstb.
2003.1319
Barsalou, L. W. (2005). Continuity of the conceptual system across species. Trends in Cog
nitive Sciences, 9(7), 309–311. doi: 10.1016/j.tics.2005.05.003
Barsalou, L. W. (2008). Grounded Cognition. Annual Review of Psychology, 59,

(p. 290)
617–645. doi: 10.1146/annurev.psych.59.103006.093639
Bernat, E., Bunce, S., & Shevrin, H. (2001). Event-related brain potentials differentiate
positive and negative mood adjectives during both supraliminal and subliminal visual pro
cessing. International Journal of Psychophysiology, 42(1), 11–34. http://dx.doi.org/
10.1016/S0167-8760(01)00133-7
Binder, J. R., Desai, R. H., Graves, W. W., & Conant, L. L. (2009). Where is the semantic
system? A critical review and meta-analysis of 120 functional neuroimaging studies. Cere
bral Cortex, 19(12), 2767–2796. doi: 10.1093/cercor/bhp055
Binder, J. R., Frost, J. A., Hammeke, T. A., Bellgowan, P. S. F., Rao, S. M., & Cox, R. W.
(1999). Conceptual processing during the conscious resting state: A functional MRI study.
Journal of Cognitive Neuroscience, 11(1), 80–93. doi: 10.1162/089892999563265
Page 16 of 28

Blakemore, S-J., Winston, J., & Frith, U. (2004). Social cognitive neuroscience: Where are
we heading? Trends in Cognitive Sciences, 8(5), 216–222. http://dx.doi.org/10.1016/
j.tics.2004.03.012
Boiger, M., & Mesquita, B. (2012). The construction of emotion in interactions, relation
ships, and cultures. Emotion Review, 4(3).221–229. doi: 10.1177/1754073912439765
Boucher, J. D. & Calson, G. E. (1980). Recognition of facial expression in three cultures.

Journal of Cross-Cultural Psychology, 11(3), 263–280. doi: 10.1177/0022022180113003
Buckner, R. L., & Carroll, D. C. (2007). Self-projection and the brain. Trends in Cognitive
Sciences, 11(2), 49–57. http://dx.doi.org/10.1016/j.tics.2006.11.004
Calabria, M., Cotelli, M., Adenzato, M., Zanetti, O., & Miniussi, C. (2009). Empathy and
emotion recognition in semantic dementia: A case report. Brain and Cognition, 70, 247–
252. doi:10.1016/j.bandc.2009.02.009
Calder, A. J., Young, A. W., Perrett, D. I., Etcoff, N. L., & Rowland, D. (1996). Categorical
perception of morphed facial expressions. Visual Cognition, 3, 81–117. doi:
10.1080/713756735
Carroll, N. C., & Young, A. W. (2005). Priming of emotion recognition. Quarterly Journal of
Experimental Psychology, 58A, 1173–1197. doi: 10.1080/02724980443000539
Clark, A. (1998). Magic words: How language augments human computation. In P. Car
ruthers & J. Boucher (Eds.), Language and thought: Interdisciplinary themes (pp. 162–
183). Cambridge: Cambridge University Press.
Clark, A. (2008). Supersizing the mind: Embodiment, action, and cognitive extension. Ox
ford: Oxford University Press.
Clore, G. L., & Ortony, A. (2008). Appraisal theories: How cognition shapes affect into
emotion. In M. Lewis, J. M. Haviland-Jones, & L. F. Barrett (Eds.), Handbook of emotions
(3rd ed., pp. 628–642). New York: Guilford Press.
Columbetti, G. (2009). What language does to feelings. Journal of Consciousness Studies:

Controversies in Science & the Humanities, 16(9), 4–26.
Corbetta, M., Patel, G., & Shulman, G. L. (2008). The reorienting system of the human
brain: From environment to theory of mind. Neuron, 58(3), 306–324. doi:10.1016/
j.neuron.2008.04.017
Corbetta, M., & Shulman, G. L. (2002). Controls of goal‐directed and stimulus‐driven at
tention in the brain. Nature Neuroscience, 3(3), 201‐215. doi: 10.1038/nrn755
Damasio, A. R. (1999). The feeling of what happens: Body and emotion in the making of
consciousness. New York: Harcourt Brace.
Page 17 of 28

Dehaene, S., Naccache, L., Le Clec, H. G., Koechlin, E., Mueller, M., Dehaene-Lambertz,
G., et al. (1998). Imaging unconscious semantic priming. Nature, 395, 597–600. doi:
10.1038/26967
Dewar, K., & Xu, F. (2009). Do early nouns refer to kinds or distinct shapes? Evidence
from 10-month-old infants. Psychological Science, 20, 252–257. doi: 10.1111/j.
1467-9280.2009.02278.x
Diehl-Schmid, J., Grimmer, T., Drzezga, A., Bornschein, S., Riemenschneider, M., Forstl,
H., et al. (2007). Decline of cerebral glucose metabolism in frontotemporal dementia: A
longitudinal18F-FDG-PET-study. Neurobiology of Aging, 28(1), 42–60. http://dx.doi.org/
10.1016/j.neurobiolaging.2005.11.002
Dimberg, U., & Petterson, M. (2000). Facial reactions to happy and anger facial expres
sions: Evidence from right hemisphere dominance. Psychophysiology, 37(5), 693–696. doi:
10.1111/1469-8986.3750693
Dimberg, U., Thunberg, M., & Elmehed, K. (2000). Unconscious facial reaction to emo
tional facial expression. Psychological Science, 11(1), 86–89. doi:
10.1111/1467-9280.00221
Edelman, G. M. (1989). The remembered present: A biological theory of consciousness.

Basic Books: New York.
Ekman, P. (1992). Are there basic emotions? Psychological Review, 99(3), 550–553. doi:
10.1037/0033-295X.99.3.550
Ekman, P., Friesen, W. V., & Ellsworth, P. C. (1972). Emotion in the human face. New York:
Pergamon Press.
Elfenbein, H. A., & Ambady, H. (2002). On the universality and cultural specificity of emo
tion recognition: A meta-analysis. Psychological Bulletin, 128(2), 203–235. doi:
10.1037/0033-2909.128.2.203
Ellsworth, P. C., & Scherer, K. R. (2003). Appraisal processes in emotion. In R. J. David

son, K. R. Scherer, & H. H. Goldsmith (Eds.), The handbook of affective science (pp. 572–
595). New York: Oxford University Press.
Etcoff, N. L., & Magee, J. J. (1992). Categorical perception of facial expressions. Cognition,
44(3), 227–240. http://dx.doi.org/10.1016/0010-0277(92)90002-Y
Fernandez-Dols, J-M., Carrera, P., Barchard, K. A., & Gacitua, M.(2008). False recognition
of facial expressions of emotion: Causes and implications. Emotion, 8(4), 530–539. doi:
10.1037/a0012724
Fernandez-Duque, D., & Black, S. E. (2005). Impaired recognition of negative facial emo
tions in patients with frontotemporal dementia. Neuropsychologia, 43(11), 1673–1687.
http://dx.doi.org/10.1016/j.neuropsychologia.2005.01.005
Page 18 of 28

Fischler, I., & Bradley, M. (2006). Event-related potential studies of language and emo
tion: Words, phrases, and task effects. Progress in Brain Research, 156, 185–209. http://
dx.doi.org/10.1016/S0079-6123(06)56009-1
Foroni, F., & Semin, G. R. (2009). Language that puts you in touch with your bodily feel
ings: The multimodal responsiveness of affective expressions. Psychological Science,
20(8), 974–980. doi: 10.1111/j.1467-9280.2009.02400.x
Freedman, D. J., Riesenhuber, M., Poggio, T., & Miller, E. K. (2001). Categorical represen
tation of visual stimuli in the primate prefrontal cortex. Science, 291, 312–316. http://
www.sciencemag.org/content/291/5502/312.short
Freedman, D. J., Riesenhuber, M., Poggio, T., & Miller, E. K. (2002). Visual categorization
and the primate prefrontal cortex: Neurophysiology and behavior. Journal of Neurophysi
ology, 88, 930–942. http://jn.physiology.org/cgi/content/abstract/88/2/929
Freeman, J. B., Rule, N. O., Adams, R. B. Jr., & Ambady, N. (2010). The neural ba
(p. 291)
sis of categorical face perception: Graded representations of face gender in fusiform and
orbitofrontal cortices. Cerebral Cortex, 20, 1314–1322. doi: 10.1093/cercor/bhp195
Frijda, N. H. (1986). The emotions. New York: Cambridge University Press.
Früholtz, S., Jellinghaus, A., & Herrman, M. (2011). Time course of implicit processing
and explicit processing of emotional faces and emotional words. Biological Psychology,
87, 265–274. doi:10.1016/j.biopsycho.2011.03.008
Fugate, J. M. B. (2013). Categorical perception of emotional faces. Emotion Review (5)1,

84–99. doi: 10.1177/1754073912451350
Fugate, J. M. B., & Barrett, L. F. (under review). Emotion words: Adding face value.
Fugate, J. M. B., Gouzoules, H., & Barrett, L. F. (2010). Reading chimpanzee faces: A test
of the structural and conceptual hypotheses. Emotion, 10(4), 544–554. doi: 10.1037/
a0019017
Gabrieli, J. D., Poldrack, R. A., & Desmond, J. E. (1998). The role of left prefrontal cortex
in language and memory. Proceedings of the National Academy of Sciences, 95, 906–913.
http://dx.doi.org/10.1006/nimg.1999.0441
Gendron, M., Lindquist, K. A., Barsalou, L. W., & Barrett, L. F. (2012). Emotion words
shape emotion percepts. Emotion, 12, 314–325. doi: 10.1037/a0026007
Gentner, D., & Goldin-Medows, S. (Eds.). (2003). Language in mind. Cambridge, MA: MIT
Press.
Gitelman, D. R., Nobre, A. C., Sonty, S., Parrish, T. B., & Mesulam, M. M. (2005). Lan
guage network specializations: An analysis with parallel task designs and functional mag
Page 19 of 28

netic resonance imaging. Neuroimage, 26, 975–985. http://www.ncbi.nlm.nih.gov.ezp-

prod1.hul.harvard.edu/pubmed/15893473
Glenberg, A. M. (2008). Toward the integration of bodily states, language, and action. In
G. R. Semin & E. R. Smith (Eds.), Embodied grounding: Social, cognitive, affective, and
neuroscientific approaches (pp. 43–70). New York: Cambridge University Press.
Glenberg, A. M., & Robertson, D. A. (1999). Indexical understanding of instructions. Dis

course Processes, 28(1), 1–26. doi: 10.1080/01638539909545067
Glenberg, A. M., & Robertson, D. A. (2000). Symbol grounding and meaning: A compari
son of high-dimensional and embodied theories of meaning. Journal of Memory and Cog
nition, 43(3), 379–401. http://dx.doi.org/10.1006/jmla.2000.2714
Goldstone, R. (1994). Influences of categorization on perceptual discrimination. Journal of

Experimental Psychology: General, 123, 178–200. doi: 10.1.1.11.6978
Gorno-Tempini, M. L., Dronkers, N. F., Rankin, K. P., Ogar, J. M., Phengrasamy, L., Rosen,
H. J., et al.(2004a). Cognition and anatomy in three variants of primary progressive apha
sia. Annals of Neurology, 55(3), 335–346. doi: 10.1002/ana.10825
Gorno-Tempini, M. L., Murray, R. C., Tankin, K. P., Weiner, M. W., & Miller, B. L. (2004b).
Clinical, cognitive and anatomical evolution from nonfluent progressive aphasia to corti
cobasal syndrome: A case report. Neurocase: The Neural Basis of Cognition, 10(6), 426–
436. doi: 10.1080/13554790490894011
Greenwald, A. G., Draine, S. C., & Abrams, R. L. (1996). Three cognitive markers of un
conscious semantic activation. Science, 273(5282), 1699–1702. doi:10.1126/science.
273.5282.1699
Halberstadt, J. B., & Niedenthal, P. M. (2001). Effects of emotion concepts on perceptual

memory for emotional expressions. Journal of Personality and Social Psychology, 81(4),
587–598. doi: 10.1037/0022-3514.81.4.587
Harré, R. (1986). The social construction of emotions. New York: Blackwell.
Harnad, S. (1987). Psychophysical and cognitive aspects of categorical perception: A criti

cal overview. In Harnad, S. (Ed.), Categorical perception: The groundwork of cognition.
New York: Cambridge University Press.
Hauk, O., Johnsrude I., & Pulvermüller, F. (2004). Somatotopic representation of action
words in human motor and premotor cortex. Neuron, 41(2), 301–307. http://dx.doi.org/
10.1016/S0896-6273(03)00838-9
Hauk, O., & Pulvermüller, F. (2004). Neurophysiological distinctions of action words in the
fronto-central cortex. Human Brain Mapping, 21(3), 191–201. doi: 10.1002/hbm.10157
Page 20 of 28

Havas, D., Glenberg, A. M., & Rinck, M. (2007). Emotion simulation during language com
prehension. Psychonomic Bulletin & Review, 14(3), 436–441. doi:10.3758/BF03194085
Herbert, C., Junghöfer, M., & Kissler, J. (2008). Event related potentials to emotional ad
jectives during reading. Psychophysiology, 45(3), 487–498. doi: 10.1111/j.
1469-8986.2007. 00638.x
Herbert, C., Kissler, J., Junghöfer, M., Peyk, P., & Rockstroh, B. (2006). Processing of emo
tional adjectives: Evidence from startle EMG and ERPs. Psychophysiology, 43(2), 197–
206. doi: 10.1111/j.1469-8986.2006.00385.x
Hochschild, A. R. (1983). The managed heart: The commercialization of human feeling.

Berkeley: University of California Press.
Hofmann, M. J., Kuchinke, L., Tamm, S., Võ, M. L. H., & Jacobs, A. M. (2009). Affective
processing within 1/10th of a second: High arousal is necessary for early facilitative pro
cessing of negative but not positive words. Cognitive, Affective, & Behavioral Neuro
science, 9(4), 389–397. doi:10.3758/9.4.389
Hornak, J., Rolls, E. T., & Wade, D. (1996). Face and voice expression identification in pa
tients with emotional and behavioral changes following ventral frontal lobe damage. Neu
ropsychologia, 34(4), 247–261. http://dx.doi.org/10.1016/0028-3932(95)00106-9
Izard, C. E. (1971). The face of emotion. New York: Appleton-Century-Crofts.
James, W. (1884). What is an emotion? Mind, 188–205.
James, W.(1890). The principles of psychology, Oxford, UK: Holt.
Keane, J., Calder, A. J., Hodges, J. R., & Young, A. W. (2002). Face and emotion processing
in frontal variant frontotemporal dementia. Neuropsychologia, 40(6), 655–665. http://
dx.doi.org/10.1016/S0028-3932(01)00156-7
Kipps, C. M., & Hodges, J. R. (2006). Theory of mind in frontotemporal dementia. Social
Neuroscience, 1(3–4), 235–244. doi: 10.1080/17470910600989847
Kissler, J., Assadollahi, R., & Herbert, C. (2006). Emotional and semantic networks in visu
al word processing: Evidence from ERP studies. Progress in Brain Research, 156, 147–
183. http://dx.doi.org/10.1016/S0079-6123(06)56008-X
Kissler, J., Herbert, C., Peyk, P., & Junghöfer, M. (2007). Buzzwords: Early cortical re
sponses to emotional words (p. 292) during readings. Psychological Science, 18(6), 475–
480. doi: 10.1111/j.1467-9280.2007.01924.x
Kissler, J., Herbert, C., Winkler, I., & Junghöfer, M. (2009). Emotion and attention in visual
word processing—An ERP study. Biological Psychiatry, 80(1), 75–83. http://dx.doi.org/
10.1016/j.biopsycho.2008.03.004
Page 21 of 28

Kober, H., Barrett, L. F., Joseph, J., Bliss-Moreau, E., Lindquist, K. A., & Wager, T. D.
(2008). Functional grouping and cortical-subcortical interactions in emotion: A meta-
analysis of neuroimaging studies. Neuroimage, 42(2), 998–1031. http://
www.ncbi.nlm.nih.gov/pmc/articles/PMC2752702
Kouider, S., & Dupoux, E. (2004). Partial awareness creates the “illusion” of subliminal se
mantic priming. Psychological Science, 15(2), 75–81. doi: 10.1111/j.
0963-7214.2004.01502001.x
Kveraga, K., Ghuman, A. S. J., & Bar, M. (2007). Magnocellular projections as the trigger
of top-down facilitation in recognition. Journal of Neuroscience, 27, 2109–2128. doi:
10.1523/JNEUROSCI.3481-07.2007
Landau, A. N., Aziz-Zadeh, L., & Ivry, R. B. (2010). The influence of language on percep
tion: Listening to sentences about faces affects the perception of faces. Journal of Neuro
science, 30¸15254–15261. doi: 10.1523/JNEUROSCI. 2046-10.2010
Lane, R. D., & McRae, K. (2004). Neural substrates of conscious emotional experience: A
cognitive-neuroscience perspective. In M. Beauregard (Ed.), Consciousness, emotional
self-regulation and the brain(pp.87–122). Amsterdam, Netherlands: John Benjamins Pub
lishing.
Lazarus, R. S. (1991). Emotion and adaptation. New York: Oxford University Press.
Lieberman, M. D., Eisenberger, N. I., Crockett, M. J., Tom, S. M., Pfeifer, J. H., & Way, B.
M. (2007). Putting feelings into words: Affective labeling disrupts amygdale activity in re
sponse to affective stimuli. Psychological Science, 18(5), 421–428. doi: 10.1111/j.
1467-9280.2007.01916.x
Lindquist, K. A., Barrett, L. F., Bliss-Moreau, E., & Russell, J. A. (2006). Language and the
perception of emotion. Emotion, 6, 125–138. doi: 10.1037/1528-3542.6.1.125
Lindquist, K. A., Gendron, M., Dickerson, B. C., & Barrett, L. F. (2014). Emotion percep
tion, but not affect perception, is impaired with semantic memory loss. Emotion, 14,
375-387.
Lindquist, K. A., Wager, T. D., Kober, H., Bliss-Moreau, E., & Barrett, L. F. (2012). The
brain basis of emotion: A meta-analytic review. Behavioral and Brain Sciences, 35(3), 121–
143. http://dx.doi.org/10.1017/S0140525X11000446
Lough, S., Kipps, C. M., Treise, C., Watson, P., Blair, J. R., & Hodges, J. R. (2006). Social
reasoning, emotion and empathy in frontotemporal dementia. Neuropsychologia, 44(6),
950–958. http://dx.doi.org/10.1016/j.neuropsychologia.2005.08.009
Lutz, C. A. (1988). Unnatural emotions. Chicago: University of Chicago Press.
Page 22 of 28

Martin, A., & Chao, L. L. (2001). Semantic memory and the brain: Structure and process.
Current Opinion in Neurobiology, 11(2), 194–201. http://dx.doi.org/10.1016/
S0959-4388(00)00196-3
Masuda T., Ellsworth P. C., Mesquita B., Leu J., Tanida S., & Van de Veerdonk, E. (2008).
Placing the face in context: Cultural differences in the perception of facial emotion. Jour
nal of Personality and Social Psychology, 94, 365–381. doi: 10.1037/0022-3514.94.3.365
Matsumoto, D., Anguas-Wong, A. M., & Martinez, E. (2008). Priming effects of language
on emotion judgments in Spanish-English bilinguals. Journal of Cross-Cultural
Psychology, 39(3), 335–342. doi: 10.1177/0022022108315489
McCullough, S., & Emmorey, K. (2009). Categorical perception of affective and linguistic
facial expressions. Cognition, 110, 208–221. doi:10.1016/j.cognition.2008.11.007
McDougall, W. (1921). An introduction to social psychology. Boston: John W. Luce. (Origi

nal work published 1908).
Miller, E. K., Freedman, D. J., & Wallis, J. D. (2002). The prefrontal cortex: Categories,
concepts and cognition. Philosophical Transactions of the Royal Society of London B.,
357(1424).1123–1136. doi: 10.1098/rstb.2002.1099
Naccache, L., Dehaene, S., Cohen, L., Habert, M-O., Guichart-Gomez, E., Galanaud, D. et
al. (2005). Effortless control: Executive attention and conscious feeling of mental effort
are dissociable. Neuropsychologia, 43(9), 1318–1328. http://dx.doi.org/10.1016/
j.neuropsychologia.2004.11.024
Niedenthal, P. M. (2007). Embodying emotion. Science, 316(5827), 1002–1005. doi:

10.1126/science.1136930
Niedenthal, P. M., Barsalou, L. W., Winkielman, P., Krauth-Gruber, S., & Ric, F. (2005). Em
bodiment in attitudes, social perception, and emotion. Personality and Social Psychology
Review, 9(3), 184–211. doi: 10.1207/s15327957pspr0903_1
Niedenthal, P. M., Halberstadt, J. B., & Setterlund, M. B. (1997). Being happy and seeing
“happy”: Emotional state mediates visual word recognition. Cognition & Emotion, 11(4),
403–432. 10.1080/026999397379863
Ochsner, K. N., Ray, R. D., Cooper, J. C., Robertson, E. R., Chopra, S., Gabrieli, J. D., et al.
(2004). For better or for worse: Neural systems supporting the cognitive down- and up-
regulation of negative emotion. Neuroimage, 23(2), 483–499. http://dx.doi.org/10.1016/
j.neuroimage.2004.06.030
Ortigue, S., Michel, C. M., Murray, M. M., Mohr, C., Caronnel, S., & Landis, T.(2004). Elec
trical neuroimaging reveals early generator modulation to emotional words. Neuroimage,
21(4), 1242–1251. http://dx.doi.org/10.1016/j.neuroimage.2003.11.007
Page 23 of 28

Patterson, K., Nestor, P. J., & Rogers, T. (2007). Where do you know what you know? The
representations of semantic knowledge in the human brain. Nature Reviews Neuro
science, 8, 976–987. doi:10.1038/nrn2277
Poldrack, R. A., Wagner, A. D., Prull, M. W., Desmond, J. E., Glover, G. H., & Gabrieli, J. D.
(1999). Functional specialization for semantic and phonological processing in the left in
ferior prefrontal cortex. Neuroimage, 10(1), 15–35. http://dx.doi.org/10.1006/nimg.
1999.0441
Raichle, M. E., MacLeod, A. M., Snyder, A. Z., Powers, W. J., Gusnard, D. A., & Shulman,
G. L. (2001). A default mode of brain function. Proceedings of the National Academy of
Sciences, 98(2): 676–82. doi: 10.1073/pnas.98.2.676
Rankin, K. P., Gorno-Tempini, M. L., Allison, S. C., Stanley, C. M., Glenn, S., Weiner, M. W.,
& Miller, B. L.(2006). Structural anatomy of empathy in neurodegenerative disease.
Brain: A Journal of Neurology, 129(11), 2945–2956. doi: 10.1093/brain/awl254
Ratner, C. (1989). A social constructionist critique of naturalistic theories of emotion.

Journal of Mind and Behavior, 10, 211–230.
Rellecke, J., Palazova, M., Sommer, W., & Schacht, A. (2011). On the automaticity of emo
tion processing in words and faces; Event-related brain potentials evidence from a super
ficial task. Brain and Cognition, 77(1), 23–32. http://dx.doi.org/10.1016/j.bandc.
2011.07.001
Roberson, D., Damjanovic, L., & Pilling, M. (2007). Categorical perception of fa
(p. 293)
cial expressions: Evidence for a “category adjustment” model. Memory & Cognition, 35,
1814–1829. doi: 10.3758/BF03193512
Roberson, D., & Davidoff, J. (2000). The categorical perception of color and facial expres
sions: The effect of verbal interference. Memory & Cognition, 28, 977–986. http://
www.ncbi.nlm.nih.gov/pubmed/11105523
Roberson, D., Davidoff, J., & Braisby, N. (1999). Similarity and categorisation: Neuropsy
chological evidence for a dissociation in explicit categorisation tasks. Cognition, 71(1), 1–
42. 10.1016/S0010-0277(99)00013-X
Rosen, H. J., Pace-Savitsky, K., Perry, R. J., Kramer, J. H., Miller, B. L., & Levenson, R. W.
(2004). Recognition of emotion in the frontal and temporal variants of frontotemporal de
mentia. Dementia and Geriatric Cognitive Disorders 17, 277–81. http://
www.ncbi.nlm.nih.gov/pubmed/15178936
Roseman, I. J. (1991). Appraisal determinants of discrete emotions. Cognition & Emotion,

5(3), 161–200. doi: 10.1080/02699939108411034
Ross, L. A., & Olson, I. R. (2010). Social cognition and the anterior temporal lobes. Neu
roimage, 49(4), 3452–3462. http://dx.doi.org/10.1016/j.neuroimage.2009.11.012
Page 24 of 28

Roy, M., Shohamy, D., & Wager, T. D. (2012). Ventromedial prefrontal-subcortical systems
and the generation of affective meaning. Trends in Cognitive Sciences, 16(3), 147–156.
http://dx.doi.org/10.1016/j.tics.2012.01.005
Russell, J. A. (1994). Is there universal recognition of emotion from facial expression? A

review of the cross-cultural studies. Psychological Bulletin, 115(1), 102–112. doi:
10.1037/0033-2909.115.1.102
Russell, J. A. (2003). Core affect and the psychological construction of emotion. Psycho
logical Review, 110(1), 145–172. doi: 10.1037/0033-295X.110.1.145
Santiago-Rivera, A. L., & Altarriba, J. (2002). The role of language in therapy with the
Spanish-English bilingual client. Professional Psychology: Research and Practice, 33(1),
30–38. 10.1037/0735-7028.33.1.30
Schacht, A., & Sommer, W. (2009). Emotions in word and face processing: Early and late
cortical responses. Brain and Cognition, 69(3), 538–550. http://dx.doi.org/10.1016/
j.bandc.2008.11.005
Scherer, K. R. (1984). Emotion as a multicomponent process: A model and some cross-cul

tural data. In P. Shaver (Ed.), Review of personality and social psychology(vol. 5) (pp. 37–
63). Beverly Hills, CA: Sage.
Schnur, T., Schwartz, M. F., Kimberg, D. Y., Hirshorn, E., Coslett, H. B., & Thompson-
Schill, S. L.(2009). Localizing interference during naming: Convergent neuroimaging and
neuropsychological evidence for the function of Broca’s area. Proceedings of the National
Academy of Sciences, 106(1), 322–327. doi: 10.1073/pnas.0805874106
Scott, G. G., O’Donnell, P. J., Leuthold, H., & Sereno, S. C. (2009). Early emotion word pro
cessing: Evidence from event-related potentials. Biological Psychology, 80(1), 95–104.
http://dx.doi.org/10.1016/j.biopsycho.2008.03.010
Semin, G. R., & Smith, E. E. (2008). Embodied grounding: Social, cognitive, affective and
neuroscientific approaches. Cambridge: Cambridge University Press.
Skrandies, W. (1998). Evoked potential correlates of semantic mapping – A brain mapping

study. Cognitive Brain Research, 6, 173–183. doi:10.1016/S0926-6410(97)00033-5
Smith, E. E. (1978). Theories of semantic memory. In W. K. Estes (Ed.), Handbook of

learning and cognitive processes(Vol. 6). Hillsdale, NJ: Erlbaum.
Smith, E. E., & Medin, D. L. (1981). Categories and concepts. Cambridge, MA: Harvard
University Press.
Strack, F., Martin, L. L., & Stepper, S. (1988). Inhibiting and facilitating conditions of the
human smile: A nonobtrusive test of the facial feedback hypothesis. Journal of Personality
and Social Psychology, 54(5), 768–777. doi: 10.1037/0022-3514.54.5.768
Page 25 of 28

Streit, M., Dammers, J., Simsek-Kraues, S., Brinkmeyer, J., Wolwer, W., & Ioannides, A.
(2003). Time course of regional brain activations during facial emotion recognition in hu
mans. Neuroscience Letters, 342(1–2), 101–104. http://dx.doi.org/10.1016/
S0304-3940(03)00274-X
Tettamanti, M., Buccino, G., Saccuman, M. C., Gallese, V., Danna, M., Scifo, P., et al.
(2005). Listening to action-related sentences activates fronto-parietal motor circuits. Jour
nal of Cognitive Neuroscience, 17(2), 273–281. doi: 10.1162/0898929053124965
Tomkins, S. S. (1962). Affect, imagery, and consciousness: Vol. 1. The positive affects.
New York: Springer.
vanHooff, J. C., Dietz, K. C., Sharma, D., & Bowman, H. (2008). Neural correlates of intru
sion of emotion words in a modified Stroop task. International Journal of Psychophysiolo
gy, 67, 23–34.
Viskontas, I. V., Possin, K. L., & Miller, B. L. (2007). Symptoms of frontotemporal demen
tia provide insights into orbitofrontal function and social behavior. Annals of the New York
Academy of Sciences, 1121, 528–545. doi: 10.1196/annals.1401.025
Wager, T. D., Barrett, L. F., Bliss-Moreau, E., Lindquist, K. A., Duncan, S., Kober, H., et al.
(2008). The neuroimaging of emotion. In M. Lewis, J. M. Haviland-Jones, & L. F. Barrett
(Eds.), Handbook of emotions (3rd ed., pp. 249–271). New York: Guilford.
Wagner, A. D., Maril, A., Bjork, R. A., & Schacter, D. L. (2001). Prefrontal contributions to
executive control: FMRI evidence for functional distinctions within lateral prefrontal cor
tex. Neuroimage, 14(6), 1337–1347. doi:http://dx.doi.org/10.1006/nimg.2001.0936
White, M., & LeVine, R. A. (1986). What is an Ii ko (good child)? In H. Stevenson, H. Azu
ma,& K. Hakuta (Eds.), Child development and education in Japan (pp. 55–62). New York:
Freeman.
Whorf, B. L. (1956). Language, thought and reality. Cambridge, MA: Technology Press of
MIT.
Widen, S. C., & Russell, J. A. (2003). A closer look at preschooler’s freely led to labels for
facial expressions. Developmental Psychology, 39(1), 114–128. doi:
10.1037/0012-1649.39.1.114
Wierzbicka, A. (1992). Talking about emotions: Semantics, culture, and cognition. Cogni
tion and Emotion, 6(3–4), 283–319. doi: 10.1080/02699939208411073
Wilson-Mendenhall, C. D., Barrett, L. F., Simmons, W. K., & Barsalou, L. W. (2011).

Grounding emotion in situated conceptualization. Neuropsychologia, 49(5), 1105–1127.
http://dx.doi.org/10.1016/j.neuropsychologia.2010.12.032
Xu, F. (2002). The role of language in acquiring object kind concepts in infancy. Cognition,
85(3), 223–250. http://dx.doi.org/10.1016/S0010-0277(02)00109-9
Page 26 of 28

Yeo, B. T. T., Krienen, F. M., Sepulcre, J., Sabuncu, M. R., Lashkari, D., Hollinshead, M., et
al. (2011). The (p. 294) organization of the human cerebral cortex estimated by intrinsic
functional connectivity. Journal of Neurophysiology, 106(3), 1125–1165. doi: 10.1152/jn.
00338.
Young, A. W., Rowland, D., Calder, A. J., Etcoff, N. L., Seth, A., & Perrett, D. I. (2005). Fa
cial expression megamix: Tests of dimensional and category accounts of emotion recogni
tion. Cognition, 63, 271–313. doi: 10.1016/S0010-0277(97)00003-6
Zahn, R., Moll., J., Krueger, F., Heuy, E. D., Garrido, G., & Grafman, J. (2007). Social con
cepts are represented in the superior anterior temporal cortex. Proceedings of the Na
tional Academy of Sciences, 104(15), 6430–6435. doi:10.1073/pnas.0607061104
Zahn, R., Moll, J., Paiva, M., Garrido, G., Krueger, F., Huey, E. D., et al. (2009). The neural
basis of human social values: Evidence from functional MRI. Cerebral Cortex, 19(2), 276–
283. doi: 10.1093/cercor/bhn080
Zhu, X., Zhang, J., Wu, T., Luo., W., & Luo, Y. (2010). Emotional conflict occurs at an early
stage: Evidence from the emotional face–word Stroop task. Neuroscience Letters, 478(1),
1–4. http://dx.doi.org/10.1016/j.neulet.2010.04.036
Notes:
(1) . It is also worth noting that the results of these meta-analyses show little evidence to
support the idea that individual categories of emotion are associated with consistent and
specific activation in the brain (as would be consistent with a natural kinds view of emo
tion) (Barrett & Wager, 2006; Kober et al., 2008; Lindquist et al., 2012; Wager et al.,
2008). Rather, these meta-analyses show that brain activity during instances of emotion
perception and experience is similar across many constructed mental states, including
(but not specific to) those seen in “emotion” tasks, “memory” tasks, “cognitive” tasks, and
other “domain-specific” tasks (Barrett & Wager, 2006; Kober et al., 2008; Lindquist et al.,
2012).
(2) . In fact, when study mode (e.g., of emotion experience vs. perception) was entered as
a factor in analysis, the experience of emotion was overall associated more with the right
lateral OFC, whereas the perception of emotion was associated more with the dmPFC/
dACC, right hippocampus, left vlPFC, right peristriate, and right occipitotemporal area
(Lindquist et al., 2012). For an explicit comparison of meta-analytic studies showing dif
ferences in brain regions associated with experience vs. perception see Wager et al.
(2008).
(3) . This debate is reflected in the controversy as to whether masked primes (unavailable
to consciousness) can be semantically processed. A few studies of masked priming
(Greenwald, Draine, & Abrams, 1996; Dehaene et al., 1998) have suggested semantic pro
cessing without conscious perception. Other studies, however, have shown that many of
these priming effects can be explained by direct motor specification (Abrams & Green
Page 27 of 28

wald, 2000), in which prior exposure to a consciously perceived word creates a direct as
sociation between this stimulus and the corresponding motor response, bypassing seman
tic analysis, or partial awareness (Kouider & Dupoux, 2004). In a rather convincing re
cent study, however, the valence of masked emotional words modulated responding in the
amygdala (Naccache et al., 2005).
Jennifer M.B. Fugate
Jennifer M.B. Fugate, Department of Psychology, Boston College, Department of Psy

chology, Northeastern University, Boston, MA
Lisa Feldman Barrett
Lisa Feldman Barrett, Psychiatric Neuroimaging Research Program, Harvard Med

ical School, Martinos Center for Biomedical Imaging, Massachusetts General Hospi
tal, Department of Psychology, Northeastern University, Boston, MA
Page 28 of 28

Discursive Social Psychology

Alexandra Kent and Jonathan Potter

This chapter introduces basic features of discursive social psychology using discursive
analyses of family mealtime interaction. It puts discursive social psychology in historical
context and distinguishes three overlapping, strands of work: (a) the use of open-ended
interviews to identify interpretative repertoires, (b) the focus on naturalist data to consid
er how versions of social life are constructed to support actions, and (c) the use of meth
ods and findings from conversation analysis to reveal the sequential and interactionally
embedded organization of social psychological phenomena. It overviews its basic theoreti
cal principles: that discourse is (a) oriented to action; (b) situated sequentially, institution
ally and rhetorically; and (c) constructed and constructive. Its systematically noncogni
tivist approach to human conduct is explained and justified. The chapter overviews re
search design, data collection and management, transcription, analysis and validation
and shows how these work together to build a systematic alternative of social psychology
Future directions for discursive social psychology are also considered.
Keywords: Discursive psychology, conversation analysis, discourse analysis, transcription, sequence organization,
action formation, qualitative methods, natural data
The following fragment comes from the very beginning of a UK family breakfast. Mum
has started the video and placed Katherine, who is 5, where she will face the camera and
be able to see Mum getting breakfast through the kitchen door. Katherine is not happy!
Page 1 of 34

Discursive social psychologists work primarily with material of this kind; that is,
(p. 296)
with interaction between people happening naturally in the settings where they live their
lives. It is not staged by the researcher in an experiment or vignette; it is life as it hap
pens. The recording is done on digital video, which allows researchers to capture differ
ent features of talk, such as prosody, as well as embodied elements of the interaction,
such as who is sitting where. Crucially, the video allows the interaction to be captured not
just as a crystalized whole, as it appears in the transcript, but as something unfolding
with precise and relevant timing. This unfolding is not just there for the researcher to ex
amine; it is central to the intelligibility of this interaction for the participants themselves.
A recording of this kind and its associated transcript allows the researcher potent access
to psychological and social psychological matters. For example, Katherine’s complaint
about her table position in lines 1 and 4 is delivered using a high-pitched, creaky voice
and stretching of the word “work” (more on transcription below). It displays upset and
perhaps pleading. In overlap with this, Mum in lines 2 and 3 uses Katherine’s name as a
“summons.” Deliberate overlap of this kind is relatively unusual in interaction. Here Mum
displays a specific disattention to the substance of Katherine’s complaint.
Mum now issues a threat. As Hepburn and Potter (2011) have noted, threats are pro
foundly social psychological in nature—they are a form of pumped up social influence—
yet they have been little studied. Mostly, they have appeared in research as variables in
relation to cognitive dissonance or power. But what makes something a threat and how it
operates has escaped social psychological study. In their study, Hepburn and Potter note
Page 2 of 34

that threats are often constructed with an if-then grammar, and this is what we see in
lines 10 to 12: =if you ↑carry on↑ ↑whingeing and whining,=during breakfas’ time I’ll
send you to the bottom step. This has an “if–then” structure that formulates both the ac
tion that is a problem (whinging and whining) and the negative upshot that will follow
(being sent to the bottom step). The latter is produced as contingent on the former.
These materials allow us to consider how the interaction unfolds and each element of the
interaction is relevant to what came before and what came after. This involves an under
standing of the role of different lexical items and grammatical organizations. For exam
ple, note Mum’s “if you carry o:n:, whingeing an whining?” on lines 3, 5, and 6. Grammat
ically, this tells us, and Katherine as an already competent speaker, that more is to come.
But rather than immediately delivering what is grammatically due, Mum delays its
progress to formulate what her own state now is “had enough of i:t,” (not willing to take
any more from Katherine) and to gloss what her (still impending) action will be: “S is your
warning now.” She holds up the action, therefore, to underline the seriousness with which
she should be taken and to characterize its nature. Conversation analytic work empha
sizes that “progressivity” is generally an important feature of interaction, so we are likely
to find something analytically interesting where that progressivity is held up (Schegloff,
2007).
Note also Mum’s use of the term “warning.” Although what she goes on to issue takes the
standard form of a threat, building it as a warning can soften her role in delivering the
noxious upshot of having to sit on the bottom step. Threats are rarely issued as explicit
speech acts; they tend to appear in the form of attributions: “are you threatening me”?
Discursive psychology builds on the basic Wittgenstein (1953) insight that mental and
psychological terms have a fundamentally practical role.
Note that by the end of line 12 Mum has fully issued the threat. At this point, Mum’s ac
tion makes relevant a range of specific next actions from Katherine. Most appropriate, at
least from Mum’s perspective, will be compliance. And Katherine does cease making com
plaints about her seating position, at least to start with. However, threats as a pumped up
attempt at social influence set up the possibility of a pumped up response, defiance. More
precisely, they generate a slot in the interaction in which Katherine can do things that will
count as defiant; Hepburn and Potter (2011) identify examples of this kind. This makes
the issuing of threats both a powerful and a risky interactional maneuver. However, with
actions such as threats and directives, there are further options. For example, Katherine
could have shown what Kent has called incipient compliance by giving some indication
that compliance is forthcoming, and yet it has not actually been offered yet (for more de
tails, see Craven & Potter, 2010; Kent, 2012a; Hepburn & Potter, 2011).
Let us note one final feature of this extract to illustrate another strand of discursive psy
chological work. Mum glosses Katherine’s complaint as “whinging and whining.” As Ed
wards (2005a) has emphasized, complaints are “object-side” descriptions; they focus on
states of affairs that are unsatisfactory—these shoes have come apart, I have been
(p. 297) given the wrong chair at breakfast. And they focus on remediation—the shoes
Page 3 of 34

should be replaced, the seating plan should be changed. However, they are open to for
mulation in ways that focus not on the states of affairs but on the complainer. Such for
mulations are “subject-side” descriptions because they focus on the subjectivity of the
complainer rather than the object. This is precisely what Mum does with Katherine’s com
plaint—she treats it as about Katherine rather than the problem seating and thereby dis
misses it.
We can see in these concrete, readily researched materials then some of the big issues of
social relations played out. We can see a practical demonstration of how social influence
and resistance go together, and we can see how objectivity and subjectivity become live,
rhetorically opposed possibilities. These are some of the basic concerns of discursive psy
chology. Analytically, we see how we need to pay attention to the way actions are built
and responded to within particular sequential positions, the role of different lexical items,
the inflection and prosody of delivery, and the role of gesture and embodied action. We
now develop some of these themes more systematically.
This chapter introduces and overviews a discursive approach to social psychological mat
ters. We refer to this as discursive social psychology (DSP). Because discursive psycholo
gy is a fundamentally social approach, we occasionally use the terms discursive psycholo
gy and discursive social psychology interchangeably. Discursive social psychology begins
with social psychological matters as they arise practically and within the settings in
which people are living their lives. Its focus is on how psychological issues and objects
are constructed, understood, and displayed as people interact in mundane and institution
al situations. Discursive social psychology looks at the ways in which psychology is impor
tant in conversations between people. This style of research starts with video or audio
recordings of people talking, in either everyday settings (family mealtimes or everyday
phone calls) or in professional settings in which the focus is on medical, legal, therapeutic
or other matters. It works directly with those materials rather than giving people ques
tionnaires, interviewing them, or putting them through experimental manipulations. Dis
cursive social psychology focuses on how psychological issues such as motivation, inten
tion or distress are revealed by speakers in interaction. For example, in a child protection
help line, the call taker may need to identify callers’ distress from what they say and tiny
tremors in their voice, and the call taker may need to say just the right thing to be sooth
ing but not encroach on the caller’s experience. Discursive psychology focuses on how
people do that and how they use basic resources that all of us use when talking to one an
other to be successful. This close focus on what goes on in its natural setting makes DSP
a particularly powerful approach for having practical impact and helping professionals of
all kinds do a better job. It focuses on questions such as:
• How is upset displayed, recognized, and received in mundane and institutional set
tings (Hepburn & Potter, 2012)?
• How do narratives in sex offender therapy sessions manage issues of blame, and how
can this be misidentified as a “cognitive distortion” (Auburn, 2005)?
Page 4 of 34

• How do “why formatted” interrogatives work as elements in complaints (Bolden &

Robinson, 2011)?
• How does one of the parties in a relationship counseling session build a description
of troubles that indirectly blames the other party and places the onus on him or her to
change (Edwards, 1997)?
• How does a speaker show that he or she is not prejudiced while developing a damn
ing version of an entire ethnic group (Wetherell & Potter, 1992)?
Questions of this kind involve a focus on matters that are social and psychological for
people as they act and interact in particular settings—in families, in workplaces, in
schools and so on. They may not all have appeared as chapter headings in the convention
al textbooks of social psychology, but they are about the way people coordinate their ac
tivities and modify the behavior of one another. Instead of starting with inner mental or
cognitive processes happening below and behind interaction, it starts with the public dis
plays, constructions and orientations that participants use with one another. These, we
suggest, are basic topics for a systematic and empirical social psychology to address.
History and Development of Discursive Social

Psychology
Charting the historical development of an intellectual field can be tricky and controver
sial because, as Potter (2010, p. 659) noted, such histories “tend to mix what philoso
phers of science distinguish as the ‘context of discovery’ (the contingent, ad hoc ways in
which a scientific field developed) from the ‘context of justification’ (the sorts of logically
(p. 298) organised arguments that are used to justify the place the scientific field has ar
rived at).” Edwards (2005b), Wetherell (2007), and Wooffitt (2005), for example, have all
offered rather different accounts of the development of a broader discursive psychology.
Discursive social psychology has a complex theoretical lineage that has selectively devel
oped strands of thinking from discourse analysis, rhetoric, sociology of science, eth
nomethodology, conversation analysis and post-structuralism (Potter & Edwards, 2001). It
is part of the broad and heterogeneous field of discourse analysis. Work in this area has
drawn on post-structuralist ideas, such as those from the Foucaultian tradition of Hen
riques, Hollway, Urwin, Venn and Walkerdine (1984), or on broader post-structuralist
thinking from Barthes, Derrida and others (e.g., Potter, Stringer & Wetherell, 1984). Of
particular significance has been the work of Wittgenstein and linguistic philosophy (see
Potter, 2001), as well as the subsequent philosophical respecification of psychology devel
oped by Harré (Harré & Gillett, 1994). Wetherell (2007) adds that discursive psychology
in particular was boosted by the social constructionist movement in social psychology
(Gergen, 1985; Shotter, 1993) and feminist psychology (Wilkinson & Kitzinger, 1995). Fi
nally, the sociology of scientific knowledge and the distinctive program of discourse ana
lytic work on science conducted by Gilbert and Mulkay (1984) and the early studies in
conversation analysis (e.g., Drew & Atkinson, 1979; Levinson, 1983) all helped to shape
Page 5 of 34

the contemporary nature of DSP. What unites all of these disparate strands of research is
an emphasis on the (a) careful empirical study of discourse, (b) ways in which discourse is
oriented to action, and (c) the way representations are built to support actions.
The present form of DSP emerged most clearly out of the form of discourse analysis out
lined in Potter and Wetherell’s (1987) Discourse and Social Psychology. Taking the publi
cation of that influential book as our starting point, we will sketch out the three main
strands of work that over the past 20 years have shaped the field of contemporary discur
sive social psychology. This is not a complete picture of the development of the field but
should illuminate some of the key issues that discourse research has engaged with.
Strand 1: Interviews and Repertoires
Starting in the mid-1980s, the major focus of discourse analytic work in psychology was
on identifying the different interpretative repertoires that are used to build social action
(most notably, Potter & Wetherell, 1987). An interpretative repertoire is a cluster of
terms, categories, and idioms that are closely conceptually organized. Repertoires are
typically assembled around a metaphor or vivid image. In most cases, interpretative
repertoires are identified by analyzing a set of open-ended interviews in which partici
pants address a set of different themes.
The repertoire notion is derived from Gilbert and Mulkay’s (1984) pioneering study of the
different repertoires that scientists use to construct their social world when they are writ
ing research papers and arguing with one another. It was further developed in Wetherell
and Potter (1992) in a major study of the way Päkehä (white) New Zealanders constructed
versions of social conflict and social organizations to legitimate particular versions of re
lations between groups. Much of the interest was in ideological questions of how the or
ganization of accounts and the resources used in those accounts could be used to under
stand the reproduction of broad patterns of inequality and privilege. Put simply, how did
white Europeans undermine Maori land claims and other grievances without appearing
self-interested or racist?
This strand of work was closely allied to and influenced by Billig’s (1996) rhetorical psy
chology and incorporated the central notion of ideological dilemmas (Billig et al., 1988),
which itself builds on the notion of interpretative repertoires from Potter and Wetherell
(1987). For example, Billig (1992) found in talk about the British royal family a web of ar
guments and assumptions that work to sustain the familiar social hierarchies and avoid
questioning privilege.
The notion of interpretative repertoires has been drawn on by many studies from across
the social sciences. It offered a picture of complex, historically developed organizations of
ideas that could be identified through research and yet were flexible enough to be re
worked within the contingencies of different concrete settings. This theorizing of the flex
ible requirements of practice offers some advantages over some neo-Foucaultian notions
of discourse that are more brittle and tectonic (Parker, 1992). Nevertheless, Wooffitt
(2005) has suggested that the notion still fails to fully accommodate the complexity of hu
Page 6 of 34

man conduct, and there are major questions as to whether the structuring of repertoires
is a consequence of preformed conceptual organizations or a by-product of the pragmatic
organization of practices (see Potter, 1996). Furthermore, the original repertoire (p. 299)
notion required a series of procedures and criteria for the reliable identification of reper
toires. Yet, many current studies offer only the vaguest idea of how the repertoires are
identified and how they relate to a corpus of data (Potter, 2009). There are important
points of principle here, illustrated in the influential exchange between Schegloff (1997)
and Wetherell (1998).
This type of work was largely based on the analysis of open-ended interviews, or group
discussions, which provided the ideal environment for generating the kinds of ideological
themes or interpretative repertoires that were a key topic of study. This has been a con
tinuing and productive theme in discourse research on such topics as gender and nation
alism (for many examples from social psychology, see Augoustinos, Tuffin, & Rapley,
1999; Condor, 2006; Reynolds & Wetherell, 2003). This work has been particularly effec
tive in tackling ideological questions that are not easily addressed by more mainstream
social cognition perspectives (Augoustinos, Walker, & Donaghue, 2006).
Strand 2: Naturalistic Data and Discursive Constructionism
A key area of difference between Potter and Wetherell’s (1987) conception of discourse
analysis and the body of work later to be identified as discursive psychology or DSP con
cerns the place of open-ended interviews in the generation of analytic materials. Much of
Potter and Wetherell’s (1987) discussion, and the majority of the very large body of subse
quent studies using interpretative repertoires that this work spawned, have used open-
ended interviews. In contrast, DSP in its modern form has almost completely abandoned
open-ended interviews as a technique for generating appropriate analytic materials. This
was partly due to profound problems with the production and analysis of open-ended in
terviews (Potter & Hepburn, 2005, 2012).
In the early 1990s, a new strand of research emerged that worked with records of natu
ralistic interactions such as conversations, newspaper reports, parliamentary debates,
news interviews, radio call-ins, and courtroom arguments. This body of work began to
look critically at how speakers could use descriptions of the world and psychological
states to perform specific actions within the interaction. One of the central themes devel
oped in discursive psychology (Edwards & Potter, 1992) was the close inferential relation
ship between versions of “reality” (things in “the world,” actions, events, history and so
on) and “mind” (things “in the head,” attitudes, dispositions, feelings, expectations and so
on). People construct versions of both in their talk and their texts, and they do so in the
service of action. For discursive psychology, this pervasive practical reasoning is an im
portant topic of study; it offers a new way of doing social psychology—hence DSP.
Whereas the earlier strand of work was, and often still is, referred to as discourse analy
sis, this strand of work was the first to explicitly call itself discursive psychology (Ed
wards & Potter, 1992, 1993). This type of discursive psychology centered on respecifying
Page 7 of 34

traditional psychological notions in relation to the function they perform in discourse

(Potter, 2009). Instead of trying to get at the internal states of participants, discursive
psychologists look at how participants draw on notions of psychological phenomena such
as emotion to accomplish an interactional goal. It engaged with major topics in traditional
social psychology but also generated major themes of its own.
Take Edwards’s (1995, 1997) analysis of “Connie and Jimmy” as an example. Within their
ostensibly factual descriptions of events during a counseling session, Connie and Jimmy
depict each other as “endemically jealous” and “flirtatious,” respectively (Edwards & Pot
ter, 2005, p. 245). By constructing each other as jealous or flirtatious, the speakers simul
taneously both provide conceptual support for their own account and discredit alternative
understandings. If Jimmy is always jealous, then he is overreacting to Connie’s behavior;
or, if Connie is being provocative toward other men, then Jimmy is justified in being un
happy with her behavior (Edwards, 1995). The example allows us to understand how Con
nie uses a psychological concept like “jealousy” not necessarily because that is how it “re
ally” happened (that is not of relevance here), but because it is interactionally useful for
her to employ that concept at that time to support her argument.
As it developed, DSP moved to a more explicit style of discursive constructionism focused

on texts and talk, with different analytic and epistemic consequences to the cognitive
form of constructionism found in Berger and Luckmann (1966) and other forms of social
construction (see Potter, 1996; Potter & Hepburn, 2008). The continuing momentum of
this work comes from its critical engagement with mainstream cognitive and social cogni
tive psychological work, shown in particular through studies that respecify core notions
such (p. 300) as memory, scripts, emotion, attribution, and perception in interactional
terms (Edwards, 1997; Hepburn, 2004; Stokoe & Hepburn, 2005).
Strand 3: Sequential Analysis
From the mid-1990s onward, discursive psychology began to draw more extensively on
the analytic principles of conversation analysis (see Schegloff, 2007; and papers in Sidnell
& Stivers, 2012). The sequential focus of conversation analysis provided valuable method
ological rigor to discursive psychology and offered a framework for a more nuanced and
precise understanding of the situated nature of discursive practices. For example, conver
sation analysis offers a way of approaching topics that psychologists would typically char
acterize in terms of a construct such as “attitude” in terms of situated practices of evalua
tion or assessment. Here, evaluations are structured events in talk; they are sequentially
organized within turn-taking and are the products of rather than the precursors to an in
teraction. The work of Pomerantz (1978, 1984, 1986) has been particularly insightful in
this area. For instance, assessments and subsequent (or second) assessments are struc
tured so as to minimize stated disagreement and maximize stated agreement between
speakers (Pomerantz, 1984). When expressing an assessment, one is therefore perform
ing an action, such as praising, insulting, complaining and so on. This action is itself
structured through the sequential organization of the talk. Research on evaluations in dis
cursive psychology reflects this emphasis, bringing to the fore the action orientation and
Page 8 of 34

sequential organization of evaluative expressions, such as the use of different types of

food evaluation to justify or account for particular courses of action (Potter, 1998; Wig
gins & Potter, 2003).
Discursive psychology has increasingly exploited the sophisticated understanding of se

quence, position, and turn design provided by conversation analysis. Indeed, there has
been a convergence of issues in both conversation analysis and DSP because both focus
on concerns with how shared knowledge is displayed and how intersubjectivity is estab
lished or contested (compare Edwards, 1999; Heritage & Raymond, 2005). This has be
come the fast developing field of mundane epistemics (Potter & Hepburn, 2008).
The convergence between discursive psychology and conversation analysis has been
highly productive and, in recent years, has produced a high volume of fine-grained, nu
anced work looking at how various psychological topics are produced through discourse.
For example, there has been an evolving concern with how categories are conversational
ly and sequentially occasioned (compare Antaki, 1998; Edwards, 1998; Potter &
Wetherell, 1987, chapter 6; Stokoe, 2009; Widdicombe & Wooffitt, 1995). This strand of
DSP is still engaged in a debate with cognitivism and its problems in different arenas
(e.g., Antaki, 2004; chapters in te Molder & Potter, 2005). However, DSP has started to
address new topics. There is a major interest in taking the administration of psychologi
cal methods as a topic in its own right, studying how particular interactional practices in
experiments, surveys, focus groups, and so on contribute to the methodical production of
psychological findings (e.g., Antaki, Houtkoop-Steenstra & Rapley, 2000; Maynard,
Houtkoop-Steenstra, Schaeffer & van der Zouwen, 2002; Puchta & Potter, 2002). There is
also a growing concern with considering how psychological matters become parts of insti
tutional practices such as therapy (Peräkylä, Antaki, Vehviläinen & Leudar, 2008), coun
seling (Kurri & Wahlström, 2001), mediation (Stokoe & Edwards, 2009), gender reassign
ment assessments (Speer & Parsons, 2006), peer evaluation (Cromdal, Tholander, &
Aronsson, 2007), and others.
Theoretical Principles of Discursive Social Psy

chology
The previous section offered a necessarily selective and brief history of three of the sig
nificant branches in the history and development of DSP out of discourse analysis. It is
important to stress that discourse research is a broad church and houses numerous dif
ferent approaches to research (see Tannen, Schiffrin & Hamilton [2001], Wood & Kroger
[2000], Edwards [2005b], and Wooffitt [2005], for longer treatments of the field from vari
ous perspectives). It is equally important to acknowledge the conversation between DSP
and mainstream psychology. The fundamental project of discursive social psychology in
volves a radical approach to and reevaluation of traditional topics of psychological en
quiry.
Page 9 of 34

Discursive social psychology is the application of ideas from discourse analysis to central
topics in social psychology. It is not a social psychology of language. Instead, it is an ap
proach to psychology that takes the action-orientated and reality-constructing features of
discourse as fundamental. Whereas the dominant social cognition paradigm gives a story
of behavior produced on the basis of information processing done on perceptual input
(see Augoustinos (p. 301) et al., 2006; Fiske & Taylor, 2008), DSP’s narrative revolves
around activities done through discourse as parts of situated practices (Edwards & Pot
ter, 1992). Whereas theory and method in social cognition presume an out-there reality
that provides input to cognitive operations, DSP focuses on the way both “reality” and
“mind” are constructed by people conceptually, in language, in the course of their execu
tion of various practical tasks (Edwards, 1997; Potter, 1996; Potter, Edwards, & Wetherell,
1993).
For theoretical, methodological and empirical reasons, discursive psychology takes dis
course to be central to social life. For example, most social activity involves or is directly
conducted through discourse. Furthermore, even where activity is
“nonverbal” (embodiment, physical actions, and their settings, etc.), its sense is often
best understood through participants’ discourse. Trying to short circuit this way to under
standing nonverbal materials is often a recipe for imposing the analysts own categories
on what is going on. Discourse is the prime currency of interaction, and if we are study
ing persons embedded in practices, then discourse will be central to that study. Discur
sive social psychology builds from three core observations about the nature of discourse
(see Potter, 2003; Potter & Edwards, 2001). These form the theoretical principles that un
derpin its research.
Principle 1: Discourse Is Action Oriented
Talk is action: our words do things. The idea of talk-as-action can be understood in a very
literal sense, as with conventional “speech acts.” For example, in the sentence “I now pro
nounce you man and wife,” the words themselves perform actual action; in this case, the
legal act of marrying two people (Austin, 1962; Searle, 1969). However, all discourse has
functions (e.g., greeting, requesting, explaining, complaining), so talk can be studied in
terms of what speakers are doing with their words rather than what their words reveal
about putative inner thoughts. Discursive social psychology uses the notion of action-ori
entation to emphasize that actions are pervasively being done even in ostensibly factual,
descriptive discourse, and to distance itself from a “speech act” approach that assumes
that some discrete set of words corresponds to a discrete act.
The key point is that discourse is studied for how action is done rather than treated as a
medium for access to a speaker’s internal mental furniture (e.g., intentions, preferences,
beliefs). This is a very different starting point from that of cognitive (and some social)
psychology. Holding discourse as the topic of study rather than just a route into the
speaker’s mind, the study of social interaction becomes an exploration into how events or
ideas come to be; whether that be how a group of friends successfully arrange to meet
Page 10 of 34

up, or how the British and American governments worked to present a legitimate case for
war in Iraq following the 9/11 attacks.
The corollary of discursive psychology’s focus on discourse is its respecification of cogni

tion. Instead of cognitive entities and processes being the principal analytic resource, as
they are in social cognition research, they are approached empirically as participants’
ways of talking. The focus is on the way cognitions are constructed in talk and how their
implications are oriented to. For example, rather than treating attitudes as inner entities
that drive behavior, for discursive psychologists, attitudes are evaluations that are stud
ied as part of discourse practices (Potter, 1998). Such an approach might consider the
way evaluations are organized interactionally, as in Pomerantz’s (1978) study of compli
ments; it might consider how attitudes are interactionally produced through social psy
chological methods (Myers, 1998; Puchta & Potter, 2002); or it might consider the way
negative evaluations of minority group members are turned from potentially accountable,
personally held attitudes into more “safely sayable” factual descriptions (e.g., Edwards,
1998; Potter & Wetherell, 1988; Wetherell & Potter, 1992).
Principle 2: Discourse Is Situated
Central to DSP is the idea that discourse is situated in three different but complementary
ways. First, actions are situated sequentially. That is, they are situated within the here
and now of unfolding conversation. They are located in time, orienting to what has just
happened and building an environment for what happens next. For example, when an in
vitation is issued, this sets up an ordered array of possible next actions, of which accept
ing or turning down are most relevant. It is not possible to simply ignore the invitation
without this being, potentially, hearable as the action of ignoring the invitation. Moreover,
when the recipient accepts or rejects an invitation, the recipient is locally displaying an
understanding that that is precisely what has been issued, so the turn-by-turn unfolding
of talk provides an ongoing check on understanding (Schegloff, 1992). The explication of
this order of (p. 302) interaction has been the central project of conversation analysis and
has highlighted an extraordinary level of specificity and organization (Schegloff, 2007).
Second, action is situated institutionally. Institutions often embody special identities that
are pervasively relevant—news interviewer, therapist, patient—such that actions will be
understood in relation to those identities. And they often involve collections of local inter
actional goals that all parties orient to (Drew & Heritage, 1992; Heritage & Clayman,
2010). These institutional goals are often themselves dependent on broader everyday
practices that are refined for the institutional setting (compare Edwards [2008] on “inten
tion” with Potter & Hepburn [2003] on “concern”). The specific analytic relevance here is
how psychological matters are introduced, constructed and made relevant to the setting’s
business (Edwards & Potter, 2001).
Third, action is situated rhetorically. Billig (1996) has emphasized the pervasive relevance
of rhetorical relations, even where there is an absence of explicit argument (e.g., he has
explicated the rhetorical underpinning of “opinion” discourse—Billig, 1989; see also My
Page 11 of 34

ers, 2004). Discursive social psychology highlights, for example, the way descriptions are
built to counter actual or potential alternatives and are organized in ways that manage
actual or possible attempts to undermine them (Potter, 1996). A major theme in DSP is
the way epistemic issues are managed using a wide range of conversational and rhetori
cal resources (Potter & Hepburn, 2008). This theme cuts right across the conventional so
cial psychological topics of memory, attribution, attitudes and persuasion.
Principle 3: Discourse Is Constructed and Constructive
Constructionism is a fundamental theme in DSP. There are two senses in which DSP is
constructionist. First, it studies the ways discourse itself is constructed. It is assembled
from a range of different resources with different degrees of structural organization.
Most fundamentally, these are words and grammatical structures, but also broader ele
ments such as categories, metaphors, idioms, rhetorical devices, descriptions, accounts,
stories and so on that are drawn on and built, in the course of interaction and in the per
formance of particular actions. For example, how is a description manufactured in a way
that presents something that has been done as orderly and unproblematic? People are ex
tremely well practiced and skilled builders of descriptions; they have spent a lifetime
learning how to do it. Part of the analytic work of DSP is to reveal the complex and deli
cate work that goes into this seemingly effortless building of discourse.
On the other hand, discourse is constructive in the sense that these assemblages of
words, repertoires and so on put together and stabilize versions of the world, of actions
and events, of mental life and furniture. For example, how does one party in a relation
ship counseling session construct a version that presents the breakdown of a long-term
relationship as primarily the responsibility of the other party, who might be the one most
in need of counseling and under most pressure to change (Edwards, 1995)? Similarly,
take for example a criminal court case: the defence and the prosecution lawyers work
with the same evidence presented to the jury but use it to build two different accounts of
events, one that condemns the defendant and one that exonerates him or her. Thus, lan
guage is used to build different versions of reality. Through this understanding, we can
see that speakers design their talk in order to accomplish specific goals in interaction.
Defining Features
Discursive social psychology falls within the broad church of discourse analysis. However,
unlike much of discourse analysis, which tended to situate itself within sociology or lin
guistics, DSP is not so much a branch of current social psychology as a discursive refor
mulation of the central problematics of social psychology. In this section, we briefly de
scribe two aspects that highlight what is distinctive about DSP.
Discourse Produced as Psychological

One thing that distinguishes the conversation analytic–influenced strand of DSP from ex
perimental work, on the one hand, and approaches based on qualitative interviews such
as the “free association narrative method” or “interpretative phenomenological analy
Page 12 of 34

sis” (Hollway & Jefferson, 2005; Smith & Osborn, 2008), on the other hand, is that it puts
participants’ own orientations at the heart of its analysis. These are the live orientations
that are practical parts of conduct unfolding in real time. This is a different order of phe
nomena to the post hoc constructions and formulations that appear in qualitative inter
views.
The focus on orientations in real time in natural interaction makes “psychological” mat
ters inescapable; such matters are a resource for participants as they coordinate their ac
tions, respond to (p. 303) expressions of liking and dislike, or as they manage incipient ac
tions such as invitations or requests. The world of discourse is psychologically imbued in
precisely the way real life is imbued. The organization of discourse with its lexical items,
prosody, categories, grammatical organizations, and plethora of different practices is
highly normative. Moreover, it unfolds in real time, with an extraordinary granularity in
which delays of less than a fifth of a second or minor changes in pitch contour can mark a
“psychological state” (Drew, 2005; Heritage, 2005).
As Edwards summarizes it:
One of DP’s tasks is to do something that psychology has not already done in any
systematic, empirical, and principled way, which is to examine how psychological
concepts (memory, thought, emotion etc.) are shaped for the functions they serve,
in and for the nexus of social practices in which we use language. (2012, p. 427)
Discursive social psychologists are focused on the way what counts as psychological is a
central concern of participants. People can construct their own and other’s dispositions,
assessments, and descriptions as subjective (psychological) or objective. For example, an
assessment of a minority group can be couched in the language of attitudes (“I am gener
ally positive about Polynesian Islanders”) or built as an objective feature of this social
group using a range of descriptive procedures (Potter & Wetherell, 1988).
As we noted in the opening paragraphs of this chapter, Edwards (2007) has distinguished
subject-side from object-side descriptions and has highlighted the way producing dis
course in either of these ways can be a central element in a range of practices. A person
can be described as having a legitimate complaint about something in the world (an ob
ject-side description) or as moaning and whining (a subject-side description that high
lights things wrong with the speaker rather than the world; Edwards, 2005a). “I love that
cake” is subject-side; “that cake is lovely” is object-side (Wiggins & Potter, 2003). This is a
member’s distinction that is regularly and methodologically deleted in the methods used
by experimental social psychologists. One of the features of the normative organization of
interaction is that it provides a baseline calibration for marking out psychological invest
ment.
Discursive psychology has a longstanding engagement with conversation analysis. One of

the novel features of Potter and Wetherell’s Discourse and Social Psychology more than
25 years ago was its use of the methods and findings of conversation analysis when re
working social psychological notions such as accounts, attitudes and social representa
Page 13 of 34

tions. In recent years, discursive psychological work has been able to draw on the in
creasing sophistication of work on repair, sequence organization and turn design as it
pursues its engagement with social psychology.
Despite the increasing synergy between the two fields, a distinctive characteristic of DSP
in its use of conversation analytic methods has been its focus on the role psychological
ideas and concepts play in everyday life. For example, Childs (2012) examines how the
mental state of “wanting” can be invoked in interaction. Through a fine-grained sequen
tial analysis, guided by the analytic canons of conversation analysis, she demonstrates
how formulations with the character “I don’t X, I want Y” enable speaker to build “de
sires” in ways that are appropriate to sequentially unfolding social interaction. Specifical
ly, she shows how they are able to “formulate an alternative sense of agency which under
mines the preceding turn and shifts the trajectory of the ongoing sequence” (Childs,
2012, p. 181). Thus, instead of focusing on it as an inner (social) psychological state,
wanting is shown to be systematically and practically deployed in a particular interaction
al sequential context in service to a specific interactional goal.
Noncognitivist
A growing body of work within discursive psychology has revealed the active, construc
tive, and rhetorical use to which speakers can (and do) employ psychological concepts
and objects. This work has reworked traditional topics such as causal attribution (Antaki,
1994; Edwards & Potter, 1992, 1993), prejudice (Edwards, 2003; Gill, 1993; Speer & Pot
ter, 2000; Wetherell & Potter, 1992), identity (Antaki, 1998; Edwards, 1998; Widdicombe
& Wooffitt, 1995), scripts (Edwards, 1994, 1997), violence and aggression (Auburn, Lea &
Drake, 1999; Hepburn, 2000; McKinlay & Dunnett, 1998), social influence (Hepburn &
Potter, 2010), and role play (Stokoe, 2011), and it has brought to the fore new topics such
as the relation between interaction and institutions (Edwards, 1995; te Molder, 1999) and
the construction and establishment of factual accounts (MacMillan & Edwards, 1999; Pot
ter, 1996; Wooffitt, 1992). More recently, classical psychological problems (p. 304) of
shared knowledge—who knows what and how is knowledge shared—have been tackled in
subtle ways (see papers in Stivers, Mondada, & Steensig, 2011). Discursive psychologists
study psychological matters in their home environment of live, unconstrained interaction,
in which the parties have a genuine and practical stake in outcomes (Edwards, 2006a).
The majority of contemporary approaches to the study of human conduct treat it as ulti
mately dependent on putative individual entities such as beliefs, attitudes, and knowl
edge. Approaches as varied as social cognition (Fiske & Taylor, 2008) and interpretative
phenomenological analysis (Smith & Osborn, 2008) have adopted some version of this
cognitivism. Where they work with discourse, they look through it in an attempt to access
the cognitive objects being described. Discursive psychologists start with discourse prac
tices—that is, people interacting with one another, in mundane and institutional settings
—and they bracket off issues of cognition. It is not that discursive psychologists do not
consider thinking, cognition, mind or feelings, but this is not something they start with or
Page 14 of 34

see as the causal underpinning of social behavior. Rather these things become a major
topic of analysis in terms of the orientations and constructions of participants.
By starting with discourse, discursive psychologists prioritize the epistemic over the onto
logical. Trying to read through participants’ talk to unitary versions of events in the world
or to singular mental events, structures, or processes draws analysis away from the busi
ness of talk and how the fashioning of versions is bound up with that business. The focus
on talk as oriented to action binds discursive psychology with conversation analysis and
some post-structuralist approaches (for more on this, see Hepburn [2003], Wetherell &
Potter [1992] and Wooffitt [2005]). At the same time, that focus separates discursive psy
chology from approaches such as interpretative phenomenological analysis, social repre
sentations, and much of experimental social cognition. However, it is important to note
that discursive psychology is not a program that suggests that social phenomena do not
have objective reality (Hammersley, 2003); to deny such things would be as realistic a
move as endorsing them (Edwards, Ashmore & Potter, 1995). Rather, discursive psycholo
gists consider the role of “phenomena” in terms of the different descriptions, glosses, cat
egories and orientations offered by social actors (Potter, 2003, 2009).
Methodological and Analytic Principles

The theoretical principles outlined earlier are central to the project of discursive psychol
ogy and they are closely aligned with the methodological practices that have been devel
oped. They support a research style that is a radical departure from the hypothetico-de
ductive method that dominates mainstream psychological and social psychological en
quiry. Crucially, the focus on sequence and context has led to the development of a radi
cally empirical approach to research, which is an alternative to the factors and variables
multivariate model underlying much experimental work. In this section, we sketch out the
key analytic and methodological principles of discursive psychology and show how they
are shaped and governed by the theoretical principles on which this type of research is
founded.
Discursive social psychology starts from the view that the careful and detailed analysis of
materials is central to making claims and developing theory. An awareness of the philoso
phy and sociology of scientific knowledge leads discursive psychologists to being cautious
about the separation of data, theory, and method (Chalmers, 1992; Woolgar, 1988). For
discursive psychologists, data are central. Ironically, given the emphasis on science and
empiricism in more orthodox cognitive social psychology, many of the insights of DSP
have come from its willingness to be open to careful records of actual people actually liv
ing their lives. Elsewhere we have noted the way the image of qualitative work as a pre
liminary stage to a full scientific social psychology is both a fiction in practice and way of
avoiding taking seriously problems of description and analysis (Potter, 2012). How this
engagement with records of interaction is accomplished will become clearer as we out
line the methodological steps involved in discursive psychological research.
Page 15 of 34

Research Design
Discursive social psychology is interested in how people use discourse to do things (e.g.,
persuade, plan, explain, argue, make decisions, show emotion, display an attitudinal
stance toward something, etc.) as well as the various normative resources that enable
those things to be done. Therefore the essential ingredient in discursive research is the
collection of interactional data. Research begins with some materials to focus on and an
interest in what might be happening in the data. Discursive research does not start with
hypotheses or use data to test them. Testing and validation is central to its analytic
(p. 305) practice but conceived very differently from orthodox experimental or survey ap
proaches.
Discursive social psychology treats social life, organized and produced through discourse,
as normative and rhetorical. However, the organization of social discourse is not deter
ministic in the sense that variables can be manipulated to produce predictable outcomes.
Discourse cannot be reduced down to the interplay of specific factors designed to pro
duce a regular pattern of outcomes. The norms of social life do not work as templates
that govern interaction. Instead, they are participant’s resources for action and under
standing, for making life accountable, describable and sanctionable (Garfinkel, 1967).
For example, greetings are typically responded to with a return greeting (Schegloff &
Sacks, 1973). Nonetheless, failure to return a greeting is not an occasion to see the norm
as somehow refuted. It the basis for a potential range of inferences about the person and
context; are they rude, hard of hearing, sulking or shy (Heritage, 1988)? Consequently,
the factors and outcomes model that underpins much social cognition research is not suit
able for this type of enquiry. Discursive social psychology projects are rarely productive if
they start from predefined research questions, particularly because those questions are
often implicitly formulated using the factors and variables assumptions of more orthodox
work. That is not to say that it may not be useful to identify broad areas of interest, such
as “how are notions of remembering invoked during arguments?” More often, re
searchers are interested in a particular interactional setting (e.g., police interviews, fami
ly mealtimes, phone calls to emergency response services) and aim to find out how psy
chological matters are produced and used in these settings.
Materials
Discourse is situated. It happens in the moment, shaped by what came before it and shap
ing what comes next (Heritage, 1984). Therefore discursive psychological research needs
to capture materials within the context in which they occur. Discursive psychologists pre
fer to analyze “naturalistic” rather than “got-up” materials. This is not a commitment to
an unsustainable philosophy of a natural world free of observer influence. Rather, it is a
preference, grounded in DSP’s conception of how discourse works, for examining records
of people living their lives, telling what happened, arguing about relationships, answering
parliamentary questions, and so on, instead of answering researcher’s questions, cooper
Page 16 of 34

ating with experimenters’ requirements, and responding to researcher’s textual vi

gnettes.
Data for discursive psychologists tend to include conversations, arguments, talk in work
settings, professional client interaction, the various situations where interaction is medi
ated and supported by technology (phones, visual displays, instruments etc.), and any oc
casion when people are doing things involving some form of interaction. Occasionally, dis
cursive psychologists will work with open-ended interviews, but these will be treated as
interactional events rather than as places where participants’ views can be excavated
(e.g., Edwards, 1998; Myers, 2004).
The focus on naturalistic materials starts to become inevitable once the importance is ful
ly recognized of discourse being situated, action-oriented and constructed. It is also a re
flection of what has become technically and analytically possible. Given that such rich
materials are increasingly tractable and can be successfully recorded, digitized, scanned,
transcribed, and rigorously studied in the wake of several decades of research, why do
anything else?
Data Collection
In terms of data collection, the main aim is to develop an archive of records of interaction
in the setting under study. There are no hard and fast rules for the size of such a collec
tion. Even small amounts of material can provide the basis for useful research, but the
more material there is and the more appropriate the sampling, the more questions will
become analytically tractable and more confidence can be placed in the research conclu
sions. In short, the better your data reflect the environment you want to study, the more
scope you will have for coming up with useful and relevant research questions, finding
enough material for analysis, checking the analysis thoroughly, and generally doing bet
ter discursive research.
The original recordings are an integral part of the research process. They are transcribed
to help facilitate analysis, but they are played again and again right through the analytic
process. As technology develops, recordings are increasingly being included in online
publications and presented at conferences as part of the dissemination of results. It is
worth spending the time and effort required to get the best data possible. There are sev
eral aspects to recording that need to be considered in advance:
Quality. The quality of recording has a powerful effect on the time taken in transcrip
tion and (p. 306) analysis. Time and resources devoted to getting high-quality record
ings will pay off handsomely when it comes to transcribing the recordings and working
with them in data sessions. Hours are wasted relistening to a key piece of talk against
a loud recording hum or using grainy video footage to work out whether a child has
taken a mouthful of food or just put the fork near her mouth.
Audio/video. If embodied activities are available to participants, then they are certain
to be a live part of the interaction. Thus, it is important to have video records of face-
to-face interaction (or non–face-to-face interaction that is technologically mediated
Page 17 of 34

with a visual modality). High-quality digital video is inexpensive, simple, and relatively
easy to manipulate, so this is not an insurmountable problem. A good rule of thumb is
that if the participants can see each other, the researcher needs to be able to see
them, too.
Quantity. Once you’ve got everything in place to record, the actual recording process
is almost always simpler and easier than analyzing and transcribing them. Because the
collection process is not overly arduous, it is worth collecting more recordings than
planned. Digital recordings can be easily stored, and they provide an important re
source for future research. Having an archive of interaction data allows cross-compar
isons to be made with similar phenomena in different settings and can greatly enhance
your work. As a cautionary note, it is important not to frivolously collect data you have
no intention of using. That is not ethical. The point we wish to make about quantity is
that you need to make sure you have enough; having more than enough is infinitely
preferable to not having enough. As a researcher, you always need to balance your
need for data with the level of imposition you place on the participants. Asking families
to video tape their normal mealtimes at home is considerably less arduous than asking
them to come into a lab to eat their dinner every evening. Therefore, you can reason
ably collect data from more meals. But if your research interest is in mealtimes, don’t
ask them to film the children doing their homework.
Who collects. A characteristic feature of contemporary discursive psychology is that
participants often collect the data themselves. This is designed to minimize the reactiv
ity generated by extended researcher involvement and allows the participants to man
age ethical issues in a way that suits them best. For example, in our recent family
mealtimes project, we left the video cameras with the families for as long as it took
them to record 10–15 meals. They quickly got used to the small video camera and
didn’t react much to it at all. If something happened that they didn’t want to submit for
analysis, they could delete the meal and record a later one.
Data storage and management. Although digital data takes up much less physical
space than questionnaires or tape recordings, it can consume a large amount of memo
ry (particularly if you use video). Make sure you have enough disc space to both store
and work with the data. It is important that your storage of original recordings is se
cure and protected, particularly if the data need to be anonymized before it is shared.
Having a coherent filing system is also necessary to help you keep track of the data
once you start selecting extracts for study, adding transcripts and analytic notations,
and sharing extracts with collaborators. Thinking about these issues in advance can
help you avoid getting in a catastrophic tangle half way through a project.
Transcription
Once the data have been collected, they need to be transcribed. Transcription is an im
portant step in the research process. Not only does it convert the data into a format that
is easy to search through and present in written documents, it is also the first step in the
analysis. The close attention required to produce an accurate transcript serves to famil
Page 18 of 34

iarize the researcher with the intricacies of the data and begins to reveal features of ana
lytic interest.
Discursive psychologists work continuously with both the original audio or video record
ings and the transcript. It is no longer the case that after transcription the recordings are
put into storage. Nevertheless, the transcript is an essential element in the research. It is
common to use two forms of transcript. A basic first-pass transcript is often generated by
a transcription service. This has just the words (not broken up by the colons, arrows, etc.,
that capture features of delivery) rendered as effectively as the service can hear them.
This kind of transcript allows the researcher to quickly go through a stretch of interaction
and get an overall feel for what is there. This can be a particularly important shortcut
when there are many hours of recordings. It is also searchable, allowing one to sift
through an entire set of materials very quickly for particular phenomena that can be iden
tified through individual lexical items or transcriber descriptions; for example, finding all
occasions where participants refer to wanting or needing something (Childs, 2012).
Discursive psychology requires an attention to the detail of interaction. The prin

(p. 307)
ciple that discourse is situated means that how something is said is just as important as
what is said (or not said). Harvey Sacks (1992) suggested that none of the detail of inter
action, whether it is pauses and repairs, the selection of particular words, or the place
ment of interruption and overlaps, should be assumed a priori to be irrelevant to interac
tion. For example, sometimes a sniff is just a sniff, the consequence of having a runny
nose; yet a sniff, in the right place, with the right kind of in-breath, could also do some
thing else, such as displaying indirect disagreement, or it might be a precursor to a more
fully fledged bout of upset (Hepburn, 2004). Therefore, the first-pass transcript is insuffi
cient as an analytic tool.
The second form of transcription is an attempt to capture on the page features of the de
livery of talk that participants treat as relevant for understanding the activities that are
taking place. The standard system used in discursive psychology was developed by Gail
Jefferson within conversation analysis and is specifically designed to support analysis of
interaction (Hepburn & Bolden, 2012; Jefferson, 2004). It was designed to be (relatively)
easy to learn and simple to produce, using standard character sets. It encodes features
such as overlaps and pauses, volume and emphasis, features of prosody such as rising
and falling intonation, and features of the speed of delivery.
Producing a Jefferson transcript is extremely labor-intensive. The ratio of record time to

transcription time can be anything above 1:20, with key factors being the quality of the
recording, the complexity of the interaction, and whether there are nonvocal elements
that need to be represented. It also takes time to learn to do quality transcription. It is
necessary to both understand the roles of the different symbols and learn to apply them
consistently. Because of the time investment required to produce quality transcripts,
there are rarely resources for completely transcribing a full set of recordings. Various cri
teria can be used to decide what to transcribe and in what order. Transcription is a
Page 19 of 34

process that continues throughout the analysis phase of research as research questions
are developed and refined with each new discovery.
Analysis
Analysis starts during transcription as the researcher begins to get to know the data and
identify potential phenomena of interest. It is at this stage that research questions begin
to emerge. Rather than posing a strict question of the type associated with hypothesis
testing and the factor–outcomes model, here the focus is often on attempting to explicate
the workings of some kind of social practice that is operating in the setting, perhaps with
the ultimate aim of making broader sense of the setting as a whole. This often means that
questions are continually refined in the course of a program of work and a study within
that program. One of the benefits of working with naturalistic materials is that they
present their own challenges that lead to novel questions. They often feature actions or
occurrences that are unexpected or not easily understood with the repertoire of explana
tory concepts available in contemporary psychology. This can provide an exciting start
point for analytic work.
A common practice in the discursive community is to use different levels of engagement

with the materials to generate questions. A key part of this method often includes data
sessions with analytically minded colleagues. Data sessions are a highly collegial activity
in which a group of colleagues study a short extract of data and offer analytic observa
tions and suggestions for further enquiry. As a researcher, this can be an extremely useful
exercise in helping to refine and develop the analytic focus.
There is no established sequence of actions that will lead to a good analysis using discur
sive psychology. However, the close methodological ties between discursive psychology
and conversation analysis mean that some of the guidelines provided for approaching
conversation analytic work are highly relevant and can prove to be a useful entry point
for discursive psychology too (see methodological chapters in Sidnell & Stivers, 2012).
Analysis can proceed based on a single case, and a detailed account of a single instance
can be worked up and published (e.g., Toerien & Kitzinger, 2007). However, analysis more
typically involves building a collection. This is a gradual process, subject to continual re
finement as you learn more about the data and discover things that interest you. Collec
tions are a feature of conversation analytic research that can also prove fruitful for dis
cursive psychological enquiry. A collection, in conversation analytic terms, is a mecha
nism to gather together a set of single cases that share common attributes (e.g., se
quences where a participant uses the modal verb would/wouldn’t to explain his or her be
havior; Edwards, 2006b). A collection is slowly built up following multiple single-case
analyses, in which each next case demonstrates the systematic commonalities (p. 308) that
exist across participants and contexts. This is an example of the “constant comparative
method,” in which each subsequent example “tests out” the hypothesis of the previous
one, leading to a continual refinement of the analysis (Silverman, 2001, p. 238). It there
fore offers an ongoing coherence check as the analysis progresses.
Page 20 of 34

In a recent study looking at children’s responses to being told what to do by their par
ents, Kent started collecting examples of children complying with their parents’ demands.
Extract 1 is an example of a Dad asking Lucy to eat nicely (line 6), and Lucy immediately
doing what she had been told to do (lines 8–9).
As the collection grew, it became clear that not all forms of compliance looked the same.
On closer analysis, it became clear that some of the extracts involved a more complicated
series action beyond just doing what one was told to do. The next stage was thus to build
a new subcollection of instances when children made it look like they were about to com
ply but then said something that defied or resisted the command.
Page 21 of 34

Extract 2 is from a different mealtime recording of the same family as Extract 1.

(p. 309)
Here, Dad also tells Lucy to eat nicely (lines 3–4). Unlike in Extract 1, where compliance
happens straight away, here it is only on line 21 that Lucy actually eats something and a
lot has happened in between. By line 21, you’d be hard pushed to prove that Lucy was
eating because Dad told her to rather than just because she felt like it.
From this subcollection of extracts a practice of incipient compliance (actions that move
toward compliance but do not in themselves constitute compliance) was identified. Chil
dren recurrently used incipient compliance to stall for time while they worked to regain
autonomy over their own behavior. Telling someone to do something is an invasive social
Page 22 of 34

action that threatens the recipient’s autonomy and self-control. Refusing to do what your
parent has told you is a risky action and typically leads to conflict and arguments (Kent,
2012b). Ultimately, parents often end up forcing children to do what they had been told to
do (Craven & Potter, 2010). Neither option is entirely satisfactory for children. Incipient
compliance was sometimes a useful technique that children could use to take back con
trol over their actions without actually defying their parents or refusing to do what they
were told (Kent, 2012a). This study is an example of how building and refining collections
can help in the early stages of analysis.
Validation
The notion of validity in DSP is different from that found in much of mainstream psycholo
gy. It is built more systematically into basic research design—the choice and presentation
of naturalistic materials in something close to their raw form, for instance—rather than
arising as a worry about extending claims and findings from a research domain into rele
vant arenas of everyday life. In DSP, analysis is made accountable to the detail of empiri
cal materials, and these are presented in a form that allows readers to make their own
checks and judgments. This form of validation contrasts with much traditional experimen
tal and content analytic work in which it is rare for anything close to “raw data” to be in
cluded or for more than one or two illustrative codings to be provided. It also permits an
accumulation of empirical data and analytic studies against which new findings can be
compared for their coherence. For example, work on fact construction builds on the in
sights about accountability from earlier studies, and its success provides a further confir
mation of the validity of those studies (Edwards & Potter, 1993).
Two specific principles of conversation analysis are useful in validating analytic claims:
deviant case analysis (checking claims against potential counter cases) and the proof pro
cedure (basing the analysis of a turn at talk on how the participants themselves treat it, in
next turns; see Heritage, 1995; Schegloff, 1992). Both principles are illustrated in studies
of television and radio news interviews, in which participants routinely avoid treating in
terviewers as accountable for views expressed in questions. That normative pattern is
supported rather than refuted by studying deviant cases in which interviewees treat their
interviewer as expressing personal views, whereupon considerable interactional trouble
ensues in subsequent turns (Heritage & Greatbatch, 1991; Potter, 1996).
Conclusion
For much of the past 100 years, psychology has developed as a hypothetico–deductive sci
ence that has conceptualized the world in terms of the effects and interactions of vari
ables on one another that can best be assessed using experiments analyzed using multi
variate statistics. This methodological apparatus has been combined with a cognitivist
form of explanation for which the causes of human action are seen to lie within individu
als. In some ways, this has been a hugely impressive and successful enterprise. Yet this
has had a number of unintended consequences that restrict its approach to human action.
Page 23 of 34

First, the search for general relationships that underlie behavior has the consequence of
moving research away from the specifics of human action. Action is typically modeled, re
stricted, or reported and transformed into the kind of counts that are amenable to multi
variate analysis. On the extremely rare occasion that records of actual interaction in nat
ural settings are used, it is quickly transformed into counts (using content analysis, say).
Second, this search for general relationships combined with the need for simple con
trolled designs means that little attention has been paid to the nature and organization of
the rich local and institutional settings in which human conduct invariably takes place.
Third, the hypothetico–deductive approach has led researchers away from careful de
scriptive studies in favor of studies that start with some kind of relationship or model to
be tested. This combines with the legacy of the distinction drawn between competence
and performance that has become (p. 310) foundational in cognitivist psychology and that
treats performance data as enormously messy and something to be bypassed by focusing,
via hypothetical models, directly on competence.
In contrast to this, discursive psychology starts with the concrete particulars of human
action recorded in specific settings with minimal researcher interference. In many ways,
it is a classically empiricist enterprise. Its analytic approach is focused on the way prac
tices are built in real time and how their organization and intelligibility depends on the
normative organization of talk. Psychological matters come into discursive psychological
study through their emergence as issues that are relevant for participants. Instead of at
tempting to capture underlying competence, it is focused on how psychological matters
are public and intelligible.
Future Directions
• Discursive Social Psychology is a radical, empirical approach to the study of social
behavior. It is continuing to develop as a discipline. Looking ahead we can see several
emerging areas for research to which DSP is well placed to contribute.
• Discursive social psychology can be applied to any setting in which interaction takes
place. This has historically encompassed phone conversations, face-to-face interac
tions, and broadcast video and radio media, as well as focusing on how written texts
are organized to build actions. However, the increasing use of technology to mediate
social interaction is opening up new lines of enquiry for researchers (see Meredith &
Potter (2013); Lamerichs & te Molder (2003) for full discussions). Internet forums and
chat rooms have already produced a significant body of DSP work on a wide range of
topics including managing accountability (Antaki, Ardévol, Núñez & Vayreda, 2005;
Sneijder & te Molder, 2004), having an “authentic” identity (Horne & Wiggins, 2009;
Sneijder & te Molder, 2009), attributing responsibility and blame (Sneijder & te Mold
er, 2005), and unsolicited advice (Vayreda & Antaki, 2009). An interest in computer-
mediated interaction is likely to continue expanding as the use of technologies for com
munication becomes more widespread. We expect to see more work looking at elec
tronic communications settings in the future.
Page 24 of 34

• Discursive social psychology will build on advances in the transcription of prosody

and developments in the understanding of sequence organization to throw further light
on the way social psychological issues are organized interactionally.
• Discursive social psychology has started to reformulate ways in which academic re
searchers and professionals can work together (Puchta & Potter, 2004; Stokoe, 2011).
This focus on successful impact is likely to be an important area of development in the
near future.
• Discursive social psychology has been making important contributions in areas of
emotion and embodiment that are often seen as particularly difficult for discourse-
based approaches (see, e.g., contrasting treatments in Hepburn & Potter [2012] and
Wetherell [2012]). Again, the strongly empirical focus has brought dividends in areas
that have often become theory heavy.
• There is likely to be a further methodological debate between DSP and approaches
that are based largely on the analysis of qualitative interviews such as free associative
narrative methods, much neo-Foucaultian work, and a now significant body of work in
interpretative phenomenological analysis. Discursive social psychology offers an ana
lytic approach that highlights the way findings in these perspectives depend on largely
implicit methodological practices used with open-ended interviews (Potter & Hepburn,
2012).
References
Antaki, C. (1994). Explaining and arguing: The social organization of accounts. London:
Sage.
Antaki, C. (1998). Identity-ascriptions in their time and place: “Fagin” and “the terminally
dim.” In C. Antaki & S. Widdicombe (Eds.), Identities in talk (pp. 71–86). London: Sage.
Antaki, C. (2004). Reading minds or dealing with psychological implications? Theory and
Antaki, C., Ardévol, E., Núñez, F., & Vayreda, A. (2005). “For she who knows who she is”:
Managing accountability in online forum messages. Journal of Computer-Mediated Com
munication, 11(1), http://jcmc.indiana.edu/vol11/issue1/antaki.html.
Antaki, C., Houtkoop-Steenstra, H., & Rapley, M. (2000). “Brilliant. Next question…”:
High-grade assessment sequences in the completion of interactional units. Research on
Language and Social Interaction, 33, 235–262.
Auburn, T. (2005). Narrative reflexivity as a repair device for discounting “cognitive dis
tortions” in sex offender treatment, Discourse & Society, 16(5), 697–718
Auburn, T., Lea, S., & Drake, S. (1999). “It’s your opportunity to be truthful”: Disbelief,
mundane reasoning and the investigation of crime. In C. Willig (Ed.), Applied discourse
Page 25 of 34

analysis: Social and psychological investigations (pp. 44–65). Buckingham: Open Universi
ty Press.
Augoustinos, M., Tuffin, K., & Rapley, M. (1999). Genocide or a failure to gel? Racism, his
tory and nationalism in Australian talk. Discourse and Society, 10, 351–378.
Augoustinos, M., Walker, I., & Donaghue, N. (2006). Social cognition: An integrated intro
duction. London: Sage.
(p. 311) Austin, J. L. (1962). How to do things with words. Oxford, UK: Clarendon Press.
Berger, P. L., & Luckmann, T. (1966). The social construction of reality. Garden City, NY:
Anchor.
Billig, M. (1989). The argumentative nature of holding strong views: A case study. Euro
pean Journal of Social Psychology, 19, 203–223.
Billig, M. (1992). Talking of the royal family. London: Routledge.
Billig, M. (1996). Arguing and thinking: A rhetorical approach to social psychology, 2nd
ed. Cambridge, UK: Cambridge University Press.
Billig, M., Condor, S., Edwards, D., Gane, M., Middleton, D. J., & Radley, A. R. (1988). Ide
ological dilemmas: A social psychology of everyday thinking. London: Sage.
Bolden, G. B., & Robinson, J. D. (2011). Soliciting accounts with Why-interrogatives in

conversation, Journal of Communication, 61, 94–119.
Chalmers, A. (1992). What is this thing called science?: An assessment of the nature and
status of science and its methods, 2nd ed. Milton Keynes: Open University Press.
Childs, C (2012). “I’m not X, I just want Y”: Formulating “wants” in interaction. Discourse
Studies, 14(2), 181–196.
Condor, S. (2006). Temporality and collectivity: Diversity, history and the rhetorical con
struction of national entitativity. British Journal of Social Psychology, 45, 657–682.
Craven, A., & Potter, J. (2010). Directives: Entitlement and contingency in action. Dis
course Studies, 12(4), 419–442.
Cromdal, J., Tholander, M., & Aronsson, K. (2007). “Doing reluctance”: Managing delivery
of assessments in peer evaluation. In A. Hepburn & S. Wiggins (Eds.), Discursive research
in practice: New approaches to psychology and interaction (pp. 203–223). London: Sage.
Drew (2005). Is confusion a state of mind? In H. te Molder & J. Potter (Eds.), Conversa
tion and cognition (pp. 161–183). Cambridge: Cambridge University Press.
Drew, P., & Atkinson, J. M. (1979). Order in court: Verbal interaction in judicial settings.
London: Macmillan.
Page 26 of 34

Drew, P., & Heritage, J. C. (1992). Analyzing Talk at Work: An Introduction. In P. Drew, & J.
C. Heritage (ed.), Talk at Work. Cambridge: Cambridge University Press, pp. 3–65.
Edwards, D. (1994). Script formulations: A study of event descriptions in conversation.

Edwards, D. (1995). Two to tango: Script formulations, disposition, and rhetorical symme
try in relationship troubles talk. Research on Language and Social Interaction, 28(4),
319–350.
Edwards, D. (1997). Discourse and cognition. London: Sage.
Edwards, D. (1998). The relevant thing about her: Social identity categories in use. In C.
Antaki & S. Widdicombe (Eds.), Identities in talk (pp. 15–33). London: Sage.
Edwards, D. (1999). Shared knowledge as a performative and rhetorical category. In J.

Verschueren (Ed.), Pragmatics in 1998: Selected papers from the sixth International Prag
matics Conference (vol. 2, pp. 130–141). Antwerp: International Pragmatics Association.
Edwards, D. (2003). Analysing racial discourse: The discursive psychology of mind-world

relationships. In H. van den Berg, H. Houtcoup-Steenstra, & M. Wetherell (Eds.), Analyz
ing interviews on racial issues: Multidisciplinary approaches to interview discourse (pp.
31–48). Cambridge: Cambridge University Press.
Edwards, D. (2005a). Moaning, whinging and laughing: The subjective side of complaints.
Discourse Studies, 7(1), 5–29.
Edwards, D. (2005b). Discursive psychology. In K. L. Fitch & R. E. Sanders (Eds.), Hand

book of language and social interaction (pp. 257–273). Mahwah, NJ: Erlbaum.
Edwards, D. (2006a). Discourse, cognition and social practices: The rich surface of lan
guage and social interaction. Discourse Studies, 8(1), 41–49
Edwards, D. (2006b). Facts, norms and dispositions: Practical uses of the modal would in
police interrogations. Discourse Studies, 8 (4), 475–501.
Edwards, D. (2007). Managing subjectivity in talk. In A. Hepburn & S. Wiggins (Eds.), Dis
cursive research in practice: New approaches to psychology and interaction (pp. 31–49).
Cambridge, UK: Cambridge University Press.
Edwards, D. (2008). Intentionality and mens rea in police interrogations: The production
of actions as crimes, Intercultural Pragmatics, 5, 177–199.
Edwards, D. (2012). Discursive and scientific psychology. British Journal of Social Psychol
ogy, 51, 425–435.
Edwards, D., Ashmore, M., & Potter, J. (1995). Death and furniture: The rhetoric, politics
and theology of bottom-line arguments against relativism. History of the Human Sciences
8(2), 25–49.
Page 27 of 34

Edwards, D., & Potter, J. (1992). Discursive psychology. London: Sage.
Edwards, D., & Potter, J. (1993). Language and causation: A discursive action model of de
scription and attribution. Psychological Review, 100, 23–41.
Edwards, D., & Potter, J. (2001). Discursive psychology. In A. W. McHoul & M. Rapley
(Eds.), How to analyse talk in institutional settings: A casebook of methods (pp. 12–24).
London: Continuum International.
Edwards, D., & Potter, J. (2005). Discursive psychology, mental states and descriptions. In
H. te Molder & J. Potter (Eds.), Conversation and cognition (pp. 241–259). Cambridge:
Fiske, S. T., & Taylor, S. E. (2008). Social cognition: From brains to culture. New York:
McGraw-Hill.
Garfinkel, H. (1967). Studies in Ethnomethodology. Englewood Cliffs, New Jersey: Pren

tice Hall.
Gergen, K. (1985). The social constructionist movement in modern psychology. American

Psychologist, 40, 266–275.
Gilbert, G. N., & Mulkay, M. (1984). Opening Pandora’s box: A sociological analysis of sci
entists’ discourse. Cambridge: Cambridge University Press.
Gill, R. (1993). Justifying injustice: Broadcasters’ accounts on inequality in radio. In E.

Burman & I. Parker (Eds.), Discourse analytic research: Repertoires and readings of texts
in action (pp. 75–93). London: Routledge.
Hammersley, M. (2003). Conversation analysis and discourse analysis: Methods or para

digms, Discourse & Society 14(6), 751–781.
Harré, R., & Gillett, G. (1994). The discursive mind. London: Sage.
Henriques, J., Hollway, W., Urwin, C., Venn, C., & Walkerdine, V. (1984). Changing the sub
ject: Psychology, social regulation and subjectivity. London: Methuen.
Hepburn, A. (2000). Power lines: Derrida, discursive psychology and the management of
accusations of school bullying, British Journal of Social Psychology, 39(4), 605–628.
Hepburn, A. (2003). An introduction to critical social psychology. London: Sage.
Hepburn, A. (2004). Crying: Notes on description, transcription and interaction. Research

on Language and Social Interaction, 37, 251–290.
Hepburn, A., & Bolden, G. B. (2012). The conversation analytic approach to tran
(p. 312)
scription. In J. Sidnell & T. Stivers (Eds.), The handbook of conversation analysis (pp. 57–
77). Oxford: Blackwell.
Page 28 of 34

Hepburn, A., & Potter, J. (2010). Interrogating tears: Some uses of “tag questions” in a
child protection helpline. In A. F. Freed & S. Ehrlich (Eds.), “Why do you ask?”: The func
tion of questions in institutional discourse (pp. 69–86). Oxford: Oxford University Press.
Hepburn, A., & Potter, J. (2011). Threats: Power, family mealtimes and social influence,
British Journal of Social Psychology, 50, 99–120.
Hepburn, A., & Potter, J. (2012). Crying and crying responses. In A. Peräkylä & M. -L. Sor
jonen (Eds.), Emotion in interaction (pp. 194–210). Oxford: Oxford University Press.
Heritage, J. C. (1984). Garfinkel and Ethnomethodology. Englewood Cliffs, NJ: Prentice

Hall.
Heritage, J. C. (1988). Explanations as accounts: A conversation analytic perspective. In

C. Antaki (Ed.), Analyzing lay explanation: A case book of methods (pp. 127–144). London,
Sage.
Heritage, J. C. (1995). Conversation analysis: Methodological aspects. In U. M. Quasthoff

(Ed.), Aspects of oral communication (pp. 391–418). Berlin: De Gruyter.
Heritage, J. C., & Greatbatch, D. (1991). On the institutional character of institutional

talk: The case of news interview interaction. In D. Boden & D. H. Zimmerman (Eds.), Talk
and social structure (pp. 93–137). Berkeley: University of California Press.
Heritage, J. C. (2005). Cognition in discourse. In H. te Molder & J. Potter (Eds.), Conversa

tion and cognition (pp. 184–202). Cambridge: Cambridge University Press.
Heritage, J. C., & Clayman, S. (2010). Talk in action. Oxford: Blackwell.
Heritage, J. C., & Raymond, G. (2005). The terms of agreement: Indexing epistemic au
thority and subordination in assessment sequences. Social Psychology Quarterly, 68, 15–
38.
Hollway, W., & Jefferson, T. (2005). Panic and perjury: A psycho-social exploration of
agency. British Journal of Social Psychology, 44, 1–18.
Horne, J., & Wiggins, S. (2009). Doing being “on the edge”: Managing the dilemma of be
ing authentically suicidal in an online forum. Sociology of Health and Illness, 31(2), 170–
184.
Jefferson, G. (2004). Glossary of transcript symbols with an Introduction. In G. H. Lerner

(Ed.), Conversation analysis: Studies from the first generation (pp. 13–23). Philadelphia:
John Benjamins.
Kent, A. (2012a). Compliance, resistance and incipient compliance when responding to di
rectives, Discourse Studies, 14(6), 711–730.
Page 29 of 34

Kent, A. (2012b). Responding to directives: What can children do when a parent tells
them what to do? In M. Theobold & S. Danby (Eds.), Disputes in everyday life: Social and
moral orders of children and young people (pp. 57–84). Bingley: Emerald Group Publish
ing Limited.
Kurri, K., & Wahlström, J. (2001). Dialogic management of conflict in domestic violence
counselling. Feminism and Psychology, 11, 187–208.
Lamerichs, J., & te Molder, H. (2003). Computer-mediated communication: From a cogni

tive to a discursive model. New Media & Society, 5(4), 451–473.
Levinson, S. C. (1983). Pragmatics. Cambridge: Cambridge University Press.
MacMillan, K., & Edwards, D. (1999). Who killed the princess? Description and blame in
the British press. Discourse Studies, 1 (2), 151–174.
Maynard, D. W., Houtkoop-Steenstra, H., Schaeffer, N. C., & van der Zouwen, J. (Eds.).
(2002). Standardization and tacit knowledge: Interaction and practice in the survey inter
view. New York: Wiley.
McKinlay, A., & Dunnett, A. (1998). How gun-owners accomplish being deadly average. In
C. Antaki & S. Widdicombe (Eds.), Identities in talk (pp. 34–51). London: Sage.
Meredith, J. & Potter, J. (2013). Conversation analysis and electronic interactions:

Methodological, analytic and technical consideration. In Lim, H. L. & Sudweeks, F. (Eds),
Innovative methods and technologies for electronic discourse analysis (pp. 370–393). Her
shey, Pennsylvania: IGI Global.
Myers, G. (1998). Displaying opinions: Topics and disagreement in focus groups. Lan
guage in Society, 27, 85–111.
Myers, G. (2004). Matters of opinion: Talking about public ideas. Cambridge: Cambridge
University Press.
Parker, I. (1992). Discourse dynamics: Critical analysis for social and individual psycholo
gy. London: Routledge.
Peräkylä, A., Antaki, C., Vehviläinen, S., & Leudar, I. (Eds.). (2008). Conversation analysis
of psychotherapy. Cambridge: Cambridge University Press.
Pomerantz, A. (1978). Compliment responses: Notes on the co-operation of multiple con

straints. In J. Schenkein (Ed.), Studies in the organisation of conversational interaction
(pp. 79–112). New York: Academic Press.
Pomerantz, A. M. (1984). Agreeing and disagreeing with assessments: Some features of

preferred/dispreferred turn shapes. In J. M. Atkinson & J. C. Heritage (Eds.), Structures of
social action: Studies in conversation analysis (pp. 57–101). Cambridge: Cambridge Uni
versity Press.
Page 30 of 34

Pomerantz, A. M. (1986). Extreme case formulations: A new way of legitimating claims.

Human Studies, 9, 219–230.
Potter, J. (1996). Representing reality: Discourse, rhetoric and social construction. Lon
don: Sage.
Potter, J. (1998). Discursive social psychology: From attitudes to evaluations. European

Review of Social Psychology, 9, 233–266.
Potter, J. (2001). Wittgenstein and Austin. In M. Wetherell, S. Taylor, & S. Yates (Eds.),
Discourse theory and practice (pp. 39–46). London: Sage.
Potter, J. (2003). Discursive psychology: Between method and paradigm. Discourse & So
ciety, 14, 783–794.
Potter, J. (2009). Discourse analysis. In A. Bryman & M. A. Hardy (Eds.), Handbook of da
ta analysis (pp. 607–624). London: Sage.
Potter, J. (2010). Contemporary discursive psychology: Issues, prospects, and Corcoran’s

awkward ontology. British Journal of Social Psychology, 49, 657–678.
Potter, J. (2012). Re-reading Discourse and social psychology: Transforming social psy
chology. British Journal of Social Psychology 51, 436–455.
Potter, J., & Edwards, D. (2001). Discursive social psychology. In W. P. Robinson & H. Giles
(Eds.), The new handbook of language and social psychology (pp. 103–118). London: John
Wiley & Sons Ltd.
Potter, J., Edwards, D., & Wetherell, M. (1993). A model of discourse in action. American
Behavioural Scientist, 36, 383–401.
Potter, J., & Hepburn, A. (2003). I’m a bit concerned—Early actions and psychological
constructions in a child protection helpline. Research on Language and Social Interaction,
36, 197–240.
(p. 313) Potter, J., & Hepburn, A. (2005). Qualitative interviews in psychology: Problems
and possibilities. Qualitative Research in Psychology, 2, 38–55.
Potter, J., & Hepburn, A. (2008). Discursive constructionism. In J. A. Holstein & J. F. Gubri
um (Eds.), Handbook of constructionist research (pp. 275–293). New York: Guilford.
Potter, J., & Hepburn, A. (2012). Eight challenges for interview researchers. In J. F. Gubri
um & J. A. Holstein (Eds.), Handbook of interview research (2nd ed., pp. 555–570). Lon
don: Sage.
Potter, J., Stringer, P., & Wetherell, M. (1984). Social texts and context: Literature and so
cial psychology. London: Routledge and Kegan Paul.
Page 31 of 34

Potter, J., & Wetherell, M. (1987). Discourse and social psychology: Beyond attitudes and
behaviour. London: Sage.
Potter, J., & Wetherell, M. (1988). Accomplishing attitudes: Fact and evaluation in racist
discourse. Text, 8, 51–68.
Puchta, C., & Potter, J. (2002). Manufacturing individual opinions: Market research focus
groups and the discursive psychology of attitudes. British Journal of Social Psychology,
41, 345–363.
Puchta, C., & Potter, J. (2004). Focus group practice. London: Sage.
Reynolds, J., & Wetherell, M. (2003). The discursive climate of singleness: The conse
quences for women’s negotiation of a single identity. Feminism and Psychology, 13, 489–
510.
Sacks, H. (1992). Lectures on conversation. Blackwell, UK: Oxford.
Schegloff, E. A. (1992). Repair after next turn: The last structurally provided defense of
intersubjectivity in conversation. American Journal of Sociology, 97, 1295–1134.
Schegloff, E. A. (1997). Whose text? Whose context? Discourse & Society, 8, 165–187.
Schegloff, E. A. (2007). Sequence organization in interaction. Volume 1, a primer in con

versation analysis. Cambridge: Cambridge University Press.
Schegloff, E. A., & Sacks, H. (1973). Opening up closings. Semiotica, 8, 289–327.
Searle, J. R. (1969). Speech acts: An essay in the philosophy of language. Cambridge, UK:
Shotter, J. (1993). The cultural politics of everyday life. Milton Keynes, UK: Open Universi
ty Press.
Sidnell, J., & Stivers, T. (Eds.). (2012). The handbook of conversation analysis. Oxford: Wi
ley-Blackwell.
Silverman, D. (2001). Interpreting qualitative data: Methods for analysing talk, text and
interaction: 2nd edition. London: Sage.
Smith, J., & Osborn, M. (2008). Interpretative phenomenological analysis. In J. Smith

(Ed.), Qualitative psychology: A practical guide to methods (2nd ed.). London: Sage.
Sneijder, P., & te Molder, H. (2004). “Health should not have to be a problem”: Talking
health and accountability in an internet forum on veganism. Journal of Health Psychology,
9(4), 599–616.
Page 32 of 34

Sneijder, P., & te Molder, H. (2005). Moral logic and logical morality: Attributions of re
sponsibility and blame in online discourse on veganism. Discourse & Society, 16(5), 675–
696.
Sneijder, P., & te Molder, H. (2009). Normalizing ideological food choice and eating prac
tices. identity work in online discussions on veganism. Appetite, 52(3), 621–630.
Speer, S., & Parsons, C. (2006). Gatekeeping gender: Some features of the use of hypo
thetical questions in the psychiatric assessment of transsexual patients. Discourse and
Society, 17, 785–812.
Speer, S., & Potter, J. (2000). The management of heterosexist talk: Conversational re
sources and prejudiced claims, Discourse & Society, 11, 543–572.
Stivers, T., Mondada, L., & Steensig, J. (Eds.). (2011). The morality of knowledge in con
versation. Cambridge: Cambridge University Press.
Stokoe, E. (2009). Doing actions with identity categories: Complaints and denials in
neighbor disputes. Text and Talk, 29, 75–97.
Stokoe, E. (2011). Simulated interaction and communication skills training: The “Conver
sation Analytic Role-play Method”. In C. Antaki (Ed.), Applied conversation analysis:
Changing institutional practices. Basingstoke: Palgrave Macmillan.
Stokoe, E., & Edwards, D. (2009). Accomplishing social action with identity categories:
Mediating neighbour complaints. In M. Wetherell (Ed.), Theorizing identities and social
action (pp. 95–115). London: Sage.
Stokoe, E., & Hepburn, A. (2005). “You can hear a lot through the walls”: Noise formula
tions in neighbour complaints. Discourse and Society, 16, 647–673.
Tannen, D., Schiffrin, D., & Hamilton, H. E. (2001). The handbook of discourse analysis
(pp. 352–371). Oxford, UK: Blackwell.
te Molder, H. (1999). Discourse of dilemmas: An analysis of communication planners’ ac

counts. British Journal of Social Psychology, 38, 245–263.
te Molder, H., & Potter, J. (Eds.). (2005). Conversation and cognition. Cambridge: Cam
Toerien, M., & Kitzinger, C. (2007). Emotional labour in action: Navigating multiple in
volvements in the beauty salon. Sociology, 41, 645–662.
Vayreda, A., & Antaki, C. (2009). Social support and unsolicited advice in a bipolar disor
der online forum. Qualitative Health Research, 19(7), 931–942.
Wetherell, M. (1998). Positioning and interpretative repertoires: Conversation analysis

and post- structuralism in dialogue. Discourse and Society, 9, 387–412.
Page 33 of 34

Wetherell, M. (2007). A step too far: Discursive psychology, linguistic ethnography and
questions of identity. Journal of Sociolinguistics, 11, 661–681.
Wetherell, M. (2012). Affect and emotion: A new social science understanding. London:
Sage.
Wetherell, M., & Potter, J. (1992). Mapping the language of racism: Discourse and the le
gitimation of exploitation. Hemel Hempstead, UK: Harvester/ Wheatsheaf.
Widdicombe, S., & Wooffitt, R. (1995). The language of youth subcultures: Social identity
in action. London: Harvester/Wheatsheaf.
Wiggins, S., & Potter, J. (2003). Attitudes and evaluative practices: Category vs. item and
subjective vs. objective constructions in everyday food assessments. British Journal of So
Wilkinson, S., & Kitzinger, C. (1995). Feminism and discourse: Psychological perspectives.
London: Sage.
Wittgenstein, L. (1953). Philosophical investigations, Oxford: Blackwell.
Wood, L. A., & Kroger, R. O. (2000). Doing discourse analysis: Methods for studying ac
tion in talk and text. London: Sage.
Wooffitt, R. (1992). Telling tales of the unexpected: The organization of factual discourse.
London: Harvester/Wheatsheaf.
Wooffitt, R. (2005). Conversation analysis and discourse analysis: A comparative and criti
cal introduction. London: Sage.
Woolgar, S. (1988). Science: The very idea. Chichester/London: Ellis Horwood/Tavistock.

(p. 314)
Alexandra Kent
Alexandra Kent, Keele University
Jonathan Potter
Jonathan Potter is Professor of Discourse Analysis and Dean of the School of Social,
Political and Geographical Sciences at Loughborough University.
Page 34 of 34

Grounding Language in Our Bodies and the World

Michael P. Kaschak and John L. Jones

Online Publication Date: Dec 2013 DOI: 10.1093/oxfordhb/9780199838639.013.004
Embodied approaches to language comprehension hold that the understanding of lan

guage is firmly rooted in the body’s systems of perception and action planning. This is
true both for the lower level processing of language (e.g., speech sounds are understood
in part through knowledge of the actions needed to produce those sounds) and for the
higher level processing that constitutes deeper comprehension (e.g., “mental models” of
the text are taken to be sensorimotor simulations of the events that are described). This
essay provides an overview of the embodied approach to language. To begin, we present
a broad theoretical discussion of the embodied perspective on cognition. Next, the grow
ing literature supporting this approach to comprehension is reviewed. Finally, further di
rections of exploration for the embodied approach are considered.
Keywords: embodied cognition, language comprehension, perception, action, evolution of language
The creation and conveyance of meaning is central to the use of language. Developing an
understanding of the means through which language conveys meaning is therefore cen
tral to the development of an account of how language is acquired and subsequently used
to accomplish a range of intra- and interpersonal goals. In the pages that follow, we tack
le the question of how language conveys meaning from the perspective of embodied cog
nition. The term embodied cognition refers to those approaches to cognitive science that
hold that our bodies’ systems of perception and action planning, and the ways that our
bodies interact with the external world, play a central role in cognition (e.g., Barsalou,
1999; Glenberg, 1997). Linguistic meaning is likewise rooted in systems of perception and
action planning. As a simple preliminary example, consider the sentence, The car ap
proached you. The embodied view proposes that the comprehension of sentences such as
this involves using one’s visual system to internally simulate what it would look like for a
car to approach you from a distance (Kaschak et al., 2005).
The outline of our essay is as follows. We begin with an overview of embodiment that con
siders the broader motivation for taking an embodied perspective on cognition in general
and for taking an embodied perspective on language in particular. With these preliminar
Page 1 of 23

ies in place, we then turn our attention to the evidence that has accrued in support of the
embodied account of linguistic meaning. There is by now a relatively large set of studies
that have attempted to address the claims of the embodied view of cognition, but we re
strict our overview to the literature that most directly addresses issues related to linguis
tic meaning. Broader overviews of the embodied account can be seen in Barsalou (2008)
and Fischer and Zwaan (2008). We conclude our essay with a discussion of the different
ways in which we might consider language and meaning to be embodied.
(p. 318) Embodied Cognition and Language

Embodiment is a relative newcomer in the field of cognitive science, but the ideas advo
cated by this approach have a long history in philosophy and psychology. For example,
components of the embodied approach can be seen in the writings of the phenomenolo
gists in philosophy (see Gallese, 2007, 2008, for a discussion), in the concern with the be
havior of organisms in their environment that was an aspect of behaviorist approaches to
psychology (e.g., Rachlin, 1991), and in ecological (Gibson, 1979) and situated (see chap
ters in Robbins & Aydede, 2009) approaches to psychology. In the field of artificial intelli
gence, Brooks (1999) was an early proponent of the embodied perspective. Finally, Lakoff
and Johnson’s (1980) seminal work on the metaphorical structure of thought was an im
portant impetus for the adoption of embodiment by cognitive psychologists (e.g., Gibbs
2011; Glenberg, 1997).
One reason to embrace an embodied approach to cognition has been discussed by several
authors under the guise of the symbol grounding problem (Glenberg & Robertson, 2000;
Harnad, 1990; Searle, 1980). The symbol grounding problem can be illustrated using
Harnad’s (1990) Chinese airport example. Imagine that you have disembarked from a
plane in China but cannot speak a word of Chinese. How will you use the signs in the air
port to find your way around? You look at a sign and observe the first logogram. Not
knowing what it means, you consult your Chinese dictionary to find the meaning of the lo
gogram. As it turns out, the definition of the logogram is also written in Chinese. You then
proceed to look up the first logogram in the definition in your dictionary. You find still
more logograms. It is clear at this point that no matter how much looking you do, the
signs and Chinese dictionary will never become meaningful. This is the point of the sym
bol grounding problem. Meaning cannot be conveyed in a system of symbols that are de
fined in terms of other symbols. For the symbols to become meaningful, they need to be
grounded in something else that is understood—such as pictorial representations on the
signs that indicate whether you are being directed to the departure gates or the baggage
claim (Glenberg & Robertson, 2000; Harnad, 1990; see de Vega, Glenberg, & Graesser,
2008, for extensive discussion of these issues). As Glenberg (1997; Glenberg & Robert
son, 2000; see also Barsalou, 1999) argues, virtually all approaches to cognition, save the
embodied approach, fail to resolve the symbol grounding problem.
The symbol grounding problem is highly relevant to thinking about how language conveys
meaning because the elements of language (e.g., words and phrases) are arbitrary sym
Page 2 of 23

bols that signify their referents. That is, the connection between the word violin and the
actual referent for that word is arbitrary in the sense that there is nothing in the form of
the word that conveys the meaning of what a violin actually is. The embodied approach
resolves the symbol grounding problem by proposing that the symbolic elements used in
language have their meaning grounded in our bodies’ systems of perception and action
planning (e.g., Glenberg, 1997). Part of this claim is the idea that we have a fundamental
understanding of our bodies and how they act in, and interact with, the external world.
This understanding provides the grounding necessary to understand language. To illus
trate, consider the simple example of what happens when one encounters a word such as
violin. Embodied approaches claim that the processing of the word violin involves retriev
ing sensorimotor information based on your experience with violins—visual information
about what a violin looks like, auditory information about what a violin sounds like, and
so on. In Barsalou’s (1999) terms, the processing of the word violin triggers the reinstate
ment of patterns of neural activity in the visual, motor, and auditory systems (for exam
ple) that were encoded during past experiences with violins. Thus, language is under
stood by constructing internal sensorimotor simulations of the objects, actions, and
events that are described to you in the linguistic input, and these simulations are directly
related to one’s past experiences interacting with the external world.
The symbol grounding problem provides one impetus for taking an embodied approach to
cognition. A second important reason for taking an embodied approach comes from an
evolutionary perspective. Why do some organisms have nervous systems, whereas others
do not? The answer to this question lies not in the complexity of the organism, but in the
need to move (Glenberg, Jaworski, Rischal & Levin 2007; Wolpert, Ghahramani, & Flana
gan, 2001). Simply put, organisms that move need to have a nervous system; organisms
that remain stationary have no need of a nervous system. The close relationship between
body morphology, nervous system characteristics, and cognitive capabilities that is cen
tral to the embodied approach to cognition has deep evolutionary roots (MacIver, 2009).
For example, the symmetrical body plan that is characteristic of virtually all animals was
an early (p. 319) evolutionary development that allowed for increased forward mobility
(Grabowsky, 1994). Body symmetry and forward mobility led to the clustering of sensory
organs around the anterior end of the organism, which in turn presaged the development
of more complex nervous systems and brains (Conway-Morris, 1998; MacIver, 2009;
Paulin, 2005). The evolution of sense organs and nervous systems was strongly influenced
by the nature of the organism’s body and the way the organism moves in the environment
(see, for example, the geometrical properties of the fly eye; Engelhaaf et al., 2002). As
MacIver (2009) suggests, the ability to receive sensory information from beyond the im
mediate surroundings of the organism opened up the possibility for more complex forms
of cognition, such as planning.
The observations in the preceding paragraph form an important basis for embodied cog
nition because they highlight the extent to which body shape, body movement, sensory
organs, and nervous systems co-evolve to allow organisms to take successful action in
their environment. This perspective also highlights the primacy of systems involved in
perception and action planning in the information processing work that is done by ani
Page 3 of 23

mals’ nervous systems. In the case of humans, evolution resulted in a complex nervous
system that has the ability not only to plan and act in the physical environment, but also
to plan and act in the social environment (e.g., Kaschak & Maner, 2009). Both of these
components were essential to the development of language within our species. Arbib
(2008) discusses how the neural pathways involved in reaching for and grasping objects
in one’s environment and in coordinated hand-to-mouth actions (pathways that are known
to involve mirror neurons; Rizzolatti, Sinigaglia, & Anderson, 2008) likely played a central
role in the evolution of language. Tomasello (1999) further argues that the development
of social cognitive abilities, such as intention reading and the ability to jointly attend to
things in your environment with conspecifics (abilities which themselves may rely on per
ception-action mechanisms such as mirror neurons; Cattaneo et al., 2011), was also es
sential to the evolution of language. As a simple example, the ability to use information
about another person’s gaze to infer what he is attending to in the surrounding environ
ment can greatly simplify the task of fixing the relationship between a word and its in
tended referent. If a person who had never seen a musical instrument before was in a
room full of instruments, and someone said, Pick up the violin while looking directly at a
violin, the person could reasonably infer that the object that was being looked at was the
intended referent of the word violin.
In contrast to the view that language has a special, unique status within the cognitive sys
tem (e.g., Pinker, 1994) or that the symbolic nature of language is divorced from (al
though connected to) the operation of perception-action systems (e.g., Louwerse, 2008),
the embodied view suggests that language is perception and action through and through.
The physical realization of language is perception (the perception of speech sounds or the
perception of orthographic symbols on the page) and action (the act of speaking or writ
ing), and these perception-action elements are linked through experience to sensorimotor
representations of the things to which the language refers (e.g., the appearance and
sound of a violin, the actions taken to draw the bow across the strings, and so on). Picker
ing and Garrod (in press) refer to these perception-action components of language as em
bodiment of form (i.e., the producing and perceiving the physical form of language) and
embodiment of content (i.e., the use of sensorimotor representations to ground linguistic
meaning). As our primary goal for this essay is to provide an overview of the evidence
suggesting that linguistic meaning is grounded in sensorimotor systems, we will devote
much of our time to discussing embodiment of content. We will return to embodiment of
form near the end of the essay, where we discuss the prospects of embodied approaches
accounting for the broad range of phenomena that fall under the rubric of linguistic
meaning.
Embodiment and Language: Empirical Find

ings
The embodied approach to language comprehension and linguistic meaning holds that
language is understood via internal sensorimotor simulations of the objects, actions, and
Page 4 of 23

events that are detailed in the linguistic input that one receives. The general assumption
is that the patterns of neural activity that are present when one encounters the described
objects, actions, and events in the world are reinstated during the comprehension process
(see Barsalou, 1999, 2008) and that linguistic meaning is grounded in this simulation
process (Glenberg & Robertson, 2000; Kaschak & Glenberg, 2000). A good deal of empiri
cal work, using both behavioral tasks and neuroimaging techniques, has been done to test
these claims over the past decade or so. Much of this work has sought to either demon
strate (p. 320) that sensorimotor representations are activated during the comprehension
of language or demonstrate that the comprehension of language affects one’s ability to
perceive and act in the world. The findings have generally been positive, providing impor
tant support for the claim that the comprehension of language is grounded in sensorimo
tor systems.
Activation of Perceptual Representations During Language Compre

hension
One particularly active area of research on embodiment in language comprehension con

cerns the use of perceptual representations during the comprehension process. Stanfield
and Zwaan (2001) present some of the earliest studies along these lines. Their experi
ments used a relatively straightforward paradigm to assess the extent to which compre
henders mentally represent the spatial orientation of objects that have been described in
a given sentence. As one example, participants in their experiments were presented with
one of two sentences involving a pencil. In one of the sentences (The pencil is on the
desk), the pencil is described in such a way that it would be laying horizontally. In the oth
er sentence (The pencil is in the cup), the pencil is described in such a way that it would
be in an upright, vertical position. Subsequent to reading these sentences, participants
were presented with either a picture of a pencil in a horizontal orientation or a picture of
a pencil in a vertical orientation. Participants needed to indicate whether the pictured ob
ject was mentioned in the immediately preceding sentence. The main finding of this work
is that participants were quicker to verify that the pictured object was mentioned in the
preceding sentence when the orientation of the object in the picture matched the orienta
tion of the object as described in the sentence (e.g., participants were quicker to respond
to the picture of an upright pencil after reading The pencil is in the cup than after reading
The pencil is on the desk). The fact that responses to the pictures were affected by the
orientation of the object, even though object orientation was irrelevant to the task at
hand, was taken as evidence that comprehenders are mentally simulating objects and
their orientations when reading sentences such as those presented in the experiment.
Stanfield and Zwaan’s (2001) paradigm, in which participants first process a sentence
and then respond to a picture presented immediately thereafter, has formed the basis of a
number of subsequent studies assessing the activation of perceptual representations dur
ing language comprehension. Zwaan, Stanfield, and Yaxley (2002) extended Stanfield and
Zwaan (2001) work by demonstrating that comprehenders also mentally simulate the
shape of objects during language comprehension. After processing a sentence such as
The eagle is in the sky, participants were quicker to respond to a picture of an eagle with
Page 5 of 23

outstretched wings than to a picture of an eagle with its wings folded in; however, partici
pants were quicker to respond to the latter picture after processing a sentence such as
The eagle is in the nest. Connell (2007) used a similar paradigm to argue that comprehen
ders activate information about object color during language comprehension (see also
Connell & Lynott, 2009; Richter & Zwaan, 2009). Yaxley and Zwaan (2007) took a step be
yond these studies to demonstrate not only that specific perceptual features (shape, ori
entation, color, and so on) were incorporated into a comprehender’s representation of the
language they were processing, but also that more contextual elements of a described sit
uation were represented. Participants were quicker to respond to a hazy picture of a
moose after reading The hiker saw the moose through foggy goggles than after reading
The hiker saw the moose through clear goggles and quicker to respond to a clear picture
of a moose after hearing the goggles described as clear as opposed to foggy.
The studies described here make that case that comprehenders activate perceptual repre
sentations of objects and situations when processing language. Another line of research
has explored the extent to which comprehenders also simulate elements of a event, such
as the motion that is being described. Zwaan, Madden, Yaxley, and Aveyard (2004) took
an initial look at this issue. They presented participants with sentences such as The pitch
er hurled the softball to you or You hurled the softball to the pitcher. After reading the
sentence, participants were presented with two successive pictures of the critical object
in the sentence (in this case, the softball). The second picture was either slightly larger
than, or slightly smaller than, the first picture, thus implying motion toward or away from
the participant, respectively. Participants were quicker to respond when the direction of
the motion in the sentence (toward or away from the participant) matched the direction
of motion implied by the pictures (toward or away from the participant). Thus, compre
henders appear to simulate motion when processing language.
Kaschak and colleagues (e.g., Kaschak et al., 2005; Kaschak, Zwaan, Aveyard, & Yaxley,
2006) have also explored the extent to which motion (p. 321) is simulated during language
comprehension. Kaschak, Madden, Therriault, Aveyard, Yaxley, Blanchard, and Zwaan
(2005) asked participants to process sentences describing motion toward them (The car
approached you), away from them (The squirrel ran away from you), upward (The rocket
blasted off into the sky), or downward (The rain fell on the spectators). At the same time,
the participants viewed black-and-white stimuli depicting motion toward them, away from
them, upward, or downward. Kaschak et al. (2005) demonstrated that making judgments
about the sentences (e.g., Is the sentence sensible? Is the sentence grammatical?) was fa
cilitated in cases where the direction of motion in the sentence mismatched the direction
of motion in the perceptual stimuli. Kaschak et al. (2005) suggest that this pattern is ob
served because the comprehension of sentences involving motion (e.g., The car ap
proached you) requires the use of the visual system (such as the middle temporal (MT)
area, which responds to motion) to simulate the motion that has been described. When
these visual resources are occupied (as they are when one is viewing a motion stimulus
that depicts motion toward the observer), the comprehender has difficulty simulating the
motion described in the sentence and therefore has more difficulty comprehending the
sentence. Kaschak et al. (2005) demonstrated this effect for visual motion (see also
Page 6 of 23

Bergen, Lindsay, Matlock, & Narayanan, 2007, and Richardson & Matlock, 2007, for fur
ther demonstrations that motion is simulated during comprehension). Kaschak et al.
(2006) employed a similar paradigm, but used auditory motion stimuli to demonstrate
that auditory motion was also simulated during language comprehension.
Meteyard et al. (2007) employed a different sort of paradigm to investigate the extent to
which the processing of language about motion involves the visual system. Participants
were asked to view random dot kinematograms that depicted coherent motion upward or
downward or that depicted random motion (i.e., neither distinctly up or down). At the
same time, they processed verbs describing upward or downward motion (e.g., lift or fall).
The processing of motion verbs affected the participants’ ability to perceive motion in the
kinematograms, such that the processing of a verb with one direction of motion (e.g., up
ward) impaired the ability to perceive motion in the opposite direction (e.g., downward).
These results suggest that the comprehension of language can affect relatively low levels
of visual processing.
A good deal of the evidence that supports the claim that linguistic processing involves
sensorimotor simulation has come from tasks that involve processing in a single perceptu
al modality (e.g., audition or vision). A smaller set of studies have demonstrated support
for the embodied simulation view by asking participants to verify object properties in dif
ferent perceptual modalities (e.g., Is a banana yellow? vs. Is a siren loud?). Just as shifting
attention between perceptual modalities produces a processing cost (Spence, Nicholls, &
Driver, 2000), it would be expected that switching between modalities in simulating ob
ject properties would also produce a processing cost. Support for this hypothesis has
been observed in several studies (e.g., Marques, 2006; Pecher, Zeelenberg, & Barsalou,
2003, 2004; Vermeulen, Niedenthal, & Luminet, 2007).
The embodied approach has also been supported by evidence from neuroimaging studies.
As reviewed in Martin (2007; see also Binder, Desai, Graves, & Conant, 2009), several
studies have demonstrated that when participants are asked to perform tasks that re
quire focusing on particular properties of an object or concept, the brain regions that are
typically associated with the processing of those properties become active. For example,
when participants are asked to verify properties related to shape, color, size, taste, touch,
and sound, functional magnetic resonance imaging (fMRI) studies have shown activity in
the cortical regions associated with the processing of those properties (e.g., Goldberg,
Perfetti, & Schneider, 2006; Gonzalez et al., 2006; Kan et al., 2003; Simmons, Martin, &
Barsalou, 2005; Simmons et al., 2007). The neuroimaging evidence has largely been con
sistent with the embodied view, although there are some exceptions to this. Rueschemey
er et al. (2010) conducted an fMRI experiment in which participants performed a task
similar to that described by Kaschak et al. (2005), processing both visual motion stimuli
and sentences describing motion. Rueschemeyer et al. (2010) found that the processing of
motion sentences activated the same cortical area that was involved in the processing of
the visual motion stimulus (area MT in the visual system). However, this was true only for
the sentences involving motion toward the participant. Sentences involving motion away
from the participant did not activate MT. These data suggest the possibility that compre
Page 7 of 23

henders do not simulate in full detail all aspects of the situation that is being described in
the linguistic input. As Rueschemeyer et al. (2010) point out, it may be that those details
of the situation that are (p. 322) simulated depend on top-down factors, such as the per
sonal relevance of the details for the comprehender. Motion toward the body may be
more behaviorally relevant than motion away from the body and thus is more likely to be
actively simulated during the comprehension process.
Activation of Motor Representations During Language Comprehen

sion
A second active area of research into the embodied approach to language comprehension
has explored the extent to which motor representations are activated during the compre
hension process. An early example of this work is reported in Kaschak and Glenberg
(2000). In their experiments, participants were asked to comprehend sentences contain
ing innovative denominal verbs (i.e., verbs created from nouns, such as Lyn crutched her
apple to Tom). To process the innovative verbs, Kaschak and Glenberg (2000) reasoned
that participants would need to derive affordances (i.e., possibilities for interaction; Gib
son, 1979) from the object named by the noun (in this case, a crutch) to simulate the way
that the object could be used to accomplish the action specified in the sentence (transfer
ring the apple from Lyn to Tom). Their results suggest that participants were more suc
cessful in comprehending sentences with innovative denominal verbs when the object
named by the verb had the proper affordances to support the action described in the sen
tence (e.g., crutches afford swinging like a baseball bat, and thus Lyn could hit her apple
to Tom).
Glenberg and Kaschak (2002) introduced a new paradigm for assessing whether action-
based representations were accessed during language comprehension. Participants were
asked to read and make sensibility judgments on sentences such as, You gave John the
pen (action away from the body) or John gave you the pen (action toward the body). To
read the sentence, participants pressed and held down a start button located in the mid
dle of a response box. After reading the sentence, the participants needed to execute an
action to indicate whether the sentence was sensible. Half of the time, participants had to
release the start button and move to push a button located farther away from their body
(i.e., to produce an arm movement away from their body). The other half of the time, par
ticipants moved from the start button to a button closer to their body (i.e., producing an
action toward their body) to respond. There were two main findings from this study. First,
the execution of the motor responses was affected by the content of the sentences. When
the sentence described action away from the body, participants were faster to release the
start button when producing the response away from their bodies. When the sentence de
scribed action toward the body, they were faster to release the start button when produc
ing the response toward their bodies. It is noteworthy that the processing of language af
fected motor planning—the time it took to prepare to execute the desired response—and
not the time it took to execute the response (i.e., the time needed to move from the start
button to the desired response key). Thus, just as language processing can affect low lev
els of visual perception (Meteyard et al., 2007), it can also affect low levels of motor be
Page 8 of 23

havior. This effect was called the action-sentence compatibility effect (ACE). Second, the
ACE was present whether the sentence described a concrete action toward or away from
the body (John handed you a pen) or an action in which the toward or away motion was
more abstract (Anna told you a story, in which there is transfer of information but no con
crete transfer).
Glenberg and Kaschak (2002) demonstrated that the comprehension of sentences about
action affects motor planning, thus suggesting that language comprehension involves
constructing internal motor simulations of the actions that are being described. Subse
quent to this initial finding, researchers have been interested in determining when and
how motor information is employed during language processing (e.g., Borreggine &
Kaschak, 2006; Bub & Masson, 2010; Glenberg, Sato, & Cantaneo, 2008; Glenberg et al.,
2008; Zwaan & Taylor, 2006). Zwaan and Taylor (2006) presented studies in which they
replicate the ACE using a different sort of motor response. Participants read sentences
such as When he walked into the room, John turned down the radio. Sentences were pre
sented phrase by phrase, and participants needed to rotate a knob clockwise or counter
clockwise to advance from phrase to phrase. Zwaan and Taylor (2006; see also Taylor &
Zwaan, 2008) report a motor compatibility effect that arises when participants read the
verb of the sentence (in this case, turned down)—participants were faster to read the
phrase when the direction of rotation for the knob matched the direction of rotation that
would be needed to complete the action indicated by the sentence. Taken together with
the results of Glenberg and Kaschak (2002; see also Borreggine & Kaschak, 2006), these
data suggest that there are at least two points in time at which motor activity can be ob
served during sentence processing—when processing the verb of the sentence (as in
Zwaan & Taylor’s 2006 rotation (p. 323) experiments; see also Boulenger et al., 2006) and
at the end of the sentence, when participants are perhaps running a simulation of the en
tire set of actions described (as in studies using the original ACE paradigm of Glenberg &
Kaschak, 2002).
When discussing the timing of motor activity during sentence comprehension, it is impor
tant to note that this activity is subject to modulation based on cues presented in the lin
guistic input. Zwaan et al. (2010) demonstrate this in experiments that employ the same
task paradigm as the Zwaan and Taylor (2006) studies. Whereas Zwaan and Taylor (2006)
found motor activity when participants were reading the verb of the sentences describing
manual rotation of an object, Zwaan et al. (2010) found that this activity was attenuated
when participants read sentences that described actions that had not yet taken place
(e.g., The boy anticipated turning up the volume on his radio). Furthermore, Taylor and
Zwaan (2008) showed that motor activity could be elicited by refocusing the comprehen
der on the nature of the action taking place. When participants were asked to read, When
he walked into the room, John turned down the radio slowly, Taylor and Zwaan (2008)
observed motor activity not only on the verb (turned down), but also on the adverb slowly.
Explorations of the ACE (and related effects) have demonstrated a range of circum
stances under which motor information is elicited during the comprehension of language.
Bub and Masson (2010; Bub, Masson, & Cree, 2008) have shed further light on the role of
Page 9 of 23

motor information in language comprehension by examining the nature of the motor rep
resentations activated during the comprehension process. They begin by making a dis
tinction between what they call functional motor gestures and volumetric motor gestures.
Functional gestures refer to actions that correspond to the way that one would use an ob
ject, such as the single-finger poking action that would be employed if one were to use a
calculator or type a text message on a cell phone. Volumetric gestures refer to actions
that correspond to the way that one would pick up and manipulate an object, such as the
open-handed grasp that would be needed to pick up a calculator or cell phone or the
pinching grasp that would be used to pick up a pencil. This distinction is akin to Johnson-
Frey and Grafton’s (2003) distinction between “acting on” an object and “acting with” an
object. Bub and Masson’s work employs tasks involving the processing of single words
naming objects (e.g., calculator) and the processing of those same words when embedded
within a sentence. The question of interest is whether the processing of nouns elicits mo
tor information during language comprehension (note that the demonstrations of motor
effects on comprehension described here tend to be based more around the meaning of
verbs), the nature of the motor information activated (acting on vs. acting with), and
whether the activation of information changes depending on whether one is processing
individual words or sentences.
To study the nature of the motor activity that arises during language comprehension (par
ticularly the distinction between functional and volumetric motor activity), Bub, Masson,
and Cree (2008) trained participants to perform several distinct motor responses (a poke,
a grasp, and so on) when cued to do so. After training, participants then executed these
responses while processing words or pictures. Bub et al.’s (2008) results demonstrate
that both functional and volumetric gestures are elicited when participants are process
ing pictures of objects or the words that label those objects and further suggest that the
processing of words may activate functional gestures more quickly than volumetric ges
tures. Later work from this lab group has shown that the activation of gestural informa
tion is affected by sentence context. Bub and Masson (2010) report that during sentence
processing, both functional and volumetric gestures are initially activated. Over time,
however, the activation of the functional gesture persists, whereas the activation of the
volumetric gestures does not. Masson, Bub, and Warren (2008) demonstrated that func
tional gestures are activated even when processing sentences that do not involve the use
of the target object for its intended purpose (e.g., The lawyer looked at the calculator).
They also demonstrated that volumetric gestures play a role in the comprehension of sen
tences that describe nontraditional interactions with an object (e.g., kicking a calculator).
The results from Bub and Masson’s studies suggest that hand-and-arm actions and pat
terns of interaction with objects play an important role in the comprehension of language,
including cases in which the hand action is needed to understand the action described
(e.g., The accountant added the numbers on the calculator, in which the functional ges
ture aids understanding of the action), where the hand action is afforded by the situation
(e.g., The lawyer looked at the calculator, in which the situation implies that the lawyer is
considering use of the calculator), or where the hand action is needed to compute a new
affordance for interaction (e.g., the volumetric gesture specifies information about the
Page 10 of 23

size and heft of an (p. 324) object and thus can be used to derive information about non
traditional object uses, such as kicking a calculator). Bub and Masson’s results combine
with other studies of motor effects (e.g., Zwaan & Taylor, 2006) to demonstrate that both
nouns and verbs elicit motor activity during language comprehension, with the implica
tion that the comprehension process involves recruiting and integrating many different
types of motor information. More specifically, it would appear that the comprehension
process involves a consideration of the action-relevant properties of the objects described
and the motor patterns associated with particular verbs and the construction of a motor
simulation that combines these two sets of motor constraints into a coherent pattern of
action (Glenberg & Robertson, 2000; Kaschak & Glenberg, 2000).
The grounding of meaning in the motor system can be shown not only in studies such as
those discussed in the preceding paragraphs, but also in studies that explore the role that
gestures play in speech. Hostetter and Alibali (2008) propose that at least some kinds of
gesture (representational gestures or gestures meant to convey content-related informa
tion when accompanying speech) arise from the spatial and motoric representations that
form the basis of linguistic meaning. Representational gestures can therefore be seen as
a type of simulated action that occurs while speaking. Hostetter and Alibali’s (2008)
suggestion fits well with Arbib’s (2008) evolutionary account of language, where it is pro
posed that neural mechanisms that were involved in the control of the hands and arms,
and in hand-to-mouth action, were co-opted for the control of speech. On this view,
speech and gesture were likely to have been very closely linked right from the develop
ment of spoken communication in humans.
The activation of motor information during language comprehension has also been ob
served using neuroimaging techniques. Pulvermuller (1999) reviews a body of evidence
demonstrating that the processing of words involving motor representations, such as
words denoting actions and words denoting tools, is linked to activity in cortical regions
known to be involved in motor planning and execution. Hauk et al. (2004) asked partici
pants to process words such as pick, kick, and lick. They found that while listening to
these verbs, the region of the motor strip that would be involved in executing the action
became active. That is, while processing kick, the part of the motor strip that controls leg
actions was active, whereas while processing pick, the part of the motor strip that con
trols hand actions was active, and while processing lick, the part of the motor strip that
controls mouth actions was active. A similar degree of specificity in motor activity has
been reported by Tettamanti et al. (2005) and Buccino et al. (2005). Taken together, the
behavioral and neuroimaging data make the case that the comprehension of language in
volves the same neural pathways that handle the planning and execution of action.
Emotion
The motoric and perceptual components of the simulations constructed during language
comprehension have received the most attention from researchers. A smaller number of
studies have shown that emotional simulation also occurs during the comprehension
process. Havas et al. (2007) report a series of experiments in which the participants’
Page 11 of 23

mood was manipulated via the pen-in-mouth procedure introduced by Strack et al. (1988).
Participants processed sentences containing emotional content (e.g., describing a happy
or sad situation) while the Strack procedure was employed to induce positive or negative
affect in them. The results show a mood congruence effect—participants were quicker to
process the sentences when the valence of their induced mood matched the valence of
the sentence. In a more striking example of this phenomenon, Havas et al. (2010)
demonstrated that injections of botulin toxin (Botox) that paralyzed the muscles involved
in frowning selectively impaired the comprehension of language about negatively va
lenced events. Interfering with the body’s ability to use facial muscles to express emotion
thus interferes with one’s ability to comprehend language that trades on those emotions.
Understanding Language About Abstract Concepts and Situations
To this point, we have largely considered experiments and data that concern the compre
hension of language about concrete, tangible actions and events (e.g., seeing moving ob
jects or taking action on objects). We now turn to consider studies that explore the role of
sensorimotor simulations in the comprehension of language about more abstract con
cepts and events, such as abstract transfer of possession or information, shifts in time,
and other such events. This is an important domain of language use for an embodied ap
proach to consider because one frequent criticism of embodiment is that, whereas the
theory provides a straightforward account of the (p. 325) comprehension of language
about concrete events, it is less clear how sensorimotor simulations underlie the compre
hension of more abstract situations. As Barsalou (1999, 2008) suggests, the notion of an
abstract concept is itself not a unitary construct—there are different types of abstraction
and thus an embodied account may need to provide multiple routes toward grounding ab
stractions in sensorimotor processes.
We have already discussed one set of findings concerning abstractions. Glenberg and
Kaschak’s (2002) study of the ACE involved abstract transfer sentences such as, Art gave
you the idea. These sentences produced the same ACE pattern that was observed for sen
tences about concrete transfer (Art gave you the mug). Following Goldberg (1995; see al
so Kaschak & Glenberg, 2000), Glenberg and Kaschak (2002) suggest that particular syn
tactic templates (such as the subject—verb—indirect object—direct object pattern used
for the dative transfer sentences in Glenberg and Kaschak’s experiments) encode basic
scenes of human experience. This is what Goldberg (1995) refers to as the scene encoding
hypothesis. The meaning of these sentence constructions is acquired by hearing and us
ing the constructions in the context of real actions and events (Goldberg, 1998). Once
these constructions have been learned for concrete situations (e.g., concrete transfer of
objects), they can be extended to ground the understanding of more abstract situations
(e.g., abstract transfer, such as the transfer of ideas or information). Thus, even though
the literal action depicted in a sentence such as Art gave you the idea has nothing to do
with motion toward or away from the body, recognition of the “transfer” template of the
event activates a general motor representation of what it means for something to be ex
changed between individuals.
Page 12 of 23

A second embodied approach to abstract concepts is grounded in Lakoff and Johnson’s

(1980) proposals about conceptual metaphors. Lakoff and Johnson (1980) propose that
many types of abstract concepts are understood via a metaphorical mapping between the
abstraction and a concrete domain of experience. As one example, the up-down axis pro
vides the spatial grounding for a number of concepts: quantity (more is up, less is down),
emotional valence (happy is up, sad is down), and power (high power is up, low power is
down), among others. As another example, the front-back axis provides the spatial
grounding for the concept of time (future is in front, past is behind). The conceptual
metaphor theory advanced by Lakoff and Johnson (1980) has received a good deal of at
tention in the literature on social (e.g., Niedenthal, Barsalou, Winkielman, Krauth-Gruber,
& Ric, 2005; Schubert, 2004, 2005; Williams & Bargh, 2008) and cognitive (e.g., Borodit
sky & Ramscar, 2002; Casasanto & Boroditsky, 2008; see Gibbs, 2011, for a review) psy
chology, and the evidence is largely supportive of the idea that conceptual metaphors
ground the understanding of particular abstract concepts (but see McGlone, 2011, for a
dissenting opinion). For the present purposes, we will restrict our discussion to those
studies that deal with the conceptual metaphor approach in the context of language com
prehension.
Santiago and colleagues (Santiago, Lupianez, Perez, & Fuenes, 2007; Santiago, Roman,
Ouellet, Rodriguez, & Perez-Azor, 2010; Torralbo, Santiago, & Lupianez, 2006) have con
ducted a series of experiments concerning the understanding of sentences about past and
future actions. Conceptual metaphor theory and embodied approaches suggest that one
should see motor and spatial compatibility effects when processing sentences about ac
tions taking place at different time points, such that responses involving action away from
the body (i.e., moving farther in front of the person) should be faster when one is process
ing language about future events, and responses involving action toward the body (i.e.,
moving toward the back of the person) should be faster when one is processing language
about past events. Although Torralbo et al. (2006) found evidence for the future-is-front/
past-is-behind metaphor, they also found a spatial mapping of time that is not found in
any linguistic metaphors: future-is-right/past-is-left. The left-to-right mapping of the time
line is reminiscent of the left-to-right mapping of the mental number line observed in par
ticipants from Western cultures (Shaki & Fischer, 2008). Subsequent research has
demonstrated both the left-right spatial grounding of time (Ulrich & Maienborn, 2010)
and the front-back mapping of time (Sell & Kaschak, 2011; Ulrich et al., 2012). Although
these findings might be considered to be mixed support for the conceptual metaphor the
ory (given that the mapping of space onto time can be found both on the metaphorical
and nonmetaphorical axes), the finding that the processing of language involving past
and future events involves spatial (e.g., Santiago et al., 2007) and motoric (Sell &
Kaschak, 2011; Ulrich et al., 2012) representations is consistent with the broader claims
of embodiment; namely, that abstract concepts and situations are understood through a
grounding in (p. 326) concrete domains of experience (see Sell & Kaschak, 2012, and
Pecher & Boot, 2011, for evidence of a similar pattern involving language about changes
in quantity).
Page 13 of 23

Kaup et al. (2006; see also Kaup, Yaxley, Madden, Zwaan, & Ludtke, 2007) address the is
sue of how a simulation-based view of comprehension would explain the understanding of
sentences involving negation. Using a paradigm similar to Stanfield and Zwaan (2001),
participants first read sentences such as The door is open or The door is not open. The
participants were subsequently presented with a picture of an open door or a closed door.
Pictures were presented 750 msec or 1,500 msec after the end of the sentence. At the
first time point, there was a match effect for the non-negated sentence—participants
were quicker to respond to a picture of an open door after reading The door is open. After
1,500 msec, however, the pattern reversed, and there was now a match advantage for the
negated sentence, such that participants were quicker to respond to a picture of a closed
door after reading The door is not open. These data suggest that negated sentences are
understood by first simulating the non-negated situation (door is open) and then simulat
ing the opposite situation (door is closed). Barsalou (1999) offers a similar take on an em
bodied account of negation.
The extant literature is far from providing a complete picture for how an embodied ap
proach to language comprehension can explain the understanding of every type of ab
stract concept or situation that might be encountered during language processing. How
ever, the positive demonstrations noted here suggest that there is promise in the view
that abstractions may be understood via their connection to more concrete forms of expe
rience (see Barsalou, 1999, for a more extensive discussion of this matter).
Conclusion
Over the past decade, a number of studies have addressed the view that the processes in
volved in language comprehension and linguistic meaning in general are grounded in sys
tems of perception and action planning. The literature reviewed in this essay helps to
make the case that the neural systems involved in perception, action planning, and emo
tional responding are involved in the comprehension of various types of language. That is,
there is embodiment in content with regard to the processing of language. Pickering and
Garrod (2013) also argue for an embodiment in form with respect to language, as the very
acts of producing and comprehending language trade heavily on and are governed by the
same principles as the systems of perception and action planning responsible for our abil
ity to successfully navigate the external environment. Pickering and Garrod (2013) draw
on Wolpert et al.’s (2003) HMOSAIC model of perception and action to argue that speak
ing and comprehending in the domain of language follows the same rules as acting and
understanding the actions of other individuals in the nonlinguistic domain. Interestingly,
Glenberg and Gallese (2011) use the structure of HMOSAIC to propose a theory of em
bodiment in content. These are interesting and needed developments in this area of re
search because it would appear that grounding language processes in perception and ac
tion planning requires showing how the known properties of those perception and action
systems shape the act of speaking, perceiving speech, and constructing sensorimotor sim
ulations in the service of language comprehension.
Page 14 of 23

Pickering and Garrod’s (2013) approach to embodiment in form may also mark an impor
tant step toward thinking about how a broader range of linguistic meaning may be con
sidered to be embodied. Much of the empirical evidence for embodiment has concerned
questions of how specific types of content are simulated. Whereas this is certainly an im
portant component of linguistic meaning, there are other layers of meaning (e.g., conver
sational implicatures and conversational pragmatics more generally) that need to be con
sidered. Pickering and Garrod’s (2013) account suggests the possibility that the physical
act of speaking may be understood in the same way that we understand other types of ac
tions that we observe. Just as the understanding of others’ actions may be grounded in
our own motor systems (e.g., Cattaneo et al. 2011; Rizzolatti et al., 2008), so, too, may the
understanding of others’ speech acts be grounded in our own motor systems and our own
knowledge of how to do things with language.
As one sketch of how embodiment in form may allow for an approach to conversational
pragmatics, consider Grice’s (1975) maxim of manner. The maxim of manner concerns
how speakers should express their meanings and indicates that speakers should (among
other things) be brief with their utterances and avoid obscure expressions. Violations of
these principles trigger a range of conversational implicatures. For example, Levinson
(2000) indicates that when a meaning is expressed in an unusual way (e.g., with obscure,
infrequent (p. 327) words, or with an overly prolix expression) it triggers an implicature
that there is something marked about the situation being described. Pickering and Gar
rod (2013) can approach the comprehension of such implicatures through the idea that
comprehenders are using their language production system to make predictions about
what they are going to hear next. That is, spoken language comprehenders are filtering
their understanding of the incoming utterance through their own knowledge and expecta
tions about how language should be produced, given their own knowledge and expecta
tions about how a given set of events should play out. It is on the basis of these expecta
tions that predictions about a speaking partner’s forthcoming utterance are made. Ob
scure expressions and overly prolix statements will initially lead to incorrect predictions
on the part of the comprehender—all things being equal, the comprehender will be ex
pecting more common, higher frequency words and shorter syntactic forms to be used.
Error in predictions may elicit the implicature generation process, which is resolved in
part by the comprehender using his or her knowledge of the potential patterns of action
that can be used to convey a particular message. In other words, if the utterance is at
odds with what was expected based on knowledge of human-relevant events and how to
communicate those in one’s language (e.g., Goldberg, 1995), then alternative interpreta
tions (i.e., implicatures) can be explored by the language comprehender. It is fair to note
that this sketch does not necessarily show (and, indeed, is not intended to show) that em
bodiment in form can provide a better account of implicatures than other extant views.
Rather, the sketch is aimed at illustrating how such aspects of linguistic meaning are not
necessarily beyond the scope of an embodied perspective.
The past decade has witnessed an explosion in knowledge about how the comprehension
of language is grounded in perception and action. The extant data make a compelling
case that at least certain components of language and linguistic meaning can be consid
Page 15 of 23

ered to be embodied. Over the course of the next decade, we hope to see much progress
in understanding the finer details of how sensorimotor simulations are constructed and
how these give rise to meaning. We also hope to see researchers concerned with embodi
ment extend their reach to consider not only how specific linguistic content can be em
bodied, but also how social and pragmatic dimensions of language use can be incorporat
ed into the embodied framework.
References
Arbib, M. A. (2008). From grasp to language: Embodied concepts and the challenge of ab
straction. Journal of Physiology, 102, 4–20.
577–660.
Barsalou, L. W. (2008). Grounded cognition. Annual Review of Psychology, 59, 617–645.
Bergen, B., Lindsay, S., Matlock, T., & Narayanan, S. (2007). Spatial and linguistic aspects
of visual imagery in sentence comprehension. Cognitive Science, 31, 733–764.
Binder, J. R., Desai, R. H., Graves, W. W., & Conant, L. L. (2009). Where is the semantic
system? A critical review and meta-analysis of 120 functional neuroimaging studies. Cere
bral Cortex, 19, 2767–2796.
Boroditsky, L., & Ramscar, M. (2002). The roles of body and mind in abstract thought.
Psychological Science, 13, 185–189.
Borreggine, K. L., & Kaschak, M. P. (2006). The action–sentence compatibility effect: It’s
all in the timing. Cognitive Science, 30, 1097–1112.
Boulenger, V., Roy, A. C., Paulignan, Y., Deprez, V., Jeannerod, M., & Nazir, T. (2006).
Cross-talk between language processes and overt motor behavior in the first 200 ms of
processing. Journal of Cognitive Neuroscience, 18, 1607–1615.
Brooks, R. A. (1999). Cambrian intelligence. Cambridge, MA: MIT Press.
Bub, D. N., & Masson, M. E. J. (2010). On the nature of hand-action representations

evoked during written sentence comprehension. Cognition, 116, 394–408.
Bub, D. N., Masson, M. E. J., & Cree, G. S. (2008). Evocation of functional and volumetric
gestural knowledge by objects and words. Cognition, 106, 27–58.
Buccino, G., Riggio, L., Melli, G., Binkofski, F., Gallese, V., & Rizzolatti, G. (2005). Listen
ing to action related sentences modulates the activity of the motor system: A combined
TMS and behavioral study. Cognitive Brain Research, 24, 355–363.
Casasanto, D., & Boroditsky, L. (2008). Time in the mind: Using space to think about time.
Cognition, 106, 579–593.
Page 16 of 23

Cattaneo, L., Barchiesi, G., Tabarelli, D., Arfeller, C., Sato, M., & Glenberg, A. M. (2011).
One’s motor performance predictably modulates the understanding of others’ actions
through adaptation of premotor visuo-motor neurons. Social Cognitive and Affective Neu
roscience, 6, 301–310.
Connell, L. (2007). Representing object color in language comprehension. Cognition, 102,

476–485.
Connell, L., & Lynott, D. (2009). Is a bear white in the woods? Parallel representation of
implied object color naming during language comprehension. Psychonomic Bulletin & Re
view, 16, 573–577.
Conway-Morris, S. (1998). The crucible of creation: The Burgess Shale and the rise of ani
mals. Oxford University Press.
de Vega, M., Glenberg, A. M., & Graesser, A. C. (Eds.). (2008). Symbols, embodiment, and
meaning. Oxford University Press.
Engelhaaf, M., Kern, R., Krapp, H. G., Kretzberg, J., Kurtz, R., & Warzecha, A. K. (2002).
Neural encoding of behaviorally relevant visual-motion information in the fly. Trends in
Neurosciences, 25, 96–102.
Fischer, M. H., & Zwaan, R. A. (2008). Embodied language—A review of the role of the
motor system in language comprehension. Quarterly Journal of Experimental Psychology,
61, 825–850.
Gallese, V. (2007). Before and below theory of mind: Embodied simulation and the neural
correlates of social cognition. (p. 328) Philosophical Transactions of the Royal Society of
London B, Biological Sciences, 362, 659–669.
Gallese, V. (2008). Mirror neurons and the social nature of language: The neural exploita
tion hypothesis. Social Neuroscience, 3, 317–333.
Gibbs, R. W. Jr. (2011). Evaluating conceptual metaphor theory. Discourse Processes, 48,
529–562.
Gibson, J. J. (1979). The ecological approach to visual perception. Boston: Houghton Mif
flin.
Glenberg, A. M. (1997). What memory is for. Behavioral and Brain Sciences, 20, 1–55.
Glenberg, A. M., & Gallese, V. (2011). Action-based language: A theory of language acqui
sition, comprehension, and production. Cortex, 1–18.
Glenberg, A. M., Jaworski, B., Rischal, M & Levin, J. R. (2007). What brains are for: Ac
tion, meaning, and reading comprehension. In D. MacNamara (Ed.), Reading comprehen
sion strategies: Theories, interventions, and technologies. New York: Lawrence Erlbaum
Associates.
Page 17 of 23

Glenberg, A. M., & Kaschak, M. P. (2002). Grounding language in action. Psychonomic

Bulletin and Review, 9, 558–565.
Glenberg, A. M., Sato, M., & Cattaneo, L. (2008). Use-induced motor plasticity affects the
processing of abstract and concrete language. Current Biology, 18, R290–R291.
Glenberg, A. M., & Robertson, D. A. (2000). Symbol grounding and meaning: A compari
son of high-dimensional and embodied theories of meaning. Journal of Memory and Lan
guage, 43, 379–401.
Glenberg, A. M., Sato, M., Cattaneo, L., Riggio, L., Palumbo, D., & Buccino, G. (2008).
Processing abstract language modulates motor system activity. Quarterly Journal of Ex
perimental Psychology, 61, 905–919.
Goldberg, A. E. (1995). Constructions. A construction grammar approach to argument

structure. Chicago: University Chicago Press.
Goldberg, A. E. (1998). Patterns of experience in patterns of language. In M. Tomasello

(Ed.), The new psychology of language (pp. 203–220). New York: Erlbaum.
Goldberg, R. F., Perfetti, C. F., & Schneider, W. (2006). Perceptual knowledge retrieval ac
tivates sensory brain regions. Journal of Neuroscience, 26, 4917–4921.
Gonzalez, J., Barros-Loscertales, A., Pulvermüller, F., Meseguer, V., Sanjuan, A., et al.
(2006). Reading cinnamon activates olfactory brain regions. Neuroimage 32, 906–912.
Grabowsky, G. L. (1994). Symmetry, locomotion, and the evolution of the anterior end: A
lesson from sea urchins. Evolution, 48, 1130–1146.
mantics, vol. 3 (pp. 41–58). New York: Academic Press.
Harnad, S. (1990). The symbol grounding problem. Physica D: Non-linear Phenomena, 42,
335–346.
Hauk, O., Johnsrude, I., & Pulvermüller, F. (2004). Somatotopic representation of action
words in human motor and premotor cortex. Neuron, 41, 301–307.
Havas, D. A., Glenberg, A. M., Gutowski, K. A., Lucarelli, M. J., & Davidson, R. J. (2010).
Cosmetic use of botulinum toxin-A affects processing of emotional language. Psychologi
Havas, D. A., Glenberg, A. M., & Rinck, M. (2007). Emotion simulation during language
comprehension. Psychonomic Bulletin & Review, 14, 436–441.
Hostetter, A. B., & Alibali, M. W. (2008). Visible embodiment: Gestures as simulated ac
tion. Psychonomic Bulletin & Review, 15, 495–514.
Page 18 of 23

Johnson-Frey, S. H., & Grafton, S. T. (2003). From “acting on” to “acting with”: The func
tional anatomy of action representation. In D. P. C. Prablanc & Y. Rossetti (Eds.) Space
coding and action production (pp. 127–139). New York: Elsevier.
Kan, I. P., Barsalou, L. W., Solomon, K. O., Minor, J. K., & Thompson-Schill, S. L. (2003).
Role of mental imagery in a property verification task: fMRI evidence for perceptual rep
resentation of conceptual knowledge. Cognitive Neuropsychology, 20, 525–540.
Kaschak, M. P., & Glenberg, A. M. (2000). Constructing meaning: The role of affordances
and grammatical constructions in language comprehension. Journal of Memory and Lan
guage, 43, 508–529.
Kaschak, M. P., Madden, C. J., Therriault, D. J., Yaxley, R. H., Aveyard, M. E., Blanchard, A.
A., et al. (2005). Perception of motion affects language processing. Cognition, 94, B79–
B89.
Kaschak, M. P., & Maner, J. K. (2009). Embodiment, evolution, and social cognition: An in
tegrative framework. European Journal of Social Psychology, 39, 1236–1244.
Kaschak, M. P., Zwaan, R. A., Aveyard, M. E., & Yaxley, R. H. (2006). Perception of audito
ry motion affects language processing. Cognitive Science, 30, 733–744.
Kaup, B., Ludtke, J., & Zwaan, R. A. (2006). Processing negated sentences with contradic
tory predicates: Is a door that is not open mentally closed? Journal of Pragmatics, 38,
1033–1050.
Kaup, B., Yaxley, R. H., Madden, C. J., Zwaan, R. A., & Ludtke, J. (2007). Experiential sim
ulations of negated text information. Quarterly Journal of Experimental Psychology, 60,
976–990.
Lakoff, G., & Johnson, M. (1980). Metaphors we live by. Chicago: University of Chicago
Press.
Levinson, S. C. (2000). Presumptive meanings. Cambridge, MA: MIT Press.
Louwerse, M. M. (2008). Embodied relations are encoded in language. Psychonomic Bul

letin & Review, 15, 838–844.
MacIver, M. A. (2009). Neuroethology: From morphological computation to planning. In P.

Robbins & M. Aydede (Eds.), Cambridge handbook of situated cognition (pp. 480–504).
Marques, J. F. (2006). Specialization and semantic organization: Evidence for multiple se
mantics linked to sensory modalities. Memory & Cognition, 34, 60–67.
Masson, M., Bub, D., & Warren, C. (2008). Kicking calculators: Contribution of embodied
representations to sentence comprehension. Journal of Memory and Language, 59, 256–
265.
Page 19 of 23

Martin, A. (2007). The representation of object concepts in the brain. Annual Review of
Meteyard, L., Bahrami, B., & Vigliocco, G. (2007). Motion detection and motion words:
Language affects low-level visual perception. Psychological Science, 18, 1007–1013.
McGlone, M. (2011). Hyperbole, homunculi, and hindsight bias: An alternative evaluation

of conceptual metaphor theory. Discourse Processes, 48, 563–574.
Niedenthal, P. M., Barsalou, L. W., Winkielman, P., Krauth-Gruber, S., & Ric, F. (2005). Em
bodiment in attitudes, social perception, and emotion. Personality and Social Psychology
Review, 9, 184–211.
Paulin, M. G. (2005). Evolutionary origins and principles of distributed neural computa

tion for state estimation and movement control in vertebrates. Complexity, 10, 56–65.
Pecher, D., & Boot, I. (2011). Numbers in space: Differences between concrete
(p. 329)
and abstract situations. Frontiers in Psychology, doi: 10.3389/fpsyg.2011.00121
Pecher, D., Zeelenberg, R., & Barsalou, L. W. (2003). Verifying different modality proper
ties for concepts produces switching costs. Psychological Science, 14, 119–125.
Pecher, D., Zeelenberg, R., & Barsalou, L. W. (2004). Sensorimotor simulations underlie
conceptual representations: Modality-specific effects of prior activation. Psychonomic Bul
letin & Review, 11, 164–167.
Pickering, M. J., & Garrod, S. (2013). An integrated theory of language production and
comprehension. Behavioral and Brain Sciences, 36, 329–347.
Pinker, S. (1994). The language instinct. New York: Harper Collins.
Pulvermüller, F. (1999). Words in the brain’s language. Behavioral and Brain Sciences, 22,
253–279.
Rachlin, H. (1991). Introduction to modern behaviorism. W. H. Freeman & Company.
Richardson, D. C., & Matlock, T. (2007). The integration of figurative language and static
depictions: An eye movement study of fictive motion. Cognition, 102, 129–138.
Richter, T., & Zwaan, R. A. (2009). Processing of color words activates color representa
tions. Cognition, 111, 383–389.
Rizzolatti, G., Sinigaglia, C., & Anderson, F. (2008). Mirrors in the brain: How our minds
share actions and emotions. Oxford University Press.
Robbins, P., & Aydede, M. (Eds.). (2009). Cambridge handbook of situated cognition. Cam
Page 20 of 23

Rueschemeyer, S-A., Glenberg, A. M., Kaschak, M. P., Mueller, K., & Friederici, A. D.
(2010). Top-down and bottom-up contributions to understanding sentences describing ob
jects in motion. Frontiers in Psychology, doi: 10.3389/fpsyg.2010.00183
Santiago, J., Lupianez, J., Perez, E., & Fuenes, M. J. (2007). Time (also) flies from left to
right. Psychonomic Bulletin and Review, 14, 512–516.
Santiago, J., Roman, A., Ouellet, M., Rodriguez, N., & Perez-Azor, P. (2010). In hindsight,
life flows from left to right. Psychological Research, 74, 59–70.
Schubert, T. W. (2004). The power in your hand: Gender differences in bodily feedback
from making a fist. Personality and Social Psychology Bulletin, 30, 757–769.
Schubert, T. W. (2005). Your highness: Vertical positions as perceptual symbols of power.

Searle, J. R. (1980). Minds, brains, and computers. Behavioral and Brain Sciences, 3,
417-457.
Sell, A. J., & Kaschak, M. P. (2011). Processing time shifts affects the execution of motor
responses. Brain and Language, 117, 39–44.
Sell, A. J., & Kaschak, M. P. (2012). The comprehension of sentences involving quantity in
formation affects responses on the up-down axis. Psychonomic Bulletin & Review19, 708–
714.
Shaki, S., & Fischer, M. H. (2008). Reading space into numbers—a cross-linguistic com
parison of the SNARC effect. Cognition, 108, 590–599.
Simmons, W. K., Martin, A., & Barsalou, L. W. (2005). Pictures of appetizing foods activate
gustatory cortices for taste and reward. Cerebral Cortex, 15, 1602–1608.
Simmons, W. K., Ramjee, V., Beauchamp, M. S., McRae, K., Martin, A., & Barsalou, L. W.
(2007). Common neural substrates for perceiving and knowing about color and action.
Neuropsychologia, 45, 2802–2810.
Spence, C., Nicholls, M. E. R., & Driver, J. (2000). The cost of expecting events in the
wrong sensory modality. Perception & Psychophysics, 63, 330–336
Stanfield, R. A., & Zwaan, R. A. (2001). The effect of implied orientation derived from ver
bal context on picture recognition. Psychological Science, 12, 153–156.
Strack, F., Martin, L. L., & Stepper, S. (1988). Inhibiting and facilitating conditions of the
human smile: A nonobtrusive test of the facial feedback hypothesis. Journal of Personality
Taylor, L. J., & Zwaan, R. A. (2008). Motor resonance and linguistic focus. Quarterly Jour
nal of Experimental Psychology, 61, 896–904.
Page 21 of 23

Tettamanti, M., Buccino, G., Saccuman, M. C., Gallese, V., Danna, M., Scifo, P., et al.
(2005). Listening to action-related sentences activates fronto-parietal motor circuits. Jour
nal of Cognitive Neuroscience, 17, 273–281.
Tomasello, M. (1999). The cultural origins of cognition. Harvard University Press.
Torralbo, A., Santiago, J., & Lupianez, J. (2006). Flexible conceptual projection of time on
to spatial frames of reference. Cognitive Science, 30, 745–757.
Ulrich, R., Eikmeier, V., de la Vega, I., Rernandez, S. R., Alex-Ruf, S., & Maienborn, C.
(2012). With the past behind and the future ahead: Back-to-front representation of past
and future sentences. Memory and Cognition, 40, 483–495.
Ulrich, R., & Maienborn, C. (2010). Left-right coding of past and future in language: The
mental timeline during sentence processing. Cognition, 117, 126–138.
Vermeulen, N., Niedenthal, P. M., & Luminet, O. (2007). Switching between sensory and
affective systems incurs processing costs. Cognitive Science, 31, 183–192.
Williams, L. E., & Bargh, J. A. (2008). Experiencing physical warmth promotes interper
sonal warmth. Science, 322, 606–607.
Wolpert, D. M., Doya, K., & Kawato, M. (2003). A unifying computational framework for
motor control and social interaction. Philosophical Transactions of the Royal Society of
London, B, Biological Sciences, 358, 593–602.
Wolpert, D. M., Ghahramani, Z., & Flanagan, J. R. (2001). Perspectives and problems in
motor learning. Trends in Cognitive Sciences, 5, 487–494.
Yaxley, R. H., & Zwaan, R. A. (2007). Simulating visibility during language comprehen
sion. Cognition, 105, 229–236.
Zwaan, R. A., Madden, C. J., Yaxley, R. H., & Aveyard, M. E. (2004). Moving words: Dy
namic mental representations in language comprehension. Cognitive Science, 28, 611–
619.
Zwaan, R. A., Stanfield, R. A., & Yaxley, R. H. (2002). Language comprehenders mentally
represent the shape of objects. Psychological Science, 13, 168–171.
Zwaan, R. A., & Taylor, L. J. (2006). Seeing, acting, understanding: Motor resonance in
language comprehension. Journal of Experimental Psychology: General, 135, 1–11.
Zwaan, R. A., Taylor, L. J., & De Boer, M. (2010). Motor resonance as a function of narra
tive time: Further tests of the linguistic focus hypothesis. Brain and Language, 112, 143–
149.
Michael P. Kaschak
Page 22 of 23

Michael P. Kaschak, Department of Psychology, Florida State University, Tallahassee,

FL
John L. Jones
John L. Jones, Florida State University.
Page 23 of 23

Literal Versus Nonliteral Language: Novelty Matters

Rachel Giora

This chapter presents evidence demonstrating that novelty rather than nonliteralness
matters, whether in terms of cost or effects. With regard to costs, highly and optimally in
novative stimuli take longer to process than familiar ones, regardless of degree of nonlit
eralness. This has been shown to be true of various populations, whether typically or
atypically developing. With regard to division of labor, familiar stimuli tend to engage left
hemisphere regions, whereas optimal innovations tend to engage right hemisphere areas,
regardless of degree of nonliteralness. In terms of effects, optimal innovations are more
pleasing than other (familiar and unfamiliar) alternatives, regardless of degree of nonlit
eralness. Some social factors play a role, too. For instance, optimally innovative sarcasm
is most enjoyable when taking the speaker’s rather than the victim’s perspective or when
outgroup members are derided by ingroup members.
Keywords: Novelty, optimal innovations, nonliteralness, sarcasm, processing, pleasure, hemispheres, group rela
tions, atypically developing
Nonliteral language has a lasting reputation for being highly unique— “the mark of ge
nius” and “above the commonplace and mean.” In contrast, literal language—“the use of
proper words”—is taken to render language “perspicuous” (Aristotle, 350 BCE). More re
cent research, however, has witnessed the pervasiveness of nonliteral language in mun
dane uses (Gibbs, 1994; Lakoff & Johnson, 1980; Lakoff & Turner, 1989, to cite but a few).
Still, the poetics of everyday language (e.g., the nonliteralness of the “legs” and the
“back” of a chair) is often lost on comprehenders on account of its lack of novelty. At the
same time, however, our passion for novelty (Berlyne, 1950; Freud, 1922; McDougall,
1923; Sheldrake, 1981/2012) forgoes the possibility that literal language may be also po
etic.
Consider, for instance, how highly conventionalized nonliteral expressions1 such as big
shoes to fill or curl up and die may become innovative by resorting to their alternative lit
eral interpretation, as in “Those are some very big shoes to fill (size 83, in fact)”2 or “curl
up and dye.”3 Consider also the literal use of “know” in KNOW Pinkwashing,4 which is
used in an innovative way, without resorting to figurativeness. While highlighting both the
Page 1 of 29

literal request to become acquainted with pinkwashing and the literal need to reject it,
this use deautomatizes or rather defamiliarizes the familiar literal meaning of “know” by
fragmenting it, thus bringing to the fore an additional sense of the word (which now in
cludes rejection).5 Note further the innovativeness of the literal “know” in the street art
of KNOW HOPE,6 which deautomatizes the sense of “no” while breaking up a familiar lit
eral collocation7 (No hope), thus enriching it with optimism.
Fig. 21.1 Sharon stone.
Consider further how (visual) background information in Lahav Halevy’s (2002) art8
deautomatizes the all too familiar name Sharon Stone (see (p. 331) Figure 21.1). Decon
structing a fixed expression by introducing an alternative referent for Sharon and a literal
meaning of Stone results in a novel literal interpretation.
Fig. 21.2 Hebron, December, 2008
(© 2014 Haaretz, Israel).
Can conventional metaphors be also revamped? Consider how, in what Levac (2008)
caught on camera, including the facial expression of the Palestinian, the all too well-
known smiley is deautomatized, rendered into an innovative “cryley” (Figure 21.2).
Page 2 of 29

Fig. 21.3 Iron and blood.
(© 2014 Artists Rights Society (ARS), New York /VG

Bild-Kunst, Bonn)
Can innovativeness be both literal and metaphorical? In Figure 21.3, using photomontage
techniques, Heartfield (1934) spelled out his anti-Nazi views by deautomatizing a conven
tional Nazi symbol—the Swastika—turning its arms into the executioners’ blood-dripping
axes. Although the novelty is “literal” in that there is no visual “semantic
anomaly” (known to typify metaphors), it nonetheless calls for a metaphorical interpreta
tion.
Page 3 of 29

Fig. 21.4 Adolf the superman: Swallows gold and

spouts junk.
(© 2014 Artists Rights Society (ARS), New York /VG

Bild-Kunst, Bonn)
Similarly, in Figure 21.4, Heartfield (1932) realizes a metaphor (“swallows gold”). As a re
sult, the novel literal reading involves a visual “semantic anomaly” that can only be inter
preted metaphorically (expressing the artist’s opposition to Hitler’s gain-based regime).
Fig. 21.5 Israeli soldiers in Hebron 2008.
(© 2014 Haaretz, Israel)
Can the realistic also be viewed as nonliteral? In addition to Figure 21.2, consider how in
Levac (2012; see Figure 21.5) real-life background information—the protest slogans “Free
Palestine” engraved on the road blocks the Israeli soldiers are leaning against—ridicule
the perceptual figure (the soldiers) by setting up a stark contrast between the two, allow
ing a novel, sarcastic reading of the percept as a whole:
Page 4 of 29

These instances suggest that it is innovativeness rather than nonliteralness that might
make a difference. It is via novelty then that both literal and nonliteral communication
could be “the mark of genius.” (p. 332) (p. 333)
Novelty Matters: The Case of Comprehen

(p. 334)
sion
Behavioral Level
According to Aristotle (350 BCE), metaphors (among other “unusual” uses) are difficult to
understand: a text or style is “a riddle, if it [mostly] consists of metaphors.” This view has
been adopted by a number of processing models, not least by the standard pragmatic
model (e.g., Grice, 1975; Searle, 1979). It has, however, been vastly questioned by various
scholars trying to show that, given various factors, especially specific contextual informa
tion or semantic space, metaphor will be no more difficult to make sense of than literal
counterparts (Gibbs, 1994; Glucksberg, 2001; Glucksberg & Keysar, 1990; Keysar, 1989;
Kintsch & Bowles, 2002, among others; for a review, see Giora, 2003).
Contextual considerations momentarily set aside (but see the section “Optimal Novelty in
Context: The Case of Sarcasm”), experimental evidence accumulated in various laborato
ries shows that it is not degree of nonliteralness that makes a difference but degree of
novelty; it is the latter that affects processing difficulties (Giora, 1997, 2003; see also
Gibbs, 1980), degree of pleasurability (Brône & Coulson, 2010; Giora, Fein, Kronrod, El
natan, Shuval, & Zur, 2004), and the neural substrates involved (Giora, 2009; Giora &
Stringaris, 2010).
“Degree of nonliteralness” pertains here to the continuum ranging between literal and
nonliteral (idiomatic, metaphoric, ironic/sarcastic, hyperbolic, etc.) language. “Degree of
novelty” pertains here to degree of salience.
According to the graded salience hypothesis (Giora, 1997, 1999, 2003), a meaning is
salient if it is coded in the mental lexicon and enjoys prominence due to exposure (e.g., fa
miliarity, frequency, conventionality) or cognitive factors (e.g., prototypicality), regardless
of degree of nonliteralness (e.g., the nonliteral curl up and die, the literal body and soul,
the literal and metaphoric sharp, the sarcastic big deal, read my lips, the literal plant
meaning of tree). A meaning is less-salient if it is coded but scores low on these variables,
regardless of degree of nonliteralness (e.g., the literal, riverside meaning of bank; the
nonliteral diagram meaning of tree). A meaning or an interpretation that is not coded is
nonsalient; it is novel or derived, regardless of degree of nonliteralness (e.g., the literal
body and sole; the nonliteral weapons of mass construction).
An interpretation based on the salient meanings of the utterance components, termed

salience-based interpretation, is not coded either. Based on constituents’ salient mean
ings, it too is indifferent to degree of nonliteralness. Consider the literal interpretation of
Page 5 of 29

“What a lovely day for a picnic,” intended sarcastically when said during a wet and wild
weather. This literal interpretation is based on the salient, here literal, meanings of the
individual constituents (e.g., lovely) that make up the utterance (see Giora et al., 2007).
Such is also the metaphoric interpretation of “This one’s really sharp,” when said of an
unintelligent student and intended sarcastically, which is based on the salient, here non
literal, meanings of the individual constituents (e.g., sharp) that make up the utterance
(see Colston & Gibbs, 2002). Such is also the sarcastic interpretation of the nonliteral
weapons of mass construction,9 which is based on the more familiar nonliteral (weapons
of mass distraction) and literal (weapons of mass destruction) collocations (Giora et al.,
2004).
With regard to processing, salient and less-salient meanings are activated unconditional
ly. However, access is ordered: more salient meanings are activated first. Salience-based
interpretations, albeit noncoded, are easier to process (“this one is really sharp” said of a
bright student) compared to nonsalient interpretations not based on the salient meanings
of their components (“this one is really sharp” said of a knife that doesn’t cut at all, see
Colston & Gibbs, 2002).
Would any degree of novelty make a difference either processing-wise or aesthetics-wise?

According to Giora et al. (2004), it is “optimal innovativeness” rather than other degrees
of innovation that is singled out in these respects. To be optimally innovative, a stimulus
should introduce a nonsalient, novel interpretation (a peace of paper) while invoking a
salient meaning of a familiar stimulus (a piece of paper). Such innovations are aestheti
cally favored over both a “familiar” alternative (a piece of paper), a “pure” innovation (a
pill of pepper), or a “variant” version (a single piece of paper). Optimal innovations may
be literal (curl up and dye), invoking a conventional collocation (here idiomatic—curl up
and die); metaphoric (pinkwashing), invoking a conventional expression (here metaphoric
—whitewashing); or sarcastic (Einstein? I don’t think so!), echoing a familiar expression
or thought (here metaphorically attributing to one Einstein’s ingenuity) while projecting a
dissociative, ridiculing stance (Du Bois, 2007; Sperber & Wilson, 1986/1995).
Optimal innovations, however, take longer to process than their familiar counterparts be
cause they involve activating additional (salient) meanings (p. 335) on top of the intended
novel one. However, they are most pleasurable, regardless of whether they are literal or
nonliteral, verbal or nonverbal (see also Hekkert, Snelders, & van Wieringen [2003]; but
see Bohrn, Altmann, Lubrich, Menninghaus, & Jacobs [2012] who show that although
pleasure ratings did not single out proverbial optimal innovations compared to other al
ternatives, their neural substrates did, indicating activation of the affect-related medial
prefrontal and medial temporal regions). Next in pleasurability are familiar expressions;
least pleasing are variant versions and (even less so) pure innovations (see Giora et al.,
2004).10
And if degree of novelty matters, would any number of evoked salient meanings matter?
According to the account of verbal wittiness as consisting in double grounding (Brône &
Coulson, 2010), pleasure derived from optimal innovations increases when optimally inno
Page 6 of 29

vative utterances are “double grounded.” What does double grounding mean? Consider,
for instance, the reference to Bridgestone—a company that manufactures car tires—in
“U.S. slowdown punctures Bridgestone’s profits.” Here the novel metaphor “puncture” is
double grounded in that its activated salient literal meaning is relevant to what is said in
more than one way. On the one hand, it is related to the profits in question. On the other
hand, it is relevant to the nature of the company, thereby triggering a number of contex
tual effects (see Sperber & Wilson, 1986/1995). In contrast, given that Cold Stone is an
ice-cream company, the same metaphor (“puncture”) in “U.S. slowdown punctures Cold
Stone’s profits” is single grounded, its salient literal meaning being relevant in only one
way.
Pleasure derived from double-grounded optimal innovations, however, comes with a cost.
Double-grounded metaphors are more taxing, taking longer to process than single-
grounded ones.
Studies focusing on nonliteral language only also demonstrate the importance of novelty.
For instance, when testing idiomatic language, findings reveal that high-salience and low-
salience idioms involve different processing routes. Whereas shorter response times and
small N400 amplitudes were observed for the last word of highly salient idioms, indicat
ing no processing difficulties, the last word of idioms low on salience prompted an in
creased N400 amplitude and longer response times, disclosing processing difficulties
(Laurent, Denhières, Passerieux, Iakimovac, & Hardy-Baylé, 2006). Similarly, studies fo
cusing on metaphor comprehension also indicate different processing routes for novel
compared to familiar metaphors. Whereas nonsalient, novel metaphors were interpreted
via comparison processes, this trend was shifted to categorization when these items un
derwent conventionalization (Bowdle & Gentner, 2005).
Differences between novel and familiar metaphors were also demonstrated for typically
developing children. Although whether the metaphoric stimuli were visual or verbal did
not make a difference, whether they were familiar or novel did (Mashal & Kasirer, 2012).
And when younger and older individuals were concerned, results showed that age-related
changes were associated with the appreciation of the novelty of the metaphors. Novelty
but not familiarity affected the appreciation of the plausibility of metaphors (Mashal,
Gavrieli, & Kavé, 2011).
Individual Differences?
Longer reading times for optimal innovations, whether literal (Tverian Horse11 invoking
Trojan horse) or metaphoric (Dying Star, invoking A Born Star—a popular Israeli TV
show), compared to conventional collocations, whether literal (wooden table) or
metaphoric (flower bed), have been also attested to for different groups (Giora, Gazal,
Goldstein, Fein, & Stringaris, 2012). In Giora et al. (2012), individuals with Asperger’s
syndrome and typically developing individuals were asked to read such optimal innova
tions and conventional collocations for comprehension. Materials were presented inside
and outside context. And although the experimental group took longer to read the targets
Page 7 of 29

overall, both groups exhibited similar patterns of behavior. Both groups took longer to
read novel items compared to familiar ones, regardless of degree of nonliteralness, thus
demonstrating that it is novelty rather than degree of nonliteralness that is effort-con
suming.
In Gold and Faust (2012), too, individuals with Asperger’s syndrome and typically devel
oping individuals took longer to process novel metaphors compared to conventional ones.
In Colich et al. (2012), who looked into sarcasm interpretation, both the experimental
group with autism spectrum disorder (ASD) and the control group exhibited similar pat
terns of behavior. Although response times to targets were significantly faster among par
ticipants with ASD compared to control participants, both groups took longer to process
optimally innovative sarcastic utterances than (salience-based) literal ones. When com
paring processing of salient versus nonsalient stimuli, different groups may exhibit simi
lar patterns of behavior, suggesting that (p. 336) it is degree of novelty rather than nonlit
eralness that matters (as will also be shown in the next section).
Biological Level
Research into the neural substrates of nonliteral language (whether with or without con
textual support) also suggests that the relevant distinction that accounts for hemispheric
division of labor has to do with degree of salience/novelty rather than with degree of non
literalness (Giora, 2003, 2007). For instance, lesion studies (e.g., Giora et al., 2000; Ka
plan et al., 1990; Winner & Gardner, 1977) and studies of individuals with Alzheimer’s
disease (Amanzio, Geminiani, Leotta, & Cappa, 2008), as well as studies involving healthy
adults (Eviatar & Just 2006; Kacinik & Chiarello, 2007) demonstrate that processing non
salient (novel) nonliteral (sarcastic, metaphoric) interpretations relies more heavily on
the right hemisphere (RH) than on the left hemisphere (LH). In contrast, processing con
ventional meanings (both literal and nonliteral) engages the LH to a greater extent than
the RH (Lee & Dapretto, 2006). This has been replicated using a variety of technologies.
Thus, results obtained from functional magnetic resonance imaging (fMRI), divide visual
field (DVF), repetitive transcranial magnetic stimulation (rTMS), and event-related poten
tial (ERP) studies demonstrate increased activation of RH areas during processing of opti
mally novel interpretations of figurative language (Arzouan, Goldstein, & Faust, 2007;
Bottini et al., 1994; Cardillo, Watson, Schmidt, Kranjec, & Chatterjee, 2012; Faust &
Mashal, 2007; Mashal & Faust, 2008; Mashal, Faust, & Hendler, 2005; Mashal, Faust,
Hendler, & Jung-Beeman, 2007, 2009; Pobric, Mashal, Faust, & Lavidor, 2008; Shamay-
Tsoory, Tomer, & Aharon-Peretz, 2005) as well as optimally novel interpretations of literal
language—for example, the compositional interpretations of idioms (Mashal et al., 2009).
And whereas RH advantage was demonstrated in processing nonsalient interpretations of
metaphors during first exposure, repeated exposure benefited the LH (Mashal & Faust,
2009) as would be predicted by the graded salience hypothesis (Giora, 1997, 2003) and
the career of metaphor model (Bowdle & Gentner, 2005). A more recent finding involving
conventionalization of novel metaphors, however, indicates a bilaterally mediated process
rather than a hemispheric shift across brain areas (Cardillo et al., 2012; for a different
Page 8 of 29

view showing that both novelty and figurativeness engage the RH, see Coulson & van Pet
ten, 2007; Diaz, Barrett, & Hogstrom, 2011).
How Optimal Innovations Shape Their Environ

ment: The Case of Production
Optimal innovations deautomatize salient meanings, whether literal or nonliteral. Their
environment may therefore “resonate” with these salient meanings (Giora, 2011a). Ac
cording to Du Bois (2012), “resonance” pertains to “the activation of affinities across ut
terances.” As shown by Giora (2007), resonance involves echoing utterances’ salient
meanings and salience-based interpretations by both prior and subsequent context. This
should be true even if these meanings and interpretations are contextually inappropriate
(e.g., the literal meanings or interpretations of metaphors).
As illustration, consider the following example (1) in which fire and its associates are used
both conventionally (here metaphorically, related to shooting; in italics for convenience)
and innovatively (here metaphorically, related to killing via food and water shortage; in
bold for convenience):
(1) Let’s be clear about this: Israel’s fire at Gaza has not ceased. There is no Is
raeli ceasefire in Gaza. There is no Israeli ceasefire even when Israel’s soldiers
aren’t shooting a single bullet in Gaza….
There are food shortages in Gaza. Israel is denying Gaza food….
Food shortages kill. Denying food is fire.
There’s a shortage of potable water in Gaza….
Water shortages kill. Denying water is fire.
Both Gaza and the West Bank will go on igniting under fire, till they kindle
Sderot [an Israeli city] again too. The bullet-less fire that Israel is shooting at
the dispossessed of Gaza is fire that it is also shooting, by proxy, at the dispos
sessed of Sderot.
(Mazali, 2006)
If we focus on the optimally innovative fire metaphors (in bold)—“Food shortages kill.
Denying food is fire….Water shortages kill. Denying water is fire”—it is obvious that they
resonate with their conventional uses in both their early and late contexts. In early con
text, fire is used conventionally to indicate using arms (against the Gazans): “Israel’s fire
at Gaza has not ceased. There is no Israeli ceasefire in Gaza.” In late context, that which
follows these novel uses, an additional novel metaphor emerges, constructed on the basis
of this and yet another salient metaphoric meaning, referring to anger as heat (see Lakoff
& Kövecses, 1987): “Both Gaza and the West Bank will go on igniting under fire.” Here
Page 9 of 29

“igniting under fire” is a double- or even triple-grounded metaphor; it has first to do with
the (p. 337) (salient, metaphoric) rage against oppression which, in turn, resonates with
its salient literal meaning (“kindle”) and then with its salient metaphoric meaning (“bul
let-less”).
Deautomatizing salient meanings via echoing them may render innovative a conventional
use. Consider the following example (2) in which prior literal information “smeared with
mud” deautomatizes the conventional metaphor “dirty” that follows the headline (and
possibly renders metaphoric the literal interpretation of a “face smeared with mud” as
well):
(2) Imagine this face smeared with mud
Until today, Foreign Minister, Tzipi Livni, managed to avoid
dirty political confrontations.
(Mualem, 2007)
In (3) (a translation of the Hebrew version), which is basically sarcastic, the metaphorical
optimal innovation “on the wings of danger” (which in Hebrew literally reads carried up
[to power] on the wings of danger) activates the salient Hebrew idiom carried up on the
wings of fame/glory. This metaphorical innovation is related to the next conventional
metaphor clip his wings (which in Hebrew literally reads “collapse/force down his wings,”
meaning “discourage”) via the salient literal meaning of the metaphoric “wings.” Where
as in the optimally innovative metaphor the “wings” are not those of whoever is carried
on them, in the (Hebrew) metaphor that follows— “clip his wings”—they are. In fact, both
metaphors rely on different affordabilities of “wings” and don’t share a metaphorical
sense. Still, juxtaposing these two metaphors deautomatizes them by foregrounding the
salient literal meaning of “wings” that they share, allowing both to be double grounded, if
not poetic:
(3) The rule of fear
So what is possible? There is no response. Because what would those who make
threats do without their threats? They may very well be able to deal with Israel’s
real problems, like its moral self-image, its refusal to make peace and its socioeco
nomic difficulties. But this is not why Netanyahu was elected. He came to power
on the wings of danger. Just don’t try to clip his wings lest, God forbid, this
stirs hope.
(Levy, March 2009)
Page 10 of 29

Can an Act Be an Optimal Innovation, Whether

Literal or Nonliteral?
Fig. 21.6 Metzer-Messer land exchange.
Can one aim at getting across an optimal innovative message, whether literal or nonliter
al, by tampering with the reality? To answer this question, let’s focus on a few examples.
Consider first the following example (Figure 21.6), titled “Metzer-Messer land
exchange.”12 This act enacts a literal interpretation of “land exchange” focusing on an al
ternative literal meaning of “land,” which in Hebrew also means “soil,” and on an alterna
tive literal meaning of “exchange.” The more familiar use of the collocation, which is
deautomatized in Figure 21.6, becomes highly frequent any time “peace” talks or discus
sions of a prospective agreement between Palestinians and Israelis are on the agenda. In
such “peace” talks, “land exchange” is used euphemistically to refer to population ex
change, or, put more bluntly, to ethnic cleansing. By enacting an optimally innovative lit
eral interpretation of the collocation, the artist Micha Ullman (1972)13 registers his objec
tion to its conventional meaning, projecting thereby a sarcastic viewpoint that ridicules
the idea.
Page 11 of 29

Fig. 21.7 Protesting the Israeli war on Gaza.
Consider further the example in Figure 21.7. While protesting the Israeli war on Gaza
(December 27, 2008—January 18, 2009), Anarchists Against the Wall—a direct action
group of Israeli activists—made use of metaphorical acts to allow Israelis an experiential
insight into the plight of the Palestinian citizens of Gaza. In Figure 21.7, they personify
dead children of Gaza scattered across the road.14 This act takes place at the entrance to
Sde Dov Air Force base (in Tel Aviv), thereby also blocking entry to the base. This action
thus sends both a literal and a metaphorical message.15
Fig. 21.8 The Anarchists’ military roadblock in Tel

Aviv.
In Figure 21.8, the Anarchists set up a “military” roadblock in Tel Aviv so that Israeli citi
zens experience, even if momentarily, what it is like to be under Israeli occupation and
control, which is what Palestinians experience on a daily if not on a moment-by-moment
basis.
Acts too, then, can be optimally innovative, both literally and nonliterally.
Page 12 of 29

Optimal Novelty in Context: The Case of Sar

casm
Sarcasm (or sarcastic irony), if not conventionalized, is usually rendered optimally innov
ative by means of context. It is in context that it also meets the conditions for optimal in
novation (Giora et al., 2004). Indeed, the interpretation of nonlexicalized sarcasm is con
structed on the fly—derived on the basis of sarcastically biasing information. Sarcastic re
marks, when understood, entertain both the novel/nonsalient sarcastic interpretation as
well as the salient meaning of the sarcastic key word(s) and the salience-based interpre
tation of the utterance (p. 338) as a whole (see Giora, 1997, 2003; Giora & Fein, 1999). Is
sarcasm, then, pleasing? Is it more pleasing than an alternative, more familiar, or
salience-based interpretation, as would follow from the optimal innovation view (Giora et
al., 2004)? Specifically, is interpreting a sarcastic remark embedded in a strongly sup
portive context more taxing than interpreting its similarly supported salience-based inter
pretation, as would be predicted by the optimal innovation view (Giora et al., 2004) and
by the graded salience hypothesis (Giora, 1997, 1999, 2003)?
Is Sarcasm Pleasing?
Sarcasm and pleasure have not been as directly assessed as were metaphors and plea
sure (for novel metaphors being more pleasing than their salience-based, often literal in
terpretation, see Giora et al., 2004; Giora, Fein, Kotler, & Shuval, forthcoming; Shuval &
Giora, 2009). However, given that sarcasm is intended and considered as humorous
(Dress, Kreuz, Link, & Caucci, 2008; Lampert & Ervin-Tripp, 2006; Leggitt & Gibbs,
2000), there is probably some way in which it is pleasing, especially if one is not the tar
get of the sarcastic remark. (p. 339)
Indeed, van Mulken, Burgers, and van der Plas (2010) found that appreciation of
(p. 340)
sarcasm depended on whether the sarcastic intent was recognized, its message accept
able, and its target excluded the present company’s group members. More recent results
(Drucker, Fein, Bergerbest, & Giora, in press) somewhat support this view. They show
that women scoring low on a sexism scale enjoyed sarcasm most when it was directed by
women against men; they enjoyed it least when it was directed by women against women.
In contrast, men scoring as low on a sexism scale did not enjoy sarcasm directed against
women as much as they did when it was directed against men, specifically by other men.
It seems that, for a powerful group, competing with ingroup members is more enjoyable
than competing with less powerful outgroup members (Drucker et al., in press). Indeed,
as shown by Bowes and Katz (2011), (self-reported) aggressors perceived their sarcastic
comments as more humorous and less aggressive than did their victims. Taking the per
spective of the aggressor, then, allows one to dismiss sarcastic comments as teasing. This
ties in well with previous findings showing that aggressive people are less sensitive to the
negativity displayed in their sarcasm (Blasko & Kazmerski, 2006; Werner & Nixon, 2005).
Page 13 of 29

Sarcasm, then, is pleasing, as would be predicted by the optimal innovation view, but on
condition that one takes the perspective of the aggressor rather than the victim’s.
Is Interpreting Sarcasm Costly?
Does (nonlexicalized) sarcasm come with a cost, as would be predicted by the optimal in
novation view? An enduring question debated in the field of nonliteral language is
whether strong contextual support may affect a smooth processing of nonsalient sarcasm
so that it is no more difficult to comprehend than its salience-based (often literal) inter
pretation when embedded in a strongly supportive context. Before trying to deal with this
question, one should first clarify the notion of strong contextual support. What context,
then, would count as strong enough to make a sarcastic remark easy to understand?
According to contextualists, a context that might facilitate sarcasm interpretation initially

is one that induces an expectation for a sarcastic remark. In one view, anticipating a sar
castic remark may be prompted by setting up a contrast between a protagonist’s expecta
tion and the reality that frustrates it (Gibbs, 1986, 2002). It may also be induced by a con
text involving multiple constraints that may cue the addressee as to the speaker’s sarcas
tic intent. For instance, in addition to failed expectation, context may include information
about a speaker’s occupation or involve proliferated uses of negations, negative emo
tions, or a victim (Campbell & Katz, 2012; Pexman, Ferretti, & Katz, 2000). Note, howev
er, that although, as shown by Campbell and Katz (2012), such hints prevail in the con
texts of sarcasm produced by their participants, the authors conclude that they do not
make up a set of necessary conditions, nor is any of them, on its own, a necessary condi
tion.
Indeed, as shown by Giora, Fein, Kaufman, Eisenberg, and Erez (2009), a context featur
ing a frustrated expectation on its own (4) neither favored a sarcastic over a nonsarcastic
remark compared to a context featuring a realized expectation (5) nor did it facilitate sar
casm comprehension compared to a context featuring a realized expectation. In fact, in
both contexts, sarcasm took longer to read than in a context featuring no expectation (6).
(4) Frustrated expectation
Context:
Sagee went on a ski vacation abroad. He really likes vacations that include sport
activities. A relaxed vacation in a quiet ski-resort place looked like the right thing
for him. Before leaving, he made sure he had all the equipment and even took
training classes on a ski simulator. But already at the beginning of the second day
he lost balance, fell, and broke his shoulder. He spent the rest of the time in a lo
cal hospital ward feeling bored and missing home. When he got back home, his
shoulder still in cast, he said to his fellow workers:
Sarcastic target sentence:
“Ski vacation is recommended for your health.”

Page 14 of 29

Final sentence:
Everyone smiled.
Comprehension question:
Do you think Sagee will go for a ski vacation again?
(5) Realized expectation
Context:
Sagee went on a ski vacation abroad. He doesn’t even like skiing. It looks danger
ous to him and staying in such a cold place doesn’t feel like a vacation at all. But
his girlfriend wanted to go and asked him to join her. Already at the beginning of
the second day he lost balance, fell, and broke his shoulder. He spent the rest of
the time in a local hospital ward feeling bored and missing home. When he got
back home, his shoulder still in cast, he said to his fellow workers:
(p. 341) Sarcastic target sentence:
Final sentence:
Everyone smiled.
(6) No-expectation
Context:
Sagee went on a ski vacation abroad. He has never practiced skiing so it was his
first time. He wasn’t sure whether he would be able to learn to ski and whether he
will handle the weather. The minute he got there, he understood it was a great
thing for him. He learned how to ski in no time and enjoyed it a lot. Besides, the
weather was nice and the atmosphere relaxed. When he got back home, he said to
his fellow workers:
Literal target sentence:
Final sentence:
Everyone smiled.
Page 15 of 29

But even when contexts do, in fact, induce an expectation for a sarcastic utterance, find
ings fail to demonstrate initial facilitation of sarcastic interpretation. Instead, most of the
findings in the field show that initial processing of sarcastic remarks involves activating
their contextually inappropriate, salience-based interpretation first.
For instance, Giora et al. (2007) showed that when contexts induced an expectation for a
sarcastic utterance, such expectation did not facilitate sarcastic interpretation initially. In
fact, it was always the contextually inappropriate, salience-based interpretation that was
facilitated first. This has been shown when contexts were dialogues featuring a sarcastic
speaker (shown to prompt an expectation for another such turn on that speaker’s behalf).
This has also been replicated in studies in which participants read only items that ended
in a sarcastic comment (the +Expectation condition) as opposed to those studies in which
half of the items ended in a sarcastic remarks and half in a nonsarcastic remark (the −Ex
pectation condition).
In the dialogues experiment, target utterances biased toward the sarcastic interpretation
took longer to read than those biased toward the nonsarcastic ones; in the +/−Expecta
tion experiments, response times to sarcastically related probes were always longer than
those to salience-based related probes, which were always activated initially, regardless
of interstimulus interval (ISI) (750–1,000 ms) or strength of expectation.
In a series of follow-up studies, in which contextual strength was reinforced by multiply

ing sarcastic cues, these results remained nonetheless constant. In, Fein, Yeari, and Giora
(2013; for an initial report, see also Giora, 2011b), dialogues similar to those used in Gio
ra, Fein, Laadan, et al. were furnished with explicit markers. Thus, in the sarcastically bi
ased dialogues (7), whenever the sarcastic speaker contributed a sarcastic utterance, an
explicit sarcastic marker (e.g., mockingly) was inserted; in the salience-based biased dia
logues (8), speakers’ turns were furnished with an explicit nonsarcastic marker (e.g., sad
ly), attesting to their sincerity. (Target utterances in bold, for convenience):
(7) Sagi: Yesterday I started working as a security guard at Ayalon shopping mall.
Yafit: Irit indeed told me she had seen you there.
Sagi (desperate): It turned out it’s quite a tough job, being on your feet all day.
Yafit: I hope that at least the pay is worth the effort.
Sagi: At the moment I get 18 shekels per hour.
Yafit (mockingly): Great salary you’re getting.
Sagi: I know that’s not enough but they promised a raise soon.
Yafit: And how much will you actually get after the raise?
Page 16 of 29

Sagi: In two weeks from now I’ll get 20 shekels per hour.
Yafit (still mockingly): Wow, a highly significant raise.
(8) Sagi: Yesterday I started working as a security guard at Ayalon shopping mall.
Yafit: Irit indeed told me she had seen you there.
Sagi (desperate): It turned out it’s quite a tough job, being on your feet all day.
Yafit: I hope that at least the pay is worth the effort.
Sagi: At the moment I get 18 shekels per hour.
Yafit (sadly): A very low salary.
Sagi: I know that’s not enough, but they promised a raise soon.
Yafit: And how much will you actually get after the raise?
Sagi: In two weeks from now I’ll get 30 shekels per hour.
Yafit (happily): Wow, a highly significant raise.
(p. 342) Results from two (subject and item) combined analyses of all 3 experiments show
that (a) reading times of target utterances were longer following sarcastically biased dia
logues than following salience-based biased dialogues. (b) Additionally they show that
sarcastically related probes took longer to respond to compared to salience-based related
probes and (marginally so compared to) unrelated probes, regardless of contextual bias.
Fein et al. (2013) further tested the contextualist approach by strengthening the contexts
in the +Expectation condition (originally in Giora, Fein, Laadan, et al., 2007) but not in
the −Expectation condition by letting participants know that sarcasm interpretation was
tested. Participants were also allowed a greater variety of delays for the lexical decision
task (750, 1,000, 1,500, 2,000, 2,500, and 3,000 ms ISIs). Results from two (subject and
item) combined analyses, however, show that strengthening the contextual bias (manipu
lated in Giora, Fein, Laadan, et al., 2007) in favor of a sarcastic interpretation by intro
ducing an additional constraint did not affect the patterns of results obtained earlier. De
spite employing multiple constraints and allowing longer processing times, sarcasm was
not facilitated. Compared to the sarcastic interpretation, however, incompatible salience-
based interpretations were made available initially in both conditions, responded to faster
than both sarcastically related and unrelated probe words.
Similarly, in Pexman, Ferretti, and Katz (2000), contexts were strongly biased in favor of a
sarcastic interpretation. They involved a contrast between what is said by a target utter
ance and the contextual information preceding it. In addition, they further featured a
speaker of high-irony occupation (e.g., comedians, factory workers; see e.g., Katz, Blasko,
& Kazmerski, 2004; Katz, & Pexman, 1997), thus raising an expectation for a sarcastic re
mark. Regardless, such strong contexts failed to facilitate sarcastic interpretations com
Page 17 of 29

pared to contexts inviting a metaphoric or neutral reading of the target. Instead, they ex
hibited increased reading times of targets relative to the neutral and no occupation con
texts.
Findings in Colston and Gibbs (2002) also showed that nonsalient (sarcastic) interpreta
tions took longer to process than interpretations based on the salient (metaphoric) mean
ings of the utterance constituents. Utterances (This one’s really sharp), whose key word
(sharp) has both salient literal and metaphorical meanings, took longer to read when em
bedded in sarcastically (9b) than in metaphorically biasing contexts (9a):
(9)
a. You are a teacher at an elementary school.
You are discussing a new student with your assistant teacher. The stu
dent did extremely well on her entrance examinations. You say to your
assistant,
“This one’s really sharp.”
b. You are a teacher at an elementary school.
You are gathering teaching supplies with your assistant teacher. Some of
the scissors you have are in really bad shape. You find one pair that
won’t cut anything.
You say to your assistant,
“This one’s really sharp.”
In addition, double-grounded optimally innovative sarcasm (as when “This one’s really
sharp” is said of a student who did extremely poorly on her entrance examinations) took
longer to process compared to single-grounded optimally innovative sarcasm (as in 9b).
Such results demonstrate the cost of processing optimal innovations which, as predicted,
is not alleviated by contextual support. Interpreting optimal innovations is taxing be
cause, on top of an intended nonsalient interpretation, they invoke salient meanings and
salience-based interpretations, even if inappropriate. Strong contextual support does not
enable bypassing these meanings and interpretations, thus allowing an aesthetic effect.
Sarcasm and Contextual Resonance: Corpus-Based Studies
If sarcasm involves activating its salient meaning and salience-based interpretation, its
environment may resonate with these meanings and interpretations. Indeed, corpus-
based studies of spoken interactions and written discourse provide support for the preva
lence of salience-based albeit incompatible interpretations in the environment of sarcas
tic remarks. For instance, in Giora and Gur (2003), 75 percent of the (Hebrew) sarcastic
remarks exchanged among friends were responded to by reference to their salience-
based but incompatible interpretations. In Kotthoff (2003), in dinner table conversations
among friends, responding to the salience-based but incompatible interpretations of the
(German) sarcastic remarks was the norm; responding to the nonsalient compatible inter
Page 18 of 29

pretation occurred significantly less frequently. However, the reverse was true of (p. 343)
interlocutors who were adversaries participating in TV talk shows.
Looking at (Hebrew) newspaper editorials and op-ed articles written by well-known sar
castic journalists during 2008–9, revealed a similar trend. On the basis of a corpus that
included 70,347 words and 1,597 sarcastic utterances (15,466 words), Giora, Raphaely,
Fein, and Livnat (2014) showed that salience-based interpretations featured dominantly
in the environment of sarcastic remarks. Although in many cases (43 percent) the envi
ronment of sarcastic utterances did not reflect any of their meanings or interpretations,
in many other cases (57 percent) it did. First, in very few cases (3 percent), the environ
ment of sarcastic utterances addressed both their sarcastic and salience-based interpre
tations; in other cases (10 percent), these sarcastic utterances induced extended sar
casm. However, in a great number of cases (36 percent), the environment (in italics, for
convenience) of the sarcastic utterances (in bold, for convenience) resonated with their
salience-based interpretations exclusively (10a–b); a significantly smaller number of cas
es (8 percent) were echoed via their sarcastic interpretations exclusively (11a–b):
(10) Environment manifesting salience-based resonance

(a) All this really could have been peachy if not for the fact that blind
ness is dangerous and the not-so-good ending is known in advance (Levy,
July, 2009).16
(b) Most of the tycoons that have borrowed, in recent years, tens of bil
lions of NIS, in bonds, from the public, find it hard to adjust to the harsh
conditions in the First Class of commercial airlines (Rolnik, 2008).17
(11) Environment manifesting sarcastic resonance

(a) A modest studio of less than 140 square meters with two enormous
rooms outfitted like the most luxurious hotel (Kashua, 2008).
(b) The man [Olmert] who made a number of courageous statements
about peace late in his tenure has orchestrated no fewer than two wars.
Talking peace and making war, the “moderate” and “enlightened”
prime minister [Olmert] has been revealed as one of our greatest fo
menters of war (Levy, January 2009).
Such results support the view that (sarcastic) optimal innovations activate their salient
meanings and salience-based interpretations that are retained even in the mind of their
producers, thus allowing them to resonate with their own contextually inappropriate in
terpretations on account of their relative accessibility.
Conclusion
A vast array of findings supports the view that it is degree of novelty rather than degree
of nonliteralness that matters. Specifically, optimally innovative stimuli (weapons of mass
construction), which, in addition to their novelty, also invoke some familiarity (weapons of
Page 19 of 29

mass destruction), affect both pleasure and mode of processing. Whereas such stimuli
have been shown to take longer to process compared to familiar ones, they have been fur
ther shown to be more pleasing compared to other alternatives. Double grounding—when
more than one familiar stimulus is prompted—increases the pleasure (Brône & Coulson,
2010).
Degree of nonliteralness, however, does not matter. Both literal and nonliteral optimal in
novations are more costly and enjoyable compared to familiar stimuli, whether literal or
nonliteral (Giora et al., 2004; Shuval & Giora, 2009; van Mulken et al., 2010). The fact
that it is degree of novelty rather than degree of nonliteralness that makes a difference
has been shown to be true of different kinds of populations, whether typically or atypical
ly developing. We all invest more in making sense of novel utterances and benefit from it
(Amanzio et al., 2008; Giora, Gazal et al., 2012; Mashal et al., 2011; Mashal & Kasirer,
2012).
With regard to enjoying optimal innovativeness, such as sarcasm, some social factors also
play a role. For instance, sarcasm is most enjoyable when taking the speaker’s rather
than the victim’s perspective (Drucker et al., in press; van Mulken et al., 2010). When un
equal relations are added to the equation, this means that the disempowered (e.g.,
women), taking their own group’s point of view, enjoy such innovations best when the
speaker is an ingroup member and the victim an outgroup member; they, however, enjoy
it least when the speaker and victim are ingroup members (Drucker et al., in press).
Context, however, does not play a role in that it does not initially facilitate processing of
optimal innovation. It does not allow bypassing salient meanings or salience-based inter
pretations even when it is strongly biased in favor of the optimally innovative target. Even
when bias is made apparent, as when a relevant occupation is mentioned (e.g., comedian;
see Pexman et al., 2000) or a speaker’s intent is made explicit (by e.g., prefacing what is
said by cues such as mockingly, by training comprehenders to (p. 344) expect sarcasm, or
by letting them know in advance that what is at stake is sarcasm comprehension; see Fein
et al., 2014; Giora et al., 2007), processing effort does not decrease compared to control
conditions. Optimal innovations are always more difficult to process than equivalent but
more familiar stimuli involving salient meanings or salience-based interpretations.
Neural substrates also support the view that it is degree of innovativeness rather than de
gree of nonliteralness that matters. Unlike familiar stimuli, which tend to engage LH re
gions (Kana, Murdaugh, Wolfe, & Kumar, 2012), optimal innovations tend to engage RH
areas, regardless of degree of nonliteralness (e.g., Amanzio et al., 2008; Arzouan et al.,
2007; Bottini et al., 1994; Cardillo et al., 2012; Eviatar & Just, 2006; Faust & Mashal,
2007; Giora et al., 2000; Kacinik & Chiarello, 2007; Kaplan et al., 1990; Mashal & Faust,
2008; Mashal et al., 2005, 2007, 2009; Pobric et al., 2008; Shamay-Tsoory et al., 2005;
Winner & Gardner, 1977, among others).
Optimal innovations, then, capture processors’ attention and make them invest in making
sense of them, while in return rewarding them with pleasure. No wonder they are used so
pervasively in propaganda and counterpropaganda of all kinds, including commercials,
Page 20 of 29

newspapers headlines, or books’ titles. Future research should examine whether beauty
we invest so much in “is truth and truth—beauty” (Keats, 1820) both in language and
thought as well as in social and political behavior.
Acknowledgments
The research reported here was supported by the Israel Science Foundation (grant No.
436/12) and the Vice President for Research and Development at Tel Aviv University En
couragement Fund. Thanks also go to Lahav Halevy’s (2002) for his permission to use
Sharon Stone, to Keren Manor/Activestills for the permission to cite Figures 21.7–21.8,
and to Micha Ullman for the permission to use Metzer-Messer- land exchange project. I
am also obliged to Vered Heruti and Noa Shuval for introducing me to some of the works
of art used here, and last but not least, to Thomas Holtgraves for all his helpful com
ments.
References
Amanzio, M., Geminiani, G., Leotta, D., & Cappa, S. (2008). Metaphor comprehension in
Alzheimer’s disease: Novelty matters. Brain and Language, 107(1), 1–10.
Aristotle. (350 BCE). Poetics. Translated by Butcher S. H. The Internet Classics Archive
http://classics.mit.edu/Aristotle/poetics.1.1.html.
Arzouan, Y., Goldstein, A., & Faust, M. (2007). “Brain waves are stethoscopes”: ERP cor
relates of novel metaphor comprehension. Brain Research, 1160, 69–81.
Berlyne, D. E. (1950). Novelty and curiosity as determinants of exploratory behavior.

British Journal of Psychology, 41(1–2), 68–80.
Blasko D. G., & Kazmerski V. A. (2006). ERP correlates of individual differences in the
comprehension of nonliteral language. Metaphor and Symbol, 21(4), 267–284.
Bohrn, I. C., Altmann, U., Lubrich, O., Menninghaus, W., & Jacobs, A. M. (2012). Old
proverbs in new skins—an fMRI (p. 345) study on defamiliarization. Frontiers in Psycholo
gy, 3(204), doi: 10.3389/fpsyg.2012.00204.
Bottini, G., Corcoran, R., Sterzi, R., Paulesu, E., Schenone, P., Scarpa, P., Frackowiak, R.
S. J., & Frith, C. D. (1994). The role of the right hemisphere in the interpretation of figura
tive aspects of language: A positron emission tomography activation study. Brain, 117,
1241–1253.
Bowdle, B. F., & Gentner, D. (2005). The career of metaphor. Psychological Review, 112,
193–216.
Bowes A., & Katz A. (2011). When sarcasm stings. Discourse Processes, 48(4), 215–236.
Page 21 of 29

Brône, G., & Coulson, S. (2010). Processing deliberate ambiguity in newspaper headlines:
Double grounding. Discourse Processes, 47(3), 212–236.
Campbell, J. D., & Katz, A. N. (2012). Are there necessary conditions for inducing a sense
of sarcastic irony? Discourse Processes, 49, 459–480.
Cardillo, E. R., Watson, C. E., Schmidt, G. L., Kranjec, A., & Chatterjee, A. (2012). From
novel to familiar: Tuning the brain for metaphors. NeuroImage, 59(4), 3212–3221.
Colich, N. L., Wang, A-T., Rudie, J. D., Hernandez, L. M., Bookheimer, S. Y., & Dapretto, M.
(2012). Atypical neural processing of ironic and sincere remarks in children and adoles
cents with autism spectrum disorders. Metaphor and Symbol, 27(1), 70–92.
Colston, H. L., & Gibbs, R. W. Jr. (2002). Are irony and metaphor understood differently?
Metaphor and Symbol, 17, 57–60.
Coulson, S., & Van Petten, C. (2007). A special role for the right hemisphere in metaphor
comprehension? ERP evidence from hemifield presentation. Brain Research, 1146, 128–
145.
Diaz, M. T., Barrett, K. T., & Hogstrom, L. J. (2011). The influence of sentence novelty and
figurativeness on brain activity. Neuropsychologia, 49, 320–330.
Dress M. L., Kreuz, R. J., Link, K. E., & Caucci, G. M. (2008). Regional variation in the use
of sarcasm. Journal of Language and Social Psychology, 27(1), 71–85.
Drucker, A., Fein, O., Bergerbest, D., & Giora, R. (in press). On sarcasm, social aware
ness, and gender. Humor: International Journal of Humor Research.
Du Bois, W. J. (2007). The stance triangle. In R. Englebretson (Ed.), Stancetaking in dis

course: Subjectivity, evaluation, interaction (pp. 139–182). Amsterdam: John Benjamins.
Du Bois, W. J. (2014). Dialogic syntax. Cognitive Linguistics.
Eviatar, Z., & Just, M. A. (2006). Brain correlates of discourse processing: An fMRI inves
tigation of irony and conventional metaphor comprehension. Neuropsychologia, 44(12),
2348–2359.
Faust, M., & Mashal, N. (2007). The role of the right cerebral hemisphere in processing
expressions taken from poetry: A divided visual field study. Neuropsychologia, 45(4), 860–
870.
Fein, O., Yeari, M., & Giora, R. (2014). On the priority of salience-based interpretations:
The case of irony. Paper submitted for publication.
Freud, S. (1922). Beyond the pleasure principle. Translated by C. J. M. Hubback. London,

Vienna: Intl. Psycho-Analytical/New York: Bartleby.com, 2010.
Page 22 of 29

Heartfield, J. (July 17, 1932). Adolf the superman: Swallows gold and spouts junk. http://
johnheartfield.tumblr.com/post/203985683/adolf-the-superman-swallows-gold-
and-spouts-junk
Heartfield, J. (March 8, 1934). Blood and iron. http://creatividadnatural.blogspot.co.il/

2010_10_01_archive.html
Hekkert, P., Snelders, D., & van Wieringen, P. C. W. (2003). Most advanced, yet accept
able: Typicality and novelty as joint predictors of aesthetic preference in industrial de
sign. British Journal of Psychology, 94(1), 111–124. doi: 10.1348/000712603762842147
Gibbs, R. W. Jr. (1980). Spilling the beans on understanding and memory for idioms in
conversation. Memory & Cognition, 8, 149–156.
Gibbs, R. W. Jr. (1986). On the psycholinguistics of sarcasm. Journal of Experimental Psy

chology: General, 115, 3–15.
Gibbs, R. W. Jr. (1994). The poetics of mind. Cambridge: Cambridge University Press.
Gibbs, R. W. Jr. (2002). A new look at literal meaning in understanding what is said and
implicated. Journal of Pragmatics, 34, 457–486.
Giora, R. (1997). Understanding figurative and literal language: The graded salience hy
pothesis. Cognitive Linguistics, 8(3), 183–206.
Giora, R. (1999). On the priority of salient meanings: Studies of literal and figurative lan
guage. Journal of Pragmatics, 31, 919–929.
Giora, R. (2003). On our mind: Salience, context, and figurative language. New York: Ox
Giora, R. (2007). “A good Arab is not a dead Arab—a racist incitement”: On the accessibil
ity of negated concepts. In I. Kecskes & L. R. Horn (Eds.), Explorations in pragmatics:
Linguistic, cognitive and intercultural aspects (pp. 129–162). Berlin: Mouton de Gruyter.
Giora, R. (2009). Irony. In: L Cummings (Ed.), Pragmatics encyclopedia (pp. 265–267).
London: Routledge.
Giora, R. (2011a). “Your baby is no longer an infant.” On metaphor as context. In M. Gluz

man & O. Lubin (Eds.), Intertextuality in literature and culture (pp. 235–247). Tel Aviv:
Hakibbuz Hameuchad. (Hebrew).
Giora, R. (2011b). Will anticipating irony facilitate it immediately? In M. Dynel (Ed.), The
pragmatics of humour across discourse domains (pp. 19–31). Amsterdam: John Ben
jamins.
Giora, R., & Fein, O. (1999). Irony: Context and salience. Metaphor and Symbol, 14(4),
241–257.
Page 23 of 29

Giora, R., Fein, O., Kaufman, R., Eisenberg, D., & Erez, S. (2009). Does an “ironic situa
tion” favor an ironic interpretation? In G. Brône & J. Vandaele (Eds.), Cognitive poetics.
Goals, gains and gaps, 383–399. Applications of Cognitive Linguistics series. Berlin/New
York: Mouton de Gruyter.
Giora, R., Fein, O., Kotler, N., & Shuval, N. (forthcoming). Know hope: Metaphor, optimal
innovation, and pleasure. In G. Brône, K. Feyaerts, & T. Veale (Eds.), Cognitive linguistics
meet humor research. Current trends and new developments. Berlin/New York: Mouton
de Gruyter.
Giora, R., Fein, O., Kronrod, A., Elnatan, I., Shuval, N., & Zur, A. (2004). Weapons of mass
distraction: Optimal innovation and pleasure ratings. Metaphor and Symbol, 19, 115–141.
Giora, R., Fein, O., Laadan, D., Wolfson, J., Zeituny, M., Kidron, R., Kaufman, R., & Sha
ham, R. (2007). Expecting irony: Context vs. salience-based effects. Metaphor and
Symbol, 22, 119–146.
Giora, R., Gazal, O., Goldstein, I., Fein, O., & Stringaris, K. A. (2012). Salience and con
text: Interpretation of metaphorical and literal language by young adults diagnosed with
Asperger’s syndrome. Metaphor and Symbol, 27(1), 22–54.
(p. 346) Giora, R., & Gur, I. (2003). Irony in conversation: Salience, role, and context ef
fects. In B. Nerlich (Ed.), Polysemy: Flexible patterns of meaning in mind and language
(pp. 297–315). Berlin: De Gruyter.
Giora, R., Raphaely, M., Fein, O., & Livnat, E. (2014). Resonating with contextually inap
propriate interpretations: The case of irony. Cognitive Linguistics.
Giora, R., & Stringaris, K. A. (2010). Neural substrates of metaphor. In: P. C. Hogan (Ed.),
The Cambridge encyclopedia of the language sciences (pp. 489–492). Cambridge, UK:
Giora, R., Zaidel, E., Soroker, N., Batori, G., & Kasher, A. (2000). Differential effects of
right and left hemispheric damage on understanding sarcasm and metaphor. Metaphor
and Symbol, 15, 63–83.
Glucksberg, S. (2001). Understanding figurative language: From metaphors to idioms.

New York: Oxford University Press.
Glucksberg, S., & Keysar, B. (1990). Understanding metaphorical comparisons: Beyond

similarity. Psychological Review, 97, 3–18.
Gold, R., & Faust, M. (2012). Metaphor Comprehension in persons with Asperger’s syn
drome: Systemized versus non-systemized semantic processing. Metaphor and Symbol,
27(1), 55–69
Grice, H. P. (1975). Logic and conversation. In P. Cole and J. Morgan (Eds.), Speech Acts.
Syntax and Semantics (Vol. 3, pp. 41–58). New York: Academic Press.
Page 24 of 29

Halevy, L. (2002). Sharon Stone. “Terrible days” [also Jewish holidays] exhibition. Tel Aviv.
Kacinik N. A., & Chiarello C. (2007). Understanding metaphors: Is the right hemisphere
uniquely involved? Brain and Language, 100(2), 188–207.
Kana, R. K., Murdaugh, D. L. Wolfe, K. R., & Kumar, S. L. (2012). Brain responses mediat
ing idiom comprehension: Gender and hemispheric differences. Brain Research, 1467,
18–26.
Kaplan, A. J., Brownell, H. H., Jacobs, R. J., & Gardner, H. (1990). The effects of right
hemisphere damage on the pragmatic interpretation of conversational remarks. Brain and
Language, 38, 315–333.
Katz, A. N., Blasko, D. G., & Kazmerski, V. A. (2004). Saying what you don’t mean: Social
Influences on language processing Current Directions in Psychological Science, 13(5),
186–189.
Katz, A. N., & Pexman, P. M. (1997). Interpreting figurative statements: Speaker occupa
tion can change metaphor into irony. Metaphor and Symbolic Activity, 12, 19–41.
Kashua, S. (23 May, 2008). Manhattan transfer. http://www.haaretz.com/manhattan-

transfer-1.246407 (retrieved on June 27, 2012).
Keats, J. (1820). Ode on a Grecian urn. http://www.eecs.harvard.edu/~keith/poems/

urn.html
Keysar B. (1989). On the functional equivalence of literal and metaphorical interpreta

tions in discourse Journal of Memory and Language, 28(4), 375–385.
Kintsch, W., & Bowles, A. R. (2002). Metaphor comprehension: What makes a metaphor
difﬁcult to understand? Metaphor and Symbolic Activity, 17, 249–262.
Kotthoff, H. (2003). Responding to irony in different contexts: Cognition and conversa

tion. Journal of Pragmatics, 35, 1387–1411.
Lakoff, G., & Johnson, M. (1980). Metaphors we live by. Chicago: University of Chicago
Press.
Lakoff, G., & Kövecses, Z. (1987). The cognitive model of anger inherent in American Eng
lish. In D. Holland and N. Quinn (Eds.), Cultural models in language and thought (pp.
195–221). Cambridge: Cambridge University Press.
Lakoff, G., & Turner, M. (1989). More than cool reason: A field guide to poetic metaphor.
Chicago: Chicago University Press.
Lampert, M. D., & Ervin-Tripp, S. M. (2006). Risky laughter: Teasing and self-directed jok
ing among male and female friends. Journal of Pragmatics, 38(1), 51–73.
Page 25 of 29

Lee, S. S., & Dapretto, M. (2006). Metaphorical vs. literal word meanings: FMRI evidence
against a selective role of the right hemisphere. NeuroImage, 29(2), 536–544.
Leggitt J. S., & Gibbs, R. W., Jr. (2000). Emotional reactions to verbal irony Discourse
Processes, 29(1), 1–24.
Laurent, J.-P., Denhières, G., Passerieux, C., Iakimovac, G., & Hardy-Baylé, M.-C. (2006).
On understanding idiomatic language: The salience hypothesis assessed by ERPs. Brain
Research, 1068, 151–160.
Levy, G. (January 5, 2009). And there lie the bodies. Haaretz http://www.haaretz.com/
print-edition/opinion/and-there-lie-the-bodies-1.267393
Levy, G. (March 8, 2009). The rule of fear. Haaretz. http://www.haaretz.com/print-edi

tion/opinion/the-rule-of-fear-1.271657
Levy, G. (July 19, 2009). Israelis don’t pay price for injustice of occupation. Haaretz.
http://www.haaretz.com:80/print-edition/opinion/israelis-don-t-pay-price-for-in
justice-of-occupation-1.280237
Levac, A. (December 12, 2008). The 21st century. Haaretz. http://www.haaretz.co.il/

misc/1.1366716
Levac, A. (April 28, 2012). Israeli soldiers in Hebron. Haaretz. http://www.haaretz.co.il/

magazine/israel64/1.1693359
McDougall, W. (1923). Outline of psychology. New York: Charles Scribner’s Sons.
Mashal, N., & Faust, M. (2008). Right hemisphere sensitivity to novel metaphoric rela
tions: Application of the signal detection theory. Brain and Language, 104(2), 103–112.
Mashal, N., & Faust, M. (2009). Conventionalisation of novel metaphors: A shift in hemi
spheric asymmetry. Laterality, 14(6), 573–589.
Mashal, N., Faust, M., & Hendler, T. (2005). The role of the right hemisphere in process
ing nonsalient metaphorical meanings: Application of principal components analysis to
fMRI Data. Neuropsychologia, 43(14), 2084–2100.
Mashal, N., Faust, M., Hendler, T., & Jung-Beeman, M. (2007). An fMRI investigation of
the neural correlates underlying the processing of novel metaphoric expressions. Brain
and Language, 100(2), 115–126.
Mashal, N., Faust, M., Hendler, T., & Jung-Beeman, M. (2009). An fMRI study of process
ing novel metaphoric sentences. Laterality, 14(1), 30–54.
Mashal, N., Gavrieli, R., & Kavé, G. (2011). Age-related changes in the appreciation of
novel metaphoric semantic relations. Aging, Neuropsychology, and Cognition, 18(5), 527–
543.
Page 26 of 29

Mashal, N., & Kasirer, A. (2012). Principal component analysis study of visual and verbal
metaphoric comprehension in children with autism and learning disabilities. Research in
Developmental Disabilities, 33, 274–282.
Mazali, R. (December 2, 2006). Talk for December 2 Demo—Coalition to Stop the Siege of
Gaza. Tel Aviv.
Mualem, M. (January 1st, 2007). Imagine this face smeared with mud (in Hebrew).
Haaretz. http://www.haaretz.co.il/misc/1.1375777
Mukařovský, J. (1932/1964). Standard language and poetic language. In P. L. Garvin (Ed.),

A Prague School reader on (p. 347) esthetics, literary structure, and style (pp. 17–30).
Washington, DC: Georgetown University Press.
Mukařovský, J. (1978). Structure, sign and function. New Haven, CT: Yale University
Press.
Pexman, P. M., Ferretti, T. R., & Katz, A. N. (2000). Discourse factors that influence on-
line reading of metaphor and irony. Discourse Processes, 29, 201–222.
Pobric, G., Mashal, N., Faust, M., & Lavidor, M. (2008). The causal role of the right cere
bral hemisphere in processing novel metaphoric expressions taken from poetry: A TMS
study. Journal of Cognitive Neuroscience, 20, 170–181.
Rolnik, G. (4 December, 2008). Taking stock / Three safety nets. http://

www.haaretz.com/print-edition/business/taking-stock-three-safety-nets-1.258808
Searle, J. R. (1979). Expression and meaning. Cambridge, UK: Cambridge University

Press.
Shamay-Tsoory, S. G., Tomer, R., & Aharon-Peretz, J. (2005). The neuroanatomical basis of
understanding sarcasm and its relationship to social cognition. Neuropsychologia, 19,
288–300.
Sheldrake, R. (1981/2012). Morphic resonance. Rochester, VT: Park Street Press.
Shklovsky, V. (1917/1965). Art as technique. In L. T. Lemon & M. J. Reis (Eds. and Trans.),
Russian formalist criticism: Four essays (pp. 3–57). Lincoln: University of Nebraska Press.
Shuval, N., & Giora, R. (2009). Figurativeness, optimal innovation and pleasure. Oryanut
ve’Safa, 2, 111–127. (Hebrew).
Sperber, D., & Wilson, D. (1986/1995). Relevance: Communication and cognition. Oxford:
Blackwell.
Ullman, M. (1972). Metzer-Messer. Land exchange between the Arab village Messer and
kibbutz Metzer. Art symposium, Menashe region, Israel. Initiated and organized by Avital
Geva, kibbutz Ein Shemer. http://en.wikipedia.org/wiki/Visual_arts_in_Israel
Page 27 of 29

van Mulken, M., Burgers, C., & van der Plas, B. (2010). Wolves, confederates, and the
happy few: The influence of comprehension, agreement, and group membership on the
attitude toward irony. Discourse Processes, 48(1), 50–68.
van Peer, W., Hakemulder, J., & Zyngier, S. (2007). Lines on feeling: Foregrounding, aes
thetics and meaning. Language and Literature, 16(2), 197–213.
Werner, N. E., & Nixon, C. L. (2005). Normative beliefs and relational aggression: An in
vestigation of the cognitive bases of adolescent aggressive behavior. Journal of Youth and
Adolescence, 34, 229–243.
Winner, E., & Gardner, H. (1977). The comprehension of metaphor in brain-damaged pa
tients. Brain, 100(71), 7–29.
Wundt, W. M. (1874). Grundzung der physiologischen Psychologie. Leipzig, Germany: En

gelmann.
Notes:
(1) . An expression becomes conventionalized, frozen, or fixed through frequent use and
habit, to the extent that it is now listed in our mental lexicon with one or more given
meanings.
(2) . http://designarchives.aiga.org/#/entries/%2Bcollections%
3A%2250%20Books%2F50%20Covers%20of%202010% 22/_/detail/relevance/asc/
3/7/21564/the-man-behind-the-nose/1. I thank Jenny Birger for this example.
(3) . http://www.curlupndye.com/
(4) . http://www.alternativenews.org/english/index.php/news/news/4366-inequality-fo
rum-2012-pinkwashing-a-palestinian-queer-community.html
(5) . Deautomatization is invited once a conventional use is questioned, thus activating a

less familiar meaning or interpretation. On deautomatization, see, e.g., Russian formalist
Shklovsky (1917/1965) and Prague linguist Mukařovský (1932/1964, 1978).
(6) . http://www.flickr.com:80/photos/idanska/247228762/
(7) . A collocation refers to a set of words that frequently co-occur so as to become a con
ventional fixed expression.
(8) . http://www.palestineposterproject.org/poster/sharon-stone
(9) . http://www.cartoonistgroup.com/store/add.php?iid=90820s
(10) . On gradual increase of aesthetic appreciation vis-á-vis gradual increase in degrees

of deviation/innovativeness, see van Peer, Hakemulder, and Zyngier (2007) who also show
that for a deviation to be noticeable it must be of a certain magnitude; for different find
Page 28 of 29

ings showing that the highly novel and the highly familiar are rated as equally least lik
able, see, e.g., Berlyne (1971), Wundt (1874), among others.
(11) . Tveria—Hebrew for Tiberias—is an Israeli city.
(12) . Metzer, an Israeli Kibbutz, and neighboring Arab villages, including Messer and
Kafin, promoted coexistence and cooperation. Many of the kibbutz members protested
the construction of the “security” fence, which robbed the Palestinians of their lands and
olive groves, especially those of Kafin. Eventually, the latter were bulldozed to make way
for the fence and the road that ran alongside it.
(13) . http://wizolibrary.wizocollege.co.il/netpub/server.np?
find&catalog=catalog&template=detail.np&field=itemid&op=matches&value=5147&site=Effervesc
(14) . http://activestills.org/search/node/blood%20on%20your%20hands
(15) . http://www.youtube.com/watch?v=WpeC7P-2LfU
(16) . On negative understatements being sarcastic, see Giora et al. (2005).
(17) . In the original Hebrew version hard and harsh were described by using the same
(polysemous) root (translated by Elad Livnat).
Rachel Giora
Rachel Giora, Department of Linguistics, Tel Aviv University
Page 29 of 29

Intentions in Meaningful Experiences of Language

Raymond W. Gibbs Jr.

This article examines how communicative intentions affect the way people use and under
stand both spoken and written language. It first provides an overview of the link between
intention and meaning and looks at evidence on the use and understanding of intentional
language. It then considers intentions in written language, some problems with the argu
ment that intentions are entirely private mental states in the minds of individual people,
and the self-organization of intentional meaning. It rejects the intentional perspective on
linguistic communication and instead views intentional behaviors as emergent products
of complex self-organization processes that characterize many physical and biological sys
tems. It also analyzes how the empirical evidence on intentions in meaningful experi
ences of language can be reconciled with the alternative perspective of self-organization
al theory.
Keywords: Intention, language, meaning, self-organization, mental states, linguistic communication, self-organiza
tional theory
Our assumptions about what people intend to accomplish seems critical to how and what
we understand of human action. For example, when people say or write something, rang
ing from single words to book-length manuscripts, our task is to infer what they intended
to communicate and not just decode the meanings of the words, sentences, or utterances.
Consider several instances of linguistic communication that illustrate the possible impor
tance of intentions in the interpretation of meaning.
A first case considers a single face-to-face conversation between two friends discussing
the presidential campaign in the United States in 2012:
Henry: “Richard, who will you vote for president in the fall election?”
Richard: “Didn’t Obama just come out in favor of gay marriage?”
Richard’s reply here does not explicitly state whom he will vote for (e.g., either Barrack
Obama or Mitt Romney). But Henry’s belief that Richard is a strong supporter of gay mar
riage easily leads him to recognize that Richard’s reply both answers the original ques
Page 1 of 29

tion and provides the main reason for his choice (i.e., Obama). Both speakers in this ex
change assume a tacit cooperation and believe that the other person has rational reason
ing abilities to infer what was intended and not simply comprehend what was said.
A second case of linguistic communication asks that you imagine yourself at work, when
during a quick check of e-mail you find the following message from your spouse:
“Would you mind stopping on the way home and pick up some wine for dinner?”
When you read this message, you assume some cooperative bond with your
(p. 349)
spouse and quickly infer that you are not being asked a question about whether you
would mind doing something, but are actually being requested to stop and purchase wine
before coming home. Interpreting indirect meanings can be difficult in cases of written
discourse, given that there are fewer face-to-face contextual cues to help you understand
what a writer really intended to communicate. Still, your interpretive aim is to recover
your spouse’s intention for you to act in a particular way, which is facilitated by your pri
or interactions with this person.
A third instance comes from the new world of sending and receiving text messages. Con
sider one example of a text message sent by a 16-year-old girl to her teachers about the
girl’s summer holidays (http://news.bbc.co.uk/2/hi/uk_news/2814235.stm):
“My smmr hols wr CWOT. B4, we used 2go2 NY 2C my bro, his GF & thr 3:- kids
FTF. ILNY, it’s a gr8 plc.”
Can you understand what this girl intended to communicate by this message? One trans
lation of the message states: “My summer holidays were a complete waste of time. Be
fore, we used to go to New York to see my brother, his girlfriend and their three scream
ing kids face to face. I love New York. It’s a great place.” Once again, interpreting this
message’s meaning likely rests with inferring the writer’s intention in creating the text in
exactly the way she did. Text messages are difficult to understand for those unfamiliar
with common text messaging conventions. Experienced text messengers even playfully
create renditions of classic lines from literature, as in “twz d bst of x, twz d wst of x,”
which was motivated by Charles Dickens’ opening statement in A Tale of Two Cities: “It
was the best of times; it was the worst of times….” More generally, as with face-to-face
conversations and personal e-mail messages, reading and understanding text messages is
intimately tied to people’s assumptions about a specific author’s presumed intentional
meanings, often in light of past personal experience.
Finally, imagine sitting down to read a volume of poetry, when you comes across the
opening lines of a poem titled “Meaning” (Milosz, 2011, p. 226):
When I die, I will see the lining of the world,

The other side, beyond bird, mountain, sunset.
The true meaning, ready to be decoded,
What never added up will add up.
What was incomprehensible will be comprehended.
Page 2 of 29

You may have various goals when reading this poetic segment, such as to understand the
poet’s thoughts about death, to learn about poetic themes in late twentieth-century poet
ry, to think deeply about your own future and fate, to experience aesthetic pleasures of
the words’ sounds, or more. Yet each of these aims may still be connected to your infer
ring something about what the poet intended to communicate. Even if you have no idea
who wrote the poem (i.e., a Nobel Prize laureate) or the circumstances in which it was
created (i.e., later in the poet’s long life), a reader’s understanding of a literary text may
commence with recovery of authorial intentions. Readers may create meaningful inter
pretations of poetry that deviate from what authors intended to communicate. Yet the
recognition that another person wrote something with some idea in mind places con
straints on how we interpret literary works.
These cases of linguistic communication differ greatly in terms of the physical and per
sonal relationships between the speaker/writer and listener/reader. But these examples
share the common feature of people expressing their thoughts in language from which
others are to interpret what they mean. This chapter explores the role that communica
tive intentions may have in people’s use and understanding of spoken and written lan
guage. There is much empirical evidence demonstrating that inferring people’s commu
nicative intentions may be critical to linguistic communication. However, many examples
of linguistic meaning are inherently vague and indeterminate, enough so to question the
common belief that understanding what another person means is primarily oriented to
ward recovering their explicitly encoded intentions. There are also cases where it is
makes little sense to suppose that speakers possess a specific set of intentions that guid
ed their language production.
I argue, as an alternative, that intentional behaviors may all be better characterized as

emergent products of complex self-organization processes. Under this perspective, rather
than being the causal starting point for certain human actions, intentional meanings are
emergent products of dynamical, self-organized processes that are the hallmark of many
physical and biological systems. The main implication of this alternative view is to reject
the classic notion that individual mental states are intentionally expressed in language
and must be (p. 350) recovered if people are to understand each other’s meanings. But
people’s linguistic actions may still be seen as purposive and justified after-the-fact as
possibly being intentional.
Intention and Meaning

Intentions are traditionally conceived of in individualistic and singular terms (Anscombe,
1963; Bratman, 1987) and refer to mental acts that operate prior to the performance of
behavioral acts. A speaker or writer, for example, must have in mind a set of assumptions
that he or she intends to make manifest, or more manifest, to an audience by what is said
or written. Thus, linguistic actions are preceded by private mental events from which an
intention to perform some behavior arises. The philosopher John Searle (1969) argued, in
Page 3 of 29

fact, that there were only five major types of “speech acts” by which people express their
underlying “illocutionary force” intentions. These include:
Representative or assertive: The speaker becomes committed to the truth of the propo
sitional content of an utterance; such as asserting “The sun is shining today.”
Directive: The speaker tries to get the hearer to fulfill what is represented by the
propositional content of an utterance, such as “Please stop talking.”
Commissive: The speaker commits to act in the way represented by the propositional
content of an utterance, such as “I’ll lend you five dollars.”
Expressive: The speaker expresses an attitude toward the propositional content of an
utterance, such as “I’m sad your wallet was stolen.”
Declarative: The speaker performs an action just representing himself or herself as
performing that action, such as “We find the defendant guilty of murder in the first de
gree.”
Successful interpretation of these linguistic statements demands that listeners or readers

go beyond the words spoken and make inferences about what the speaker/writer meant.
This inference process must produce some understanding of the illocutionary force or
general intention underlying what a speaker said. Grice (1957, 1968) argued that recogni
tion of a speaker’s intention is of a special kind, called an m-intention, or “meaning inten
tion.” An m-intention is a speaker’s recognition to produce an effect in a listener by
means of that person’s recognition of that intention. For example, when a speaker says
“The sun is shining today,” she intends that (a) you believe that the sun is shining today,
(b) that you recognize her intention, and (c) that this recognition is part of your reason
for believing that the sun is shining today.
Several accounts have been advanced for specifying exactly how listeners infer speakers’
m-intentions. Recall the conversation between Henry and Richard about whom Richard
will vote for in the presidential election. Understanding Richard’s answer to Henry’s
question requires a chain of reasoning regarding Richard’s intentions because his answer
does not logically follow from Henry’s question. Grice called the intended message be
hind Richard’s utterance a “conversational implicature,” which is a natural outcome of
speakers’ and listeners’ tacit adherence to the “cooperative principle.” This principle
states that a speaker must “make your conversational contribution such as is required, at
the stage at which it occurs, by the accepted purpose or direction of the talk exchange in
which you are engaged” (Grice, 1975, p. 45). The cooperative principle carries with it
four maxims:
Maxim of Quantity: Make your contribution as informative as is required, but not more
so, for the current purposes of the exchange.
Maxim of Quality: Do not say anything you believe to be false or for which you lack ad
equate evidence.
Maxim of Relation: Say only what is relevant for the current purposes of the conversa
tion.
Maxim of Manner: Be brief, but avoid ambiguity and obscurity of expression.
Page 4 of 29

Grice noted that speakers do not always uphold these maxims. As long as speakers gener
ally adhere to the overall cooperative principle, they can “flout” any of these maxims to
produce certain implicatures. For example, Richard’s response to Henry’s question flouts
the maxim of manner by implicating that he will vote for Obama because of his support
for gay marriage. According to Grice’s analysis, Henry would not consider Richard’s re
sponse to be uncooperative because Henry would seek an interpretation given what he
assumes about Richard, and what he believes Richard assumes about him, in order to de
rive an acceptable and “authorized” interpretation.
Both Searle and Grice’s views of meaning rest on the idea that speakers communicate
their intentions via language, and listeners must infer speakers’ intentions to understand
their meanings. A different proposal, from relevance theory, on utterance interpretation
assumes that speakers aim to be optimally relevant in saying what they do (Sperber
(p. 351) & Wilson, 1995). Under this view, every act of ostensive behavior communicates a
presumption of its own optimal relevance, such that it will be relevant enough to warrant
the listener’s attention and compatible with the speaker’s own goals and preferences (the
communicative principle of relevance). Speakers design their utterances to maximize the
number of cognitive effects listeners infer while minimizing the amount of cognitive effort
to do so. Listeners understand speakers’ pragmatic messages via the “relevance-theoretic
comprehension procedure” by following a path of least effort in computing cognitive ef
fects (Sperber & Wilson, 2002), They do this by testing interpretive hypotheses (e.g., dis
ambiguations, reference resolutions, implicatures) in order of accessibility, then stopping
when their expectations of relevance are satisfied.
For example, consider the following exchange between two university professors (Sper
ber & Wilson, 2002, p. 319):
Peter: “Can we trust John to do as we tell him and defend the interests of the Lin
guistics department in the University Council?”
Mary: “John is a soldier!”
How does Peter understand Mary’s metaphorical assertion about John? Peter’s mentally
represented concept of a soldier includes many ideas that may be attributed to John.
Among these are: (a) John is devoted to his duty, (b) John willingly follows orders, (c) John
does not question authority, (d) John identifies with the goals of his team, (e) John is a pa
triot, (f) John earns a soldier’s pay, and (g) John is a member of the military. Each of these
ideas may possibly be activated to some degree by Mary’s use of “soldier” in relation to
John. However, certain of these attributes may be particularly accessible given Peter’s
preceding question, in which he alludes to trust, doing as one is told, and defending inter
ests. Following the relevance-theoretic comprehension procedure, Peter considers these
implications in order of accessibility, arrives at an interpretation that satisfies his expec
tations of relevance at (d), and stops there. He does not even consider further possible
implications such as (e–g), let alone evaluate and reject them. In particular, Peter does
not consider (g), the literal interpretation of Mary’s utterance, contrary to what is ad
Page 5 of 29

vanced by the Gricean view, and consistent with the psychological evidence on inferring
metaphorical meaning (Gibbs, 1994; Gibbs & Colston, 2012).
Relevance theory assumes, then, that understanding speakers’ utterances is not dedicat
ed toward determining speakers’ individual mental states, but is oriented toward infer
ring contextual implications from what is said, yielding various possible cognitive effects.
In this way, the relevance of any utterances is assessed in terms of cognitive effects and
processing effort, not just in terms of inferring speakers’ private mental intentions.
Evidence on Intentional Language Use and Un

derstanding
There is a tremendous body of psycholinguistic research supporting the idea that people
regularly seek what others intend to communicate when understanding language, as well
as nonverbal gestures and various kinds of human actions. Many studies have shown that
people typically remember a pragmatic implication of an utterance rather than the utter
ance itself or what it directly asserts (Bransford, 1979). For example, people hearing sen
tences such as “He dropped the delicate glass pitcher” and “The housewife spoke to the
manager about the increase in meat prices” often remember them according to their
pragmatic force, as in “He broke the delicate glass pitcher” and “The housewife com
plained to the manager about the increase in meat prices” (Schweller, Brewer, & Dahl,
1976). Listeners encoded inferences that went beyond the explicit information contained
in the sentences, but which were in accord with speakers’ probable intentions in making
these utterances. The intention behind a speaker’s utterance is encoded and represented
in memory, not the sentence or utterance meaning.
The strong impulse that people have toward inferring speakers’ communicative intentions
is nicely illustrated in a large body of research on figurative language understanding
(Gibbs, 1994; Gibbs & Colston, 2012). Many experimental studies demonstrate that peo
ple are able to quickly infer the intended pragmatic meanings of many verbal metaphors
(e.g., “Surgeons are butchers”), idioms (e.g., “John spilled the beans about his marriage
to Mary”), different types of irony (e.g., “A fine friend you are” [implying that you are not
a good friend]), and proverbs (e.g., “Don’t put all your eggs in one basket”). In fact, peo
ple often understand figurative meanings for these expressions in contexts far more
quickly than they do so-called literal uses of these expressions or nonfigurative equivalent
statements (e.g., “John revealed the secret about his marriage to Mary”). The speed with
which people can infer figurative and indirect meanings demonstrates how they are read
ily disposed to understand what speakers (p. 352) intended to communicate by their
speech acts and are not immediately biased toward what speakers’ utterance literally
mean apart from their pragmatic interpretations.
The empirical evidence that people often interpret utterances according to speakers’ pre
sumed intentions does not necessarily indicate whether there is an automatic process at
work. An important imperative within experimental psycholinguistics is to uncover differ
Page 6 of 29

ent cognitive and linguistic processes that operate in the very moments of speaking and
listening (i.e., online processing). One attempt to investigate automatic intention recogni
tion during linguistic processing examined if people immediately inferred the specific
speech act motivating a speaker’s utterance (Holtgraves, 2008). Consider a scenario that
ends with a speaker uttering a statement with a particular speech act in mind:
Jenny and Emily had been close friends since grade school.
Now they were rooming together in college.
Emily tended to be very forgetful.
Today, Jenny was sure Emily didn’t remember her dentist appointment.
Jenny: “Don’t forget to go to your dentist appointment today.”
Jenny’s comment aimed to remind Emily of her dentist appointment, and listeners should
implicitly recognize the speech act remind as part of their understanding of Jenny’s inten
tions. A set of experiments tested this hypothesis by having participants read scenarios
like this, one line at a time, at their own pace on a computer screen. After reading the fi
nal statement (i.e., Jenny’s utterance), participants were presented with a probe word
such as “remind” and made a speeded judgment as to whether or not this word described
the speaker’s action in saying his or her utterance.
Across a series of experiments, people generally agreed that the probe words were rele
vant to the speakers’ aims significantly more often than when people read control con
texts that changed the speech act performed (e.g., when Jenny said, “I’ll bet you forgot to
go to your dentist appointment today”). Participants were also significantly faster making
their probe recognition judgments when they read contexts that primed the speech act
utterance (e.g., “Don’t forget to go to your dentist appointment today”) than they were to
make the same judgments when reading slightly altered contexts that changed an earlier
statement (e.g., “Today, Jenny was sure Emily had forgotten her dentist appointment”)
and the speaker’s final utterance (e.g., “I’ll bet you forgot to go to your dentist appoint
ment today”).
A similar pattern of results was found when participants actively engaged in “computer
chat” with an experimenter (posing as another student) using instant messaging. In these
studies, however, participants had to say whether a probe word appeared in the last re
mark sent by their partner (e.g., did “Blame” appear in the remark “I did terrible in
Harmon’s class and it was all his fault”). People were slower to state that the probe word
was not in the speaker’s utterance than when they read a remark that changed the
speech act conveyed (e.g., “I did terrible in Harmon’s class but it really wasn’t anybody’s
fault”). Once again, people appeared to understand speakers’ remarks by inferring some
thing about the specific illocutionary forces motivating these utterances.
Page 7 of 29

Holtgraves (2008) takes considerable care to suggest that these experimental findings do
not necessarily indicate that people comprehend every utterance according to some un
derlying speech act, noting that some utterances do not convey specific speech acts and
that many speech acts unfold across several conversational turns. Nonetheless, these re
sults may be interpreted as demonstrating “that conversationalists are clearly oriented to
speech act performance, and that speech act recognition will frequently play a role in
conversation comprehension” (Holtgraves, 2008, p. 638; also see Holtgraves, 2002).
People’s impulse toward inferring intentional meanings can sometimes lead them to make
interpretive errors. For instance, people often misunderstand speakers’ intentions be
cause of superficial processing, such as answering the question “How many of each type
of animal did Moses take on the ark?” by quickly responding “two” because they fail to
notice that it was Noah, and not Moses, who rescued animals from the Biblical great flood
(Erickson & Mattson, 1981). Studies also show that people often fail to understand gar
den-path sentences, such as “While Mary bathed the baby played in the crib.” When peo
ple read this sentence and were then asked, “Did Mary bathe the baby?” most replied,
“yes,” even though the sentence does not say that Mary bathed the baby (i.e., she only
bathed herself) (Ferrerira, Christianson, & Hollingsworth, 2001). Moreover, people some
times judge speakers’ utterances as being cooperative and meaningful even when these
violate Gricean maxims, such as the maxim of Quantity. For example, people find the
statements (a) “Put the apple on the towel in the (p. 353) box” and (b) “Put the apple
that’s on the towel in the box” to be equally concise and meaningful, despite the fact that
(a) is ambiguous in the way that (b) is not (Engelhardt, Bailey, & Ferreira, 2006). Overall,
ordinary language comprehension may therefore be structured around the use of fast and
frugal heuristics to infer speakers’ intentions, which occasionally leads listeners to make
interpretive errors but also allows people to not be sidetracked by utterances that are
less than felicitous in discourse.
Psychologists and sociologists have long argued that speakers design each utterance so
that their addressees can figure out what they intend by considering the utterance
against their current common ground (Clark, 1996). Some common ground information is
cultural (i.e., information broadly shared by members of a community), and some informa
tion is personal (i.e., information uniquely shared by two or more speakers). When people
converse, they typically design their utterances to take into account the perspective of
the listener, which facilitates addressees understanding of speakers’ communicative in
tentions.
A simple demonstration of this is seen in a study looking at people’s assumptions about

mutually known beliefs and knowledge when speaking with others (Kraus, 1987). This
study had an experimenter stopping people on the street in downtown Boston, Massachu
setts, where he asked for directions to Jordan Marsh, a large department store about six
blocks away. To a third of the people, the experimenter asked “Can you tell me how to get
to Jordan Marsh?” To another third, the experimenter said, “I’m from out of town. Can
you tell me how to get to Jordan Marsh?” To the remaining third, the experimenter asked,
Page 8 of 29

“Can you tell me how to get to Jordan Marsh?” but did so employing a rural Missouri ac
cent, representative of a speech style in a different part of the United States.
The addressees’ responses were secretly recorded and analyzed for the number of words
spoken and the number of places en route that were referred to. When the experimenter
prefaced his question with “I’m from out of town,” people responded with significantly
more words and more place names than when asked this question without the preface.
Alerting the addressee to the fact that the speaker does not share the same community
knowledge clearly gets respondents to design their answers differently. But the respon
dents also gave longer, more detailed answers when the experimenter asked his question
without the “I’m from out of town” preface but spoke with a Missouri accent. Again, peo
ple designed their answers given their assumptions about how well their addressee may
most easily infer their communicative intentions.
The accrual of common ground enables speakers and listeners to more readily coordinate
their intentional meanings in discourse. Consider, for example, an experiment in which
two persons talk to each other, but cannot see each other (Clark & Wilkes-Gibbs, 1986).
Both sit before schematic drawings of cartoon figures, new to both parties. One conver
sant describes a specific figure from her set of figures, and the other identifies the cor
rect picture from his set using the heard description alone. Unsurprisingly, participants
get better at this task over time. Speakers initially provide detailed descriptions of the
figures to make initial identifications possible, but over time each pair of dialog partners
eventually evolves a shared idiosyncratic lingo specific to the given task environment,
thus allowing them to pick out figures more quickly. Thus, on a first trial, one speaker re
ferred to a figure by saying, “All right, the next one looks like a person who is ice skating,
except that they’re sticking two arms out in front.” But on the sixth trial in this study, the
same speaker simply said, “The ice skater.” These results suggest that understanding
what a speaker intends to communicate, and the criteria by which listeners judge that
they have understood, is a joint product requiring coordination and cooperation between
listeners and speakers.
Participants in real-life conversations sometimes design their utterances with the inten
tion of excluding some person from understanding their pragmatic meaning. One set of
experiments explored this type of situation. In this particular case, participants in a card-
arranging task had to communicate the ordering of photographs of Stanford University
scenes, but there was a third person in the room, provided with the same set of pictures,
and the two conversants had to try to ensure that the third person did not succeed in the
task (Clark & Shaefer, 1987). Thus, the speaker had to ensure that the addressee under
stood but simultaneously conceal his or her meaning from the overhearer. All three par
ticipants were Stanford University students and thus “experts,” but the two conversants
were friends and the overhearer was a stranger.
Because the three participants had the same community membership, it was expected
that the conversant would use “private keys” or information that was part of their partic
ular common ground but that was unknown to the overhearer. Although there were cer
Page 9 of 29

tain instances when the speakers (p. 354) would slip up and uttered the name of a scene,
the vast majority of references contained these private keys. For example, a speaker re
ferred to a fountain on campus as “where someone wanted to put my teddy bear.” Over
all, the addresses were twice as successful in correctly arranging the photographs as
were the overhearers, suggesting that speakers and listeners can often successfully hide
their communicative intentions from some people.
Multiple studies also suggest that, under some circumstances, such as stress or high lev
els of cognitive burden, speakers can be more egocentric in their productions than the
traditional common ground view would predict. Listeners do not consistently consider
common ground in their comprehension (Barr & Keysar, 2005). People frequently mis
judge the effectiveness of their own communication precisely because they do not cor
rectly understand what is and is not part of their common ground with others. Speakers
who have learned the meaning of opaque phrases, for instance, sometimes overestimate
the likelihood that other people know those meanings (Keyser & Bly, 1995). Speakers also
think their own utterances are less ambiguous and more effective than they actually are
(Keysar & Henly, 2002). This series of empirical findings do not claim that recovery of
speakers’ intentions is unimportant, but only that speakers and listeners sometimes act
more egocentrically and fail to automatically keep people’s communicative intentions
forefront in mind.
Finally, listeners may recognize speakers’ intentions without necessarily responding to

their implied messages. Imagine a debate between political candidates in which one of
them is asked a question about his or her views on the “war on terror.” This sensitive top
ic may be one that a candidate hoped to avoid given past his or her previous controversial
statements on the topic. Studies show that listeners better recognize that a speaker is
dodging the intended meaning of a question when they are oriented toward the relevance
of the response than when they are asked to make a social evaluation of the speaker (e.g.,
“Do you like the person?”) (Rogers & Norton, 2011). People even rated speakers positive
ly when they dodged the question asked as when they directly answered the question. Of
course, there are situations in which dodging a direct question is polite (e.g., when an
swering “Do you like the dress I am wearing?”), and people often ignore questions or
their implications in many conversations that touch on a wide range of topics. Still, it is
clear that people engage in various strategies in conversation in regard to whether or not
they focus directly on speakers’ apparent communicative intentions. This conclusion does
not imply that people are unable to infer intentional meanings, only that their responses
need not always directly address the import of speakers’ communicative intentions.
Intentions in Written Language

There is a long history of psycholinguistic research suggesting that understanding of a
speaker’s utterances depends, to a significant degree, on making assumptions about that
person’s intentions and mental states. But are intentions relevant to interpreting written
language? Literary theorists, philosophers, and psychologists have long debated whether
Page 10 of 29

readers’ assumptions about authorial intentions should play a role in how written dis
courses are interpreted. Scholars have provided a variety of responses to the question of
“intentionalism,” each of which entails different methods of reading and interpreting writ
ten discourse (Gibbs, 1999).
“Subjective intentionalism” is a longstanding approach to narrative reading that focuses

on biographies of authors who create texts. This view asserts that an author’s successful
ly realized intentions determine a text’s meaning. Authorial intentions, if properly execut
ed, are publically manifest in the work. Moreover, the intentions behind a narrative need
not be consciously present in the author’s mind when he or she wrote the text. An author
might even have difficulty articulating a work’s main intentions if later asked, just as peo
ple often have difficulty in describing what they did in performing a familiar, yet complex
action.
It is almost impossible for authors to write in such as way that all their intentions will be
transparent to all readers. At the same time, authors don’t necessarily intend to commu
nicate only unambiguous meanings in their texts. Many authors also intend to achieve
various aesthetic goals, some of which are best met by leaving certain messages unclear.
Furthermore, it is not always reasonable to ascribe a single intention to the author of a
text, especially a literary work. Longer narrative works are generally produced over a pe
riod of time, and it is unlikely that the complex structure of goals and intentions that mo
tivate the final text will remain perfectly consistent throughout the writing process.
For these reasons, the New Critics in the 1930s through 1960s rejected subjective inten
tionalism as both impractical and misguided. They argued for a fundamental separation of
subject and object (reader and text), rendering the text separate from both its (p. 355) his
torical context and its readers. The “intentional fallacy” suggested that the original de
sign or intention of the author is not the appropriate standard either for judging the aes
thetic merits of a text or for determining anything about its formal characteristics (Wim
satt, 1954). Intentions may be relevant to what a writer was trying to do, but they are not
always indicative of what the author has actually done in making a work. Readers need
not consider authors’ intentions and can instead restrict their interpretations of what tex
tual segments mean by appealing to the contexts in which such segments appear.
In the 1960s, a significant backlash emerged against the strict views of New Criticism. E.
D. Hirsch (1967) argued that a text has to represent someone’s meaning and that no ade
quate proposal exists for judging the validity of an interpretation unless the author’s nar
rative intentions are taken into account. Authorial meanings must be tied to public, share
able meanings as evident in the text itself. Readers should begin by positing a genre for a
particular work (e.g., scientific prose, lyric poetry, a novel, epic narrative, or a play) and
use this as a general framework for interpreting the author’s typical meaning. Most gen
erally, without consideration of authorial intentions, there is simply too much indetermi
nacy and instability in the public linguistic conventions governing meaning and no stable
object for narrative analysis.
Page 11 of 29

The pendulum swung back toward rejection of authorial intentions with the rise of post
structuralist and deconstructionist approaches to literary interpretation. Roland Barthes,
for example, heralded the “death of the author” as a precondition for the “birth of the
reader” in literary criticism. Although authors may think they know what they intend,
their thoughts and language are at the mercy of socioeconomic, psychological, and histor
ical forces that cause them to write something other than what they actually intended.
This blindness makes authors’ intentions far less interesting than the operation of these
external forces as revealed in their work.
One proposal from the 1990s attempted to reconcile the openness of the literary text with
considerations of intentionality. According to this idea, what a work means is expressed
by a “hypothetical intention” (i.e., a complex intention that a member of the intended or
an ideal audience would be most justified in ascribing to the text) (Levinson, 1992). Read
ers’ interpretations of texts depend on their inferences about a hypothetical author found
ed in the linguistic conventions and artistic practices at the time the author wrote the
work, as well as publically available knowledge of how the text was created. A work may
display a multiplicity of meanings, given the large set of intentions readers can hypothe
size about an author and the conditions under which a work was written.
These debates over intentionalism have focused primarily on whether appeal to an

author’s real or hypothetical intentions should be used in justifying specific interpreta
tions of written texts (including literary, legal, religious, and scientific texts). Yet many an
ti-intentionalists, such as the New Critics and deconstructionists, confuse the activity of
self-consciously arguing about authors’ intentions as the final product of interpretation
with whether assumptions about intentions have some automatic role in the processing of
written meaning. Too many literary debates assume that readers have a choice in ac
knowledging or rejecting consideration of an author’s intentions in determining a text’s
meaning. Critics of intentionalism ignore the possibility that readers tacitly assume that
meaning is the product of some sort of intentional agent, an assumption that shapes, to
some extent, how texts are presumed to be understood and appreciated. No theory can
eliminate the cognitive impulse to assume that someone wrote a text for some intentional
purpose, which, at the very least, constrains fast, unconscious reading processes, similar
to those that occur in understanding spoken language.
Consider one set of studies that show how readers’ assumptions about authors determine
the amount of cognitive effort they put into understanding written figurative statements
(Gibbs, Kushner, & Mills, 1991). Participants were presented with various comparison
statements (e.g., “Cigarettes are time bombs”) and were told that these were written ei
ther by famous twentieth-century poets or randomly constructed by a computer program
lacking intentional agency. The participants’ task in one study was to rate the “meaning
fulness” of the different comparisons and in another study simply to read and push a but
ton when they had comprehended these statements.
Readers found metaphorical comparisons, such as “Cigarettes are time bombs,” to be

more meaningful when supposedly written by famous twentieth-century poets than when
Page 12 of 29

these same metaphors were seen as random constructions of a computer program. Peo
ple also took much less time to comprehend these comparisons when they were told the
poets wrote the statements. Moreover, they (p. 356) took longer to reject anomalous utter
ances (e.g., “A scalpel is a horseshoe”) as “meaningful” when the poets supposedly wrote
these statements. Readers assume that poets have specific communicative intentions in
designing their utterances, an assumption that does not hold for the output of an unintel
ligent computer program. Consequently, people put more effort into trying to understand
anomalous phrases when they were supposedly written by poets. They more quickly re
jected as “meaningless” these same anomalous expressions when told that an unintelli
gent computer program wrote them because computers are assumed to lack communica
tive intentions. Thus, people’s assumptions about the basic fact of human intentionality
(i.e., hypothetical intentionalism) shape their immediate processing and even in some cas
es their ultimate interpretations of written language.
A different set of studies explored whether readers typically infer something about autho
rial intentions when reading literary texts (Claassen, 2012). Some experiments suggest
that people inferred that some human agent wrote a text with some purpose in mind as
they interpreted segments from James Joyce and Jorge Luis Borge, for example, even if
they knew nothing specific about these authors. When readers were provided with bio
graphical information in advance about authors, they appeared to incorporate this knowl
edge into the text representations they created. However, these findings in favor of the
general intentionalist position depended, to some extent, on the exact empirical task used
to assess readers’ understandings. Asking people to “talk aloud” as they interpreted texts
was not as effective in generating impressions about the author as were methods that
probed people’s more immediate and local understandings of textual elements, such as
asking them to explicitly recall previously given information about authors. In general,
people typically inferred aspects of authorial intentions both when they knew something
about the actual empirical author and even when no such information was known or giv
en. Knowing something about an author may matter greatly, for example, when readers
draw inferences about the moral point of a literary text or when determining whether a
text should be interpreted as expressing ironic or satirical meanings (Pfaff & Gibbs,
1997).
Some Problems with the Intentional View

Spoken and written language appear to be strongly but not always exclusively centered
on the expression and recovery of people’s communicative intentions. Yet the idea that in
tentions are solely private mental states in the minds of individual people can be criti
cized in various ways. For example, some philosophers take issue with the Gricean view
and suggest that linguistic understanding is analogous to direct perception rather than
inferential reasoning (Millikan, 2005). Under this perspective, linguistic understanding is
not a matter of inferring speakers’ mental states, but is more focused on determining the
functions that an utterance has in context. Thus, we ascribe meaningful content to lin
guistic utterances via social and environmental structures (e.g., conventions, functions,
Page 13 of 29

practices) that guide a speaker’s or writer’s linguistic actions. Speakers can be conceived
as behaving purposefully in producing linguistic utterances, just as hearts and kidneys
behave purposefully, without having to make assumptions about listeners’ underlying in
tentional beliefs (Gregoromichelaki et al., 2011). Psychological and computational stud
ies, in fact, show that various low-level linguistic, social, and cognitive mechanisms pro
vide explanations of how people appear to achieve audience-designed productions with
out the need for constructing explicit models of interlocutors (Horton & Gerrig, 2005) or
the need to rationalize speakers’ actions in terms of mental state ascriptions (Barr, 2004;
Pickering & Garrod, 2004).
The meanings understood in conversation rest not just in the minds of individual speakers
but emerge from the collaborative process of interaction between conversational partici
pants. For instance, speakers often deliberately offer their listeners a choice of constru
als, so when listeners make their choice, they help to determine what the speaker is tak
en to mean (Clark, 1996). A speaker may present an utterance with one intention in mind,
but when the listener misconstrues it, the speaker then changes his or her mind and ac
cepts the new construal.
An example of this interactive perspective, taken from the radio, is one in which a base
ball announcer is being interviewed about the game about to be played and is asked by a
journalist, “Who is behind the plate today?” referring to which player is taking the posi
tion of catcher. But the announcer misconstrues this to be a question about whom the um
pire is for the game (who stands behind the catcher) and begins talking about the umpire.
The journalist accepts this answer and maintains this topic by asking other questions
about the umpire. Later on, the (p. 357) journalist admits that he wanted to know who was
playing catcher today. Speakers sometimes don’t correct listeners’ misunderstandings be
cause they deem it too trivial, disruptive, embarrassing to correct, or because what listen
ers infer and respond to somehow works better in the situation. Still, once an interpreta
tion is grounded, it is taken to be what the speaker meant. Nonetheless, speakers’ inten
tions can partly emerge from the process of negotiating meaning in conversational situa
tions.
Many face-to-face situations also do not come with single, prespecified intentional mean
ings that originate in the minds of speakers (Gibbs, 1999). Consider a situation in which a
teacher attempts to get a student to do his homework by saying “We’re going to have a
big test next week. You better catch up doing all of your overdue homework.” Although
the teacher’s primary goal may be to get the student to do his homework, she may also be
concerned with other things as well, such as motivating the student to take greater per
sonal responsibility for his education. In this case, getting the student to catch up with his
homework assignments would only be one of a broader set of goals.
But what specific intentions must we infer to adequately understand the teacher’s utter
ance to her student? This question assumes, however, that intentions are directly linked
to one individual, in this case the teacher. But as Sperber and Wilson (1995) argue, a fun
damental mistake is to suppose that pragmatics “should be concerned purely with the re
Page 14 of 29

covery of an enumerable set of assumptions, some explicitly expressed, others implicitly

conveyed, but all individually intended by the speaker…There is a continuum of cases,
from implicatures where the hearer was specifically intended to recover implicatures
which were merely manifest, and to further modification of the mutual cognitive environ
ment of speaker and hearer that the speaker only intended in the sense that she intended
her utterance to be relevant and hence to have rich, and not entirely foreseeable cogni
tive effects” (Sperber & Wilson, 1995, p. 201).
Under Sperber and Wilson’s view, the teacher had more than one explicitly held commu
nicative intention in mind when saying what she did. Thus, she may have made her utter
ance to specifically make sure that her student’s homework was done in preparation for
next week’s exam, but she may also have aimed to make manifest a wide array of other
meanings that she didn’t explicitly have in mind when speaking to her student (e.g., that
her student needed to do his homework more regularly, that he should take more respon
sibility in general about doing his school work, that he should take more responsibility for
his education and life actions in general). The teacher might have in mind a “cognitive
content” without explicitly having in mind everything that might be entailed by that cog
nitive content in question. Even if the student asked the teacher what she meant by her
utterances, her response would not necessarily be the one that she had explicitly in mind
precisely because thinking about what she meant elaborates and possibly changes what
may have been initially in her mind when speaking. One may claim that the teacher’s in
tentions, despite their vagueness, are still in her mind in some manner. But saying that
the teacher had some set of intentions in her head that caused her to speak as she did
does not explain how the teacher’s intentions are refined and elaborated upon as a social
product of her interaction with her student.
Part of the difficulty in attributing specific intentions to speakers is that many utterances
in discourse are incomplete, with other speakers quickly chiming in with their own contri
butions, some of which take the talk in a new direction. Consider this brief excerpt from a
discussion among people attending a “Friends of the Earth” club meeting (Gre
goromichelaki et al., 2011, p. 208):
A: “So what is that? Is that er…booklet or something?”
B: “It’s a book.”
C: “Book.”
B: “Just…talking about al…you know alternative”
D: “Om erm, renewable yeah”
B: “energy really I think”
A: “Yeah.”
Page 15 of 29

This set of exchanges illustrates how people move in and out of the parsing and produc
tion roles so that it is not possible to ascribe specific intentions to particular utterances.
The give and take of many conversational exchanges touches on a related problem for at
taching intentionality to single minds. When people use language, it is, in its simplest
forms, like shaking hands, playing a piano duet, or paddling a two-person canoe, collec
tive behavior. When we engage in collective activity, individuals are guided by collective
intentions, sometimes called “we-intentions” (Searle, 1990). Consider an example of a
dean at a university asking the psychology department how it (p. 358) intends to fill a va
cant faculty position (Velleman, 1997). When the dean does this, she seems only to be
asking the twenty-six members of the department to form a single intention. She is cer
tainly not envisioning that the different department members will arrive at twenty-six in
dividual intentions that somehow converge. What rules out this possibility is not that con
vergence among the particular people is impossible; it’s that none of them is in a position
to have an individual intention about what the department should do. Filling the faculty
vacancy is up to the department, and so any intention on the topic must be formed and
held by the group as a whole. The department’s collective intention is not the same as the
sum of the faculty member’s individual intentions. This situation illustrates how inten
tions can be viewed as social products that emerge through the interaction of mutual be
liefs between relational participants.
Figuring out who intended what meaning in discourse is even more complex in cases of
interpreting written texts authored by many people. In some academic fields, like physics,
scholarly papers literally have several hundred authors, only some of whom communicate
with one another about the writing of the text; in many cases, the authors do not know
one another or may have never met before (Gibbs, 1999). What set of speaker or authori
al intentions are to be recovered in situations like this? One could suppose that whomev
er ends up writing these texts must come to some agreements about what they intend to
communicate. Yet studies show that many collaborative writing efforts reveal conflicts
and unresolved tensions between different ideas and ways of expression (Gibbs, 1999).
This makes it impossible to assume that writers have some predetermined thoughts in
mind that they wish for others to infer if these people are to properly interpret the mean
ings of their texts. Modern literary texts (i.e., hypertexts) composed by, in some cases,
thousands of authors, may express emergent interpretive possibilities that simply cannot
be reduced to sets of “intended meanings” generated by individuals’ mental states.
Finally, the intentional view of linguistic communication may be culturally specific. Al
though many cultures appear to focus on individuals’ subjective mental states (i.e., their
thoughts, desires, beliefs, and intentions), other cultures focus on the consequences of a
person’s actions or talk, not on what a person may have possibly intended to communi
cate. Studies of Samoan discourse, for example, show that a speaker’ understanding of an
event and his or her personal motivations and intentions in saying what is stated are
deemed irrelevant (Duranti, 1988). What an utterance means depends on what other
takes it to mean. Samoans ignore the orator’s alleged intentions and concentrate on the
social consequences of a speaker’s words. Consequently, a Samoan speaker will never re
Page 16 of 29

claim the meaning of his words by saying “I didn’t mean it.” The audience is more likely
to be asked to say more about what something means than is the original speaker.
Samoans practice interpretation as a way of publically controlling social relationships
rather than as a way of figuring out what a given person intended to communicate.
The Self-Organization of Intentional Meaning

The number of problematic cases for the intentional view suggests that there may be no
single norm or set of norms that drives all people’s complex linguistic behaviors. How do
we account for the evidence on intentional language use yet also acknowledge the prob
lems with the intentional perspective on linguistic communication?
One answer to this question assumes that linguistic communication need not originate in
a single cause, such as private intentional mental states, and that recovery of speakers’
prior mental intentions is not the basis for language understanding. The genesis for this
alternative comes from the large body of research demonstrating how human perfor
mance emerges as products of human self-organizing systems. Any system whose struc
ture is not imposed from outside forces or from internal blueprints (e.g., internal mental
representations) alone can be said to self-organize (Bak, 1996; Kelso, 1995). Self-organiz
ing systems are capable of creating new structures because their dynamics are dominat
ed by these interactions instead of by the activity of isolated components. Emergent
mechanisms are temporary or “soft-assembled” because they do not endure as passively
stored representations within the system’s dynamics. Soft-assembly processes operate in
highly context-sensitive ways within particular environmental niches to create the very
specific physical patterns and behaviors within each system (Gibbs & Van Orden, 2010).
A wide variety of physical, biological, and human behaviors have now been described as
emergent products of self-organizational processes. These include the formation of galax
ies, termite nests, snowflakes, the foraging patterns of bees and ants, the dynamic shapes
of flocks of birds, the (p. 359) symmetrical patterns on butterfly wings, the regular spots
on a leopard’s skin, the formation of whirlpools in rivers, the formation of bacterial cul
tures, dynamics of traffic jams on freeways, the performance of stock markets, and neu
ronal activity in the human brain.
How does this self-organizational perspective explain linguistic communication? Consider

first several examples of nonlinguistic action that may be understood as intentional but
are emergent products of self-organizing processes. For instance, imagine that you are
taking a tennis swing when engaged in a match against another player. If you are a begin
ner, you may find yourself making an effort to adequately track the ball, keep your racket
perpendicular to the court, hit the ball squarely, and the like. But an expert is absorbed in
the game and experiences something quite different. The expert’s focus is on the oppo
nent and the oncoming ball. All the expert feels is her arm being drawn to the appropri
ate position and the racket forming the optimal angle with the court—an angle that she
need not be aware of. The cumulative experience is one of satisfying these different con
straints. Your actions are brought forth by perceived circumstances in such a way as to
Page 17 of 29

bring about some satisfactory result (e.g., hitting the ball so that it is difficult for your op
ponent to return).
In the same moment, the tennis shot simultaneously satisfies a stack of goals. The same
tennis swing keeps you physically active, furthers your enjoyment of tennis, wins the
match, wins the set, breaks an opponent’s serve, wins the rally, and makes the shot, for
example. All these aspects of the swing are intentional, but the various intentions are sus
tained on different time scales, all of which are longer than the motor coordination that
makes up a swing. Coordinative structures on shorter time scales, closer to real time, are
expressed in your swing at a particular tennis ball, which comes on a specific trajectory,
requiring an exact return force to reach the opponent’s baseline (but not a centimeter
further!). The actual swing is never explicitly “represented” nor are the goals it satisfies.
It is as though the required swing just happened to be available and was solicited by cir
cumstances to satisfy inherent circumstantial constraints.
Consider another example in which you bobble the soap in the shower. Somehow, your
grasp of the soap itself propels the slick bar up and out of your hand. You react immedi
ately and although you may juggle a bit, quick inexpert juggling brings the soap under
control—you catch it before it gets way. Like the tennis swing, catching the soap satisfies
a hierarchy of goals. Also like the tennis swing, catching the soap is an intentional act,
with aspects of intentionality on many different time scales—including time scales of cul
tural change in the practice of hygiene, for example, and time scales of movement trajec
tories that catch soap or apply soap to your body. Most importantly, the explicit goal to re
trieve the soap could not have existed in detail before your juggling intention began.
There is a sense in which your intention was to control the soap, but the intention must
circumscribe a very broad category of actions to anticipate juggling. Likewise, the pre
cise movements that bring the soap back into your control could not have been anticipat
ed before they were enacted. The idea that we must discover a single-cause bobble-the-
soap representation is more obviously ludicrous in this case. And yet every enacted move
ment of the body is as uniquely situated as the juggling trajectories that catch the soap.
Skilled tennis swings and fumbled bars of soap illustrate how we act in the world without
first representing our actions as single causes (Juarrero, 2000). Skillful coping in the
world does not require explicit mental representation of goals or the bodily movements
that would achieve those goals. Instead, coping simply requires constraints that may cir
cumscribe the possibilities for action. In other words, skillful coping is purposive without
entertaining a purpose.
Producing and understanding linguistic utterances have many similarities to what hap
pens when one is swinging at a tennis ball. Linguistic utterances may be enacted in an in
tentional way without there being an underlying intention driving the action. Speakers’
words (and paralinguistic and bodily actions) are produced to simultaneously achieve a
stack of goals, including coordinating with others in the moment, making a specific com
ment in light of what else has just been said or occurred, relieving a sense of incompati
bility between what you expected and what occurred, and perhaps re-establishing some
Page 18 of 29

equilibrium with others in context (e.g., the teacher talking to her student about doing his
homework). These different goals act as constraints on what words are uttered and bodily
acts undertaken, with each of these being sustained on different time scales. Some goals
emerge from longer time scales (e.g., general coordination with others), some on shorter
time scales (e.g., the immediate desire to understand that he needs to his homework),
(p. 360) with others operating along even rapid time scales (e.g., saying specific words in
particular ways to convey one’s emotional state).
To illustrate this more concretely, consider a telephone conversation between Jill and Jeff,
who are girlfriend and boyfriend, in which the topic of Jill’s possible pregnancy comes up
(Du Bois & Englebretson, 2004; greatly redacted from conversation SBC027, leaving out
many nonlinguistic but verbal actions, such as laughter, response cries, and paralinguis
tic information). One point to notice is the specific twists and turns in the conversation as
it unfolds, which makes it difficult to assume that each utterance and each understanding
of what another has said is directly linked to the speakers’ inner mental states. Another
important feature of this conversation, as with all discourse, is that it emerges from a hi
erarchy of constraints and not just from the intentional mental states of the speakers.
JILL: “Hello.”
JEFF: “How’s my favorite girl in the world?”
JILL: “ Hey…baby.”
JEFF: “Who’s, who’s the girl that…I love? Who’s the girl that I’ll do anything
for?…I’ll wash her feet with my mouth.”
JILL: “Aw honey…how are you?”
JEFF: “After a ten mile run.”
JILL: “Ew”
JEFF: “I would”
JILL: “Ew”
JEFF: “I would”
JILL: “Ew! How are you…honey?”
JEFF: “I’m doing good.”
JILL: “Miss you”
JEFF: “I miss you too. How’s my favorite girl?”
JILL: “I’m…good”
JEFF: “Is she…Is she okay?”

Page 19 of 29

JILL: “ Unhuh”
JEFF: “Are you guys having fun?”
JILL: “Yes”
JEFF: “I’m so glad.”
JILL: “Aw…”
JEFF: “What are you going into purr mode?”
JILL: “Oh…gosh…we kept missing you…. Hunh.”
JEFF: “I know…it’s been a while”
JILL: “Last night and today”
JEFF: “It’s been a long time”
JILL: “It’s been so long…I know…You missed all the drama here.”
JEFF: “No…drama?”
JILL: “Yeah, there was such draa. There was drama and there was suspense and
then there was suspense. And then there was relief and then there was ecstasy.”
JEFF: “No”
JILL: “Well, I’ll”
JEFF: “Not the ecstasy”
JILL: “Ecstasy of sorts”
JEFF: “Wait…tell me what…We’ll see I …The reason…I have…I kinda…I wasn’t in
clined call you”
JILL: “Uhuh”
JEFF: “Because I just wanted to give you and Beth your time”
JILL: “Yeah”
JEFF: “You know…Like I felt like”
JILL: “Totally”
JEFF: “You know, so that’s…I let…There were a couple of times I wanted to call
you and I didn’t for that reason.”
Page 20 of 29

JILL: “Unhunh…Aw thanks. Yeah, well the thing was I hadn’t told you? But my pe
riod is really late. You know? And I guess we would talk and then we’d hang up
and then I thought…oh…go…I forgot to tell you. But”
JEFF: “I’ve been meaning to ask you about that”
JILL: “Really? You were getting nervous.”
JEFF: “You haven’t had it yet”
JILL: “No. I mean…and it’s been so…long. And so it’s like eight days late? And I
thought I’m sure I’m pregnant. I’m positive. And I was waiting till Jill got here?”
JEFF: “That’s bad”
JILL: “To even to…try to check?”
JEFF: “Oh…yeah”
JILL: “And um…so we went to go get…an EPT…at…pregnancy test”
JEFF: “Yeah”
JILL: “And then…we’d…That’s why we were calling you last night, you know?
And…”
JEFF: “What happened?”
JILL: “And then…Of course it came out fine. So that’s the
JEFF: “It came out negative?”
JILL: “Yeah so that’s the relief and the ecstasy”
JEFF: “So…you can’t be pregnant.”
JILL: “No. I don’t…No. Unless it was a false negative, which I doubt. Because…”
JEFF: “That’s rare…Right?”
JILL: “Right…I think that would be really rare”
JEFF: “So…You know what I mean…like can…would it come out negative…let’s

say I, you were pregnant in its early st early early stages”
JILL: “Mmm”
JEFF: “Like, wou…like would a negative, you know what I’m saying?”
JILL: “Mmm”
Page 21 of 29

JEFF: “Like could you get…Is it possible that you could still be pos…positive?”
JILL: “I think…there’d be a slight chance of it being a false…negative. But I don’t

think so
Cause I’m pretty late. And I think I’m late enough where I could have like…enough
of…
The hormone that…the pregnancy test tests for?” I think I would have enough of
that in my urine that…of course it would show up…if I had any in there?”
JEFF: “Yeah”
JILL: “You know? So I …I’m prêt…I took that as a good sign”
JEFF: “Oh my God”
JILL: “You know”
JEFF: “Honey, how come you’ve been…keeping all this inside
JILL: “I know…I didn’t mean to keep it in…I mean…I didn’t mean”
JEFF: “Don’t you love me?”
JILL: “Oh”
JEFF: “Aw”
JILL: “I didn’t mean to not tell you at all. It was just…we would hang up the
phone, then I thought…oh yeah, I forgot to mention that I’m kinda late…Cause I
didn’t think it was a huge deal, but then…I guess I wasn’t trying…I think I was in a
bit of denial. Like I wasn’t really think about it much, ya know. Cause I knew once
Jill got here then…I would go get the test.”
JEFF: “So is that all the drama?”
JILL: “That was the drama.”
This conversation demonstrates several points favoring a self-organizational per

(p. 361)
spective on linguistic communication. First, a closer analysis of the audiotape of this con
versation suggests tremendous coordination between people as they manage their turn-
taking using a variety of linguistic and nonlinguistic devices. For example, Jill and Jeff’s
discourse exhibited coordination on many different levels, including their use of certain
words, syntax, intonation, prosody, response cries (e.g., “Ew”), and topics touched on.
One difficulty with many theories of linguistic pragmatics is their failure to acknowledge
the full, embodied actions that speakers engage in during talk, such that language is only
part of the large complex of interpersonal coordination. A self-organization perspective
claims that interpersonal coordination, including that taking place in conversation, is all
Page 22 of 29

about emergent coordinative structure in which all aspects of a person’s dynamics may
come together to more closely mimic the other as the two systems come to change and
behave as one (Fowler, Richardson, Marsh, & Shockley, 2008). The resulting behavior
cannot be reduced to alignment of different mental representations, including those relat
ed to each person’s presumed intentional mental states. Instead, self-organization in be
havior arises from the interplay of brain, body, and environment as a single “context-sen
sitive” system that continuously changes over time as the conversation proceeds. Paying
attention to purposive action as a holistic context-sensitive system removes the urge to
simply reduce linguistic meanings to internal mental states.
A second salient feature of Jill and Jeff’s conversation is that the speakers’ utterances are
rarely complete statements of previously conceived thoughts. Instead, the words spoken,
and the ways they are articulated, emerge in the moment of speaking and are surely
shaped by many forces that may not have much to do with an individual’s prior mental in
tentions. Consider just a few of these forces:
• Evolutionary forces to maintain cooperation with ingroup individuals to enhance indi

vidual and group welfare, including survival.
• Culturally specific ideas about appropriate forms of discourse with one’s romantic
partner, including when and how to discuss potentially problematic issues such as
pregnancy.
• Historical forces related to Jill and Jeff’s past experiences of speaking with one an
other, including those regarding intimate matters such as Jill’s periods.
• Social forces, such as beliefs about when it is appropriate for Jeff to call Jill, especial
ly when she has a friend visiting.
• Social forces related to the “drama” that occurred when Jill and her friend sought to
find out if she was indeed pregnant.
• Present bodily experience, such as Jeff’s sense of physical fatigue, his desire to have
intimate (p. 362) contact with Jill, and Jill’s own fear about possibly being pregnant giv
en that her period was late.
• Immediate motivations to respond appropriately in the conversation given the touchy
subjects of pregnancy and what expectations Jill and Jeff may have had about being
able to talk about this before the point they did on the phone.
• Previously stated words and nonlinguistic behaviors that primed each speaker’s use
of certain terms (e.g., “drama,” “keeping it inside,” and sequences of laughing).
• Brain and neural activity, including that which emerges as brain systems become
coupled during interpersonal interaction.
These are just a few of the forces that serve to constrain Jill and Jeff’s ongoing conversa
tional behaviors. Each force operates along a different time scale, with some of these
crawling along at very slow speeds, such as evolutionary and historical forces, while oth
ers zip along at very fast speeds, such as immediate linguistic processes and the firing of
neurons in human brains. The various time scales are not independent but are hierarchi
Page 23 of 29

cally organized and nested within one another such that different forces affecting lan
guage experience are coupled in complex, nonlinear ways.
Characterizing the ways people use language in interpersonal communication must take
into account all of these interacting forces and not simply focus on only what are tradi
tionally conceived as more immediate or proximate factors, such as privately held mental
intentions. There is no overarching mechanism that decides the process of formulating or
interpreting linguistic utterances. Instead, the system as a whole (i.e., our brains, bodies,
and world interactions), given its very specific past history and present circumstances,
will settle or relax into certain areas of stability or even instability that will constitute
what is ultimately said, written, or understood. People may find themselves in situations
requiring communication with others in which they can simply let their recent thought
processes and the environmental constraints take care of the fine-grained details of how
linguistic and nonlinguistic behaviors are manifested. Both Jill and Jeff’s speech results
from self-organizing tendencies, within and across individuals, even before any specific
intention to speak in certain ways ever reaches awareness. There may be moments when
reflective thought processes help us to decide what to do or say. But these processes,
which operate along slower time scales, are by no means the single causal basis for what
we do. People’s abilities to coordinate their interactions when communicating emerges
from self-organizational processes operating along multiple, linked time scales without
the need to align internal mental representations between individuals. Coordination in
linguistic communication, once more, simply arises from self-organizational processes
within and across individuals.
Re-evaluating the Empirical Evidence

How can one reconcile the empirical evidence on intentions in meaningful experiences of
language with the alternative perspective of self-organizational theory? Most generally,
the self-organizational view suggests that there are multiple constraints affecting linguis
tic production and understanding. People’s choices of words and how they are expressed
and listeners’ interpretations of linguistic utterances unfold from dynamical processes
that should not be reduced to the inner mental states of individual language users. Lan
guage emerges from the interaction of many subsystems, none of which dominates or
controls what is stated. Understanding what speakers mean by what they say is con
structed given a wide assortment of forces that also self-organize to produce an emergent
meaning product. What people say and understand, therefore, is not singly grounded in
assumptions about their putative, private intentional acts.
Our judgment that some interpretations of what people say or write reflect intentional
meanings is an after-the-fact assessment that we can sometimes make. But this assess
ment is not necessarily an automatic part of both linguistic production and understanding
processes. This conclusion parallels the findings from many other areas of psychological
research showing how our judgments about intentionality, and whether single private
Page 24 of 29

thoughts truly precede our subsequent actions, may be after-the-fact narratives rather
than reflections of the underlying causal bases for what we do or say (Wilson, 2004).
The psychological research showing that people are often focused on recovery of commu
nicative intentions is by no means incorrect. People do coordinate and cooperate to vary
ing degrees and across many behavioral levels when speaking and writing. Research also
correctly notes how human goals in speaking and listening dominate the actual linguistic
meanings of words, utterances, sentences, and larger discourse entities. Furthermore,
people may also, in some contexts, be able to quickly judge the (p. 363) type of meaning or
speech act that some utterance conveys in discourse. Speakers and listeners may also ex
hibit tremendous coordination during conversational and written interactions, which is
greatly enhanced given their past experiences with each other. But all of these experi
mental findings do not demand positing of inner intentional mental states that act as the
main causal force behind what people say or write. As noted earlier when considering the
situation of trying to grab a slippery bar of soap in the shower, people can appear to act
in purposive ways without each of their actions being causally linked to some explicit
mental representation referring to their “intentions.”
Finally, the divergent experimental findings showing that people can sometimes speak
egocentrically, ignore people’s conversational goals, or fail to adequately coordinate may
also be readily explained by a dynamical, self-organizational perspective. Once again, the
emergent product of multiple constraints will often exhibit moments of stability as people
appear to talk in a coordinated manner and seem to be closely following each other’s pre
sumed intentional behaviors. But most aspects of language and cognition, including those
observed in experimental situations, reflect unstable behavior in which people are driven
to in-between dynamical states. A crucial feature of a self-organized dynamical system is
its balance of stability and instability.
This mixture of stability and instability in language performance is not due to entirely dif
ferent mechanisms (e.g., one explaining the stability in people’s coordinated language
use and another to explain why people sometimes behave egocentrically). Psycholinguists
have profitably employed a self-organizational perspective to explore some of the dynam
ics associated with online sentence processing and other forms of linguistic behavior,
again without the need to postulate specific intentional mechanisms as part of their ac
counts of psychological data (Gibbs & Van Orden, 2010; Spivey, 2007).
Conclusion
The key to understanding how intentional meaning emerges is to acknowledge how hu
man behavior of all sorts comes into being via dynamical, self-organizational processes.
These processes operate along multiple, interacting time scales such that any human act
is constrained by a complex set of forces that never reduce to single causes or, more im
portantly, single representations such as an intentional mental states. We can surely note
that someone acted in some intentional manner, but we should attribute this intentionali
ty to the entire person (i.e., as an emergent product of the interaction between a brain, a
Page 25 of 29

body, and the world, including a world of other people) and not to some prior, privately
held mental belief or desire. All of the psycholinguistic evidence favoring the intentional
view of meaning, reviewed earlier in this chapter, is completely consistent with the self-
organizational view of intentional meaning as an emergent product. People are continual
ly oriented toward the purposive actions of others and will make inferences about what is
said with the implicit goal of successfully coordinating with others. Yet the analytic mis
take is to assume that linguistic communication is primarily devoted to the expression
and recovery of inner mental states.
References
Anscombe, E. (1963). Intentions. Ithaca, NY: Cornell University Press.
Bak, P. (1996). How nature works: The science of self-organized criticality. New York:
Copernicus.
Barr, D. (2004). Establishing conventional communication systems: Is common knowledge

necessary? Cognitive Science, 28, 937—962.
Barr, D., & Keysar, B. (2005). Mindreading in an exotic case: The normal adult human. In
B. F. Malle & S. D. Hodges (Eds.), Other minds: How humans bridge the divide between
self and other. (pp. 271–283). New York: Guilford Press.
Bransford, J. (1979). Human cognition: Learning, remembering, and understanding. Bel

mont, CA: Wadsworth.
Bratman, M. (1987). Intentions, plans, and practical reason. New York: Cambridge Uni
versity Press.
Clark, H. H. (1996). Using language. New York: Cambridge University Press.
Clark, H. H., & Schaefer, E. (1987). Concealing one’s meaning from overhearers. Journal
of Memory and Language, 26, 209–225.

22, 1–39.
Claassen, E. (2012). Author representations in literary reading. Amsterdam: Benjamins.
Du Bois, J., & Englebretson, R. (2004). Santa Barbara corpus of spoken American English,
Part 3. Philadelphia: Linguistic Data Consortium.
Duranti, A. (1988). Intentions, language and social action in a Samoan context. Journal of
Pragmatics, 12, 13–33.
Engelhardt, P., Bailey, K., & Ferreira, F. (2006). Do speakers and listeners observe the
Gricean maxim of quantity? Journal of Memory and Language, 54, 554–573.
Page 26 of 29

Erickson, T., & Mattson, M. (1981). From words to meaning: A semantic illusion. Journal
of Verbal Learning and Verbal Behavior, 20, 540–552.
Ferreira, F., Christianson, K., & Hollingworth, A. (2001). Misinterpretations of garden-

path sentences: Implications for models of sentence processing and reanalysis. Journal of
Psycholinguistic Research, 30, 3–20.
Fowler, C., Richardson, M., Marsh, K., & Shockley, K. (2008). Language use, coordination,
and the emergence of cooperative action. In A. Fuchs & V. Jirsa (Eds.), (p. 364) Coordina
tion: Neural, behavioral, and social dynamics (pp. 161–180). Berlin: Springer-Verlag.
Gibbs, R. W. Jr. (1994). The poetics of mind: Figurative thought, language, and under
standing. New York: Cambridge University Press.
Gibbs, R. W. Jr. (1999). Intentions in the experience of meaning. New York: Cambridge
University Press.
Gibbs, R. W. Jr., & Colston, H. L. (2012). Interpreting figurative meaning. New York: Cam
Gibbs, R. W. Jr., Kushner, J., & Mills, R. (1991). Authorial intentions and metaphorcompre
hension. Journal of Psycholinguistic Research, 20, 11–30.
Gibbs, R. W. Jr., & Van Orden, G. (2010). Adaptive cognition without massive modularity.
Language & Cognition, 2, 149–176.
Gregoromichelaki, E., Kempson, R., Purver, M., Mills, G., Cannn, R., Meyer-Viol, W., &
Healey, P. (2011). Incrementality and intention-recognition in utterance processing. Dia
logue and Discourse, 2, 199–233.
Grice, H. P. (1957). Meaning. Philosophical Review, 64, 377–388.
Grice, H. P. (1968). Utterer’s meaning, sentence-meaning, and word-meaning. Founda

tions of Language, 4, 225–242.
mantics, vol. 3 (pp. 41–58). New York: Academic Press.
Hirsch, E. (1967). Validity in interpretation. New Haven, CT: Yale University Press.
Holtgraves, T. (2008). Automatic intention recognition in conversation processing. Journal

of Memory and Language, 58, 627–645.
Horton, W. S., & Gerrig, R. J. (2005). The impact of memory demands on audience design
during language production. Cognition, 96, 127–142.
Page 27 of 29

Juarrero, A. (2000). Dynamics in action: Intentions action as complex system. Cambridge,

MA: MIT Press.
Kelso, J. A. S. (1995). Dynamic patterns: The self-organization of brain and behavior. Cam
bridge, MA: MIT Press.
Keysar, B., & Bly, B. (1995). Intuitions of the transparency of idioms: Can one keep a se
cret by spilling the beans? Journal of Memory and Language, 34, 89–109.
Keysar, B., & Henly, A. (2002). Speakers’ overestimate of their effectiveness. Psychologi
Krauss, R. M. (1987). The role of the listener: Addressee influence on message formula
tion. Journal of Language and Social Psychology, 6, 81–98.
Levinson, J. (1992). Intention and interpretation: A last look. In G. Iseminger (Ed.), Inten
tion and interpretation (pp. 221–256). Philadelphia, PA: Temple University Press.
Millikan, R. (2005). Language: A biological model. Oxford: Oxford University Press.
Milosz, C. (2011). Selected and last poems: 1931–2004. New York: Harper Collins.
Pfaff, K., & Gibbs, R. W. Jr. (1997). Authorial intentions in understanding satirical texts.
Poetics, 25, 45–70.
Pickering, M. J., & Garrod, S. (2004). Towards a mechanistic psychology of dialogue. Be
Rogers, T., & Norton, M. (2011). The artful dodger: Answering the wrong question the
right way. Journal of Experimental Psychology: Applied, 17, 139–147.
Schweller, K., Brewer, W., & Dahl, D. (1976). Memory of illocutionary and perlocutionary
effects of utterances, Journal of Verbal Learning and Verbal Behavior, 15, 325–337.
Searle, J. R. (1969). Speech acts. New York: Cambridge University Press.
Searle, J. R. (1990). Collective intention and action. In P. Cohen, J. Morgan, & M. Pollack
(Eds.), Intentions in communication (pp. 401–416). Cambridge, MA: MIT Press.
Sperber, D., & Wilson, D. (1995). Relevance: Cognition and communication (2nd ed.). Ox
ford: Blackwell.
Sperber, D., & Wilson, D. (2002). Pragmatics, modularity and mindreading. Mind & Lan
guage, 17, 3–23.
Spivey, M. (2007). The continuity of mind. New York: Oxford University Press.
Velleman, J. (1997). How to share an intention. Philosophical and Phenomenological Re

search, 57, 29–50.
Page 28 of 29

Wilson, T. (2004). Strangers to ourselves: Discovering the adaptive unconscious. Cam

bridge, MA: Harvard University Press.
Wimsatt, W. (1954). The verbal icon: Studies in the meaning of poetry. Louisville: Univer
sity of Kentucky Press.
Raymond W. Gibbs Jr.
Raymond W. Gibbs, Jr. (PhD 1980) is professor of psychology at the University of Cali
fornia, Santa Cruz. He received his PhD from the University of California, San Diego,
and did postdoctoral research in cognitive science at Yale and Stanford Universities
before joining the faculty at UC Santa Cruz. Gibbs's research focuses on language,
thought, and embodied experience, especially in relation to pragmatics and figura
tive language. He is author of The Poetics of Mind: Figurative Thought, Language,
and Understanding (1994), Intentions in the Experience of Meaning (1999), and Em
bodiment and Cognitive Science (2006), all published by Cambridge University Press.
He is also coeditor of Metaphor in Cognitive Linguistics (with Gerard Steen, 1999)
and is currently editor of the journal Metaphor and Symbol. Gibbs became interested
in Cognitive Linguistics because of his studies on idiom and metaphor processing,
and more general concern with the relations between thought, language, and the
body. Raymond W. Gibbs, Jr., can be reached at [email protected].
Page 29 of 29

Electrophysiological Research on Conversation and Discourse Processing
Electrophysiological Research on Conversation and Dis

course Processing
John C.J. Hoeks and Harm Brouwer

Research into the electrophysiology of language comprehension has essentially been

“speakerless.” This has left three vital aspects of communication—it is social, pragmatic,
and dynamic—severely underresearched. This chapter makes a case for the investigation
of language users involved in active conversation and describes the problems and possi
bilities that accompany this choice of subject. It gives an overview of what is currently
known about how the social, pragmatic, and dynamic puzzles of communication are
solved by the brain, and it describes the well-filled toolbox of language-related event-re
lated brain potential (ERP) components (e.g., Nref, N400, P600) that are at our disposal.
Keywords: Language, communication, conversation, ERPs, pragmatics, N400, P600, Nref
Research into the electrophysiology of language comprehension has essentially been

“speakerless.” For more than 30 years, participants in language experiments have read
isolated and unconnected sentences, word-by-word, from a computer screen. In the past
decade or so, there has been increased, although certainly not overwhelming interest in
presenting language stimuli auditorily, together with a subtle shift from studying isolated
sentences toward looking at multisentence discourse or even minidialogues. However, in
all of these studies, participants are, at best, detached observers of short, unrelated frag
ments of conversations between people they don’t know and don’t care about. A number
of recent studies (e.g., Dumas, Nadel, Soussignan, Martinerie, & Garnero, 2010; Dumas,
Chavez, Nadle, & Martinerie, 2012; Stephens, Silbert, & Hasson, 2010; Suda et al., 2010;
for an overview, see Babiloni & Astolfi, 2012) took issue with that and started to investi
gate participants actively taking part in an interaction. Unfortunately, these studies had
to restrict themselves to looking at very global measures of contiguity and synchronicity
of processing; none of them takes into account what participants say to each other, when
they say it, and how they say it.
This state of affairs is not as strange as it seems. For investigating brain activity through
the analysis of event-related brain potentials (ERPs), the experimenter needs full control
over the experimental stimuli: what they are, how they are realized, and exactly when
Page 1 of 37

they appear. As yet, there is no way to get this kind of control over utterances in free-run
ning interactions (Schilbach et al., 2013; see also Van Berkum, 2012). However, as we ar
gue later, there are ways to approximate real conversation while maintaining strict exper
imental control over content, form, and timing (e.g., the dialogue immersion paradigm;
Hoeks, Schoot, Neijmeijer, & Brouwer, in preparation). But, in addition to this rather
technical reason, there is also a more historical reason for the relative neglect of conver
sation as a research topic. Psycholinguistic investigations of language processing started
out with (p. 366) the implicit adage that if you want to know more about language process
ing, you should start with its basic building blocks: words. Hence, psycholinguistics set
off with the investigation of written words, often presented in isolation. It was soon recog
nized, however, that word recognition and meaning activation crucially depend on the
sentential context in which the words occur and eventually also that the processing of a
sentence can only be properly understood if the larger context of the sentence is taken in
to account. What we are seeing now is the beginning of a third paradigmatic shift in lan
guage processing research, one that carries us from investigating reader’s cognitive re
sponses to a sentence or a text to having participants actively interact in conversation.
Why Active Conversation?

Once a participant enters into an actual conversation, three vital aspects of communica
tion come to the fore that are severely underresearched: (1) communication is social—for
instance, speakers and hearers use a variety of communication strategies in order not to
damage the public image or “face” each puts forward. (2) Communication is pragmatic—
communication is ultimately a means toward a goal, and precisely the appreciation of this
goal-directedness makes our interlocutors try to make sense of what we say (or do). This
“sense making” can sometimes be quite fast, but it often requires considerable inferenc
ing because the words and sounds uttered by our conversational partners (or other non
verbal signals that they emit) are at best pieces of evidence (cf., Sperber & Wilson, 1986,
1995) or instructions to make us think, do, and feel things. Understanding an utterance,
we assume, always requires inferencing, or pragmatic processing, on the basis of the
communication at hand and of what is known about the conversational context, including
the speaker. And (3) communication is dynamic; it is a process that must be monitored
and managed (Clark & Brennan, 1991). Turns must be negotiated and signaled, and feed
back is necessary to indicate that the listener understood what the speaker meant to com
municate. This “backchannel information” can consist of head nodding, smiling, making
small sounds such as “uhuh” or “hmhm,” but can also be fully linguistically expressed and
take the form of occasionally completing an utterance of the speaker (with one or few
words) or even with brief, but complete turns (“yes,” “that’s odd!,” “incredible!,” “Is that
so?”). Leaving out such backchannel information may lead socially sensitive speakers to
stop talking altogether (Kraut, Lewis, & Swezey, 1982).
It is important to note that these social, pragmatic, and dynamic dimensions are highly in
terrelated. For instance, frequency and timing of turn-taking in a conversation is regulat
ed by social factors such as power distance and familiarity between the interlocutors. Al
Page 2 of 37

so, diverging from socially appropriate turn-taking and backchannel scripts will lead to an
increase in pragmatic processing aimed at ascertaining the possible meaning of these
transgressions. Furthermore, the “indirectness” of an utterance, which is strongly deter
mined by social rules of engagement (here, a specific set of rules is applicable, which we
refer to as “linguistic politeness”), will affect the pragmatic processing necessary to un
cover the intended meaning of an utterance. Despite the commonalities among the three
dimensions, it is fruitful to distinguish them and study them separately.
It could be argued that there is no real need to study active conversation, with all its intri
cacies, because essentially every important process and representation will also be
present when conversations are overheard. And, indeed, we all are very familiar with the
role of “eavesdropper,” if not in real situations, then surely when reading a novel or
watching a play or a movie. On each of these occasions, we observe other people interact
and converse, and we can readily predict the feelings and mental representations of the
protagonists. However, it is also intuitively clear that being an observer is something
quite different from being a real actor. This is, of course, especially true for language pro
duction, but it is likely to be equally applicable to language comprehension. For one
thing, you understand a conversation better if you are a real interactant than if you are
merely overhearing (Schober & Clark, 1989). Another important concept is “involve
ment.” When we watch movies, we are motivated to understand what the characters
mean by their utterances and behavior. We can also be emotionally involved, at least to
some degree. However, if you are a real actor (instead of an observer), it becomes essen
tial that you understand what the other person is saying because your face is at stake,
which is clearly not the case when you, for instance, watch a conversation on television.
Lots of things may go wrong during an actual conversation; you might not be able to find
a good answer to a question; you stutter; your conversational partner is not very consid
erate, leading you to experience face loss, which, although often partial and temporary,
can lead to (p. 367) several kinds of negatively valenced emotions (especially embarrass
ment, but also fear, guilt, sadness, etc.). In addition, as an actor, you may not reach your
personal goals if you do not understand what the other person is saying and you cannot
act on that appropriately. Hence, you will be trying your best to make sense of what the
other person wants to communicate. Thus, it seems likely that the processes and repre
sentations involved in language use are different in degree and possibly also in type when
we actively interact versus when we observe the interactions of others.
For these reasons, we propose that the main object of study for the electrophysiology of
language use should be active conversation. We believe it is not enough to look at partici
pants as observers of interaction—although that is certainly a considerable improvement
over the word-by-word rapid serial visual presentation (RSVP) method that is still current
—because active conversation may involve different processes and different representa
tions. This conclusion is shared by Schilbach et al. (2013), who add: “The state of the art
in neuroimaging provides severe limitations to studying free-running interactions using
the full range of verbal and nonverbal channels. Most likely successful studies will need
to identify and isolate salient communicative subsystems” (Schilbach et al., 2013, p. 405).
Moving toward free-running interaction presents a formidable challenge to experi
Page 3 of 37

menters, and looking at specific methodological setups focusing on subsystems seems to

be a necessary step. One approach to bring conversations under experimental control is
through scripting. Hoeks et al. (in preparation) use what they call the dialogue immersion
paradigm (DIP) to have participants take part in a conversation. What participants do not
know is that the responses they get from their conversational partner are prerecorded,
which permits reliable presentation of language stimuli with millisecond precision.
Participant’s own contributions are scripted to match with these responses. Although this
can hardly be seen as free-running interaction, it does permit the crucial involvement of
the participant in what she thinks is a real exchange, in which the participant becomes
sensitive to face-related issues, to pragmatics, and to the dynamics of backchannel infor
mation and turn-taking—the participant is immersed in the conversation. It is not impossi
ble that advances in ERP recording techniques and innovations in experimental design
will lead to interactions that are freer than the very restrictive DIP. However, for now, it
seems to be a viable paradigm to study aspects of active conversation.
Why Event-Related Potentials?

The DIP was put forward as a solution to technical and methodological problems that
arise when a researcher wants to use ERPs to investigate the neural correlates of lan
guage processing. But why would we want to use ERPs in the first place? In a recent
overview, Van Berkum (2012) deplores the state of affairs in the electrophysiology of lan
guage and paints a gloomy picture of a field with lots of scattered findings and little con
sensus on what they mean. However, since he wrote his chapter, there have been new de
velopments. Brouwer, Fitz, and Hoeks (2012), for instance, offer a parsimonious account
of the two most prominent ERP components, the so-called N400 and P600, which we dis
cuss in more detail later. Furthermore, Brouwer and Hoeks (2013) have combined this ac
count of language-related ERP components into a core anatomical model of language pro
cessing. But let us start by describing ERPs and explain how they can inform theories of
language processing.
The Electrophysiology of Language Processing
What Are Event-Related Potentials?

The human brain contains billions of interconnected neurons (or nerve cells), which are
effectively tiny information processors. Through cooperation, networks of these informa
tion processors are able to carry out the complex computations underlying higher cogni
tive functions, such as navigating the physical and social world and satisfying goals. Elec
troencephalography (EEG)—the technique used to measure ERPs—provides a method to
study these neural networks online (during real-time processing) by making use of the
physiological properties of computation in the brain.
Cooperative computation among neurons requires interneuron communication, and, in

the brain, this is handled by electrical and chemical signaling. A signal sent between neu
rons can be either excitatory or inhibitory. If a neuron is excited to a large enough de
Page 4 of 37

gree, it will send out electrical signals, called action potentials, which, upon leaving the
neuron are transformed into chemical signals that either excite or inhibit a next neuron.
The excitation or inhibition of neurons is manifested by tiny voltage fluctuations (in the
order of microvolts; micro =1 per million). If enough of these voltage fluctuations take
place simultaneously and with the (p. 368) same polarity (i.e., positive or negative, relative
to a neutral reference point), they can be measured using electrodes placed on the scalp,
which is what is referred to as EEG. A single EEG recording typically reflects thousands
of neural processes in parallel. Most of these processes may be involved in sensory, cogni
tive, and motor activities unrelated to language processing. When averaging over numer
ous similar EEG recordings time-locked to a specific (e.g., a linguistic) stimulus, this
background activity is filtered out, and what is left reflects the neurophysiological activity
as elicited by the stimulus. This stimulus-related activity is referred to as an event related
brain potential or ERP.
A typical ERP signal is a temporal sequence of negative and positive voltage deflections
(relative to a prestimulus baseline), which are called components. A single ERP compo
nent is taken to index the neural activity underlying a specific computational operation
carried out in a specific neuronal ensemble (Luck, 2005; Näätänen & Picton, 1987). Com
ponents vary in polarity (i.e., they are more positive or more negative than electrodes
chosen as a reference), amplitude, latency, duration, and distribution over the scalp, sug
gesting that different components reflect distinct functional processes, possibly carried
out in distinct cortical regions. This means that if two stimuli are processed differently to
some degree, this difference might be apparent in the ERP responses that they produce.
Kutas and Hillyard (1980) were the first to see the significance of ERPs for the study of
language processing. They presented subjects with sentences containing a final word that
rendered the overall sentence meaning anomalous (“He spread the warm bread with
socks”) and contrasted the ERP signals evoked by these sentence-final words with those
evoked by nonanomalous ones (“He spread the warm bread with butter”). This contrast
revealed that, relative to controls, semantically anomalous sentence-final words produce
an increase in the amplitude of the so-called N400 component, a negative deflection in
the ERP signal that starts around 200–300 ms post-word onset and peaks at about 400
ms. This N400 effect has been interpreted to reflect increased semantic integrative pro
cessing for anomalous relative to nonanomalous sentence-final words. An overwhelming
number of studies have since investigated the factors that modulate the presumed seman
tic processes underlying the N400 component (see Federmeier & Laszlo, 2009; Kutas &
Federmeier, 2011, for overviews). The general picture that has emerged from this work is
that N400 amplitude not only increases in response to semantically anomalous sentence
final words, but is in fact evoked by every content word in a sentence, reflecting how well
this word fits with its prior context (Kutas & Hillyard, 1984; Kutas, Lindamood, & Hill
yard, 1984).
Van Berkum (2012), in his overview on the electrophysiology of discourse and conversa
tion, focused primarily on the N400 component. However, we believe that we should di
rect our attention to the other main language component, the P600, which we think is
Page 5 of 37

more pertinent to the investigation of communication and higher cognitive processing. To

appreciate this shift in focus, we first provide a brief introduction to the “semantic illu
sion” phenomenon, which has received a considerable amount of interest over the past
decade (Bornkessel-Schlesewsky & Schlesewsky, 2008; Brouwer et al., 2012; Chow &
Phillips, 2013; Hoeks, Stowe, & Doedens, 2004; Kim & Osterhout, 2005; Kolk, Chwilla,
Van Herten, & Oor, 2003; Kuperberg, 2007; Kuperberg, Sitnikova, Caplan, & Holcomb,
2003; Nieuwland & Van Berkum, 2005; Stroud & Phillips, 2012; Van Herten, Kolk, &
Chwilla, 2005; Van Herten Chwilla, & Kolk, 2006).
The Semantic Illusion Phenomenon

For a long time, N400 amplitude has been assumed to reflect semantic integration (Brown
& Hagoort, 1993; Chwilla, Brown, & Hagoort, 1995; Hagoort, Baggio, & Willems; Hagoort
& Van Berkum, 2007; Lotze, Tune, Schlesewsky, & Bornkessel-Schlesewsky, 2011; Oster
hout & Holcomb, 1992; Wang, Bastiaansen, Yang, & Hagoort, 2009). On this view, the
N400 component indexes the compositional, or combinatory processes involved in inte
grating the meaning of an incoming word with the semantic representation of its prior
context. However, recently, a different perspective on the N400 component, the memory
retrieval hypothesis, has gained influence (Kutas & Federmeier, 2000; Van Berkum, 2009).
Under the retrieval view, N400 amplitude does not reflect integration (or any other kind
of combinatorial processing) but rather the retrieval from long-term memory of the lexical
and semantic knowledge associated with a word. Retrieval is facilitated if the (lexical and
conceptual) knowledge associated with a word is preactivated by its prior context, lead
ing to a reduction of N400 amplitude. Preactivation occurs through lexical priming (ex
plaining N400 modulations in word pairs), as well as through (p. 369) message-level and
script priming from the context (e.g., Ditman & Kuperberg, 2007; Otten & Van Berkum,
2007; St. George, Mannes, & Hoffman, 1994; Van Berkum, Hagoort, & Brown, 1999; Van
Berkum, Zwitserlood, Brown, & Hagoort, 2003). According to the memory retrieval ac
count, the N400 effect observed by Kutas and Hillyard (1980) reflects the fact that the
conceptual knowledge associated with the nonanomalous sentence-final word butter is
preactivated through both lexical priming (butter fits well with the words bread and
spreading) and message-level or script priming (for instance, spreading and bread
together easily evoke a “breakfast” scenario stored in long-term conceptual memory, in
which things like butter, but also jam, eggs, coffee, etc. are very likely to occur), whereas
the conceptual knowledge associated with the anomalous word socks is not preactivated.
As a consequence of this difference in preactivation, the conceptual knowledge associat
ed with butter is more easily retrieved (because it is preactivated) than that associated
with socks.
It has been quite difficult to decide between these competing interpretations of the N400
component. However, a set of studies investigating so-called semantic illusion sentences
eventually proved crucial for this decision. Hoeks et al. (2004), for instance, tested the
processing of sentences in which two semantically plausible arguments for a verb appear
in the wrong order (e.g., “De speer heeft de atleten geworpen”; lit: “The javelin has the
athletes thrown”), rendering the overall sentence meaning semantically anomalous. Rela
Page 6 of 37

tive to a sentence in which the arguments appear in canonical order (e.g., “De speer werd
door de atleten geworpen”; lit: “The javelin was by the athletes thrown”), the semantic in
tegration and memory retrieval hypotheses make opposite predictions regarding N400
amplitude at the critical verb thrown. The semantic integration hypothesis predicts an
N400 effect at the critical word thrown, reflecting increased difficulty in integrating its
meaning with the preceding context. The memory retrieval hypothesis, conversely, pre
dicts no N400 effect at the critical verb because, in both the reversal and the control con
dition, the conceptual knowledge associated with the critical verb thrown should be
equally easy to retrieve; in both conditions, thrown is primed lexically through javelin and
athletes, as well as through world knowledge (throwing a javelin is a normal thing to do
for an athlete). Consistent with the memory retrieval hypothesis, Hoeks et al. observed no
N400 effect at the critical verb.
Proponents of the semantic integration hypothesis have argued that the absence of an
N400 effect can be explained by assuming that participants made a coherent semantic
representation of the sentence without reference to its syntactic structure. That is, they
assume that through an independent semantic analysis readers were effectively tricked
into a semantic illusion leading them to believe that the stimuli made perfect sense. A few
hundred milliseconds later, however, the processor did somehow find out that this inter
pretation is incorrect. This is reflected in the increase of P600 amplitude, a positive de
flection in the ERP signal that, on average, reaches its maximum around 600 ms. In some
of these “multistream” models—all of these models include an independent semantic ana
lyzer and an algorithmic/syntactic stream and sometimes additional processing streams
as well—the P600 effect is associated with effortful syntactic processing in order to make
sense of the utterance (see Gouvea et al., 2010, for an overview of syntactic interpreta
tions of the P600). Other multistream models view the P600 as indexing the resolution of
a conflict between syntactic and semantic streams (Bornkessel-Schlesewsky & Schle
sewsky, 2008; Kim & Osterhout, 2005; Kolk et al., 2003; Kos, Vosse, van dem Brink, & Ha
goort, 2010; Kuperberg, 2007; van de Meerendonk, Kolk, Chwilla, & Vissers, 2009; van de
Meerendonk, Kolk, Vissers, & Chwilla, 2010; Van Herten et al., 2005, 2006; Vissers,
Chwilla, & Kolk, 2006, 2007; Vissers, Chwilla, Egger, & Chwilla, 2013; Ye and Zhou,
2008). However, in an extensive review, Brouwer et al. (2012) showed that none of the
complex multistream models can actually explain all of the relevant data. They argue for
a single stream architecture, the retrieval-integration account, which is based on differ
ent assumptions of the functional significance of the N400 and the P600.
Brouwer et al. argue that the retrieval account of the N400 component offers the most
parsimonious explanation because the absence of an N400 effect in reversal anomalies
can easily be explained as lexical and message-level priming; no special semantic stream
is necessary (see Chow & Phillips, 2013; Stroud & Phillips, 2012, for similar arguments).
However, if the N400 component does not reflect compositional or combinatorial seman
tic integration processes, then where in the ERP signal do these processes show up? Giv
en that semantic integration is essential for the comprehension system, one would expect
it to be reflected in the electrophysiology of language processing. Brouwer et al. suggest
that these integrative semantic processes are (p. 370) indexed by the P600 component.
Page 7 of 37

They hypothesize that the P600 is not a single component but actually a family of late
positivities that reflect the word-by-word construction, reorganization, or updating of a
mental representation of what is being communicated—the “MRC.”
An MRC is very similar to a “mental model” (Johnson-Laird, 1983) or “situation

model” (Kintsch, 1998; Zwaan & Radvansky, 1998). The difference with these models is
only of degree: whereas mental models and situation models represent interpretations of
all types of situations, be they linguistic or otherwise, an MRC is a representation of how
a person interprets his or her current communicative situation. A communicative situa
tion, such as a conversation, can be seen as containing a number of different aspects: a
representation of the current utterance, a representation of the discourse/conversation
until now, a representation of the “face” of self and of the other interactants, or a repre
sentation of the language user’s private representation of self and others, including their
long-term and short-term (conversational) goals. All of these aspects are assumed to be
represented in the current MRC of an individual involved in a conversation. And adapting
this MRC leads to a increase in P600 amplitude.
Summarizing from the perspective of the retrieval-integration account, the P600 compo
nent is modulated by each word in a sentence as the lexical information activated by a
word is integrated into the current mental representation, resulting in an updated repre
sentation of the input given thus far. The effort involved in this integrative processing is
reflected in P600 amplitude; that is, the more effort it takes to arrive at a coherent MRC,
the higher the amplitude of the P600 component (see Molinaro, Carreiras & Duñabeitia,
2012, for a clear example of such graded P600 modulations evoked by non-anomalous ex
pressions). An important consequence of this view is that language comprehension pro
ceeds in biphasic N400–P600 cycles. The retrieval of the conceptual information associat
ed with a word modulates the N400 component, and the integration of this information
with an unfolding MRC into an updated representation is reflected in P600 amplitude. Im
portantly, we propose that these two components are responsive to general cognitive pro
cessing, and that neither of them is language-specific: every meaningful stimulus will
cause the retrieval of those features in long-term memory that are associated with that
stimulus, and thus, every meaningful stimulus will evoke an N400. Likewise, creating a
new representation of a (linguistic or nonlinguistic) situation, or any change in that repre
sentation, is hypothesized to lead to an increase in P600 amplitude.
The Brain Basis of the Retrieval-Integration Account
Retrieval-integration cycles provide a general and parsimonious account of the elicitation

patterns of the N400 and the P600 component. This sheds light on the how and the when
of comprehension, but not on the where, which is indispensable information for creating
and supporting a viable model of language comprehension. Event-related potentials have
a poor spatial resolution, which makes it very difficult to pinpoint where an observed ERP
signal was generated in the brain. Hemodynamic-based (“blood-based”) neuroimaging
techniques, by contrast, such as functional magnetic resonance imaging (fMRI) and
positron emission tomography (PET), can be used to localize cognitive processing but
Page 8 of 37

have a poor temporal resolution. This might suggest that electrophysiological and hemo
dynamic methods provide complementary information, which we would ideally like to
combine to arrive at a viable neurobiological model of language comprehension. Howev
er, due to the fundamentally different nature of electrophysiological and hemodynamic
measurements, it is not immediately clear how they can be combined; that is, due to their
differences in spatial and temporal resolution, it is often impossible to simply align them
for a given experimental paradigm (cf. Lau, Phillips, & Poeppel, 2008).
Fig. 23.1 Schematic overview of a typical retrieval-

integration cycle in the left hemisphere. Depending
on whether the linguistic input is spoken or written,
an incoming word reaches the left posterior middle
temporal gyrus (pMTG) via either the auditory cortex
(ac) or the visual cortex (vc), respectively. The pMTG
then retrieves the lexical information associated with
a word from the association cortices, which gener
ates the N400 component. Via the dorsal pathway
(dp) or the ventral pathway (vp), the retrieved infor
mation is connected to the left inferior frontal gyrus
(IFG), where it is integrated with its prior context
(the prior mental representation of what is being
communicated; MRC) into an updated representation
of what is being communicated (an updated MRC).
This integrative processing generates the P600 com
ponent. Finally, the updated MRC is connected back
to the pMTG via one of the pathways, providing a
context for upcoming words, upon which a new re
trieval-integration cycle may start.
A more promising strategy, therefore, is to start from the processes assumed to be re
flected in the different language-related ERP components and map these onto cortical ar
eas or networks that could host them. On the basis of this “process alignment strategy,”
Brouwer and Hoeks (2013) have recently proposed a minimal functional-anatomic map
ping of the retrieval-integration account that focuses on the anatomical and computation
al epicenters (cf. Mesulam, 1990, 1998) or hubs (Buckner et al., 2009) of the language
network. Building on several large-scale reviews on the cortical organization of the com
prehension system (Bookheimer, 2002; Dronkers, Wilkins, Van Valin, Redfern, & Jaeger,
2004; Friederici, 2002, 2011; Hickok & Poeppel, 2004, 2007; Lau et al., 2008; Shalom &
Poeppel, 2008; Turken & Dronkers, 2011; Vigneau et al., 2006), they identified the left
Page 9 of 37

posterior middle temporal gyrus (lpMTG; Brodmann area [BA] 21; the area represented in
darker grey in Figure 23.1) as (p. 371) an epicenter that mediates the lexical retrieval
processes underlying the N400 component. That is, the lpMTG is assumed to mediate the
mapping of word form to conceptual representations, which are stored in a distributed
manner across the association cortices (cf. Elman, 2004, 2009; Pulvermüller, 1999, 2001;
Rogers & McClelland, 2004). As such, the focus of activity reflected in N400 amplitude is
assumed to take place in the lpMTG, but the full range of activity reflected in the N400
component also includes the activation of conceptual features in the association cortices.
The integrative processes underlying the P600 component, in turn, are assumed to be me
diated by the left inferior frontal gyrus (lIFG; BA 44/45/47; the lighter grey area in Figure
23.1), which is a neuroarchitecturally complex, highly parcellated area (Amunts et al.,
2010; Amunts & Zilles, 2012; Friederici, 2011). Brouwer and Hoeks suggest that this
anatomical parcellation may underlie a fine-grained functional topology of the lIFG (see
also Friederici, 2011; Hagoort, 2005), in which different parcels of the area may carry out
different subprocesses involved in MRC construction, thereby unifying ongoing debates
about the functional role of the lIFG (see Grodzinsky & Santi, 2008; Rogalsky & Hickok,
2011, for overviews). On this view, the mental representation of what is being communi
cated is maintained in (or its activity is coordinated by) the lIFG. Critically, linking the
P600 to the lIFG suggests that characteristically different P600s in terms of electrophysi
ological properties like onset, duration, amplitude, and scalp distribution, which Brouwer
et al. (2012) assume to reflect different subprocesses of MRC construction, may have an
anatomical basis in different parcels of the lIFG. Like the lpMTG and the N400, the lIFG
is assumed to serve as an epicenter that is host to the focus of activity reflected in P600
amplitude, but the full range of activity underlying the P600 component may include ac
tivity from adjacent areas as well.
The retrieval epicenter (lpMTG) and the integration epicenter (lIFG) are wired together
by means of two white matter pathways, a dorsal pathway (including the classic arcuate
fasciculus and the superior longitudinal fasciculus) and a ventral pathway (including the
extreme fiber capsule system, the uncinate fasciculus, and the inferior and medial longi
tudinal fasciculi), the precise functions of which are still much debated (Baggio and Ha
goort, 2011; Friederici, 2009, 2011; Hickok & Poeppel, 2004, 2007; Tyler et al., 2011;
Weiller, Musso, Rijntjes, & (p. 372) Saur, 2009). Consequently, it is as yet not possible to
choose which of these pathways supports the connection of information retrieved in the
lpMTG to the lIFG for integration and which pathway serves to connect the updated MRC
back from the lIFG to the lpMTG to provide a context for upcoming words. The presence
of these pathways, however, indicates that there are at least two pathways supporting re
trieval-integration cycles.
Putting this all together, we can walk through a typical functional-anatomic retrieval-inte
gration cycle. An incoming word reaches the lpMTG via either the auditory cortex (ac; see
Figure 23.1) or the visual cortex (vc), depending on whether the linguistic input is spoken
or written. The lpMTG then mediates the retrieval of the lexical information associated
with this word from the associations cortices, where it is stored in a distributed manner.
This process generates the N400 component of the ERP signal, the amplitude of which
Page 10 of 37

corresponds to the ease of retrieval. The retrieved lexical information is linked up to the
lIFG via either the dorsal (dp) or ventral pathway (vp), where it is integrated with its pri
or context (the prior MRC) into an updated representation of what is being communicat
ed (an updated MRC). The extent of work required to get this updated representation of
what is communicated is reflected in P600 amplitude. Meaning aspects of the updated
representation are fed back from the lIFG to the lpMTG via the dorsal or ventral pathway,
which provide a context for possible upcoming words on which a new retrieval-integra
tion cycle may start. Figure 23.1 provides a schematic overview of a retrieval-integration
cycle.
The functional-anatomic mapping of the retrieval-integration account proposed by Brouw

er and Hoeks (2013) is not intended to be a full-fledged neurobiological model of lan
guage comprehension but the extendable core of the comprehension system (see Brouw
er, 2014, for a proof of concept in terms of neurocomputational simulations). As such, it
can serve as a starting point for research into the coupling of brain anatomy and electro
physiology, as a step toward a neurocognitive model of communication.
The Electrophysiology of Discourse and Con

versation
Conversation is without doubt the most natural form of human interaction. Some have
likened it to a dance, in which two partners together carry out a complex pattern of coor
dinated movement. However, this metaphor is perhaps a bit too friendly and also a bit
misleading. The dance metaphor suggests that the two partners in the interaction are
equal and that the interaction is fun. Although that may be true in some cases, it certainly
is not in others. If conversation is to be considered a dance, it is one in which, often
enough, one of the dancers is dominant and makes the other dancer go somewhere, in
stead of just giving a good show. But what is more, conversation isn’t fun—it’s serious
business, an essential part of human social functioning.
Social Aspects of Communication
As biological beings, we humans are born with a set of goals, the primary goal being pro
creation (some would formulate it slightly differently: our genes have goals, namely to
spread; we are “survival machines” for our genes; cf. Dawkins, 2006). To reach the goal of
procreation, evolution has provided us with internal drives that keep us alive long enough
to procreate and take care of our offspring. Brain research has suggested that these moti
vators are implemented in specific areas deep in the brain, such as the hypothalamus and
other parts of what researchers have sometimes referred to as the limbic system. These
motivators (hunger, thirst, sex drive, aggression, etc.) create a web of distal and proximal
goals, hierarchically ordered in subgoals, sub-subgoals, and so forth. To reach these
goals, we have to negotiate the physical world, but because we are also social animals, we
encounter our conspecifics while trying to satisfy our goals, so we need to negotiate the
social world as well. Furthermore, some of our goals lie entirely in the social realm, goals
Page 11 of 37

like being recognized and valued. What separates us from more solitary animals is that
we are not alone in the world, we have a public, we are “seen,” every moment of the day,
if we like it or not. And that is where communication comes in: we can get what we want
only by interacting with our fellow human beings because they are everywhere.
The ubiquity of fellow human beings means that there is a pressure to adapt our behavior
to certain social rules that have developed to minimize within-group competition while
maintaining high chances of individual survival. Being a social animal often means that
we have to inhibit our first, impulsive original plans and adapt them to suit the require
ments of the social situation. These impulsive behaviors may involve impulses to mate
with a certain individual, to grab food that is available, to drink, to hit someone who
comes too close or who (p. 373) shows interest in your piece of food. The social rules that
have been created to regulate these impulses to minimize conflict are sometimes highly
complex and take a long time to learn (you need to be with children only a couple of min
utes to appreciate the latter fact). Some of these social rules can be considered a sort of
“baseline” and preclude, among other things, showing behavior in public that has to do
with bodily functions (belching, farting, scratching, nose picking, grunting, etc., with the
exception of eating), but also include rules regarding clothing, posture, hairdo, gait, bodi
ly proximity, duration of eye contact, and more. The majority of social rules, however,
seem in place to govern active interaction. The whole set of social rules regulating behav
ior, which varies from culture to culture, can be called “politeness,” subsuming “linguistic
politeness.” Linguistic politeness rules specifically concern what to say and how to say it.
Brown and Levinson (1987) argue that the rules for formulating, for instance, requests to
another person are not random and haphazard but are centered around “directness.” As a
general rule, extreme directness (“bald on record”), as in “Leave this room,” is only al
lowed under very specific circumstances, such as when there is an emergency or when
the power distance between interactants is very high. In all other cases, more indirect ut
terances are required, in which requests are embedded in lexical material, in which rea
sons are given and the speaker apologizes (“I am really sorry, but could you please leave
the room, we have to clean the floor”).
We have many social goals, such as getting what we want from people (through persua
sion). However, one of our most important social goals is what is sometimes called “im
pression management” (Schlenker, 1980), a concept derived from the notion of “face” in
troduced by Erving Goffman (1959, 1967)—how to create the desired image of yourself in
the public that is currently present. This impression can be said to have an upper limit,
which is the representation we desire—that is, how we want to seem (to this end, we also
use postures, clothes, and other accessories, such as cars, houses, etc.; see Goffman,
1959)—and a perhaps more general lower limit (that we want to avoid), which is the
“neutral” baseline mentioned earlier. According to Goffman (1959), each time we are in a
social situation, we choose a “stance” or “role,” for instance, one in which we are very
successful in our job, our family is doing well, and we are happy, but also friendly and
sympathetic—all of these are of course idealizations of our real mammalian selves. Effec
tively, we thus have at least two types of interactant representations that have to be taken
into account in a given interaction— our own public face and the public face of the inter
Page 12 of 37

locutor—but also the two corresponding private representations of self and other at play
that may overlap with the public face representations but that are certainly not identical.
Most people are surprisingly good at face management during conversation. This is no
small feat because it requires handling at least four different representations (public face
of the interactants, plus the private representations including the interactants’ goals), a
representation of the current situation, and a set of complex social rules, including the
specific and quite intricate rules for conversation. But we all know people who either
seem to lack the social knowledge required to take the face of the interactant into ac
count or who simply do not care too much about the consequences that face damage
might have, either regarding their own face or the face of others. So there are individual
differences in social sensitivity and social knowledge (which will most often be correlat
ed) that can increase the chance of face-threatening situations. Sometimes face damage
is the result of clumsiness in you yourself or your fellow interactant. However, it may also
be intentional, motivated by aggression, anger, disappointment, and other emotions (cf.
Bousfield, 2008; Culpeper, 1996). This illustrates that the rules of social engagement are
essentially normative: they can be broken, although there will be consequences.
Despite the importance of the social aspects of language use we just discussed, there
have been no electrophysiological studies on how social aspects of conversation are
processed. There is only one study by Hoeks, Schoot, Taylor, and Brouwer (in prepara
tion), who looked at participants “overhearing” visually presented minidialogues. One of
the protagonists in the minidialogue asks the other person for a favor or asks a mere
knowledge question. Social rules of conversation tell that a blunt “No,” without giving
reasons or apologizing is not acceptable, even when people know each other really well
(in the experiment, participants were told that the miniconversations were between two
people who were in a relationship).
Request condition
partner 1: “Could you please put out the garbage tomorrow?”
partner 2 (blunt): “No.”
partner 2 (softened): “No, I’m sorry, I have to leave early”
Knowledge Question condition
partner 1: “Did you know that John will put out the garbage tomorrow?”
partner 2 (blunt): “No.”
partner 2 (softened): “No, I’m sorry, I didn’t know”
Hoeks et al. found a significant positivity for the blunt NO (“No.” with a period),
(p. 374)
as compared to the softened NO (“No,” with a comma, later followed by an apology) for
both requests and knowledge questions. In the request condition, however, the positivity
started earlier, had a broader scalp distribution, and was of a larger magnitude than in
Page 13 of 37

the knowledge question condition. These positivities were interpreted as P600 effects re
flecting the reorganization of the MRC as a result of less than polite linguistic behavior.
Part of the processing may have involved the updating of speaker representation—“this
person saying ‘no’ is rude, I better be on my guard.”
To what extent this knowledge about the speaker is represented as an integral part of the
developing discourse representation, or is relatively separate, is a matter for future re
search. We do know from fMRI research that the application of social rules most likely in
volves areas in the frontal cortex (i.e., medial prefrontal cortex, see Barbey & Grafman,
2011). In addition, studies have identified the bilateral temporal parietal junction (i.e., in
both hemispheres) as an area that may be specifically involved when reasoning about the
intentions of persons in interaction, which makes it a good candidate brain area support
ing a separate speaker representation (Bara, Ciaramidaro, Walter, & Adenzato, 2011; De
cety & Lamm, 2007). It will have to be established how a representation of speaker fits in
to the anatomical core network of language processing that we described earlier.
Returning to the experiment on politeness just described, part of the processing difficulty
that was found for bluntly saying “No” may in some sense also be pragmatic; what did the
person want to communicate by being blunt? That is, why did the person refrain from of
fering an apology or some kind of a motivation for the refusal? This kind of inference,
where the intended meaning of the speaker is recovered, is further discussed in the next
section.
Pragmatic Aspects of Communication
Only very few experiments on language comprehension featured real speakers. Generally,
participants simply read isolated sentences or unrelated texts from a screen and never
have to wonder what the speaker would actually mean by these sentences or texts be
cause, of course, there was no one there to mean something. However, in real life, lan
guage is uttered with a goal—speakers want to get a certain message across—and listen
ers try their best to get at that message. Utterances that are fully unambiguous do not ex
ist. This means that listeners will have to engage in effortful processing to get a grip on
what the speaker meant to communicate. This processing can stop once the listener man
ages to create a coherent representation of the conversation so far. Coherence in conver
sations is achieved by three means: (1) establishing reference, (2) using information
structural cues to identify the new information and how and where it fits in the larger dis
course, and (3) using inference to establish what the speaker meant to communicate or
when there are problems with (1) or (2). Precisely because we know that a speaker has a
goal and therefore wants to be coherent, listeners operate on the basis of the evidence
presented in the utterance, on the basis of the representation of the situation, and on the
basis of the knowledge of the speaker (including what is known or might be surmised
about his or her goals) to figure out what is meant.
Page 14 of 37

Establishing Coherence: Reference

Every time a noun phrase (the baker), a name (John), or a pronoun (he) is encountered,
the listener must determine the entity this expression refers to: its antecedent. Retrieving
the conceptual features associated with this referring expression seems to be a prerequi
site for finding the right antecedent. Full nouns (e.g., “baker”) will require a more exten
sive retrieval process than either names or pronouns because these latter are rather
“knowledge lean.” For instance, in isolated sentences such as “John hit Mary,” names
such as “John” and “Mary” contain very little information. Similar to pronouns, such
names will hardly activate more features than those regarding number (singular vs. plur
al) and gender (grammatical and/or biological). However, if the name refers to a known
figure, either from reality (“Bram Stoker”) or from fiction (“Count Dracula”), there is
more information about this person in long-term memory that can be retrieved. Note that
these retrieval processes can be expected to be especially extensive when nouns or “real”
names are mentioned in a discourse for the first time; retrieval effort will be much re
duced for repeated mentions. After this meaning activation, the processor can look for a
matching antecedent.
Van Berkum, Brown, and Hagoort (1999) were the first to look at the electrophysi
(p. 375)
ological correlates of processing reference. They presented participants with sentences

that contained a singular definite noun phrase (NP), like “The dean told the lecturer…,”
embedded in a story that had either introduced a single referent (one lecturer) or two
equally plausible referents (two lecturers). Relative to the one-referent condition, the crit
ical NP in the ambiguous, two-referents condition was found to produce a sustained nega
tivity that was broadly distributed over the scalp but with a distinct frontal focus starting
at about 200 ms post word onset. The increase in the amplitude of this sustained negativi
ty has been labeled the Nref effect and has been taken to reflect the increased difficulty
of establishing reference when there are multiple possible antecedents. This interpreta
tion is supported by the finding that the Nref effect disappears if the ambiguity is elimi
nated before the critical NP, for instance, by having one of the referents leave the scene
(Nieuwland, Otten, & Van Berkum, 2007).
Van Berkum (2009) suggests that the Nref effect is not a response to an anomaly, but “the
brain’s natural inclination to immediately relate every shred of new information to what is
known already” (p. 287). In line with that, we would like to hypothesize that each refer
ring expression elicits an Nref, a negative deflection of the ERP signal that reflects the
search for the right antecedent. We need more research to provide a detailed view of the
Nref as a component, but we already know a number of things about the Nref as an effect
(see Van Berkum, Koornneef, Otten, & Nieuwland, 2007, for review). For one, it is often
broadly distributed over the scalp, although quite pronounced frontally. Also, the Nref ef
fect often appears as a sustained negativity, although, as we argue below, it can some
times be very difficult to judge whether we are dealing with a sustained effect or merely a
shifted waveform that did not return to baseline. Furthermore, the extent of the Nref ef
fect may depend on the ease with which the reference problem can be resolved. In the
case of a global ambiguity (e.g., “Robert and Daniel were talking, and he…”), the proces
Page 15 of 37

sor may have a hard time finding the right antecedent for “he,” leading to sustained pro
cessing difficulty; in other cases, the negative shift may appear to be more “phasic.” And,
finally, as far as latency is concerned, it has been claimed that the Nref appears no earlier
than 300–400 ms post onset, but the difference waves presented in Van Berkum et al.
(1999; p. 157, fig. 1b) suggest that the original Nref effect already starts around 200 ms
after the onset of the ambiguous word. It is highly likely that the Nref as a component
starts even earlier.
It seems plausible that the processor first activates salient knowledge about a given refer
ring expression before being able to find the most plausible antecedent. As we argued
above, this activation of relevant knowledge will be reflected in the amplitude of the
N400 and referent resolution in the Nref. If we compare the latencies of the ERP compo
nents involved (i.e., the N400 for meaning activation and the Nref for establishing refer
ence), these two processes seem, for a large part, to be running in parallel. Reference as
signment thus does not seem to wait for meaning activation to have completed. This
makes it difficult to separate the two components (and even raises the question whether
they are all that different, except for the representational system—conceptual memory or
current discourse representation—that is consulted).
Summarizing, we assume that as soon as referring expressions are encountered, their

lexical semantic features will be activated, giving rise to an N400, and the right an
tecedent will be sought, giving rise to an Nref. We do expect considerable differences be
tween nouns, names, and pronouns. As we pointed out earlier, nouns (and “real” names)
will require the most extensive retrieval from long-term memory and will thus elicit the
largest N400 amplitude. The associated Nref is most probably running in parallel with or
slightly behind the N400 and will thus overlap with it. In contrast, names and pronouns
have relatively little lexical and semantic information associated with them, so the re
trieval stage will probably be rather restricted; thus, the processing of names and pro
nouns will most likely be most pronounced in the Nref component. These assumptions
mesh with the retrieval-integration account of language processing presented earlier. The
retrieval-integration account has identified two core systems that are both functionally
and anatomically distinct: the semantic memory system containing all lexical and world
knowledge, with a hub in the temporal lobe; and a working memory system involved in
creating, adapting, and maintaining a representation of what is currently communicated,
with a hub in the inferior frontal lobe. It makes sense that in the interpretation of nouns,
both of these representations are accessed to a certain degree. The same is likely true for
the interpretation of names and pronouns but, in this case, the balance is skewed toward
accessing the situation model/MRC. We propose that searching the (p. 376) situation mod
el leads to an Nref, whereas searching the semantic memory system leads to an N400.
This is in line with Kotchoubey’s (2006) account of the relation between component polar
ity and cognitive processing. On that account, negative shifts (N400, Nref) reflect
preparatory processing and search, whereas positive shifts (P600) reflect combinatory
processing and synthesis. An fMRI study by Nieuwland, Petersson, and Van Berkum
(2007) showing that referential processing leads to activation of frontal brain areas pro
vides further support of this hypothesis. So, for content nouns (and “real” names), we ex
Page 16 of 37

pect both N400 and Nref; for pronouns and “empty” names, we expect an Nref and hard
ly any modulation of the N400. This is consistent with findings from a recent study by
Taylor and colleagues (Taylor, 2013; Taylor, Stowe, Redeker & Hoeks, in preparation).
They presented participants with spoken minidialogues like the following:
A: “Lisa lost the chess tournament”
B: “Oh?”
A: “The teacher congratulated her.”
Compared to dialogues in which Lisa won the tournament, there was no N400 effect on
the pronoun, even though its antecedent was unexpected, because it is strange to con
gratulate Lisa if she just lost a tournament; a P600 effect was found instead.
To be clear, there have been reports of “N400 effects” for names and pronouns. Indeed,
there is a comprehensive literature concerned with the so-called repeated name penalty.
Ledoux, Gordon, Camblin, and Swaab (2007), for instance, found longer reading times for
a repeated name in an eye-tracking study using sentences such as: “At the office, Daniel
moved the cabinet because Daniel needed room for the desk.” Repeating the name
(Daniel) to refer to the same antecedent is infelicitous and leads to processing difficulty
(hence the “penalty”). In a replication of this experiment with ERPs reported in the same
paper, Ledoux et al. found a negative shift for the repeated name as compared to a felici
tous control sentence (“At the office, Daniel and Amanda moved the cabinet, because
Daniel needed room for the desk”). They interpret this negativity as an N400 effect, but,
in the light of our previous discussion, it is more likely to have been an Nref effect. In
deed, a close look at the data from Ledoux et al. reveals that the negativity is broadly dis
tributed over the scalp and clearly visible at frontal electrodes, a pattern that is very simi
lar to the repeated name penalty effect found in an earlier ERP experiment (Swaab, Cam
blin, & Gordon, 2004). The auditory version of the same experiment also produced a neg
ative shift that was actually largest at frontal electrodes (Camblin, Ledoux, Boudewyn,
Gordon, & Swaab, 2007). Thus, it seems that if a referring expression is “marked” in
some way (using the full form “Daniel” when a reduced form, the pronoun “he,” is called
for), it is more difficult to find the correct referent, leading to an increase in Nref ampli
tude. Interestingly, the Camblin et al. data show that if the name is not a marked form, as
in the felicitous control sentence, the negativity is actually reduced as compared to a con
trol sentence, indicating that finding the antecedent has become easier. This repetition ef
fect on the Nref mirrors the well-known repetition effect on the N400 (Kutas & Feder
meier, 2011).
Whether we can categorize a given effect as an Nref or N400 of course depends crucially
on the estimated onset and scalp distribution observed. Unfortunately, though, onset and
scalp distribution are also known to vary with materials used, mode of delivery (auditory
versus visual), and the individual differences in functional and anatomical organization of
Page 17 of 37

language in the participants. We return to this issue in the discussion section because it is
a general problem that affects the interpretation of other components and effects as well.
Until now, we have discussed cases in which the antecedents are present in the discourse
model/MRC of the listener. When there is not yet an antecedent, the listener must create
a brand new discourse entity. In case of definite referring expressions, the definiteness
acts as a signal to listeners that they are supposed to have an antecedent for this expres
sion (i.e., the “definiteness presupposition”; Russell, 1905). If that is not the case, listen
ers may have to accommodate the antecedent in the current discourse representation.
Accommodation of discourse entities has been shown to lead to P600 effects. For in
stance, Burkhardt (2006) compared the following discourse fragments:
Tobias was having a chat with Nina. He said that the conductor was…
Tobias visited a conductor in Berlin. He said that the conductor was…
Only in the second fragment is the definite expression “the conductor” from the second
sentence properly introduced in the first sentence, namely as an indefinite NP (“a con
ductor”). In the first discourse fragment, the conductor from (p. 377) the second sentence
is not introduced, leading to a P600 effect, presumably due to accommodation.
Sometimes, the antecedent for a referring expression can only be identified through so-
called reference transfer (cf., Nunberg, 1979). Schumacher (2011) examined the process
ing of mini-discourses in which the second sentence contains a semantic anomaly (e.g.,
“The doctor asks his assistant again who called that early. The assistant responds that the
hepatitis had called that early”). In its literal sense, the critical NP “the hepatitis” is a
poor agent for the verb “call,” rendering the sentence anomalous. However, this anomaly
disappears if the reference of “the hepatitis” is transferred from its literal meaning (e.g.,
liver inflammation) to a referent that is contextually associated with it (e.g., a patient suf
fering from liver inflammation). Relative to a minidiscourse in which no transfer of refer
ence was required (e.g., “The doctor asks his assistant again who called that early. The
assistant responds that the therapist had called that early”), Schumacher found that the
critical NP “the hepatitis” produced a P600 effect, which she interpreted as reflecting en
riched composition.
There is also the possibility that antecedents are present, but none of them is available
for the referring expression that is used. For instance, Van Berkum et al. (2007) review a
number of studies looking at sentences such as “Anna shot at Linda as he…,” where there
is no matching antecedent. Compared to the control sentence “David shot at Linda as
he…,” the failure to find an antecedent led to a P600 effect. Thus, the failure to find a
matching referent leads to immediate interpretation problems.
In conclusion then, we propose that each referring expression elicits an Nref from search
ing the discourse representation or MRC. Marked delivery, for instance by using a full NP
where a pronoun is required, will lead to an increase in the Nref amplitude, reflecting the
extra effort needed to find the correct antecedent. In addition to an Nref, all referring ex
Page 18 of 37

pressions will elicit an N400 due to the activation of lexical semantic information—con
tained in the conceptual memory system—although this is most pronounced for nouns
and “real” names. If establishing reference turns out to be impossible or leads to an im
plausible interpretation, a P600 will ensue, reflecting the reorganization of the MRC. A
P600 may also appear in the absence of serious problems like those mentioned above
(i.e., persistent reference failure or semantic anomaly), namely, when a discourse entity
needs to be accommodated or when the referring expression needs some “preprocessing”
before the antecedent can be successfully identified.
Establishing Coherence: Information Structure

In discourse, and especially in conversation, coherence is achieved not only through ref
erential links, but also via the information structure of an utterance, and more specifically
through the information structural dependencies between an utterance and its preceding
context. Information structure can be broadly described as the way the information con
tained in an utterance is segmented (cf. Chafe, 1979). In general terms, the informational
content of an utterance can be divided into information that is backgrounded, also termed
topic or “given,” and information that is in focus, also termed “comment” or “new.” The
way in which the informational status of these different parts of the message is signaled
varies per language. For instance, in languages like English and Dutch, prosody plays an
important role in the marking of topic and focus: the topic is deaccented, and the focus
part of the message receives a pitch accent. Other languages, such as Italian and French,
predominantly use more structural means, such as clefting and word-order variations to
indicate focus structure (Lambrecht, 1994). Marking the information structure of an ut
terance enables the listener to know which aspect of the message is meant as the “real”
information and what part of the current discourse representation should serve as the
“attachment site” for this information.
The importance of information structure becomes apparent if information structure cues

are misplaced. Quite a few studies have looked at what happens if backgrounded informa
tion is treated as new or new information is treated as if it were already known (e.g.,
Bornkessel et al., 2003; Cowles, Kluender, Kutas, & Polinsky, 2007; Dimitrova, Stowe, Re
deker, & Hoeks, 2012; Heim & Alter, 2006; Hruska & Alter, 2004; Hruska, Alter, Stein
hauser, & Steube, 2001; Li, Hagoort, & Yang, 2008, Li, Yang, & Hagoort, 2008; Magne,
Astésano, Lacheret-Dujour, Morel, Alter, & Besson, 2005; Toepel & Alter, 2004; Toepel,
Pannekamp, & Alter, 2007; Wang, Hagoort, & Yang, 2009; Wang et al., 2011). For in
stance, Dimitrova, Stowe, Redeker, and Hoeks (2012) looked at spoken minidialogues in
which, in some conditions, the topic of the target sentence sometimes received a pitch ac
cent (superfluous accent), whereas in other conditions, the focus element sometimes re
mained unaccented (missing accent). (p. 378)
Q: Did the club give a bonus or a fine to the player?
#A: The club gave a bonus to the PLAYER. (missing accent on bonus)
A: The club gave a BONUS to the player.
Page 19 of 37

Q: Did the club give a bonus to the player or to the trainer?
#A: The club gave a BONUS to the player. (superfluous accent on bonus)
A: The club gave a bonus to the PLAYER.
Both missing and superfluous accents were shown to lead to P600 effects, indicating
problems in creating a coherent MRC. In addition, the superfluous accent condition also
showed a long-lasting right-lateralized negativity. Importantly, in this study, and also oth
er studies reviewed here, treating background information as new was operationalized by
adding a pitch accent to the noun of the topical NP (here: “a bonus”). As we have seen in
the studies on the repeated name penalty discussed earlier, marked delivery in the form
of an overspecification (“Daniel” instead of “he”) leads to increased search in the situa
tion model, causing an increase in Nref amplitude. Accenting a noun that should not be
accented according to information structure rules could be seen as a marked delivery as
well. This could explain why Dimitrova et al., but also other studies, found a negativity.
Dimitrova et al. suggested that this negativity may have been an N400 effect, but we
would like to entertain the possibility that the negativity was not an N400 effect, but an
Nref effect, instigated by the marked delivery of the referring expression. If we look at
the relevant waveforms in Dimitrova et al. (2012, fig. 5), we can see that the negativity
starts around 200 ms post onset and goes on for quite some time (at some electrodes
even until after 1,000 ms post onset) and is also present at frontal electrodes.
It is important to note that, in contrast to many of the earlier studies, Dimitrova et al. did
not use a prosodic judgment task precisely because they hypothesized that such a task
might interfere with the normal processing of the minidialogues. They were concerned
that, instead of focusing on the meaning of the exchange, participants’ attention would be
directed to categorizing the stimulus and making a decision as to whether it is prosodical
ly and contextually well-formed. These kinds of task-induced cognitive processes involve
controlled processing, decision making, and keeping the decision in memory until it is
time to respond, plus the associated motor preparation for pushing the button. Thus, the
actual comprehension processes of spoken minidialogues will be confounded with deci
sion-related processing (see Magne et al., 2005, for similar reasoning).
If the application of a superfluous pitch accent on a referring expression should be seen

as a form of marked delivery, giving rise to Nref effects rather than N400 effects, this
should hold for accented nouns (see Dimitrova et al., 2012, for an overview) as well as for
accented names and pronouns. The experiment by Taylor et al. (in preparation; Taylor,
2013) that we discussed in the section on reference included conditions in which the criti
cal pronoun was accented.
A: “Lisa won the chess tournament”
B: “Oh?”
A: “The teacher congratulated HER.”
Page 20 of 37

Taylor et al. found that accenting the pronoun in this case, as compared to the condition
in which the pronoun was unaccented, led to a prolonged negativity with a frontal maxi
mum (especially on the right). This is very likely also an Nref effect, indicating the in
creased effort expended to find the correct referent. So, in general, it seems that a mis
match between the expected and actual information structure leads to increases in P600
amplitude; in addition, superfluous marking gives rise to an increase in Nref amplitude.
An example of an information structural mismatch that does not involve marked delivery
is a study by Hoeks, Stowe, Hendriks, and Brouwer (2013). They have recently looked at
violations of projected information structure in question–answer pairs.
Q1: What happened?
A: The producer fired the actor.
Q2: What did the producer and the director do?
#A: The producer fired the actor.
In the context of question 1, answer A is perfectly fine. However, question 2 sets up an ex
pectation of both the producer and the director to become topics in the answer, either to
gether (e.g., “They went shopping”) or separately, as contrastive topics (“The producer
fired the actor, and the director hired him again”). This expectation of contrastive topics
is not met by the answer. At “actor,” Hoeks et al. found a positivity that started quite ear
ly (350 ms post onset), had a broad distribution, and became more frontal as time elapsed
(from (p. 379) 600 ms onward). Despite this early onset, they interpreted this positivity as
a P600 due to MRC reorganization to create a coherent representation, for instance, one
including the computation of the pragmatic implication that the director did not do any
thing because he was powerless or uninterested. This latter kind of processing or infer
ence is the third source of coherence in the mental representation of what is communicat
ed.
Establishing Coherence: Inference

Speakers strive to deliver a coherent message and listeners to create a coherent repre
sentation from the input they get. The links between referring expressions and their an
tecedents provide a scaffolding for putting into place the new information that is con
veyed by an utterance. Cues in the input that indicate where to find topic and focus are
extremely helpful in that process. All the holes that are left in the representation are then
filled by inferencing, to get at what the speaker might have wanted to say. A specific type
of inference only occurs in conversational settings, and these are called conversational
implicatures (Grice, 1975). These arise, for instance, when an answer does not match the
requirements of the question regarding relevance (i.e., the answer should be relevant),
quantity (i.e., the answer should contain just enough information, not more, not less),
quality (i.e., the answer should be truthful), and manner (i.e., the answer should be clear
and orderly) (see also Sperber & Wilson, 1986, 1995, who propose that “relevance” is the
overarching concept). The listener will use inferencing to solve the problems that result
Page 21 of 37

from a mismatch, including the computation of the real implicature, or “What did the
speaker mean to communicate by creating this mismatch in the first place?” The study by
Hoeks et al. (2013) mentioned in the previous section can be considered an example of an
exchange in which a conversational implicature could have been computed. In the mis
match condition, the answer violated the quantity requirements, possibly giving rise to in
ferencing about its true meaning.
Another type of conversational implicature can be seen in ironic utterances. Regel,

Gunter, and Friederici (2011) looked at short texts presenting a mismatch between what
the protagonist said and what he believed was actually true (i.e., a mismatch regarding
quality requirements; see Spotorno, Cheylus, Van der Henst & Noveck, 2013, for a similar
experiment). They presented participants with story fragments, for example, about some
one hearing many mistakes in the performance of a Bach sonata. The person in question
would then look at the orchestra playing the sonata and say to his or her conversational
partner: “These artists are gifted.” In the control condition, the person would be listening
in ecstasy and again say “These artists are gifted,” but this time be sincere. Regel et al.
found that in the irony condition the critical verb produced a P600 effect relative to the
control condition. This increase in P600 amplitude reflects that the hearer has to com
pute the ironic meaning to make his interpretation become meaningful. That is, in the
irony condition, where the sonata is played with a lot of mistakes, the literal interpreta
tion of the sentence “These artists are gifted” is at odds with the situation and does there
fore not lead to an overall coherent and meaningful interpretation of the speaker’s utter
ance.
These two types of inference can be categorized under the particularized conversational
implicatures because they critically depend on the specific context in which they occur.
The other category is the one of generalized conversational implicatures that are more in
dependent of specific aspects of the situation. One prime example of this second category
is the so-called scalar implicature. This term refers to the speaker’s scaling of an utter
ance by using a weaker, less informative term on a scale (e.g., “some”) instead of a
stronger, more informative one (e.g., “all”) to communicate that none of the stronger,
more informative quantifications on that scale holds. It is not always clear what exactly
speakers would like to communicate by using scalar implicatures. For instance, one can
say “Some students are hard working” to mean that not all students are hard working.
However, it would have been just as easy to use the utterance: “Not all students are hard
working.” Maybe the use of scalar implicatures is related to the use of irony, in which the
speaker says things he does not believe himself to draw attention to his opinion.
Certain types of scalar implicatures induce world knowledge violations, as in “Some peo
ple have lungs” (Nieuwland, Ditman, & Kuperberg, 2010), or “Some turtles have
shells” (Noveck & Posada, 2003) because it can be assumed that all people have lungs
and all turtles have shells. Under the retrieval-integration account, both the calculation of
a scalar implicature and the possibly resulting contextual or world knowledge violations
should lead to reorganization of the MRC and thus cause an increase in P600 amplitude.
However, neither of the two studies mentioned above reported any statistics for the P600
Page 22 of 37

window, so we cannot evaluate the MRC hypothesis on the basis of these results. In
(p. 380) addition, Nieuwland et al. (2010) and Noveck and Posada (2003) found contradic
tory results regarding N400 amplitude, but because their materials were not matched on
Cloze probability (either within or between experiments), these findings may not be very
informative.
A recent study using sentence–picture verification did report analyses in the P600 time
window (Hunt III, Politzer-Ahles, Gibson, Minai, & Fiorentino, 2012). Hunt III et al. used
sentences such as “The student has cut some of the brownies” accompanied by pictures
where (a) none of the brownies had been cut (false), (b) all of the brownies had been cut
(underinformative), or (c) some of the brownies had been cut. They found a P600 for both
false and underinformative conditions relative to the true condition, just as predicted by
the MRC hypothesis. In addition, they found modulation of the N400, which matched the
Cloze probabilities for the different sentence–picture pairings (e.g., when hearing “the
student sliced some of the…,” a picture in which some brownies are sliced will preacti
vate “brownies,” a picture without any sliced brownie will not, and a picture in which all
brownies are sliced will be somewhere in between).
A final category of inferences are the well-known bridging inferences. Bridging inferences
are said to occur when information from the current sentence has to be integrated with
the previous context, as in “Horace got some picnic supplies out of the car. The beer was
warm” (Clark & Haviland, 1977). Here, readers must infer that beer was included in the
picnic supplies, otherwise they would not be able to understand the meaning of the two
sentences. In terms of ERP effects, we would predict that the NP “the beer” would evoke
an increased Nref because “the beer” is a definite expression that does not yet have an
antecedent; also, an N400 will be evoked because semantic features of beer must be acti
vated; and, in addition, a P600 will ensue, both as a consequence of the introduction of a
new discourse entity and as a reflection of the inference processes that eventually led to
the assumption of beer being part of the picnic supplies. To our knowledge, this specific
kind of bridging inference has not yet been investigated with ERPs, but there is a related
experiment by Burkhardt (2007) that may be pertinent to this question. Burkhardt looked
at minidiscourses such as
Yesterday a Ph.D. student was {shot / killed / found dead} downtown.
The press reported that the pistol…
By using the verb “shot” in the preceding context sentence, the definite NP “the pistol” is
not a completely new discourse entity because it has already been implied as an instru
ment of shot. The formulation “found dead” by contrast, does not provide any such intro
duction because it does not imply an action or an instrument for that action, so it must be
inferred that the pistol had something to do with the student’s death. The verb “killed”
seems to be somewhere in between because it indicates that the student was murdered,
but doesn’t reveal the precise instrument. The P600 effects that were found are consis
tent with this interpretation: a P600 effect was found for “found dead,” as well as for
“killed,” relative to “shot.” Burkhardt did not find effects on the N400 or report effects on
Page 23 of 37

any other component. This may mean that the search for an antecedent and the preacti
vation of the word “pistol” was more or less the same for the three conditions.
In summary, then, listeners strive to have a coherent representation of what is communi

cated. All three central ways to achieve coherence: establishing reference, creating infor
mation structural links, and inferencing, can be investigated with ERPs. Until now, partic
ipant were made—at best—to eavesdrop on small and relatively uninvolving texts or, in a
few cases, conversations. The presence of a real speaker may change the relative impor
tance of the three cohesive devices but also change the processing more radically as
more representations come into play, among which is the representation of the speaker:
who is he or she, and, by extension, what does he or she mean by articulating a given ut
terance? Some researchers have been skeptical about the importance of the representa
tion of our conversational partners for language use (see, e.g., Barr, this volume; Keysar,
2007). They claim that language use, especially language production, is first and fore
most rather egocentric and automatic and that any audience design that people manage
to do eventually is effortful and is the first to go when task demands increase. This may to
some extent be true, but it is of course obvious that we do adapt to the person that is be
fore us because our personal and social goals in each situation differ with the person (or
persons) we are encountering: is it our boss, our child, our husband or wife, our col
league, our friend, our neighbor, a stranger at a bus stop? Every single person and every
encounter will be “assessed” in terms of your personal and social goals at that time, in re
lation to what you know about them, ranging from superficially visible or inferable char
acteristics to memories of shared events. All these (p. 381) example persons stand in dif
ferent (hierarchical) social relations to us and also differ in familiarity. What we want
from our boss is different from what we want from our children or our students. And that
is also different from what we want from our colleagues when we meet them at the coffee
machine. So, in what we say and how we say it, we adapt constantly, although on a lower
linguistic level—we may be less proficient in tailoring our contributions and frequently re
sort to automatized routines. It is very likely that the same holds for language compre
hension: interpretation is guided by what we know about our interlocutors. A nice illus
tration of that fact is seen in a study by Van Berkum, Van den Brink, Tesink, Kos, and Ha
goort (2008). Participants listening to sentences such as “I have a large tattoo on my
back” spoken with an upper class accent showed an increased N400 amplitude to the crit
ical word “tattoo” as compared to the same sentence uttered in a low-class accent. This
finding shows that listeners take the characteristics of the speaker into account and that
they are able to do so very quickly. So, we would like to conclude that speakers matter, al
though perhaps not always on all levels of language use.
Dynamic Aspects of Communication
When the speaker is present, the language situation changes from static to dynamic. We
have to negotiate turns and the frequency of turns. We have to establish whether it is al
lowed to interrupt the speaker or whether we let ourselves be interrupted. We have to
give and receive backchannel information and decide on the sort of feedback and the tim
ing of it. Kraut et al. (1982) showed that feedback serves to regulate the content of
Page 24 of 37

speech. It also has an effect on more structural aspects, such as the form of referring ex
pressions, the occurrence of ellipsis, and the like. But backchannel information also indi
cates that the listener is interested and committed to engaging in the conversation. Fur
thermore, feedback reflects the state of the relationship between speaker and listener:
smiles and friendly nods signal a positive relation. Unfortunately, no single study on elec
trophysiology of language seems to have looked at these aspects of human interaction.
The DIP may play a role in filling this gap.
Conclusion
With ERPs we have a valid and reliable toolkit at our disposal to study language use in
discourse and conversation. We have identified three components that are linked to three
basic linguistic operations: establishing reference, meaning activation, and integration.
The Nref is sensitive to referential processing—searching for the antecedent in the dis
course representation or MRC. The N400 is sensitive to preactivation and reflects the re
trieval of semantic (and lexical) features of incoming stimuli. And the P600 reflects the
construction or revision of an MRC. In the meantime, there are still some problems that
need to be solved. For instance, one important problem is how we can design experimen
tal methods that allow us to measure ERPs in active conversation. A first step may be the
DIP, in which participants are under the illusion that they are talking via an intercom to a
person in another room, although the utterances of this person are in fact prerecorded
(Hoeks et al., 2013). To make this work, the participant’s own contributions are also
scripted, so that they match the responses of their conversational partner. The DIP pro
vides a means, be it a rather restrictive one, to study the social, pragmatic, and dynamic
aspects of communication that come into play when participants are engaged in an actual
conversation.
Another issue is a bit more technical. Until now, much research has focused on the N400
and how it can be modulated by the context. In this chapter, we argued that the P600 is
the most interesting component for the investigation of discourse and dialogue. However,
the processes reflected in N400 and P600 amplitude may overlap in time. Because ERPs
are additive, this means that the amplitude of any P600 depends on the amplitude of the
preceding N400. In other words, a specific P600 waveform of a given magnitude will turn
out more positive if for any reason the preceding N400 is more positive, and P600 ampli
tude will turn out less positive if the preceding N400 is less positive. Thus, P600 ampli
tude is modulated by N400 amplitude, and possible P600 effects may become obfuscated.
As a consequence, we may conclude that a given contrast between an experimental and a
control condition only produces an N400 effect and no P600 effect, whereas, in reality,
there is a difference in P600 amplitude. This issue has been noted before in the literature
(e.g., Hagoort, 2003; Kutas, Van Petten, & Kluender, 2006) but has of yet not led to a
change in how ERPs are analyzed. One reason for this is probably that the P600 has often
been seen as an effect that is only generated when something goes wrong during process
ing and not as a component that is always present in the ERP signal. Conversely, compo
nent overlap is, in fact, the reason that the P600 has been taken to be an effect because
Page 25 of 37

its overlap with the (p. 382) N400 component fosters the impression that it is sometimes
present (e.g., no preceding N400 effect or rather large P600 effect) and sometimes ab
sent (e.g., large N400 effect or rather small P600 effect). It is not easy to find a good way
to estimate “true” P600 amplitude that takes into account preceding differences in the
ERP signal. One approach to this would be to baseline on the N400 time window (cf. Ha
goort, 2003). If we can find a way to adjust for such prior differences between conditions,
we might explain why a P600 effect sometimes appears to be absent where it is in fact ex
pected (e.g., see Van Petten & Luka, 2012).
Not only component overlap, but also other differences in the ERP signal preceding the
target word are to be reckoned with, especially where utterances appear in a context, as
is the typical case in discourse and conversation. Every change in the context may affect
the expectations that listeners have about the information contained in the upcoming ut
terances. In addition, in spoken language, especially when languages are used in which
information structure is indicated by means of prosody, prosodic cues are sometimes
available well in advance of the target word. For instance, prosodic marking of topic and
focus is not strictly local, but has an effect on the entire prosodic contour of an utterance.
This is also a problem that we need to take seriously and one for we which we do not yet
have a satisfactory solution. Further research in this area is definitely necessary.
Addressing these issues is critical for the successful investigation of the internal function
al structure of the P600. Brouwer et al. (2012) suggested that the P600 may not reflect
one single process, but may subsume many different subprocesses, which can be differen
tiated and labeled on the basis of differences in onset, duration, and scalp distribution. In
the model proposed by Brouwer and Hoeks (2013), the P600 component is generated in
the left lIFG (BA 44/45/47). They argued that the complex neuroarchitecture of this area
supports a fine-grained functional topology, hosting different subprocesses of MRC con
struction and reorganization. As such, different P600s in terms of scalp distribution and
other electrophysiological properties like onset, duration, and amplitude arise because
they are generated in different, but potentially overlapping subareas of the lIFG. Howev
er, the functional and anatomical neurocognitive organization that underlies the scalp dis
tribution of an effect is different for each individual brain. Thus, the derivation of a sound
functional categorization of P600 is not as straightforward as it might seem. Possibly, this
issue may be circumvented by comparing anatomy and ERP components on an individual,
within-participant level and not (only) by looking at between-participants data.
Once we know more about what the different P600 subtypes mean (and by which brain
areas they are subserved), we may start to answer questions regarding the nature of the
representations involved in conversation. We have suggested that successful conversation
requires an interactant to manage a number of representations: two representations for
himself: his private goals and knowledge and his public “face,” two representations (pub
lic and private) of his main conversational partner, a representation of the content or the
upshot of the conversation as a whole, and a representation of the more situational as
pects of the conversational setting. The question is what these representations look like,
how they interact, and how they are accessed. Are they all part of a single representation,
Page 26 of 37

or are they stored and maintained somewhat separately, perhaps in totally different brain
areas? It would make sense to separate private and public representations physically, to
prevent leakage, although in the brain, distance doesn’t seem to matter much.
In conclusion, then, if we want to arrive at a plausible neurocognitive model of communi

cation, we cannot afford to neglect the social, pragmatic, and dynamic aspects of lan
guage use. The actual or perceived presence of an interactant may crucially shape the
representations and the processes involved in communication. Thus, we should turn to in
vestigating active conversation.
Acknowledgments
We would like to thank Noortje Venhuizen for comments on an earlier version of this man
uscript. Writing of this chapter was in part supported by the Netherlands Organization for
Scientific Research (NWO) PGW grant 10–26.
References
Amunts, K., Lenzen, M., Friederici, A. D., Schleicher, A., Moroson, P., Palomero-Gallagher,
N., & Zilles, K. (2010). Broca’s region: Novel organizational principles and multiple recep
tor mapping. PLoS Biology, 8(9).
Amunts, K., & Zilles, K. (2012). Architecture and organizational principles of Broca’s
area. Trends in Cognitive Sciences, 16(8), 418–426.
Babiloni, F., & Astolfi, L. (2012). Social neurosciences and hyperscanning techniques:
Past, present and future. Neuroscience (p. 383) & Biobehavioral Reviews. http://
www.sciencedirect.com/science/article/pii/S0149763412001194
Baggio, G., & Hagoort, P. (2011). The balance between memory and unification in seman
tics: A dynamic account of the N400. Language and Cognitive Processes, 26, 1338–1367.
Bara, B. G., Ciaramidaro, A., Walter, H., & Adenzato, M. (2011). Intentional minds: A
philosophical analysis of intention tested through fMRI experiments involving people with
schizophrenia, people with autism, and healthy individuals. Frontiers in Human Neuro
science, 5(7), 1–11.
Barbey, A. K., & Grafman, J. (2011). An integrative cognitive neuroscience theory of social
reasoning and moral judgment. Wiley Interdisciplinary Review: Cognitive Science, 2 (1),
55–67.
Barr, D. J. (this volume). Perspective taking and its impostors in language use: Four pat
terns of deception. In Thomas Holtgraves (Ed.), The Oxford Handbook of Language and
Social Psychology (pp.). New York: Oxford University Press.
Page 27 of 37

Bookheimer, S. (2002). Functional MRI of language: New approaches to understanding

the cortical organization of semantic processing. Annual Review of Neuroscience, 25,
151–188.
Bornkessel, I., Schlesewsky, M., & Friederici, A. D. (2003). Contextual information modu
lates initial processes of syntactic integration: The role of inter- versus intrasentential
predictions. Journal of Experimental Psychology: Learning, Memory, and Cognition, 29(5),
871–882.
Bornkessel-Schlesewsky, I., & Schlesewsky, M. (2008). An alternative perspective on se

mantic P600 effects in language comprehension. Brain Research Reviews, 59, 55–73.
Bousfield, D. (2008). Impoliteness in interaction. Amsterdam/Philadelphia: John Ben

jamins.
Brouwer, H. (2014). The Electrophysiology of Language Comprehension: A Neurocompu

tational Model. PhD thesis, University of Groningen.
Brouwer, H., Fitz, H., & Hoeks, J. C. J. (2012). Getting real about Semantic Illusions: Re
thinking the functional role of the P600 in language comprehension. Brain Research,
1446, 127–143.
Brouwer, H., & Hoeks, J. C. J. (2013). A time and place for language comprehension: Map
ping the N400 and the P600 to a minimal cortical network. Frontiers in Human Neuro
science, 7(758), 1–12.
Brown, C. M., & Hagoort, P. (1993). The processing nature of the N400: Evidence from
masked priming. Journal of Cognitive Neuroscience, 5, 34–44.
Buckner, R. L., Sepulcre, J., Talukdar, T., Krienen, F. M., Liu, H., Hedden, T., Andrews-Han
na, J. R., Sperling, R. A., & Johnson, K. A. (2009). Cortical hubs revealed by intrinsic func
tional connectivity: Mapping, assessment, of stability, and relation to Alzheimer’s disease.
The Journal of Neuroscience, 29, 1860–1873.
Burkhardt, P. (2006). Inferential bridging relations reveal distinct neural mechanisms: Ev
idence from event-related potentials. Brain and Language, 98(2), 159–168.
Burkhardt, P. (2007). The P600 reflects cost of new information in discourse memory.
NeuroReport, 18(17), 1851–1854.
Camblin, C. C., Ledoux, K., Boudewyn, M., Gordon, P. C., & Swaab, T. Y. (2007). Process
ing new and repeated names: Effects of coreference on repetition priming with speech
and fast RSVP. Brain Research, 1146, 172–184.
Page 28 of 37

Chafe, W. L. (1979). Givenness, contrastiveness, definiteness, subjects, topics and point of

view. In C. N. Li, C. N. (Ed.), Subject and topic (pp. 27–55). New York: Academic Press.
Chwilla, D. J., Brown, C. M., & Hagoort, P. (1995). The N400 as a function of the level of
processing. Psychophysiology, 32, 274–285.
Chow, W.-Y., & Phillips, C. (2013). No semantic illusion in the “semantic P600” phenome
non: ERP evidence from Mandarin Chinese. Brain Research, 1506, 76–93.
Clark, H. H., & Brennan, S. E. (1991). Grounding in communication. Perspectives on So

cially Shared Cognition, 13, 127–149.
Clark, H. H., & Haviland, S. E. (1977). Comprehension and the given-new contract. In R.
O. Freedle (Ed.), Discourse processes, Vol. 1: Advances in research and theory (pp. 1–40).
Norwood, NJ: Ablex.
Cowles, H. W., Kluender, R., Kutas, M., & Polinsky, M. (2007). Violations of information
structure: An electrophysiological study of answers to wh-questions. Brain and Language,
103(3), 228–242.
Culpeper, J. (1996). Towards an anatomy of impoliteness. Journal of Pragmatics, 25(3),

349–367.
Dawkins, R. (2006). The selfish gene (30th anniversary edition). New York: Oxford Univer
sity Press.
Decety, J., & Lamm, C. (2007). The role of the right temporoparietal junction in social in
teraction: How low-level computational processes contribute to meta-cognition. The Neu
roscientist, 13(6), 580–593.
Dimitrova, D. V., Stowe, L. A., Redeker, G., & Hoeks, J. C. J. (2012). Less is not more:
Neural responses to missing and superfluous accents in context. Journal of Cognitive
Neuroscience, 24(12), 2400–2418.
Ditman, T., & Kuperberg, G. R. (2007). The time course of building discourse coherence in
schizophrenia: An ERP investigation. Psychophysiology, 44, 991–1001.
Dronkers, N. F., Wilkins, D. P., Van Valin, R. D., Jr. Redfern, B. B., & Jaeger, J. J. (2004). Le
sion analysis of the brain areas involved in language comprehension. Cognition, 92, 145–
177.
Dumas, G., Chavez, M., Nadel, J., & Martinerie, J. (2012). Anatomical connectivity influ
ences both intra- and inter-brain synchronizations. PLoS ONE, 7(5), 1–11.
Dumas, G., Nadel, J., Soussignan, R., Martinerie, J., & Garnero, L. (2010). Inter-brain syn
chronization during social interaction. PLoS ONE, 5(8), 1–10.
Elman, J. L. (2004). An alternative view of the mental lexicon. Trends in Cognitive

Sciences, 8, 301–306.
Page 29 of 37

Elman, J. L. (2009). On the meaning of words and dinosaur bones: Lexical knowledge
without a lexicon. Cognitive Science, 33, 547–582.
Federmeier, K. D., & Laszlo, S. (2009). Time for meaning: Electrophysiology provides in
sights into the dynamics of representation and processing in semantic memory. Psycholo
gy of Learning and Motivation, 51, 1–44.
Friederici, A. D. (2002). Towards a neural basis of auditory sentence processing. Trends

in Cognitive Sciences, 6, 78–84.
Friederici, A. D. (2009). Pathways to language: Fiber tracts in the human brain. Trends in
Cognitive Sciences, 13, 175–181.
Friederici, A. D. (2011). The brain basis of language processing: From structure to func
tion. Physiological Reviews, 91, 1357–1392.
Goffman, E. (1959). The presentation of self in everyday life. New York: Doubleday.
(p. 384) Goffman, E. (1967). Interaction ritual: Essays on face-to-face behavior. New York:
Pantheon.
Gouvea, A. C., Phillips, C., Kazanina, N., & Poeppel, D. (2010). The linguistic processes
underlying the P600. Language and Cognitive Processes, 25(2), 149–188.
semantics, Vol. 3: Speech acts (pp. 41–58). New York: Academic Press.
Grodzinsky, Y., & Santi, A. (2008). The battle for Broca’s region. Trends in Cognitive
Sciences, 12, 474–480.
Hagoort, P. (2003). Interplay between syntax and semantics during sentence comprehen
sion: ERP effects of combining syntactic and semantic violations. Journal of Cognitive
Hagoort, P. (2005). On Broca, brain, and binding: A new framework. Trends in Cognitive
Sciences, 6, 416–423.
Hagoort, P., Baggio, G., & Willems, R. M. (2009). Semantic unification. In: Gazzaniga, M.
S. (Ed.), The cognitive neurosciences (4th ed., pp. 819–836). Cambridge, MA: MIT Press.
Hagoort, P., & Van Berkum, J. J. A. (2007). Beyond the sentence given. Philosophical
Transaction of the Royal Society, Series B: Biological Sciences, 362, 801–811.
Heim, S., & Alter, K. (2006). Prosodic pitch accents in language comprehension and pro
duction: ERP data and acoustic analysis. Acta Neurobiologiae Experimentalis, 66, 55–68.
Hickok, G., & Poeppel, D. (2004). Dorsal and ventral streams: A framework for under
standing aspects of the functional anatomy of language. Cognition, 92, 67–99.
Page 30 of 37

Hickok, G., & Poeppel, D. (2007). The cortical organization of speech processing. Nature
Reviews Neuroscience, 8, 393–402.
Hoeks, J. C. J. Stowe, L. A., Hendriks, P., & Brouwer, H. (2013). Questions left unan
swered: How the brain responds to missing information. PLoS ONE, 8(10), 1–9.
Hoeks, J. C. J., Schoot, L., Taylor, R. C., & Brouwer, H. (in preparation). How to say “no”
politely. An ERP investigation. Manuscript in preparation.
Hoeks, J. C. J., Schoot, L., Neijmeijer, M. I. A., & Brouwer, H. (in preparation). The Dia
logue Immersion Paradigm (DIP) as a tool to investigate active conversation. Manuscript
in preparation.
Hoeks, J. C. J., Stowe, L. A., & Doedens, G. (2004). Seeing words in context: The interac
tion of lexical and sentence level information during reading. Cognitive Brain Research,
19, 59–73.
Hruska, C., & Alter, K. (2004). Prosody in dialogues and single sentences: How prosody
can influence speech perception. In A. Steube (Ed.), Information structure: Theoretical
and empirical aspects (pp. 211–226). Berlin: Walter de Gruyter.
Hruska, C., Alter, K., Steinhauer, K., & Steube, A. (2001). Misleading dialogs: Human’s
brain reaction to prosodic information. In C. Cave, I. Guaitella, & S. Santi (Eds.), Orality
and gestures (pp. 425–430). Paris: L’Hartmattan.
Hunt III, L., Politzer-Ahlers, S., Gibson, L., Minai, U., & Fiorentino, R. (2012). Pragmatic
inferences modulate N400 during sentence comprehension: Evidence from picture-sen
tence verification. Neuroscience Letters, 534, 246–251.
Johnson-Laird, P. N. (1983). Mental models. Cambridge, UK: Cambridge University Press.
Keysar, B. (2007). Communication and miscommunication: The role of egocentric process

es. Intercultural Pragmatics, 4(1), 71–84.
Kim, A., & Osterhout, L. (2005). The independence of combinatory semantic processing:
Evidence from event-related potentials. Journal of Memory and Language, 52, 205–225.
Kintsch, W. (1998). Comprehension: A paradigm for cognition. Cambridge, UK: Cam

Kolk, H. H. J., Chwilla, D. J., Van Herten, M., & Oor, P. J. W. (2003). Structure and limited
capacity in verbal working memory: A study with event-related potentials. Brain and Lan
guage, 85, 1–36.
Kos, M., Vosse, T., van den Brink, D., & Hagoort, P. (2010). About edible restaurants: Con
flicts between syntax and semantics as revealed by ERPs. Frontiers in Psychology, 1, 1–
11.
Page 31 of 37

Kotchoubey, B. (2006). Event-related potentials, cognition, and behavior: A biological ap

proach. Neuroscience and Biobehavioral Reviews, 30, 42–56.
Kraut, R. E., Lewis, S. H., & Swezey, L. W. (1982). Listener responsiveness and the coordi
nation of conversation. Journal of Personality and Social Psychology, 43(4), 718–736.
Kuperberg, G. R. (2007). Neural mechanisms of language comprehension: Challenges to

syntax. Brain Research, 1146, 23–49.
Kuperberg, G. R., Sitnikova, T., Caplan, D. N., & Holcomb, P. J. (2003). Electrophysiologi
cal distinctions in processing conceptual relationships within simple sentences. Cognitive
Brain Research, 17, 117–129.
Kutas, M., & Federmeier, K. D. (2000). Electrophysiology reveals semantic memory usage
in language comprehension. Trends in Cognitive Sciences, 4, 463–470.
Kutas, M., & Federmeier, K. D. (2011). Thirty years and counting: Finding meaning in the
N400 component of the event related brain potential (ERP). Annual Review of Psychology,
62, 621–647.
Kutas, M., & Hillyard, S. A. (1980). Reading senseless sentences: Brain potentials reflect
semantic incongruity. Science, 207, 203–205.
Kutas, M., & Hillyard, S. A. (1984). Brain potentials during reading reflect word expectan
cy and semantic association. Nature, 307, 539–550.
Kutas, M., & Lindamood, T. E., & Hillyard, S. A. (1984). Word expectancy and event-relat
ed brain potentials during sentence processing. In S. Kornblum & J. Requin (Eds.),
Preparatory states and processing (pp. 217–317), Hillsdale, NJ: Erlbaum.
Kutas, M., Van Petten, C., & Kluender, R. (2006). Psycholinguistics electrified II: 1994–
2005. In M. J. Traxler & M. A. Gernsbacher (Eds.), Handbook of psycholinguistics (2nd
ed., pp 659–724). New York: Elsevier.
Lambrecht, K. (1994). Information structure and sentence form. Cambridge, UK: Cam
Lau, E. F., Phillips, C., & Poeppel, D. (2008). A cortical network for semantics:
(de)constructing the N400. Nature Reviews Neuroscience, 9, 920–933.
Ledoux, K., Gordon, P. C., Camblin, C. C., & Swaab, T. Y. (2007). Coreference and lexical
repetition: Mechanisms of discourse integration. Memory and Cognition, 4, 801–815.
Li, X., Hagoort, P., & Yang, Y. (2008). Event-related potential evidence on the influence of
accentuation in spoken discourse comprehension in Chinese. Journal of Cognitive Neuro
science, 20(5), 906–915.
Li, X., Yang, Y., & Hagoort, P. (2008). Pitch accent and lexical tone processing in Chinese
discourse comprehension: An ERP study. Brain Research, 1222, 192–200.
Page 32 of 37

Lotze, N., Tune, S., Schlesewsky, M., & Bornkessel-Schlesewsky, I. (2011). Meaningful
physical changes mediate lexical-semantic integration: Top-down and form-based (p. 385)
bottom-up information sources interact in the N400. Neuropsychologia, 49, 3573–3582.
Luck, S. J. (2005). An introduction to the event-related potential technique. Cambridge,

MA: MIT Press.
Magne, C., Astésano, C. Lacheret-Dujour, A., Morel, M., Alter, K., & Besson, M. (2005).
On-line processing of “pop-out” words in spoken French dialogues. Journal of Cognitive
Mesulam, M. M. (1990). Large-scale neurocognitive networks and distributed processing

for attention, language, and memory. Annals of Neurology, 28, 597–613.
Mesulam, M. M. (1998). From sensation to cognition. Brain, 121, 1013–1052.
Molinaro, N., Carreiras, M., & Duñabeitia, J. A. (2012). Semantic combinatorial process
ing of non-anomalous expressions. NeuroImage, 59, 3488–3501.
Näätänen, R., & Picton, T. (1987). The N1 wave of the human electric and magnetic re
sponse to sound: A review and an analysis of the component structure. Psychophysiology,
24, 375–425.
Nieuwland, M. S., Ditman, T., & Kuperberg, G. R. (2010). On the incrementality of prag
matic processing: An ERP investigation of informativeness and pragmatic abilities. Jour
nal of Memory and Language, 63(3), 324–346.
Nieuwland, M. S., & Otten, M., & Van Berkum, J. J. A. (2007). Who are you talking about?
Tracking discourse-level referential processing with ERPs. Journal of Cognitive Neuro
science, 19, 228–236.
Nieuwland, M. S., Petersson, K. M., & Van Berkum, J. J. A. (2007). On sense and refer
ence: Examining the functional neuroanatomy of referential processing. NeuroImage, 37,
993–1004.
Nieuwland, M. S., & Van Berkum, J. J. A. (2005). Testing the limits of the semantic illusion
phenomenon: ERPs reveal temporary semantic change deafness in discourse comprehen
sion. Cognitive Brain Research, 24, 691–701.
Noveck, I. A., & Posada, A. (2003). Characterizing the time course of an implicature: An
evoked potentials study. Brain and Language, 85(2), 203–210.
Nunberg, G. (1979). The non-uniqueness of semantic solutions: Polysemy. Linguistics and

Philosophy, 3(2), 143–184.
Osterhout, L., & Holcomb, P. J. (1992). Event-related brain potentials elicited by syntactic
anomaly. Journal of Memory and Language, 31(6), 785–806.
Page 33 of 37

Otten, M., & Van Berkum, J. J. A. (2007). What makes a discourse constraining? Compar
ing the effects of discourse message and scenario fit on the discourse-dependent N400-ef
fect, Brain Research, 1153, 166–177.
Pulvermüller, F. (1999). Words in the brain’s language. Behavioral and Brain Sciences, 22,
252–279.
Pulvermüller, F. (2001). Brain reflections of words and their meaning. Trends in Cognitive
Sciences, 5, 517–524.
Regel, S., Gunter, T., & Friederici, A. D. (2011). Isn’t it ironic? An electrophysiological ex
ploration of figurative language processing. Journal of Cognitive Neuroscience, 23(2),
277–293.
Rogalsky, C., & Hickok, G. (2011). The role of Broca’s area in sentence comprehension.
Journal of Cognitive Neuroscience, 23, 1664–1680.
Rogers, T., & McClelland, J. (2004). Semantic cognition: A parallel distributed processing
approach. Cambridge, MA: The MIT Press.
Russell, B. (1905). On denoting. Mind, 14(56), 479–493.
Schilbach, L., Timmermans, B., Reddy, V., Costall, A., Bente, G., Schlicht, T., & Vogeley, K.
(2013). Toward a second-person neuroscience. Behavioral and Brain Sciences, 36(4), 393–
414.
Schlenker, B. R. (1980). Impression management: The self-concept, social identity, and in
terpersonal relations. Monterey, CA: Brooks/Cole.
Schober, M. F., & Clark, H. H. (1989). Understanding the addressees and overhearers.
Cognitive Psychology, 21(2), 211–232.
Schumacher, P. (2011). The hepatitis called…. In J. Meibauer & M. Steinbach (Eds.), Ex
perimental pragmatics/semantics (pp. 199–219). Amsterdam/Philadelphia: John Ben
jamins.
Shalom, D. B., & Poeppel, D. (2007). Functional anatomic models of language: Assem
bling the pieces. The Neuroscientist, 30, 119–127.
Sperber, D., & Wilson, D. (1986). Relevance. Oxford: Blackwell Publishers.
Sperber, D., & Wilson, D. (1995). Relevance (2nd ed.). Oxford: Blackwell Publishers.
Spotorno, N., Cheylus, A., Van der Henst, J-B., & Noveck, I. A. (2013). What’s behind a
P600? Integration operations during irony processing. PLoS ONE, 8(6), 1–10.
Stephens, G. J., Silbert, L. J., & Hasson, U. (2010). Speaker—Listener neural coupling un
derlies successful communication. Proceedings of the National Academy of Sciences,
107(32), 14425–14430.
Page 34 of 37

St. George, M., Mannes, S., & Hoffmann, J. E. (1994). Global expectancy and language
comprehension. Journal of Cognitive Neuroscience, 6(1), 70–83.
Stroud, C. M. A., & Phillips, C. (2012). Examining the evidence for an independent seman
tic analyzer: An ERP study in Spanish. Brain and Language, 120, 108–126.
Suda, M., Takei, Y., Aoyama, Y., Narita, K., Sato, T., Fukuda, M., & Mikuni, M. (2010).
Frontopolar activation during face-to-face conversation: An in situ study using near-in
frared spectroscopy. Neuropsychologia, 48(2), 441–447.
Swaab, T. Y., Camblin, C. C., & Gordon, P. C. (2004). Electrophysiological evidence for re
versed lexical repetition effects in language processing. Journal of Cognitive Neuro
science, 16(5), 715–726.
Taylor, R. C. (2013). Tracking referents: Markedness, world knowledge and pronoun reso
lution. PhD thesis, University of Groningen.
Taylor, R. C., Stowe, L. A., Redeker, G., & Hoeks, J. C. J. (in preparation). Accented pro
nouns and unexpected referents. An ERP study. Manuscript in preparation.
Toepel, U., & Alter, K. (2004). On the independence of information structure processing
from prosody. In A. Steube (Ed.), Information structure: Theoretical and empirical aspects
(pp. 227–240). Berlin: Walter de Gruyter.
Toepel, U., Pannenkamp, A., & Alter, K. (2007). Catching the news: Processing strategies
in listening to dialogs as measured by ERPs. Behavioral and Brain Functions, 3(53), 1–13.
Turken, U., & Dronkers, N. F. (2011). The neural architecture of the language comprehen
sion network: Converging evidence from lesion and connectivity analyses. Frontiers in
Systems Neuroscience, 5, 1–20.
Tyler, L., Marlsen-Wilson, W., Randal, B., Wright, P., Devereux, B., Zhuang, J., Papoutsi,
M., & Stamatakis, E. (2011). Left inferior frontal cortex and syntax: Function, structure,
and behaviour in patients with left hemisphere damage. Brain, 134, 415–431.
Van Berkum, J. J. A. (2009). The “neuropragmatics” of simple utterance compre

(p. 386)
hension: An ERP review. In U. Sauerland & Yatsushiro (Eds.), Semantics and pragmatics:
From experiment to theory (pp. 276–316). New York: Palgrave MacMillan.
Van Berkum, J. J. A. (2012). The electrophysiology of discourse and conversation. In M.

Spivey, M. Joanisse, & K. McRae (Eds.), The Cambridge handbook of psycholinguistics
(pp. 589–614). Cambridge, UK: Cambridge University Press.
Van Berkum, J. J. A., Brown, C. M., & Hagoort, P. (1999). Early referential context effects
in sentence processing: Evidence from event-related brain potentials. Journal of Memory
and Language, 41, 147–182.
Page 35 of 37

Van Berkum, J. J. A., Hagoort, P., & Brown, C. M. (1999). Semantic integration in sen
tences and discourse: Evidence from the N400. Journal of Cognitive Neuroscience, 11(6),
657–671.
Van Berkum, J. J. A., Koornneef, A. W., Otten, M., & Nieuwland, M. S. (2007). Establishing
reference in language comprehension: An electrophysiological perspective. Brain Re
search, 1146, 158–171.
Van Berkum, J. J. A., Van den Brink, D., Tesink, C, M, J, Y., Kos, M., & Hagoort, P. (2008).
The neural integration of speaker and message, Journal of Cognitive Neuroscience, 20 (4),
580–591.
Van Berkum, J. J. A., Zwitserlood, P., Brown, C. M., & Hagoort, P. (2003). When and how
do listeners relate a sentence to a wider discourse? Evidence from the N400 effect. Cog
nitive Brain Research, 17, 701–718.
Van de Meerendonk, N., Kolk, H. H. J., Chwilla, D. J., & Vissers, C. T. W. M. (2009). Moni
toring in language perception. Language and Linguistics Compass, 3, 1211–1224.
Van de Meerendonk, N, Kolk, H. H. J., Vissers, C. T. W. M., & Chwilla, D. J. (2010). Moni
toring in language perception: Mild and strong conflicts elicit different ERP patterns.
Journal of Cognitive Neuroscience, 22, 67–82.
Van Herten, M., Kolk, H. H. J., & Chwilla, D. J. (2005). An ERP study of P600 effects elicit
ed by semantic anomalies. Cognitive Brain Research, 22, 241–255.
Van Herten, M., Chwilla, D. J., & Kolk. H. H. J. (2006). When heuristics clash with parsing
routines: ERP evidence from conflict monitoring in sentence perception. Journal of Cogni
tive Neuroscience, 18, 1181–1197.
Van Petten, C., & Luka, B. J. (2012). Prediction during language comprehension: Benefits,
costs, and ERP components. International Journal of Psychophysiology, 83(2), 176–190.
Vigneau, M., Beaucousin, V., Herv, P., Duffau, H, Crivello, F., Houd, O., Mazoyer, B., Tzou
rio-Mazoyer, N. (2006). Meta-analyzing left hemisphere language areas: Phonology, se
mantics, and sentence processing. NeuroImage, 30, 1440–1422.
Vissers, C. T. W. M., Chwilla, D. J., & Kolk, H. H. J. (2006). Monitoring in language percep
tion: The effect of misspellings of words in highly constrained sentences. Brain Research,
1106, 150–163.
Vissers, C. T. W. M., Chwilla, D. J., & Kolk, H. H. J. (2007). The interplay of heuristics and
parsing routines in sentence comprehension: Evidence from ERPs and reaction times. Bio
logical Psychology, 75, 8–18.
Vissers, C. T. W. M., Chwilla, U. G., Egger, J. I. M., & Chwilla, D. J. (2013). The interplay
between mood and language comprehension: Evidence from P600 to semantic reversal
anomalies. Neuropsychologia, 51(6), 1027–1039.
Page 36 of 37

Wang, L., Bastiaansen, M., Yang, Y., & Hagoort, P. (2011). The influence of information
structure on the depth of semantic processing: How focus and pitch accent determine the
size of the N400 effect. Neuropsychologia, 49, 813–820.
Wang, L., Hagoort, P., & Yang, Y. (2009). Semantic illusion depends on information struc
ture: ERP evidence. Brain Research, 1282, 50–56.
Weiller, C., Musso, M., Rijntjes, M., & Saur, D. (2009). Please don’t underestimate the
ventral pathway in language. Trends in Cognitive Sciences, 13, 369–370.
Ye, Z., & Zhou, X. (2008). Involvement of cognitive control in sentence comprehension:
Evidence from ERPs. Brain Research, 1203, 103–115.
Zwaan, R. A., & Radvansky, G. A. (1998). Situation models in language comprehension

John C.J. Hoeks
John C. J. Hoeks, Center for Language and Cognition Groningen, University of

Groningen
Harm Brouwer
Harm Brouwer, BCN Neuro-Imaging Center, University of Groningen, The Nether

lands
Page 37 of 37

Politeness and Reasoning: Face, Connectives, and Quantifiers
Politeness and Reasoning: Face, Connectives, and

Quantifiers
Jean-Francois Bonnefon

Connectives and quantifiers are the building blocks of judgment and reasoning. Pragmat
ic factors affect their interpretation, which in turn affects inferences and decisions. To ac
count for these effects, the psychology of judgment and reasoning imported theoretical
frameworks from linguistic pragmatics. These frameworks focused on conversational effi
ciency but neglected politeness. Efficiency can explain a large range of interpretative ef
fects (e.g., the scalarity of “some” and “or,” the narrow range of numerical translations
for verbal uncertainty terms, the propagation of uncertainty from antecedent to conse
quent in conditional reasoning), but it does not explain the interpretation of connectives
and quantifiers that apply to face-threatening content such as criticisms, requests, impo
sitions, or bad news. When face-threatening content is involved, efficiency and politeness
are antagonistic, and classic efficiency effects disappear for politeness reasons. Thus, po
liteness explains why people sometimes revert to the logical interpretation of “some” and
“or,” why they assign greater probability to severe events, and why they sometimes can
cel causal inferences.
Keywords: politeness, reasoning, pragmatics, connectives, quantifiers
The psychology of judgment and reasoning explores the processes that allow one to men
tally manipulate information in order to produce new conclusions. The information that
serves as the input of the reasoning process can come from a variety of sources. It may
be the end product of some previous reasoning; it may come from memory or perception;
and, most importantly for this chapter, it can come from discourse. As we will see, this
last case is quite special because it entails that interpretation enters the reasoning
process: what is reasoned from is, in that case, the interpretation that one made of what
was communicated. Without a model of this interpretative process, it is simply impossible
to account for the conclusions that will ultimately be reached by the reasoner.
To handle this difficulty, psychologists of judgment and reasoning turned to conceptual

frameworks borrowed from linguistic pragmatics. As we will see, the bulk of this research
appealed to the notion that people used language efficiently to explain why they reasoned
Page 1 of 31

illogically. In a recent twist, though, data built up that showed that people were as likely
to use language inefficiently and that this gave their reasoning the appearance of logicali
ty. This chapter tells the story of this paradigmatic change whose instrumental concept
was that of politeness.
Defining politeness can feel like a “combat with a many-headed hydra” (Watts, 2003, p.
xi). I will not engage the hydra in this chapter, mostly because this chapter only provides
a glimpse of what politeness can be or do. Indeed, we will mostly be concerned with a
sub-sub-subset of politeness: linguistic, negative, hedging politeness strategies. Defini
tions and examples will come later, but readers can readily (p. 388) perceive that there is
much more to politeness than what we will address in this chapter. We will still have plen
ty to consider, though, because negative politeness and hedging strategies have a large
and multifaceted impact on reasoning.
Reasoning consists, for a large part, in the manipulation of connectives and quantifiers.
Accordingly, this chapter focuses on how politeness considerations can impact the inter
pretation of these connectives and quantifiers and how this interpretation, in turn, im
pacts the output of the reasoning process. The section “Connectives and Quantifiers” in
troduces the connectives and quantifiers that will be studied in this chapter. The section
on “Efficiency and Politeness” then introduces the two general constraints that bear on
their interpretation in everyday conversation: efficiency and politeness. The main objec
tive of this section is to show how efficiency and politeness can exert opposite influences
on interpretation, as soon as the message that must be communicated is potentially up
setting. This insight is then applied to the interpretation of connectives and quantifiers in
the last three sections of the chapter.
Connectives and Quantifiers

To narrow things down to a tractable scope, we restrict ourselves in this chapter to one
(important) subfield of judgment and reasoning, the manipulation of connectives and
quantifiers. Connectives (if, and, or, not…) and quantifiers (some, most, probable, possi
ble…) form the backbone of reasoning, and the way they are interpreted is especially im
portant when reasoning from discourse.
Connectives
One could argue that every instance of reasoning relies, implicitly or explicitly, on the use
of connectives. The four basic connectives (displayed in Table 24.1) are negation, conjunc
tion, disjunction, and the conditional.
Page 2 of 31

Table 24.1 The four basic connectives. Disjunctions and Conditionals

are given special consideration in this chapter.
Connective Symbol Read as
Negation ¬x not x
Conjunction x∧y x and y
Disjunction x∨y x or y
Conditional x⊃y if x then y
• Negation. I have little to say about negation in the rest of the chapter, mostly be
cause it has not been targeted yet with experiments studying its polite use in reason
ing (if any). This would most certainly be a fruitful area for future research.
• Conjunction. The logical meaning of x ∧ y (x and y) is simply that both x and y are
true, and no more. In everyday language, conjunction can convey further information,
typically temporality and causation (Hertwig, Benz, & Krauss, 2008). For example, the
sentence in (1a) would typically be understood as conveying temporal order (I changed
clothes and then I had a workout), and the sentence in (1b) would typically be under
stood as conveying both temporal order and causation (she got pregnant and, as a con
sequence of that, they married). Conjunction-based reasoning has not been yet studied
through the lenses of politeness theory, and (as a consequence) will not reappear in
this chapter.
(1a) I changed clothes and had a workout.
(1b) She got pregnant and they got married.
• Disjunction. The logical meaning of x ∨ y (x or y) is that x is true, or y is true, or

both are true. In other words, logical disjunction is inclusive. It includes the possibility
that both x and y are true at the same time. Later in this chapter, disjunction will be a
prime example of the paradigmatic shift in the pragmatics of reasoning: the tradition
al, efficiency-based pragmatics of reasoning explains why people deviate from the in
clusive, logical meaning—and the politeness-based pragmatics of reasoning explains
why people revert back to the logical meaning.
• Conditionals. The conditional is easily the most studied connective in the psycholo
gy of reasoning. The logical meaning of x ⊃ y (if x then y) is that it is impossible to ob
serve a situation where x is true and y is false. This definition is deceptively simple and
allows for very odd inferences (Bonnefon & Politzer, 2011). Later in this chapter, we
will see that politeness can have divergent effects on the interpretation of conditionals,
pulling it closer or pushing it away from their logical meaning.
Page 3 of 31

Quantifiers
Words that express percentages or probabilities feature prominently in the psychology of

judgment (p. 389) and reasoning (Politzer, van der Henst, Delle Luche, & Noveck, 2006;
Sanford & Moxey, 2003). These quantifiers come in much greater numbers than logical
connectives (see Table 24.2 for examples borrowed from Moxey & Sanford, 1997; Wall
sten, Budescu, & Zwick, 1993), and it would be utterly impossible to cover them all in this
chapter. What I will do instead is focus on the most studied of quantifiers (i.e., some) and
on the most typical terms for expressing verbal uncertainty (i.e., possible and probable).
Table 24.2 Examples of quantifiers commonly used for judgment and

reasoning. This chapter focuses on some and possible/probable.
Percentages Probabilities
(almost, nearly, virtually) all impossible
some a (slight, good) chance
many (pretty) sure
(quite a, very) few Doubtful
lots Probable
mostly Unlikely
a (slight, large, vast) majority Tossup
• Some. Logically speaking, the existential quantifier “some” in sentences of the form
(2a and 2b) means that at least one x is also a y, or that at least one x did y, respective
ly. Critically, this logical meaning is broad. Broadness here means that (2a) does not
rule out the possibility that all xs are also ys and that (2b) does not rule out the possi
bility that all xs did y. This is the reason why sentences like (3a and 3b) are logically
correct, even though they strike us as bizarre. Later in this chapter, we will see how
the efficiency-based approach to pragmatics explains why people deviate from this log
ical interpretation (with predictable effects on reasoning; Schmidt & Thompson, 2008)
and how politeness-based pragmatics explain why they often revert back to the logical
interpretation.
(2a) Some xs are ys.
(2b) Some xs did y.
(3a) Some squares have four equal sides.
(3b) Some of my readers can read.
Page 4 of 31

• Possible and probable. The terms “possible” and “probable” (or, equivalently, “pos
sibly” and “probably”) are unique among those studied in this chapter in that they do
not have a logical meaning proper. Their everyday meaning is very imprecise (Budescu
& Wallsten, 1995; Wallsten, Budescu, Rapoport, Zwick, & Forsyth, 1986), although
they do share at least one stable characteristic: they are not used for very high proba
bilities. That is, it would be very unusual for the word “probably” in sentence (4) to be
interpreted as meaning “a 95 percent chance,” let alone “a 100 percent chance.” We
will see that there is a good and well-understood pragmatic reason for that—but also
that, as often seen in this chapter, politeness can turn this standard pragmatic phe
nomenon on its head.
(4) I will possibly see Bob on Monday.
The connectives and quantifiers we surveyed in this section (if, or, some, possibly, etc.)
feature prominently in the psychology of judgment and reasoning: it is by manipulating
them that people derive inferences and reach conclusions. One critical factor in this re
spect is how people interpret connectives and quantifiers. People do not rely on logic
alone when interpreting connectives and quantifiers: they also take into account the prag
matic constraints that bear on the use of these terms in everyday conversation. As we will
now see, these constraints can be roughly classified in two families: efficiency and polite
ness.
Efficiency and Politeness

In the 1960s, and for several decades, psychologists accumulated data that showed that
people did not follow the strictures of logic when manipulating connectives and quanti
fiers. Very soon, observers of the field raised the concern that these results did not tell
much about the ability to reason logically (Henle, 1962). Rather, the fact that people did
not reach the logical conclusions afforded by connectives and quantifiers meant that they
did not interpret words such as “if” and “or” the same way logic defined ⊃ and ∨. And,
thus, conversational pragmatics started to inform the psychology of reasoning. In this
section, we consider in turn the efficiency-based pragmatics that has long been main
stream in reasoning research and the politeness-based pragmatics whose implications we
will focus on in the rest of this chapter.
Efficiency
To understand how people interpreted connectives and quantifiers, scholars of reasoning

sought theoretical tools from linguistics, and more (p. 390) specifically from Gricean prag
matics. Gricean pragmatics, as well as neo- and post-Gricean reformulations (Grice, 1989;
Horn, 1984; Levinson, 2000; Sperber & Wilson, 1986/1995), roughly consider that speak
ers use language the way experienced travelers use luggage: they pack as much as possi
ble in as little as possible. In other words, speakers are expected to use language effi
ciently, conveying the maximal amount of relevant information in the most economic way.
Page 5 of 31

Relevance theory (Sperber & Wilson, 1995) offers an elegant formulation of this principle,
which has often been used to explain reasoning data (e.g., Sperber, Cara, & Girotto, 1995).
According to relevance theory, we adopt the interpretation of a statement that makes it
optimally relevant in context because we automatically assume that the speaker aimed at
optimizing relevance. The relevance of a statement refers to the balance between its cog
nitive effects (the beliefs we acquire or revise in its light) and the cognitive effort re
quired to process the statement. An interpretation is optimally relevant when its cogni
tive effects could not have been achieved by any simpler (typically, shorter) statement.
Thus, listeners should adopt the interpretation of a statement that is maximally informa
tive because they expect that an efficient speaker would indeed have aimed at conveying
the greatest possible amount of information with her statement. Let us consider again the
example of conjunction, and more specifically the statement in (5), which might be inter
preted as (6a), (6b), or (6c):
(5) She got pregnant and they got married.

(6a) She got pregnant and (in no particular order) they got married.
(6b) She got pregnant and (in that order) they got married.
(6c) She got pregnant and (in that order and because of this) they got married.
The three interpretations (6a–c) are arranged in increasing order of informativeness. Be
cause (6c) is the maximally informative interpretation, and because there would be no
simpler way than (5) to convey (6c), listeners should tend to interpret (5) as (6c).
This approach to pragmatics has had great influence on the psychology of reasoning (see,
for a review, Politzer, 2004) because it readily applies to many reasoning experiments.
The idea is that participants to reasoning experiments import to the lab their expectation
of linguistic efficiency: they assume that experimenters behave as efficient speakers who
use words in a maximally informative way. As a consequence, participants encountering
words like “if” and “some” would adopt their maximally informative interpretation. There
fore, deviations from logical reasoning should be expected as soon as the maximally infor
mative interpretation of the connective (or quantifier) is not its logically sanctioned inter
pretation.
We will see many examples illustrating the success of this efficiency-based approach, as it
explains quite well what goes on in many reasoning experiments. However, the efficiency-
based approach struggles to explain what goes on in many common contexts, namely, the
contexts in which people have a reason not to seek efficiency. Let us now consider what
could be their reasons not to.
Politeness
The built-in limitation of the efficiency-based approach is its assumption of a mutually un
derstood willingness to be sincere and informative. There are many contexts in everyday
life when one cannot make this assumption, though. To get an idea of what these contexts
Page 6 of 31

are, let us consider what would happen if people were always sincere and maximally in
formative:
(7a) What do you think of my new haircut?

(7b) It’s awful. Easily your worst haircut ever.
(8a) Can we do lunch? Are you free on Monday?
(8b) I’m free, but we cannot do lunch because you are too boring.
(9a) How can I get promoted?
(9b) You can’t. Not a chance.
And let us see now what people are more likely to say:
(10a) What do you think of my new haircut?
(10b) It’s original. It’s very much you.
(11a) Can we do lunch? Are you free on Monday?
(11b) Ah, let me check this and get back to you; I might be busy on Monday.
(12a) How can I get promoted?
(12b) Promotions are rare in this company, they are for the very best.
The first series of examples features sincere and efficient responses, but we would not ex
pect people to express themselves that way. We would rather expect them to provide the
inefficient responses featured in the second series of examples, and the reason we do so
is that we generally expect people to be polite to each other.1
(p. 391) Goffman (1967) argued that individuals who are engaged in social interaction are
strongly motivated to protect their face and that of the other persons they are interacting
with. “Face” is the positive social identity that individuals project and claim for them
selves, and it is negotiated in every daily interaction: one can give face to others or cause
others to lose face, and this occurs reciprocally. When face is endangered, or threatened,
speakers might resort to politeness strategies in order to mitigate the face threat (Brown
& Levinson, 1987).
A great many speech acts can threaten face, either that of the speaker or that of the lis
tener, but the listener’s face can typically and commonly be threatened by impositions
(13a); hostile statements, such as criticism or disapproval (13b); and bad news or bad
prospects (13c).
(13a) Wait for me here.

(13b) This is a stupid thing to say.
(13c) You are not getting a raise this year.
The surest way not to threaten the listener’s face would be never to impose, criticize, or
announce bad news. This is, in fact, the default and most polite strategy listed by polite
ness theory (Brown & Levinson, 1987): don’t say it at all. There are situations, though, in
which it is necessary to perform a face-threatening act. Giving useful feedback to stu
dents, for example, might require pointing out weaknesses in their work. Helping pa
Page 7 of 31

tients understand their treatment options may require announcing bad news about their
current condition and bad prospects attached to side effects of possible treatments.
The simple everyday act of requesting something from someone is itself a face-threat (an
imposition). Accordingly, when requests must be made and yet face must be protected,
the requester may use one of several politeness strategies. The simplest strategy would
be to add a conventional politeness marker such as “please” as in (14a). A more elaborate
strategy is to use an indirect yet conventional form of request, such as (14b). The sen
tence (14b) is literally a question, rather than a request, but it is so conventional a re
quest (just as “Can you pass the salt?”) that it is hard to construe it otherwise. An even
more elaborated strategy is to use a nonconventional, indirect form of request, such as
(14c). Depending on the context, the sentence (14c) could be an indirect request or a can
did comment, and its meaning has to be inferred from the particulars of the situation (De
meure, 2010; Demeure, Bonnefon, & Raufaste, 2008; Holtgraves, 1994; Holtgraves &
Yang, 1990, 1992).
(14a) Wait for me here, please.

(14b) Can you wait for me here?
(14c) It would be really convenient to meet here when I’m done.
It is immediately apparent from examples (14a–c) that increasingly elaborate politeness

strategies2 make for increasingly inefficient statements. Granted, the additional process
ing cost implied by (14a) as compared to (13a) is not much to speak about, and the addi
tional processing cost implied by (14b) is not much either, given the conventional nature
of the pragmatic inference from “Can you…?” (Levinson, 2000). Even so, these two sen
tences are slightly less efficient than (14a) with respect to conveying the desire of the
speaker that the listener wait for her. The nonconventional formulation (14c), however,
clearly requires greater processing, in order for the listener to retrieve the contextual el
ements that will allow her to understand the intention of the speaker.
This tension between efficiency and politeness is also present for other face-threatening
acts, such as negative remarks or replies (Holtgraves, 1998, 1999). Imagine, for example,
that a student is asking his supervisor for feedback (15) and receives one of the replies
shown in (16):
(15) What did you think of my paper?

(16a) It needs to be clearer.
(16b) Perhaps it needs a little more clarity.
(16c) The topic you chose is hard to address clearly.
The reply (16a) most efficiently conveys what needs to be conveyed but is also the least
tactful. The reply (16b) employs a politeness strategy known as hedging, which will fea
ture prominently in this chapter. It is more tactful than (16a) but also more difficult to
process. Finally, the indirect and very tactful reply (16c) is the least efficient way to con
vey the meaning that was neatly encapsulated in (16a). As shown by Holtgraves (1998),
the very fact that (16c) is not a direct reply to the question (15) makes people search for a
Page 8 of 31

reason why the speaker might have chosen to express herself in such a Gricean-ineffi
cient way, and they typically come to the conclusion that she was trying to save the face
of the listener by not bluntly asserting (16a). In fact, when the context makes this inter
pretation implausible (e.g., when it is explicitly stated that the paper was very clear), peo
ple find it hard to make sense of (16c).
Fig. 24.1 When the news is good, one shows consid

eration by being accurate: in this case, there is no
tension between politeness and efficiency. But when
the news is bad, consideration is bought at the cost
of reduced accuracy: in this case, a tradeoff must be
made between politeness and efficiency. The data for
this plot were taken from Experiment 3 in Bonnefon
et al. (2009). They were reanalyzed for the purpose
of this chapter. The plot did not appear in the 2009
article, nor were the data analyzed this way in the ar
ticle.
As illustrated by these examples, it would seem that people expect speakers to hit
(p. 392)
the right trade-off between efficiency and politeness. That is, they realize that speakers
cannot always express themselves in the most efficient manner without upsetting the lis
tener. The use of a politeness strategy makes it less likely that the listener will be upset,
but this increased concern for the listener implies sacrificing some degree of efficiency. It
is important to note right away that politeness and efficiency do not always impose antag
onistic constraints on communication. Indeed, when the news is good, it would show less
consideration to the listener to communicate such tidings inefficiently or inaccurately.
Figure 24.1 illustrates this point with a reanalysis of data published in Bonnefon, Feeney,
and Villejoubert (2009). Participants in this experiment rated how accurately and how
considerately a speaker communicated good or bad news in various situations. Only when
the news was bad (and thus face-threatening) did a tension appear between accuracy and
consideration, taken here as crude proxies for efficiency and politeness.
Page 9 of 31

To summarize, previous research in experimental pragmatics identified two broad as

sumptions about speakers that guide the interpretation listeners make of their state
ments. The first assumption is that speakers seek to maximize the informational efficien
cy of their statements. The second assumption is that speakers seek to minimize the face-
threatening effects of their statements. Unfortunately, these two goals (maximization of
efficiency and minimization of face-threat) are not always compatible, in which case
speakers have to optimize the tradeoff between efficiency and politeness.
In the remainder of this chapter, we consider the implications of this insight for the inter
pretation of connectives and quantifiers in judgment and reasoning tasks. As we will see,
the analysis of connectives and quantifiers in these subfields has been characterized by a
strong focus on efficiency-based pragmatics. This strong focus is not a concern when con
nectives and quantifiers qualify nonthreatening content because, in this case, efficiency
and politeness are not in conflict. However, as soon as connectives and quantifiers qualify
face-threatening content (e.g., bad news or criticisms), the efficiency-focused approach
runs into trouble for failing to capture the effect of politeness.
The Scalarity of “Some” and “Or”

Some
In the reasoning literature, the quantifier “some” is most often studied in the context of
logical syllogisms (17) or immediate quantified inferences (18):
(17a) Some of my clients are not in the fanfare;

(17b) Everyone in the fanfare is my friend;
(17c) Therefore, some of my friends are my clients.
(18a) Some of my students are male;
(18b) Therefore, some of my students are female.
As a matter of fact, conclusions (17c) and (18b) are both logically invalid. That is, (17c) is
not necessarily true given (17a and 17b), and (18b) is not necessarily true given (18a). As
we already saw in the section on quantifiers, the logical interpretation of “some xs are ys”
is that at least one x is also a y, and nothing more. In particular, “some xs are ys” is consis
tent with the possibility that all xs are ys. As a consequence, the premise (18a) means that
at least one of the students is male, but does not rule out the possibility that all students
are male. People typically think otherwise, though (e.g., (p. 393) Evans, Handley, Harper,
& Johnson-Laird, 1999). They find the inference from (18a) to (18b) to be strongly com
pelling and do not usually realize that it is logically invalid. The standard explanation for
this phenomenon is that people adopt the everyday, conversational interpretation of
“some” rather than its logical interpretation (Newstead, 1995; Schmidt & Thompson,
2008). According to this account, although the logical interpretation of “some xs are ys” is
broad (encompassing the case in which all xs are ys), its conversational interpretation is
Page 10 of 31

narrow (excluding the case in which all xs are ys). If people adopt a narrow interpretation
of “some,” both conclusions (17c) and (18b) become valid.
The rationale for adopting a narrow interpretation of “some” is rooted in efficiency-based

pragmatics and appeals to the notion of scalarity (Horn, 1984). Scalar terms are such that
they can be placed on a scale that respects logical implication from right to left, but not
from left to right; for example:
(19a) < trying, succeeding >

(19b) < mediocre, good, great >
(19c) < some, all >
The scale (19a) reflects the semantic facts that (a) one cannot succeed without trying,
and (b) one can try without succeeding. Similarly, a movie cannot be great if it is not
good, but it can be good without being great. According to efficiency-based pragmatics,
speakers should always use the strongest possible term on the scale that is consistent
with their beliefs. If I have succeeded in uploading a file, it would be subinformative to
say that I “tried,” and if I think that a movie was great, it would be subinformative to say
that it was “good.” As a consequence, an efficient speaker who says that a movie was
“good” can be assumed not to believe that it was great. Similarly, an efficient speaker
who says that “some of the students in the class are male” can be assumed not to believe
that all students are male.
Intensive research efforts are being carried out to characterize the cognitive processes
underlying the scalar inference from “some” to “not all” (e.g., Bott, Bailey, & Grodner,
2012; Breheny, Katsos, & Williams, 2006; Huang & Snedeker, 2009). We will not delve in
to this literature in this chapter, and we will instead focus on the way politeness consider
ations can push people to adopt the broad, logical meaning of “some.” To repeat the con
clusion of a previous section: as soon as a quantifier qualifies face-threatening content,
efficiency and politeness considerations are no longer aligned but are rather antagonistic.
Therefore, face-threatening content should pull people away from the efficiency-based in
terpretation of the quantifier. Compare now the following scenarios:
(20a) You cooked a chili for a group of guests.

Later, one of the guests tells you that:
“Some people loved your chili.”
(20b) You cooked a chili for a group of guests.
Later, one of the guests tells you that:
“Some people hated your chili.”
The quantifier “some” in (20a) does not qualify face-threatening content. In this case, effi
ciency and politeness are aligned, and people should reach the narrow interpretation “not
all.” In contrast, “some” in (20b) qualifies face-threatening content. In this case, efficien
cy and politeness push interpretation in opposite directions: people should be in doubt
about whether “some” is meant narrowly or broadly. This is exactly what happened when
Bonnefon et al. (2009) offered this and similar scenarios to samples of undergraduates.
Page 11 of 31

When asked whether it was possible that everyone loved the chili, most participants
adopted a narrow reading: 83 percent said it was not possible. But when asked whether it
was possible that everyone hated the chili, participants were split: 42 percent said that it
was actually possible.
This result provides a first demonstration of the general claim made in this chapter.3
When the quantifier “some” qualifies nonthreatening content, its interpretation is
straightforwardly based on efficiency: it is then interpreted as implying “not all,” which
explains why reasoners give nonlogical responses to quantified syllogisms. However,
when this same quantifier qualifies face-threatening content, politeness considerations
counter efficiency considerations, and reasoners are more likely to give “some” its logical
interpretation of “at least one and possibly all.” In more general terms, efficiency prag
matics explains deviation from logical reasoning, but politeness pragmatics explains why
people appear to be more logical when they reason from face-threatening content.
Or
Consider the following argument:
(21a) The box contains both a pencil and an eraser.

(21b) Therefore, either there is a pencil in the box or there is an eraser in the box.
The conclusion (21b) is logically valid, since logical disjunction x y is inclusive: it is consis
tent with the possibility that both x and y are true. However, (p. 394) 40 to 50 percent of
adult reasoners disagreed with similar conclusions in a classic study by Braine and Ru
main (1981). This deviation from standard logic can be explained on pragmatic grounds,
appealing again to the notions of efficiency and scalarity. Just as the pair < some, all >,
the connectives “or” and “and” can be placed on a scale:
(22) < or, and >
For scalarity reasons, a maximally informative speaker would not use a disjunction if she
could have used a conjunction: if I know that we will take both a bus and a taxi to arrive
at our destination, it would be subinformative to say that “we will take a taxi or a bus to
arrive at our destination.” As a consequence, listeners who assume speakers to be effi
cient should adopt an exclusive interpretation of disjunction and consider that “x or y”
means “either x or y but not both” (Breheny et al., 2006; Chevallier et al., 2008; De Neys
& Schaeken, 2007; Noveck, Chierchia, Chevaux, Guelminger, & Sylvestre, 2002).
As long as “or” does not connect face-threatening propositions, we can expect reasoners
to adopt the efficient interpretation of disjunction. Our central claim in this chapter,
though, is that face-threatening content introduces politeness constraints on language
that push interpretation away from efficiency. In the case of disjunction, we should be
able to show that face-threatening content decreases reasoners’ tendency to give “or” an
exclusive interpretation.
Page 12 of 31

This claim was tested by Feeney and Bonnefon (in press), who presented reasoners with
scenarios in which the disjunction “or” either connected two pieces of good news (23b) or
two pieces of bad news (24b):
(23a) Imagine that Clare is a children’s author. Over lunch her publisher tells her
that the sales of her last book have been so…
(23b) good that she will receive increased royalties or she will be entitled to an up
front payment for her next book.
(24a). Imagine that Clare is a children’s author. Over lunch her publisher tells her
that the sales of her last book have been so…
(24b) poor that she will receive decreased royalties or she will be denied an upfront
payment for her next book.
The disjunction in (23b) does not connect face-threatening contents and should be given
its efficient, exclusive interpretation. The disjunction in (24b), in contrast, connects face-
threatening contents: its interpretation should therefore reflect the tension between po
liteness and efficiency and be drawn to inclusiveness. This is exactly what happened. A
standardized measure of interpretation showed that the disjunctions in (23b) and (24b)
were given exclusive and inclusive interpretations, respectively. Reasoners did not expect
successful Clare to get increased royalties and an upfront payment, but they could very
well envision poor-sales Clare to get decreased royalties and no upfront payment.
In this experiment, the contents connected by “or” were either neutral or face-threaten
ing as a function of whether they constituted good or bad news. Other manipulations are
possible that would generate face-threatening content; for example, criticisms (as in 25a)
or impositions (as in 25b):
(25a) You are an idiot, or you are not paying attention.

(25b) You must work early in the morning or leave late at night.
The statement (25a) does seem to irresistibly invite a gloss such as “and you’re most
probably both,” although materials of this type have not been tested yet. Whether (25b)
invites an inclusive interpretation is also an open question. The pragmatics of statements
similar to (25b)—i.e., disjunctions connecting two impositions—were studied in at least
one experiment (Elqayam, Ohm, Evans, & Over, 2010), but the focus of Elqayam et al.’s
work was not the exclusive or inclusive interpretation of “or.” Rather, the work focused on
the implications conveyed by the order of the disjuncts, showing that the disjunct that is
introduced first (“work early in the morning”) is perceived as the one preferred by the au
thority asserting the disjunction. Future work will say whether criticisms and impositions
have the same effects as bad news; that is, do they make reasoners adopt the logical, in
clusive interpretation of “or” rather than its efficient, exclusive interpretation?
Page 13 of 31

Verbal Uncertainty: The Severity Effect

Verbal expressions of uncertainty such as “it is possible/likely/probably the case that…”
play an important role in judgment and reasoning, in particular because they can inform
decisions in complex situations, in which the decision-maker is not in possession of all the
relevant information and must turn to an expert for advice. At the individual level, health
and finance are good examples (p. 395) of domains wherein a decision-maker is likely to
need expert input. At a global level, climate change is largely recognized as a domain
wherein careful communication of uncertainty is required to inform national and interna
tional policies, as well as individual awareness of scientific data (Budescu, Broomell, &
Por, 2009).
There is considerable inter- and intraindividual variance in the interpretation of vague

quantifiers such as “possibly,” and this interpretation is heavily dependent on context
(Budescu & Wallsten, 1995; Karelitz & Budescu, 2004). Nonetheless, the general princi
ple of scalarity, which we already encountered for logical connectives, can also apply to
verbal probability terms. Indeed, probability terms can be ordered on informativeness
scales such as (26):
(26) < possible, probable, certain >
As a consequence, an efficient speaker would not say that something is “possible” or

“probable” if he or she was sure that this thing was true. Thus, even though the assertion
that x is “possible” would be consistent with any probability in [0, 1], efficiency-based
considerations allow one to infer that the probability of x cannot be very high. As often oc
curs, though, considerations of politeness complicate that simple picture.
Instances of the Severity Effect
One of the foremost pragmatic phenomenon in the interpretation of probability phrases

relates to an early puzzle (Weber & Hilton, 1990) that has come to be known as the sever
ity effect (Bonnefon & Villejoubert, 2006). The severity effect manifests itself when rea
soners must provide a numerical translation of a probability term that is applied to out
comes of varying badness. For example, in the context of doctor–patient communication,
not all diagnoses or prognoses are equally frightening:
(27a) It’s probably a gastric disturbance.

(27b) It’s probably skin cancer.
Page 14 of 31

In (27), the numerical probability of skin cancer would be perceived as higher than the
numerical probability of a gastric disturbance, even though the two conditions are simi
larly qualified as “probable.”4 The severity effect was observed in a variety of contexts in
addition to doctor–patient communication. Villejoubert, Almond, and Alison (2009)
demonstrated a severity effect in a forensic context, in which reasoners had to interpret
statements made about offenders’ characteristics. For example:
(28a) The offender will possibly return to the crime scene.

(28b) The offender will possibly kill again.
(29a) The offender is likely a postman.
(29b) The offender is likely a psychopath.
The more severe version of each statement, such as (28b) or (29b), was interpreted as 8
percentage points more probable than its less severe version, on average. Stronger ma
nipulations of severity can produce larger severity effects. Harris and Corner (2011)
investigated severity effects in the context of environmental risks linked to climate
change, nuclear power, and nanotechnology, using especially strong manipulations of
severity. For example, one scenario featured a traffic accident involving supplies for a nu
clear power station. In the less severe condition, the supplies were drinking water for the
employees, the release of which would cause a minor traffic delay. In the more severe
condition, the supplies were toxic uranium, the release of which would kill thousands of
people. In both conditions, the accident was described as “unlikely” or “likely” by a senior
analyst. In this and similar scenarios, the probability of the accident was perceived as 10
to 15 percentage points higher in the more severe condition.
The Politeness Account
Bonnefon and Villejoubert (2006) offered a politeness-based explanation of the severity

effect, grounded in the pragmatic interpretation of terms such as “possibly” and “proba
bly.” This explanation capitalizes on the same analysis we already encountered for terms
such as “some” and “or”—that is, it contrasts situations in which these terms qualify neu
tral versus face-threatening content. When a probability term qualifies neutral content,
its pragmatic interpretation is governed by efficiency considerations: something that is
only qualified as “possible” cannot have a very high probability (adjusting for base rate)
because if it had, an efficient speaker would have used a stronger term. In contrast, a
probability term that qualifies face-threatening content (such as bad news or bad
prospects) could be used by a polite speaker to attenuate the face-threat by sugar-coating
the bad news. In this case, politeness would trump efficiency, and the fact that an event is
only qualified as “possible” would not imply that its probability is low. To test this expla
nation, Bonnefon and Villejoubert (2006) asked reasoners (p. 396) to imagine that their
family doctor would make statements such as:
(30a) You will possibly become insomniac within the year to come.
(30b) You will possibly become deaf within the year to come.
Page 15 of 31

Reasoners provided their numerical interpretation of the probability that they would be
come insomniac or deaf based on these statements, but they also indicated what they
thought to be the intention of the doctor when she used the word “possibly”: conveying
uncertainty about her prognosis or tactfully breaking the bad news. It was found that the
critical factor was the perceived intention of the doctor, rather than the severity of the
prognosis per se. The perceived probability of the prognosis was about. 55 when the doc
tor was assumed to express uncertainty, and above. 70 when she was assumed to be tact
ful. Remarkably, these figures were broadly the same for insomnia and deafness. Howev
er, reasoners were much more likely to assume the doctor to be tactful when making the
deafness prognosis, hence the severity effect: the more severe the outcome, the greater
the chance that reasoners shift from an efficiency interpretation to a polite interpretation.
This politeness-based explanation received further support from a series of experiments

reported in Juanchich, Sirota, and Butler (2012), in which the numerical interpretation of
“possibly” was 10 to 25 percentage points higher when reasoners assumed that it was
used for politeness purposes.5 Furthermore, Juanchich et al. (2012) observed that this
change in interpretation drove a change in decision-making. For example, reasoners who
were told that their (imaginary) stocks would “possibly” drop in value soon were 20 per
cent more likely to sell if they considered that this term was used politely.
Visual and Numerical Severity Effects
Although these data strongly suggest that politeness accounts to some extent for the
severity effect, it is likely that other mechanisms play a role as well. For example, it has
been observed that underestimating the probability of a severe outcome could result in a
costly tendency not to take protective measures (Harris & Corner, 2011; Harris, Corner,
& Hahn, 2009; Weber, 1994). According to this asymmetric loss account, it would be ratio
nal for a reasoner to overestimate the probability of severe outcomes but not that of mild
ly negative outcome—and the severity effect would reflect that rational (mis)calibration.
The best evidence in favor of the asymmetric loss account came from studies reported by
Harris et al. (2009), in which reasoners estimated probabilities from visual displays. In
one experiment, reasoners observed a large matrix of randomly distributed yellow and
black cells (2,236 cells in total) and were told that each cell represented an apple tree in
an orchard. Yellow cells in particular represented trees whose apples were ill-suited for
consumption. In the mild severity condition, they were sour and inedible; in the high
severity condition, they were fatally poisonous. Reasoners were asked to estimate the
probability of choosing a bad apple when randomly picking from a tree. Quite remarkably,
reasoners perceived this probability as being higher (by 3 to 7 percentage points) when
the apples were fatally poisonous.
It is hard to see how politeness could play a role in this visual version of the severity ef
fect. It is thus likely that the severity effect is a combination of politeness and asymmetric
loss—or perhaps that asymmetric loss is the kernel of the effect, which is magnified by
politeness expectations when the appropriate conditions are met.
Page 16 of 31

In addition to being demonstrated with verbal and visual representations of probability,

the severity effect can be observed with numerical representations of probability. This
striking version of the effect was obtained while investigating pregnant women’s under
standing of statistical information related to prenatal diagnosis. Pighin, Bonnefon, and
Savadori (2011) presented pregnant women with hypothetical prenatal diagnoses such
as:
(31) Elisa is a 30-year-old pregnant woman. Her gynecologist says: There is a 1 in

307 chance that your child will be affected by Down syndrome.
The scenarios varied the numerical probability of the diagnosis, as well as its severity: in
the mild severity condition, the diagnosis was about insomnia rather than Down syn
drome. The pregnant women participating in the study indicated how high was the nu
merical probability mentioned in the diagnosis, on a 7-point scale anchored at extremely
low and extremely high. This measure was meant to capture the subjective magnitude of
the probability that was communicated numerically (in fraction form). If numerical proba
bilities fall prey to the severity effect, then a given probability applied to Down syndrome
should “feel” bigger than when it is applied to insomnia. This is exactly what happened:
the probability of the severe outcome felt 1 (p. 397) to 2 points larger on the subjective 7-
point scale, even though the numerical probability of the two outcomes was objectively
the same.
This result left open the question of whether politeness might be responsible for the sub
jective overestimation of numerical probabilities in the context of severe outcomes. An
other paper, though, suggested that this was indeed the case. Sirota and Juanchich (in
press) presented reasoners with numerical assessments of various risks, in percentage
(32a) or ratio formats (32b), and asked them to provide their subjective assessments of
the probability that the speaker had in mind.
(32a) There is a 50 percent chance that your job application be rejected.

(32b) There is 1 chance in 3 that your stocks will lose value.
Critically, the experiments manipulated whether politeness was to be expected from the
speaker. For example, in one experiment, the speaker was either described as “a polite
person, who tends to soften any bad news or criticism” (politeness condition) or “a person
who tends to give his opinion bluntly, even bad news or criticism” (efficiency condition).
In another experiment, the speaker was described as perceiving that “the hearer could be
easily upset” (politeness condition) or that “the hearer expected clear
information” (efficiency condition). It was expected that in the polite condition, reasoners
would adjust upward the numerical probability that was communicated—and they did.
Whatever the manipulation or the probability format, reasoners interpreted the probabili
ties upward in the politeness condition. For example, the statement (33b) translated as a
33 percent chance in the efficiency condition (unsurprisingly), but as a 39 percent chance
in the politeness condition.
Page 17 of 31

In sum, the politeness explanation of the severity effect, which was developed in the con
text of verbal probability expressions, appears to also account for the variant of the effect
observed for numerical probability expressions. So far, it seems rather unlikely that po
liteness might account for visual variants of the severity effect—but this is certainly a tan
talizing possibility, worthy of future research.
Polite Conditionals
In this section, we consider how politeness issues affect reasoning with the conditional
connective “if x then y.” We begin with a brief evocation of a rather atypical use of the
conditional, whose purpose is not to express causality but conventional politeness. We
then move on to causal conditionals and consider in turn two politeness effects. The first
effect follows straightforwardly from the results on verbal uncertainty that were reviewed
in the previous section. The second effect relates to the way we combine several causal
conditionals to form a representation of causal structure.
Noncausal Conditionals
Although we will mostly concentrate on causal conditionals in the rest of this chapter, a
section on politeness and conditionals would not be complete without a reference to com
mentative conditional speech acts (Van Canegem-Ardijns & Van Belle, 2008). Commenta
tive conditional speech acts use the conditional form to attenuate face-threats. For exam
ple, compare:
(33a) You’re wrong.

(33b) If I understand correctly, you’re wrong.
Or compare:
(34a) Take out the garbage before 6 p.m.

(34b) If you think about it, take out the garbage before 6 p.m.
There is no clear causal connection between antecedent and consequent in (33b) or

(34b). For example, the fact that the listener would think of taking out the garbage does
not cause the speaker’s request to do it. In these examples, the conditional form is a po
liteness device (Tedeschi, 1977; van der Auwera, 1986) the function of which is to provide
plausible deniability.
Indeed, after having asserted (33a), the speaker has no way to claim that she did not in
fact criticize the listener; and after having asserted (34a), the speaker has no way to
claim that she did not in fact order the listener around. In contrast, (34b) allows the
speaker to deny that she was issuing an order and to claim that she was merely making a
suggestion. Likewise, (33b) allows the speaker to deny that she corrected the listener and
to claim instead that she was not understood correctly.
Page 18 of 31

This polite use of the conditional is not commonly encountered in reasoning experiments
(but see Bonnefon & Politzer, 2011) where conditionals commonly feature a causal rela
tion between antecedent and consequent (be it predictive or diagnostic). To the best of
my knowledge, linguistic pragmatics did not identify polite uses of causal conditionals in
the way it identified polite uses of (noncausal) commentative conditionals. In the next sec
tion, we will see that politeness concerns do (p. 398) nonetheless matter when interpret
ing (and reasoning from) causal conditionals.
Uncertainty Propagation
The conditional “if x then y” is by far the most studied of logical connectives (Politzer,
2007). Even the most elementary of all conditional inferences (35) already accounts for a
fair share of reasoning research:
(35a) If x is true then y is true;

(35b) x is true;
(35c) Therefore, y is true.
The interest for this inference is largely due to its simplicity and apparent automaticity.
Upon seeing this inference in abstract form, close to 100 percent of reasoners agree that
it is logically valid; that is, that (35c) is necessarily true when (35a and b) are true. This
does not mean, though, that they are likely to use this inference much in real life. Indeed,
much of the information we reason from in everyday life is vague, uncertain, or unreli
able, rather than precise and guaranteed to be true.6 For example, reasoners might be
unable to assume that x is true, but be limited instead to the assumption that x is possibly
true, probably true, etc. In that case, reasoners tend to propagate the uncertainty of x to
their conclusion about y (George, 1995, 1997):
(36a) If x is true then y is true;

(36b) x is possibly true;
(36c) Therefore, y is possibly true.
As it has often been the case in this chapter, though, this claim must be nuanced to incor
porate politeness concerns. Let us compare two scenarios, in which x differs in terms of
whether it constitutes a face-threat:
(37a) If your application is approved, then you will hear from me tomorrow;
(37b) It is possible that your application is approved.
(38a) If your application is rejected, then you will hear from me tomorrow;
(38b) It is possible that your application is rejected.
We would expect reasoners to infer from (37) that the listener will “possibly” hear from
the speaker the day after (as a result of standard uncertainty propagation from premise
to conclusion). However, we also know (see the previous section of this chapter) that the
term “possible” in (38b) can be interpreted as a politeness device: the speaker is tactfully
communicating that chances are the application will be rejected. Reasoners who adopt
Page 19 of 31

this interpretation should adjust upward the probability of the conclusion and infer that
the listener should very probably hear from the speaker the day after. Pighin and Bonne
fon (2011) tested this hypothesis in a medical context, presenting reasoners with hypo
thetical dialogues between doctors and patients, such as:
(39a) While talking to his doctor during a visit, Christian is informed that if pain de
creases, he will change therapy.
(39b) The doctor remarks that: “It is possible that the pain decreases.”
(40a) While talking to his doctor during a visit, Christian is informed that if pain in
creases, he will change therapy.
(40b) The doctor remarks that: “It is possible that the pain increases.”
As predicted, reasoners tended to propagate uncertainty for dialogues such as (39) and to
adjust probability upward for face-threatening dialogues such as (40). That is, most par
ticipants concluded from (39) that Christian would “possibly” change therapy, and most
participants concluded from (40) that Christian would “probably” or “certainly” change
therapy. This result provides yet another demonstration that inferences can depend on
whether the premises feature face-threatening content. The next section introduces a
slightly different perspective by attending to the conditions under which reasoners infer
that a premise is face-threatening in the first place.
Causal Structure
A basic problem of conditional reasoning is to form a causal model out of two rules “if c 1
then e” and “if c 2 then e,” where c 1 and c 2 are causes and e is an effect. For example:
(41a) If I dive in the pool, I’ll get wet;

(41b) If I take a shower, I’ll get wet.
(42a) If the store is opened, I’ll buy some groceries;
(42b) If I have money, I’ll buy some groceries.
For the most part, the literature on causal conditionals has focused on two different mod
els that can be built out of such pairs of rules (e.g., Bonnefon & Hilton, 2002; Byrne,
1989; Oaksford & Chater, 2003; Politzer, 2005; Politzer & Bonnefon, 2006; Politzer &
Bourmeau, 2002; Stevenson & Over, (p. 399) 1995). In the Collider model (see Figure
24.2), c 1 and c 2 are construed as two alternative causes of the common effect e: each of
them is sufficient to produce e. In contrast, in the Enabler model, c 1 and c 2 are no longer
construed as alternative, sufficient causes of e: only their joint occurrence produces e. In
some cases, the linguistic content of c 1, c 2, and e make it easy to determine which causal
model should be formed out of the two rules. For example, the two rules in (41) suggest a
Collider model, whereas the two rules in (42) suggest an Enabler model.
In other cases, the content of c 1, c 2, and e are ambiguous enough that they could equally
well suggest Collider or Enabler models (Stenning & van Lambalgen, 2005). As an illus
Page 20 of 31

tration, consider the following problem. A patient (Bill) has been diagnosed by two doc
tors:
(43a) The first doctor says: “If Bill is given rest, he will make a quick recovery.”
(43b) The second doctor says: “If Bill is given treatment, he will make a quick recov
ery.”
Fig. 24.2 Two causal models can be formed from the

causal rules “if c 1 then e” and “if c 2 then e.” In the
Collider model (left), c 1 and c 2 are alternative, indi
vidually sufficient causes of the common effect e. In
the Enabler model (right), c 1 and c 2 are jointly but
not individually sufficient to cause e.
There is a potential ambiguity here as to how these two rules might be combined. One
may understand that either rest or treatment will cause a quick recovery (the Collider
model), or one may understand that rest and treatment jointly but not independently
cause a quick recovery (the Enabler model). Politeness, however, complicates the situa
tion by introducing a third possible model. The first hint of that model came from a study
by Stevenson and Over (2001), which manipulated the purported medical expertise of the
two doctors in problems similar to (43). The authors observed that when the second doc
tor had more expertise than the first, reasoners formed a causal model in which treat
ment alone (but not rest) was sufficient for a quick recovery. The authors claimed that the
higher expertise of the second doctor prompted reasoners to consider that her statement
was meant as a correction and to discard the first statement as a mistake from the inexpe
rienced doctor.
Note that (43b) is a rather contrived way to phrase a correction. Instead of directly telling
the first doctor that she was wrong, the second doctor is asserting a conditional that is
consistent with three causal models: Collider, Enabler, and Correction. This is, of course,
a typical move for a polite speaker: instead of bluntly correcting the listener (a face-
threatening act), the speaker asserts a statement that the listener is free to understand as
a correction. From this analysis, Demeure, Bonnefon, and Raufaste (2009) derived the fol
lowing prediction: the probability for a statement such as (43b) to be interpreted as a cor
rection, in a given context, should be proportional to the threat level of a correction in
that context. In other words, if the context is such that the listener does not mind being
corrected, (43b) should not be interpreted as a correction (because the speaker could
have afforded to be more direct). In contrast, if the context is such that the listener would
find a correction very upsetting, (43b) should be interpreted as a (politely phrased) cor
rection.
Page 21 of 31

To test their hypothesis, Demeure et al. (2009) presented reasoners with hypothetical dia
logues between workers in a factory. The workers had gone through training with a new
machine and were discussing what they remembered about the machine:
(44a) The first worker says: “If the water level is low, then the machine stops.”
(44b) Another worker says: “If the oil level is low, then the machine stops.”
The experiment manipulated the threat level of a correction through a visual and verbal
description of the first worker. In one condition, the worker was drawn smiling and with
his arms opened and described as an open-minded employee who listened to others and
cared about their opinions and ideas. In another condition, the worker was drawn frown
ing and crossed-armed and described as a very touchy employee who disliked being con
tradicted, liked to be in control, and liked to impose his point of view.
This manipulation had a large effect on interpretation and reasoning. In the “frowning
worker” condition, reasoners were more likely to perceive (44b) as a correction, assigned
a lower probability to the conditional in (44a), and were less inclined to predict that the
machine would stop when told that the water level was low. A second experiment replicat
ed these effects with another manipulation (p. 400) of threat level, which simply consisted
in telling the participants that the first speaker liked or hated the second speaker. The ra
tionale for this manipulation was to increase the affective distance between the speakers,
a parameter that is known to increase the need for politeness (Slugoski & Turnbull, 1988;
Spencer-Oatey, 1996).
These data demonstrated that when politeness was unnecessary (either because of the
personality of the first speaker or because of his affection for the second speaker), pairs
of statements such as (44) were combined in a Collider model. But when the need for po
liteness was high, reasoners preferentially adopted a Correction model, with predictable
effects on the inferences they were ready to make.
Future Directions
Connectives (if, not, or, and, etc.) and quantifiers (some, most, all, likely, probably, etc.)
are the building blocks of judgment and reasoning. They are the words that people use to
combine and qualify propositions in order to reach an understanding of the world and of
the consequences of their decisions. The psychology of judgment and reasoning commit
ted early on to the notion that pragmatic factors could affect the interpretation of connec
tives and quantifiers, which could in turn affect inferences and decisions. To account for
these effects, the psychology of judgment and reasoning imported theoretical frameworks
from linguistic pragmatics. These theoretical frameworks had a strong focus on conversa
tional efficiency, but disregarded facework and politeness.
Considerations of efficiency can explain a large range of interpretative effects; for exam
ple, the scalarity of “some” and “or,” the narrow range of numerical translations for a
word such as “probable,” and the propagation of uncertainty from antecedent to conse
Page 22 of 31

quent in conditional reasoning. Considerations of efficiency fail, though, as soon as con

nectives and quantifiers apply to face-threatening content such as criticisms, requests,
impositions, or bad news. When face-threatening content is involved, considerations of ef
ficiency and politeness are antagonistic and classic efficiency effects disappear, for polite
ness reasons. Thus, politeness considerations can explain (for example) why people some
times revert to the logical interpretation of “some” and “or,” why they assign greater
probability to severe events, and why they sometimes cancel causal conditional infer
ences.
At this point, though, one might ask how important these politeness effects are at the the
oretical and practical levels. Could they merely be acceptable exceptions to the general
rules of efficiency? That is, can we be happy with using efficiency as the general explana
tory principle of interpretation effects and consider politeness as a subsidiary principle to
be called upon in some special cases? I believe that there are two reasons why we should
rather give equal footing to politeness and efficiency when studying interpretative effects
in judgment and reasoning. The first reason is that face-threatening matters are too fre
quent, in everyday life, to be considered mere exceptions to a rule of efficient communi
cation. The second reason is that face-threatening matters are involved, almost by defini
tion, in higher-stakes situations than neutral matters.
The frequency with which we use connectives and quantifiers in a polite way, in everyday
life, is not easy to quantify. Some corpus studies would indicate that this frequency is sub
stantial. For example, Youmans (2001) observed that about 40 percent of probability
terms (in an American English corpus) were used for face-management purposes. In par
allel, Fortantet (2004) reported on the various functions of “I think” in academic dis
course, observing one polite occurrence (management of face threat) for two occurrences
where “I think” was used to introduce uncertainty. Even without hard quantitative evi
dence, it seems a compelling intuition that we very often need to offer criticisms, make
requests, make impositions, and announce bad prospects—all situations in which polite
ness considerations are required to account for the interpretation of connectives and
quantifiers.
The importance of politeness for judgment and reasoning does not only derive from the
frequency with which we need to convey threatening information—we also need to take
into account the practical importance of conveying threatening information. Note that the
default choice of speakers, in many situations, is simply to opt out of communicating up
setting information: only if the stakes are sufficiently high do speakers decide to (politely)
point out mistakes, bad choices, or bad prospects to listeners. The fact that they use less
efficient language in these situations than they would otherwise paves the way for an un
fortunate circumstance that we may call the high-stakes paradox (Bonnefon, Feeney, & De
Neys, 2011):
• The higher the stakes of a decision, the more important it is to obtain clear informa
tion;
Page 23 of 31

• The higher the stakes of a decision, the greater the chance that bad prospects must
be evoked and mistakes must be corrected;
• The greater the chance that bad prospects must be evoked and mistakes
(p. 401)
must be corrected, the greater the chance that informants will sacrifice clarity for po
liteness;
• As a consequence, the more important it is that we obtain clear information, the less
er the chance that informants will communicate clearly.
If we want to limit the damage that the high-stakes paradox can inflict, we need to inves
tigate the factors that prompt speakers to use connectives and quantifiers politely—but
also the accuracy with which listeners detect that speakers are using (or not) a polite
strategy. We need to pay special attention to situations in which listeners might not be
able to detect the polite intention of the speaker. This inability may result from neurologi
cal conditions such as Parkinson disease (Holtgraves & McNamara, 2010; McNamara,
Holtgraves, Durso, & Harris, 2010), from stressors such as concurrent cognitive load
(Bonnefon, De Neys, & Feeney, 2011), from intercultural misunderstandings, from limita
tions linked to developmental trajectories, or even from interindividual differences in per
sonality (Feeney & Bonnefon, in press). As suggested from this inventory, a full under
standing of polite misunderstandings will require a joint effort of cognitive, social, devel
opmental, and personality psychology, together with the help of linguistics and the neuro
sciences. Politeness has long been a natural topic for the cognitive sciences, studied in
linguistics, anthropology, artificial intelligence, psychology, and the neurosciences—but
interdisciplinary research programs have been rare, perhaps because each field has fo
cused on its own set of politeness problems. The problem of avoiding polite misunder
standings in high-stakes decision situations will perhaps be relevant enough to all the
cognitive sciences to engage in a fully integrated research effort.
References
Blum-Kulka, S. (1987). Indirectness and politeness in requests: Same or different? Journal
of Pragmatics, 15, 131–146.
Bonnefon, J. F. (2009). A theory of utility conditionals: Paralogical reasoning from deci

sion-theoretic leakage. Psychological Review, 116, 888–907.
Bonnefon, J. R, De Neys, W., & Feeney, A. (2011). Processing scalar inferences in face-
threatening contexts. In L. Carlson, C. Hölscher, & T. Shipley (Eds.), Proceedings of the
33rd Annual Conference of the Cognitive Science Society. Austin, TX: Cognitive Science
Society.
Bonnefon, J. F, Feeney, A., & De Neys, W. (2011). The risk of polite misunderstandings.
Current Directions in Psychological Science, 20, 321–324.
Bonnefon, J. F, Feeney, A., & Villejoubert, G. (2009). When some is actually all: Scalar in
ferences in face-threatening contexts. Cognition, 112, 249–258.
Page 24 of 31

Bonnefon, J. F, & Hilton, D. J. (2002). The suppression of modus ponens as a case of prag
matic preconditional reasoning. Thinking and Reasoning, 8, 21–40.
Bonnefon, J. F, & Politzer, G. (2011). Pragmatics, mental models, and one paradox of the
material conditional. Mind and Language, 26, 141–155.
Bonnefon, J. F, & Villejoubert, G. (2006). Tactful or doubtful? Expectations of po

(p. 402)
liteness explain the severity bias in the interpretation of probability phrases. Psychologi
Bott, L, Bailey, T. M., & Grodner, D. (2012). Distinguishing speed from accuracy in scalar
implicatures. Journal of Memory and Language, 66, 123–142.
Braine, M. D. S., & Rumain, B. (1981). Development of comprehension of “or”: evidence

for a sequence of competencies. Journal of Experimental Child Psychology, 31, 46–70.
Breheny, R., Katsos, N., & Williams, J. (2006). Are generalised scalar implicatures gener
ated by default? An on-line investigation into the role of context in generating pragmatic
inferences. Cognition, 100, 434–463.
bridge: Cambridge University Press. (Original work published 1978)
Budescu, D. V., Broomell, S., & Por, H. H. (2009). Improving communication of uncertain
ty in the reports of the intergovernmental panel on climate change. Psychological
Science, 20, 299–308.
Budescu, D. V., & Wallsten, T. S. (1995). Processing linguistic probabilities: General prin
ciples and empirical evidence. In J. Busemeyer, D. L. Medin, & R. Hastie (Eds.), Decision
making from a cognitive perspective (pp. 275–318). San Diego, CA: Academic Press.
Byrne, R. M. J. (1989). Suppressing valid inferences with conditionals. Cognition, 31, 61–
83.
Chevallier, C., Noveck, I. A., Nazir, T., Bott, L., Lanzetti, V., & Sperber, D. (2008). Making
disjunctions exclusive. Quarterly Journal of Experimental Psychology, 61, 1741–1760.
Culpeper, J. (2011). Impoliteness: Using language to cause offence. Cambridge: Cam

De Neys, W., & Schaeken, W. (2007). When people are more logical under cognitive load—
Dual task impact on scalar implicatures. Experimental Psychology, 54, 128–133.
Degen, J., & Tanenhaus, M. K. (2012). Processing scalar implicature: A constraint-based

approach (Tech. Rep.). Retrieved fromhttp://www.bcs.rochester.edu/people/jdegen/
docs/DegenTanenhaus_underreview.pdf
Page 25 of 31

Demeure, V. (2010). Facework and utilitarian relevance in the disambiguation of state

ments with two indirect interpretations. Journal of Language and Social Psychology, 29,
443–457.
Demeure, V., Bonnefon, J. F., & Raufaste, E. (2008). Utilitarian relevance and face man
agement in the interpretation of ambiguous question/request statements. Memory and
Cognition, 36, 873–881.
Demeure, V., Bonnefon, J. F., & Raufaste, E. (2009). Politeness and conditional reasoning:
Interpersonal cues to the indirect suppression of deductive inferences. Journal of Experi
mental Psychology: Learning, Memory, & Cognition, 35, 260–266.
Elqayam, S., Ohm, E., Evans, J. S. B. T., & Over, D. E. (2010). First things first: Order bias
in deontic disjunctions. Language and Cognitive Processes, 25, 375–401.
Evans, J. S. B. T. (2002). Logic and human reasoning: An assessment of the deduction par
adigm. Psychological Bulletin, 128, 978–996.
Evans, J. S. B. T. (2012). Questions and challenges for the new psychology of reasoning.
Thinking and Reasoning, 18, 5–31.
Evans, J. S. B. T., Handley, S. J., Harper, C., & Johnson-Laird, P. N. (1999). Reasoning about
necessity and possibility: A test of the mental model theory of deduction. Journal of Exper
imental Psychology: Learning, Memory, and Cognition, 25, 1495–1513.
Feeney, A., & Bonnefon, J. F. (in press). Politeness and honesty contribute additively to the
interpretation of scalar expressions. Journal of Language and Social Psychology, doi:
10.1177/0261927X12456840.
Fortantet, I. (2004). I think: opinion, uncertainty or politeness in academic spoken Eng

lish? Revista Electrónica de Lingüística Aplicada, 3, 63–84.
George, C. (1995). The endorsement of the premises: Assumption-based or belief-based

reasoning. British Journal of Psychology, 86, 93–111.
George, C. (1997). Reasoning from uncertain premises. Thinking and Reasoning, 3, 161–
240.
Goffman, E. (1967). Interaction ritual: Essays on face-to-face behavior. Garden City, NJ:
Anchor Books.
Grice, H. P. (1989). Studies in the way of words. Cambridge, MA: MIT Press.
Harris, A. J. L., & Corner, A. (2011). Communicating environmental risks: Clarifying the
severity effect in interpretations of verbal probability expressions. Journal of Experimen
tal Psychology: Learning, Memory and Cognition, 6, 1571–1578.
Harris, A. J. L., Corner, A., & Hahn, U. (2009). Estimating the probability of negative
events. Cognition, 110, 51–64.
Page 26 of 31

Hartshorne, J., & Snedeker, J. (2012). The speed of inference: Evidence against rapid use
of context in calculation of scalar implicatures (Tech. Rep.). Retrieved from http://
www.gameswithwords.org/Hartshorne/papers/HartshorneSnedeker_SIRT.doc
Henle, M. (1962). On the relation between logic and thinking. Psychological Review, 69,
366–378.
Hertwig, R., Benz, B., & Krauss, S. (2008). The conjunction fallacy and the many mean
ings of and. Cognition, 108, 740–753.
Heyman, T., Schaeken, W., & Pipijn, K. (2012). Logical or pragmatic, as long as it suits our
convenience: Scalar inferences in a pro-and contra-attitudinal context. In N. Miyake, D.
Peebles, & R. Cooper (Eds.), Proceedings of the 34th annual conference of the cognitive
science society (pp. 456–461). Austin, TX: Cognitive Science Society.
Holtgraves, T. (1994). Communication in context: effects of speaker status on the compre

hension of indirect requests. Journal of Experimental Psychology: Learning, Memory, and
Cognition, 20, 1205–1218.
Holtgraves, T. (1998). Interpreting indirect replies. Cognitive Psychology, 37, 1–27.
Holtgraves, T. (1999). Comprehending indirect replies: When and how are their conveyed
meaning activated? Journal of Memory and Language, 41, 519–540.
Holtgraves, T., & McNamara, P. (2010). Parkinson’s disease and politeness. Journal of
Language and Social Psychology, 29, 178–93.
Holtgraves, T., & Yang, J. (1990). Politeness as universal: Cross-cultural perceptions of re
quest strategies and inferences based on their use. Journal of Personality and Social Psy
chology, 59, 719–729.
Holtgraves, T., & Yang, J. (1992). Interpersonal underpinnings of request strategies: Gen
eral principles and differences due to culture and gender. Journal of Personality and So
Horn, L. R. (1984). Toward a new taxonomy for pragmatic inference: Q-based and R-
based implicature. In D. Schiffrin (Ed.), Meaning, form, and use in context (pp. 11–42).
Washington, DC: Georgetown University Press.
Huang, Y. T., & Snedeker, J. (2009). Online interpretation of scalar quantifiers: In
(p. 403)
sight into the semantics-pragmatics interface. Cognitive Psychology, 58, 376–415.
Juanchich, M., Sirota, M., & Butler, C. L. (2012). Effect of the perceived functions of lin
guistic risk quantifiers on risk perception, severity and decision making. Organizational
Behaviour and Human Decision Processes, 118, 72–81.
Page 27 of 31

Karelitz, T. M., & Budescu, D. V. (2004). You say “probable” and I say “likely”: Improving
interpersonal communication with verbal probability phrases. Journal of Experimental
Psychology: Applied, 10, 25–41.
Lee, J. J., & Pinker, S. (2010). Rationales for indirect speech: The theory of the strategic
speaker. Psychological Review, 117, 785–807.
Levinson, S. C. (2000). Presumptive meanings: The theory of generalized conversational

implicature. Cambridge, MA: MIT Press.
McNamara, P., Holtgraves, T., Durso, R., & Harris, E. (2010). Social cognition of indirect
speech: Evidence from Parkinson’s disease. Journal of Neurolinguistics, 23, 162–171.
Moxey, L. M., & Sanford, A. J. (1997). Choosing the right quantifier: usage in the context
of communication. In T. Givon (Ed.), Conversation: cultural, cognitive, and communicative
perspectives (pp. 207–231). Amsterdam: J. Benjamins.
Newstead, S. E. (1995). Gricean implicatures and syllogistic reasoning. Journal of Memo

ry & Language, 34, 644–664.
Noveck, I. A., Chierchia, G., Chevaux, F., Guelminger, R., & Sylvestre, E. (2002). Linguis
tic-pragmatic factors in interpreting disjunctions. Thinking and Reasoning, 8, 297–326.
Oaksford, M., & Chater, N. (2003). Conditional probability and the cognitive science of
conditional reasoning. Mind and Language, 18, 359–379.
Oaksford, M., & Chater, N. (2007). Bayesian rationality: The probabilistic approach to hu
man reasoning. Oxford: Oxford University Press.
Pighin, S., & Bonnefon, J. F. (2011). Facework and uncertain reasoning in health commu
nication. Patient Education and Counseling, 85, 169–172.
Pighin, S., Bonnefon, J. F., & Savadori, L. (2011). Overcoming number numbness in prena
tal risk communication. Prenatal Diagnosis, 31, 809–813.
Pinker, S., Nowak, M. A., & Lee, J. J. (2008). The logic of indirect speech. Proceedings of
the National Academy of Sciences, 105, 833–838.
Politzer, G. (2004). Reasoning, judgement, and pragmatics. In I. A. Noveck & D. Sperber

(Eds.), Experimental pragmatics (pp. 94–115). London: Palgrave.
Politzer, G. (2005). Uncertainty and the suppression of inferences. Thinking and Reason
ing, 11, 5–33.
Politzer, G. (2007). Reasoning with conditionals. Topoi, 26, 79–95.
Politzer, G., & Bonnefon, J. F. (2006). Two varieties of conditionals and two kinds of de
featers help reveal two fundamental types of reasoning. Mind and Language, 21, 484–
503.
Page 28 of 31

Politzer, G., & Bourmeau, G. (2002). Deductive reasoning from uncertain conditional.
British Journal of Psychology, 93, 345–381.
Politzer, G., van der Henst, J-B., Delle Luche, C., & Noveck, I. A. (2006). The interpreta
tion of classically quantified sentences: A set-theoretic approach. Cognitive Science, 30,
691–723.
Sanford, A. J., & Moxey, L. M. (2003). New perspectives on the expression of quantity.
Current Directions in Psychological Science, 12, 240–243.
Schmidt, J., & Thompson, V. (2008). “At least one” problem with “some” formal reasoning
paradigms. Memory and Cognition, 36, 217–229.
Sirota, M., & Juanchich, M. (in press). To what extent do politeness expectations shape
risk perception? Even numerical probabilities are under the spell! Acta Psychologica.
Slugoski, B., & Turnbull, W. (1988). Cruel to be kind and kind to be cruel: sarcasm, ban
ter, and social relations. Journal of Language and Social Psychology, 7, 101–121.
Spencer-Oatey, H. (1996). Reconsidering power and distance. Journal of Pragmatics, 26,

1–24.
Sperber, D., Cara, F., & Girotto, V. (1995). Relevance theory explains the selection task.
Cognition, 57, 31–95.
Sperber, D., & Wilson, D. (1995). Relevance, communication and cognition. Oxford: Black
well. (Original work published 1986)
Stenning, K., & van Lambalgen, M. (2005). Semantic interpretation as computation in

nonmonotonic logic: the real meaning of the suppression task. Cognitive Science, 29,
919–960.
Stevenson, R. J., & Over, D. E. (1995). Deduction from uncertain premises. Quarterly Jour
nal of Experimental Psychology, 48A, 613–643.
Stevenson, R. J., & Over, D. E. (2001). Reasoning from uncertain premises: Effects of ex
pertise and conversational context. Thinking and Reasoning, 7, 367–390.
Tedeschi, P. (1977). Some aspects of conditional sentence pragmatics. In R. W. Fasold &

R. W. Shuy (Eds.), Studies in language variation: Semantics, syntax, phonology, pragmat
ics, social situations, ethnographic approaches (pp. 136–151). Washington, DC: George
town University Press.
Van Canegem-Ardijns, I., & Van Belle, W. (2008). Conditionals and types of conditional
perfection. Journal of Pragmatics, 40, 349–376.
van der Auwera, J. (1986). Conditionals and speech acts. In E. C. Traugott, A. ter Meulen,
J. S. Reilly, & C. A. Ferguson (Eds.), On conditionals (pp. 197–214). Cambridge: Cam
Page 29 of 31

Villejoubert, G., Almond, L., & Alison, L. (2009). Interpreting claims in offender profiles:
The role of probability phrases, base rates and perceived dangerousness. Applied Cogni
tive Psychology, 23, 36–54.
Wallsten, T. S., Budescu, D. V., Rapoport, A., Zwick, R., & Forsyth, B. (1986). Measuring
the vague meanings of probability terms. Journal of Experimental Psychology: General,
115, 348–365.
Wallsten, T. S., Budescu, D. V., & Zwick, R. (1993). Comparing the calibration and coher
ence of numerical and verbal probability judgments. Management Science, 176–190.
Watts, R. J. (2003). Politeness. New York: Cambridge University Press.
Weber, E. U. (1994). From subjective probabilities to decision weights: The effect of asym
metric loss functions on the evaluation of uncertain outcomes and events. Psychological
Bulletin, 115, 228–242.
Weber, E. U., & Hilton, D. J. (1990). Contextual effects in the interpretation of probability
words: perceived base-rate and severity of events. Journal of Experimental Psychology:
Human Perception and Performance, 16, 781–789.
Youmans, M. (2001). Cross-cultural differences in polite epistemic modal use in American

English. Journal of Multilingual and Multicultural Development, 22, 57–73. (p. 404)
Notes:
(1) . That is not to say that we expect all individuals to be polite all the time. Impoliteness
does happen and is an interesting phenomenon in its own right (Culpeper, 2011).
(2) . As a caveat, I should note here that the most elaborated, contrived, or indirect for
mulations are not necessarily perceived as the most polite (Blum-Kulka, 1987; Holtgraves
& Yang, 1990). It might be that the very inefficiency of the most indirect formulations is
annoying for listeners: very indirect requests would therefore, quite ironically, show some
lack of respect for the listener’s time and attentional resources (Blum-Kulka, 1987). An
other possibility is that very indirect formulations come across as manipulative, a per
spective that meshes well with the speaker-as-strategist approach of indirectness (Lee &
Pinker, 2010; Pinker, Nowak, & Lee, 2008).
(3) . The findings reported in Bonnefon et al. (2009) have mostly been discussed so far in
the context of experimental pragmatics, rather than in the context of reasoning. They be
long to the growing body of research showing that the derivation of scalar inferences de
pends on multiple contextual cues whose integration has to be more complex than what
has been assumed so far by dominant pragmatic theories (Bott et al., 2012; Degen &
Tanenhaus, 2012; Hartshorne & Snedeker, 2012; Heyman, Schaeken, & Pipijn, 2012).
(4) . A tricky issue in demonstrations of the severity effect is to control for the base rate
of the different outcomes (e.g., gastric disturbance is much more prevalent than skin can
Page 30 of 31

cer). People find that “it will probably rain on London tomorrow” denotes a high probabil
ity, whereas “it will probably rain on San Francisco tomorrow” denotes a lower probabili
ty, simply because of the basic frequency of rain in the two cities. Although researchers have
to be careful to untangle the effect of base rate and severity in their demonstrations, it
now seems clear that the severity effect is independent of base rate considerations (Har
ris & Corner, 2011).
(5) . This same paper featured a replication of the politeness effect in a Slovak popula
tion. So far, and to my knowledge, politeness effects on reasoning from connectives and
quantifiers have been reported in France, England, Ireland, Dutch-speaking Belgium,
Italy, the United States—and Slovakia.
(6) . This observation has played a critical role in the ongoing revolution of reasoning re
search. The field focused on abstract deduction for decades (Evans, 2002), in spite of
growing concerns about the importance of abstract deduction for everyday reasoning.
One paradigm change later, abstract deduction has been relegated to being a peripheral
problem. Research now focuses on the way people reason from uncertain, value-laden
premises (Bonnefon, 2009; Evans, 2012; Oaksford & Chater, 2007). This paradigm change
naturally paves the way for research on politeness and reasoning because uncertain, val
ue-laden messages are precisely the kind that may call for tactful communication.
Jean-Francois Bonnefon
Jean-François Bonnefon is a French behavioral scientist working at the Centre Na

tional de la Recherche Scientifique. He is the director of the CLLE research center in
Toulouse.
Page 31 of 31

Language Variation in the Classroom

John Edwards

Subject: Psychology, Social Psychology, Educational Psychology
Schools have often made little or no provision for the linguistic heterogeneity of their
pupils, and assumptions of “best” or “correct” usage continue to complicate classroom
dynamics. Linguistic scholarship has convincingly demonstrated that, just as there are no
intrinsically “better” or “worse” languages, so there are no substandard dialects.
Nonetheless, since social perceptions and prejudices have always been able to turn lin
guistic difference into deficit, it seems clear that the life chances of many children will
improve to the extent that they expand their language repertoires—and it is entirely rea
sonable to expect that schools will assist in bilingual and bidialectal adaptations. The piv
otal psychological difficulty, then, is to provide such assistance as part of inclusive class
room practice while avoiding any condemnation of the varieties that children bring with
them from home.
Keywords: Bidialectalism, bilingualism, classrooms, language deficit, language difference, language variation
The school provides a most interesting language setting. Mingled there are official
and unofficial languages, mother tongues and foreign tongues, student slang and
street vernacular. Perhaps it is the only place where all these varieties can meet
and interact. We should take proper advantage of this wonderful diversity.
– Steiner and Ladjali, 2003, pp. 83–84; author’s translation
Most school systems must deal with linguistic heterogeneity. One may think immediately
of immigrant contexts—in the “receiving” countries of the New World, for instance—but
most states also contain substantial indigenous diversity. Furthermore, there are often
significant within-group dialect variations, most notably along the standard–nonstandard
continuum. “Standard” dialects achieve pride of place in society because they are the va
rieties spoken by those who are (currently) in positions of influence and power; they are
the forms used by educated people, in the serious media, in print, and so on. All other di
alects are necessarily and logically “nonstandard.” It is important to understand, howev
Page 1 of 22

er, that the distinction here has absolutely nothing to do with any inherent linguistic supe
riority or inferiority; rather, it rests entirely on the social status of speakers. It is, in a
sense, accidental. If the power center of the United States had emerged in New Orleans
rather than in New England, we can be sure that prestige forms of American English
would now sound rather different. In scholarly eyes, then, dialects can be either standard
or nonstandard—none, however, can be considered “substandard,” a term that has no
meaning for the linguist. Unfortunately, as we all know, it retains a great deal of meaning
in society at large.1
The language variation that exists in all but the tiniest of unstratified communities neces
sarily (p. 408) becomes a feature of interest in all educational settings, too. In his lan
guage “law,” the late Peter Nelde pointed out that “there is no contact without
conflict” (see Edwards, 2008) and, although the latter word may be a little strong in some
contexts, we could surely give the maxim universal applicability if we replaced it with
“tension.”
In the not-too-distant past—not in the past at all, indeed, in many places—classroom plu
ralism usually came up against a rather rigid and intolerant system imbued with a belief
in the wisdom and correctness of “mainstream” values. This was of course particularly
obvious at the level of linguistic variation. The main issue concerning the educational
treatment of “non-mainstream” speakers—whether immigrant or indigenous—was typical
ly the rapidity and efficiency with which they could be brought into some standardized
fold. Whatever the classroom practices of individual teachers may have been, the neces
sary adaptations here were formally seen as a responsibility of the pupils and their fami
lies. It was the school’s clear duty to maintain existing educational arrangements, ones
that reflected majority values and that were of course constructed to maintain them.
Educational systems nowadays are more likely to make at least some adaptations, and ex
pressions of intolerance founded on the assimilative belief that it is wrong to cater over
much to “foreigners” have lost their force in many contemporary settings. Looking among
the great multicultural and multilingual cities of the world, we find more than 300 lan
guages spoken in London and more than 200 in New York. In Toronto, there are also
about 200 language groups represented, and some of them are very large: more than
100,000 in each of the Chinese, Italian, and Portuguese communities, for instance. This
means, incidentally, that about half of all school children in Toronto come from homes
where neither French nor English is the parental language. And yet, in these and many
other cities across the globe, immigrants soon become citizens—their “foreign” status is
only temporary, and their impact (and, of course, their voting power) makes itself increas
ingly felt.
It would be wrong, however, to imagine that adherence to rather narrow linguistic

regimes has disappeared, even in what we like to believe are advanced or “developed” so
cieties. This is particularly evident when we turn from languages to dialects—varieties of
the same language—from the foreign varieties that arrive in the classroom to the nonstan
dard variants of the “mainstream” languages that dominate on both sides of the school
Page 2 of 22

gates. (A connection between the two is worth pointing out here. Speakers of “foreign”
languages often become speakers of nonstandard forms of the school language rather
than of more standard and, hence, more socially acceptable variants.) I will return shortly
to some of the enduring problems associated with inappropriate or inaccurate percep
tions of language.
Difficulties remain, but at least we are now more likely to think it fairer and more socially
enlightened to recognize and build on linguistic variation in the classroom, to see it as a
strength rather than a problem to be remedied. The stated aim in many contexts has to
do, then, with repertoire expansion rather than language replacement.2 This is linked
with the relatively new literature that documents the important linkages between lan
guage and group solidarity (see Edwards, 2009). Identities at risk of assimilation are now
to be protected, cross- and subcultural respect is to be integral, and some have called for
schools to “empower” minority group children. Part of this thrust involves both the recog
nition that all languages and dialects—however nonstandard or low in social status—are
vehicles of group culture and the desire to clarify this, particularly for speakers who have
a low regard for their own usage; see the “minority group reaction,” discussed later.
Making the classroom a more sensitive and more “inclusive” environment seems entirely
unexceptionable and, indeed, the evolving solution to those linguistic difficulties already
mentioned. Like many advances, however, it brings with it further problems. It may, for
instance, create a tension which Bullivant (1981) touched on when discussing the interac
tion between “civism” and “pluralism”—that is, between the transmission of general
knowledge and core skills and the more or less active promotion of cultural values, in
cluding linguistic variants.3 Education is not always the ally of enduring linguistic and
cultural diversity: it often has intrusive qualities, championing literacy over orality (for
example) and, more generally, presenting methods and points of view that are not always
those of “smaller” cultures. There is, in short, a sense of cultural and linguistic bullying,
which those sensitive to multicultural variation find distressing. Indeed, it is not difficult
to sympathize with (some) laments about intrusive and culturally insensitive educational
paradigms.
Two complications immediately suggest themselves. The first is that members of “small
er” groups often want their children to gain access to a more (p. 409) “mainstream” soci
ety. They may therefore be suspicious of what well-meaning others would see as sensitive
responses to cultures that have traditionally been of subaltern status. As one English
commentator put it a generation ago, “Black parents don’t want Black studies or multicul
tural education for their children—that is for white children; Black pupils need to be good
at science, history, geography—at what society thinks of as things of worth” (see Tomlin
son, 1984, p. 149).4 Black parents and educators have voiced similar objections in the
United States, most notably in rejecting the use of Black English at school; see Edwards
(2010) on the “Ebonics” controversy, for example. One can, of course, argue that these
are unnecessarily restrictive views, perhaps the result of internalizing longstanding and
destructive social prejudices. One of the most poignant aspects of the social evaluation of
language is the widely reported tendency for nonstandard speakers to accept and agree
Page 3 of 22

with unfavorable stereotypes of their speech styles; this is what Lambert and his col
leagues—describing French–English relationships in Quebec more than a generation ago
—termed the “minority group reaction” (Lambert, Hodgson, Gardner, & Fillenbaum,
1960). It is, of course, merely the linguistic manifestation of a much broader interplay be
tween social dominance and subordination.
It ought to be possible for an education that leads to broader social access and mobility to
coincide with—rather than ride roughshod over—a multicultural sensitivity that respects
and enhances more local values, practices, and modes of expression. This may lead, how
ever, to the second complication. Formal education necessarily involves going beyond
what is culturally immediate or local, opening windows to other values and cultures, both
past and present. In an unequal world—one whose disparities create risks for “smaller”
or less prestigious languages and dialects, in fact—education will perforce become yet an
other evidence of those disparities. Those concerned with gaining places in the media for
minority languages have learned that they are double-edged swords: although it is clear
that access to them is important, they also facilitate the transmission of those larger influ
ences on decline. Similar risks may occur in educational settings. Thus, there are many
historical examples of authorities promoting education through the medium of local ver
naculars—not with the expectation or wish that they may be sustained in this way but,
rather, in the belief (quite often borne out) that the practice will actually hasten language
shift. Suitably educated people would surely come to realize the limitations of their own
language and the great advantages of a “bigger” one (see Edwards, 1995; 2009). About
200 years ago, Daniel Dewar wrote the following:
The cultivation of either the Irish or the Gaelic is the most effectual, as well as the
most expeditious plan that can be adopted for their extinction. Make any people
intelligent and rational, and they will gradually lose their prejudices; many of
them will acquire a taste for general knowledge, and they will seek for it in the
general tongue of the empire.
(Dewar, 1812, p. 97)
No one can doubt, surely, that the practice of using a native variety to accelerate the shift
toward another language has continued since Dewar’s time and in many contexts beyond
Celtic ones. Too few people, both then and now, have considered a policy that supports
both languages—a shame, since contact between the language of home and hearth, on
the one hand, and a “bigger” variety for use beyond the garden gate, on the other, seems
an ideal context for the emergence and maintenance of bilingualism. Of course, linguistic
stability here is rather rare, and one language (no prizes for guessing which!) tends to
steadily encroach on the other. Again, this is a very common global phenomenon: one
need only think of the typical language trajectory followed by most immigrants to the “re
ceiving” countries of the New World, for example.
Page 4 of 22

Preparing for Language Diversity at School

Given what I have already touched on, it is abundantly clear that the responses of teach
ers to linguistic heterogeneity are of the greatest importance. It follows that their accu
rate understanding of it is crucial. Whether it is a matter of accepting or rejecting non
standard dialects or foreign languages, of adapting classrooms to language diversity or
attempting to maintain a strict monolingual regimen, of seeing school as a contributor to
social change or as a supporter of some “mainstream” status quo, of arguing the merits of
“transitional” versus “maintenance” programs of bilingual education, of embracing multi
culturalism or recoiling from it—in all these matters, the knowledge, sensitivities, and at
titudes of teachers are of no small importance.
Nonetheless, as I have recently pointed out elsewhere (Edwards, 2010), teachers general
ly remain woefully ill-prepared to meet linguistic and cultural variation in the classroom.
This has always been the case, in fact. Many jurisdictions essentially denied (p. 410) a di
versity that was always there: thus, for example, when I began research work in Nova
Scotia, it was immediately clear that prevailing perspectives made little room for long
standing groups of low social status. In some schools, there were sizeable groups of
African-Canadian pupils, descendants of those who came to Canada on the “underground
railroad” during the American Civil War or of those who had been given land grants in re
turn for service in the British army. In other schools, there were Acadian children of
French-speaking background. An inability or an unwillingness to see such groups as any
thing other than minor aberrations in an English/Celtic mainstream had the predictable
consequences; see Edwards and McKinnon (1976).
Poor teacher preparation is becoming even more worrisome than it once was. This is sim
ply because linguistic heterogeneity is increasingly present in all sorts of settings. Even
schools in “traditional” and rural areas whose populations were historically both local and
stable are now increasingly confronted with children from many different backgrounds.
The geographical spread of Spanish speakers throughout the United States—far beyond
the initially adopted settlement areas of Florida and the Southwest—is a case in point.
Another is the widening distribution and permanence of “guest workers” in western Eu
rope.
However, even if there were no great likelihood of teachers encountering social or lin
guistic diversity in their classrooms (increasingly implausible as this would seem), I think
that a good case could still be made for giving much more attention to such diversity. As I
implied above, all education worthy of the name is multicultural; a logical implication is
that any heightening of teachers’ cross-cultural and cross-subcultural sensitivities must
be a good thing. Andersson and Trudgill (1990, p. 179) make the simple but powerful
point that “teachers who are prepared to take an open-minded, unprejudiced attitude to
wards the varieties of language spoken by their pupils will be the ones who also succeed
best in fostering and developing children’s linguistic interests and abilities.” We could ex
Page 5 of 22

pand this, in fact, and say that such teachers are likely to succeed best in developing all of
their pupils’ potentials.
In this connection—going beyond, that is, classrooms containing substantial language

and/or dialect variation—teachers should be aware of a wider world in which English is
increasingly dominant, but where linguistic globalization often involves the emergence
and stability of local “Englishes.” Educational sensitivity here could “replace absolutist
conceptions of what is proper and correct in language with greater flexibility and princi
pled pragmatism regarding norms and models” (Ferguson, 2006, p. 174).
It seems obvious enough that the most up-to-date information bearing on linguistic and
cultural issues should be presented to teachers—more pointedly, to teachers-in-training—
as a matter of course. There is strong and convincing evidence, for example, that Ameri
can Black English is just as valid as any other English variant, that it is just as rule-gov
erned, that its patterns of pronunciation and emphasis are just as regular, and that it
serves the cognitive needs of its speakers just as well as any other form of speech does.
Black English makes an excellent “test case” for dialect validity because it had for so long
been seen as a particularly deformed and unsystematic approximation to “good” English
and because its speakers were victims of a prejudice that went well beyond language
alone.
Powerful evidence here has been provided, most notably in the now-classic work of
William Labov and his associates (Labov, 1976; 1994). An example is found in his demon
stration of the regularity of use of the copula verb (the verb “to be”). Where standard
English allows contraction of this verb (so that “he is going to work” can become “he’s
going to work”), Black English allows its deletion (“he going to work”—or, to cite the most
likely verb ending in this context, “he goin’ to work”). This is a regular and not a random
feature. The point is cemented when we note that, whereas standard English doesn’t
allow contraction (“he was going to work” cannot be simplified to “he’s going to work”),
Black English doesn’t permit deletion. The reason is clear and has to do with avoiding
confusion and ambiguity where past and present action is involved. But permitting dele
tion in one variant, where contraction is the norm in the other, is simply to employ a dif
ferent rule. The reasonable extrapolation from the Black English “test case” reinforces
the point I made earlier: no dialect is substandard.
Information bearing on the inherent validity of all dialects is not rocket science. It can be
presented to and understood by anyone who has an open mind. I know, because I have
made presentations to teachers and teachers-in-training, as well as to hundreds of senior
students in language seminars. So much the more tragic, then, that many are still left to
labor under stereotyped, inaccurate, and potentially harmful illusions. Many of the chil
dren who come to school with nonstandard dialects have very real burdens to bear—bur
dens of poverty, of immobility, (p. 411) of prejudice and discrimination. It is quite unneces
sary for these difficulties to be exacerbated further because children’s linguistic capabili
ties (and, often by extension, their cognitive skills) are misunderstood or denigrated. The
Page 6 of 22

consequences, in terms of impediments to learning, early curtailment of formal educa

tion, and reduced chances in the world beyond the school gates, can be enormous.
Michael Halliday (1968, p. 165) once succinctly observed that “A speaker who is made
ashamed of his own language habits suffers a basic injury as a human being; to make any
one, especially a child, feel so ashamed is as indefensible as to make him feel ashamed of
the colour of his skin.” If some of those “language habits’ were in fact deficient, then
there would be a case for attempting to remedy them. But the sort of evidence just
touched on has established that there are no “substandard” varieties, that all are “cor
rect,” and that it is social history, practice, and convention—rather than any intrinsic lin
guistic quality or lack of it—that have elevated some varieties and downgraded others.
Enduring Problems with Teacher Education

Teachers are not malicious, of course, in their attempts to deal with what they so often
see as language “problems.” As well, they—and at least some members of society at large
—are probably more accepting of language variants than once was the case. Nonetheless,
although teachers view language differences as emerging from varied home backgrounds
(true), they also often continue to see them as necessitating some sort of compensatory
action (dubious). In my Nova Scotia research, teachers frequently commented on their
pupils’ “poor” English and their own obligation to replace it with the “correct” version.
Some felt, indeed, that the gap between the two was wide enough to make their work tan
tamount to teaching a new language altogether. These sentiments were most forcefully
expressed when Black and Acadian children were under discussion. Articulating a fairly
general opinion, one teacher told me that “Blacks have a slang language all their own.
They will not use proper English when opportunity arises” (Edwards, 2010, p. 142).
These findings are unfortunately consistent with those reported in other settings. In her
discussion of the educational situation of Puerto Rican children in America, Walsh (1991,
p. 107) reported the following opinion:
These poor kids come to school speaking a hodge podge. They are all mixed up
and don’t know any language well. As a result, they can’t even think clearly. That’s
why they don’t learn. It’s our job to teach them language—to make up for their de
ficiency. And, since their parents don’t really know any language either, why
should we waste time on Spanish? It is “good” English which has to be the focus.
And, in a European context, Hélot and Young (2005, pp. 242–244) have shown that, since
the French educational system is still largely “envisaged from a monolingual point of
view…it is difficult for most teachers to view the different languages and cultural back
grounds of their pupils as other than problematic.” These are clear expressions of what
has been termed the “language deficit” perspective, one that remains regrettably com
mon despite the overwhelming scholarly evidence against it.
Page 7 of 22

Furthermore, we cannot assume that apparently appropriate teacher-training initiatives

always produce the desired results. Important work by McDiarmid (1992) and McDiarmid
and Price (1990) revealed that multicultural training programs in the United States ap
parently made little impact on teachers-to-be; in some cases, they seem to have actually
reinforced minority group stereotyping. Later studies confirm the continuing difficulties.
Zientek (2007) discusses the inadequate information provided about cultural and linguis
tic diversity in the classroom, and Tenenbaum and Ruck (2007) outline some teacher be
liefs. Asian pupils were expected to do best at school, followed by “European-American”
children, then Hispanics and, finally, black children. Teachers were also found to be more
encouraging in their interactions with those of whom they expected the most. The dan
gers of such stereotypic preconceptions are obvious, contributing as they so easily can to
self-fulfilling prophecies.
Even in the more focused area of bilingual education, there is evidence that teacher train
ing has often been less than adequate. Grinberg and Saavedra (2000) cite some represen
tative comments that demonstrate how university courses leading to teacher certification
are often of “little relevance.” One trainee noted that, “in my preparation as a bilingual
educator I was not prepared for the reality in the school” (p. 433). Another observation:
Living here in the heart of New Mexico, we have very fertile grounds to develop
strong, effective bilingual programs…[but] the university does not have a good
program to prepare teachers…there is no rigor…the content of the classes is mini
mal, at a low level. (p. 434)
A few years ago, Burtonwood and Bruce (1999) reminded us that the Swann Re
(p. 412)
port (1985), which had argued for greater attention to linguistic and cultural pluralism in
British schools, was soon overtaken by educational reform legislation that stressed the
importance of a national curriculum—and thus, in many eyes, made less room for diversi
ty in the classroom. In like manner, in the United States, President Bush’s No Child Left
Behind (NCLB) initiative of 2001 replaced bilingual education legislation and buttressed
the emphasis on English-language schooling. (Indeed, the Office of Bilingual Education
and Minority Language Affairs changed its name, in 2002, to the Office of English Lan
guage Acquisition.) Rolling back bilingual education initiatives, NCLB also undercuts sec
ond-language acquisition for native English speakers, most of whose foreign-language in
struction (such as it is) will not begin until secondary school. In a world in which—for all
the power of English—multilingualism remains a powerful asset, this most recent empha
sis has an isolating tendency for America and Americans. There is a very strong reliance
on standardized assessment devices—Crawford (2004) notes that NCLB might well stand
for “No Child Left Untested’—and the results of testing are to be made public by group
(black children, Hispanic children, English as second language [ESL] pupils, and so on).
The program rewards narrow “teaching for the tests” in a context in which the lion’s
share of time and attention is given to reading and arithmetic.
For present purposes, we can focus on NCLB’s impact on minority group children (see
Valenzuela, Prieto, & Hamilton, 2007). The increased emphasis on English proficiency,
Page 8 of 22

coupled with the intensive testing ethos of the new arrangements, certainly does mean
more difficulties for children who are not native speakers of English (see also Fine, Jaffe-
Walter, Pedraza, Futch, & Stoudt, 2007). As Evans and Hornberger (2005) have written,
the incentive for schools to do well on the annual testing exercises could—and should—
theoretically mean that more attention will be paid to linguistic diversity and its educa
tional ramifications. But such a theoretical consequence is often swamped by the practi
cal demands for increasing time to be given to the two key themes: literacy and mathe
matics. There are several good treatments of the matter—see Edwards (2010) for an
overview; see also the specific treatments of ESL pupils and native non-English-speaking
children whose mother tongues are at risk by Abedi (2004) and Grenoble and Whaley
(2006), respectively.
Contemporary Currents
When discussing the “culture” of the foreign-language classroom, the best of the recent
books and articles give some attention to the language–dialect distinction and, more par
ticularly, to the appropriate understanding of the validity of nonstandard dialect varieties.
A good example is provided by Siegel (2007, p. 76), whose discussion shows just how lit
tle ground has been gained in this area. Describing creoles and nonstandard dialects in
education, he points out that, despite several decades of sociolinguistic insight, accurate
depictions of such varieties “have not filtered down to many educators and administra
tors, or to the general public”; see also Zéphir (1997; 1999), who draws explicit parallels
between the educational reception of creole and that of Black English. (A creole usually
emerges from a pidgin—a variety, often arising in contexts of trade, that involves only the
limited vocabulary and grammar needed for basic cross-cultural transactions—when chil
dren born in pidgin-speaking communities begin to develop [to “creolize”] their linguistic
inheritance. Nobody’s mother tongue thus becomes somebody’s mother tongue.)
At twenty-year intervals (1943, 1964, and 1984; also Preston, 2003), the American Dialect
Society published works outlining “needed research” in dialect studies. In the latest of
these, several authors write about the important linguistic demonstrations of the validity
of Black English and other nonstandard dialects and about the useful developments in
language-attitude research, “perceptual dialectology,” and “folk linguistics.” They also ac
knowledge, however, that unenlightened stereotypes continue their baleful course. It is a
tragedy of the first magnitude that almost half a century’s worth of linguistic, sociologi
cal, and psychological research has yet to be presented appropriately to those in the front
lines of language and dialect contact; see Edwards (2010) for fuller details. Kautzsch
(2006) points to the necessity for more open-minded and well-informed teachers, and for
educational systems committed to “difference” rather than “deficit” stances on cultural
and dialect variations. Godley et al. (2007, p. 124) reveal the persistence of assessments
that equate “standard” with “correct,” and Black English with “incorrect, ungrammatical
English.” Bluntest and most general of all, perhaps, are the comments made by Cochran-
Smith (1995, p. 493): the American educational system is “dysfunctional for large num
bers of children who are not part of the racial and language mainstream.” We would, of
Page 9 of 22

course, be naïve indeed to imagine that (p. 413) problems here are unique to the United
States. The relatively large literature in this geographical context says much more about
research opportunities there than it does about the scope of the issues in other settings.
There is some evidence to suggest that younger teachers are initially more inclined to ac
cept a “different-but-not-deficient” argument about nonstandard varieties. However, as
Brouwer and Korthagen (2005, p. 153) point out, “occupational socialization in schools is
a known factor counteracting attempts at educating innovative teachers.” This is simply a
specific example of a general phenomenon: we are all very susceptible to the cognitive
and emotional tone of our surroundings, and the issues on which we are least well in
formed are those most open to influence. A corollary is that attempts to dispel misinfor
mation may act as inoculations against later vulnerability; thus, providing new teachers
with an accurate understanding of the language competence of their pupils may prove of
the greatest value. Brouwer and Korthagen go on to write that the “effects of teacher ed
ucation on the actual practices of teachers are generally meager” (p. 153). The implica
tions are that the provision of linguistic and psychological fact should be done early,
preferably before teachers first enter the classroom and that it must be done well in or
der to have any chance of becoming that “inoculation.” Fortunately (as already noted), it
is not especially difficult to present relevant findings in a succinct, appropriate, and
“user-friendly” manner.
Adding his voice to the complaint that “formal educational policies for the treatment of
non-standard varieties in schools are conspicuous by their absence in most educational
systems,” Corson (2001, p. 68) also points to the benefits of having more nonstandard di
alect speakers and more minority group members as teachers. Quiocho and Rios (2000)
note that there are fewer such teachers in America, Britain, and elsewhere than we might
like to see. The argument here, of course, is simply that these teachers will be more likely
to demonstrate linguistic and cultural sensitivity in the classroom. It is important to point
out, however, that “ethnic” or nonstandard-speaking individuals who become teachers
may, by that fact alone, be or become atypical of the group. Relatedly, the process of
teacher training may tend to accelerate their middle-class socialization. Grinberg and
Saavedra (2000, p. 436) are blunter, noting that once Latino and other minority group
members “enter the system, internal processes of colonization take over.” It is by no
means clear, then, that increasing the number of teachers from particular cultural or so
cial class groups will lead to a commensurate increase in multicultural sensitivities in the
classroom.
And there is a further matter that seems never to be mentioned. When Quiocho, Rios, and
many other like-minded scholars call for educators and institutions to encourage more mi
nority group students to take up the profession of teaching, they may be encouraging a
sort of self-imposed restriction that does not apply to members of the social “main
stream.” Over the years, I have had a number of Canadian native students in my universi
ty seminars, many of whom told me that they intended to become teachers. As I came to
know them a little better, it was apparent that—as the educated minority within a socioe
conomically depressed minority—they felt a duty to “give back” to their community. Such
Page 10 of 22

altruistic motives are, of course, highly commendable, but I came to realize that at least
some of my students were charting their career course out of a sense of obligation, rather
than on the basis of personal preference. And it struck me that this was, in some sense,
yet another burden that they carried, yet another limitation that their white counterparts
rarely had to consider at all.
Assessment and Remediation

Educational and social disadvantage underpinned by (among other things) inaccurate lin
guistic opinion and prejudice remains a great problem. Popular speech and language atti
tudes continue to hold certain dialects and accents as better or worse than others. There
is no real difficulty, of course, in the possession of personal preferences, in the fact that I
think Italian to be the most beautiful and mellifluous of languages, whereas you find the
greatest music and poetry in Scottish Gaelic. But there are dangers when we imagine
that we are arguing about substantive linguistic issues, about the inherent properties of
one variety or another, or—worse still—about the cognitive attributes thought to accom
pany certain language forms. In all these ways, the so-called deficit theory of nonstan
dard dialects and, indeed, of certain foreign languages continues to hold wide sway.
Indeed, it is not so very long ago that some children were considered victims of linguistic
and cultural “deprivation”—even though a moment’s thought reveals that the term must
bizarrely imply groups of individuals who are deprived of someone else’s social attribut
es! After all, it makes no sense to imagine a community deprived of itself. (Unless,
(p. 414) that is—and this must surely have been the unspoken assumption in some quar
ters—we are to think that some groups of people have no culture at all or no language
worthy of the name.) People may not be able to articulate such “deficit” conceptions, may
indeed be quite unaware that they have such ideas, but it is the easiest thing in the world
to demonstrate their influence: just ask people about “correct” and “incorrect” language
(see Lippi-Green, 1997; Trudgill, 1975). Teachers are hardly immune from the sway of
deficit theories, but their position vis-à-vis children—from an early age and over a long
and formatively important period—makes classroom atmospheres and postures of central
importance.
“Deficit” points of view may have been thoroughly discredited by modern scholarship, but
there is no doubt that some individuals and some groups remain at a disadvantage in so
cieties where widely held attitudes regularly translate difference into deficiency—which
is to say, of course, in all stratified societies. As a term, then, “disadvantage” is entirely
accurate. It has been necessary, however, to defend its use from time to time, largely be
cause it has suffered from guilt by association. It ought not, however, to be seen as a eu
phemism for “deficit” or “deprivation,” terms that this chapter—and my previous work in
the area—has, I hope, helped to see off; see the discussion in Edwards (1981). No, “disad
vantage” remains a tersely accurate description—when, of course, it is properly under
stood as a reflection of group difference that is, itself, a product of social inequality, dis
continuity, and comparison. What could be more obvious than that the knowledge, atti
Page 11 of 22

tudes, inclinations, and usages that emerge in one setting—and that arise naturally, rea
sonably, and, as we might say, in a “socially logical” manner—will often prove inappropri
ate and maladaptive in another? Thus, children who are going to school, who are leaving
their home environments for the first important time, may well be at a disadvantage; this
need have absolutely nothing to do with any intrinsic or substantive cognitive-linguistic
problems.
Conceptions of linguistic deprivation and deficit are profoundly misguided. They ignore
the fact, for instance, that no human grouping has ever been found with a language inad
equate for its own social purposes. But we can easily see, again, that an acceptance of
this point of view hardly rules out judgments of disadvantage when members of a particu
lar community move beyond their own borders: the variety that serves them perfectly
well at home may not best equip them in altered circumstances. For this reason, it is en
tirely possible to wholeheartedly reject deficit perspectives while simultaneously endors
ing programs of language repertoire expansion. To remake an important point: these
need not imply for a moment any intrinsic defect in the maternal variety; they simply ac
knowledge that different settings may call for different responses. No one will accuse a
Francophone executive of having an inherently flawed mother tongue just because she
learns English when posted to Detroit. No one should consider American Black English as
a flawed dialect just because it is not the sort of English that she will acquire when she
gets there.
This brings us squarely to questions of the appropriate treatment of classroom language

variation. On the one hand, we realize that there are no inferior or inadequate forms; on
the other, we must accept that—in all socially stratified societies—power, prestige, and
social dominance very typically convert difference into deficit and perceive nonstandard
language as substandard language. So, while we await the linguistic millennium, we re
main well aware that “only before God and the linguist are all languages equal. Everyone
knows that you can go further with some languages than you can with others” (Mackey,
1978, p. 7).
Accepting that scholarly enlightenment will not penetrate all corners of society for some
little time yet, and wishing to assist those whose languages and dialects mark them out in
unfavorable ways, should we then move where we can, however regretfully, to replace
stigmatized varieties with more socially acceptable ones? This seems an unreasonable
course. Not only would it require us to grit our teeth and ignore the weight of relevant
academic research, but it would also interfere with the important connections between
language and identity. All varieties, however negatively marked they may be outside the
group, and however disadvantageous to the mobility of their speakers, are nonetheless
vehicles of solidarity and belonging. Like all dialects, “disadvantaged” forms anchor
group members in their collective. The attachment and affection felt for particular vari
eties may in fact be greater when they and their users are low in social status. After all, a
sense of—and a desire for—the security that group membership affords is normally
heightened in circumstances of external threat or discomfort. This goes some way toward
Page 12 of 22

answering the question posed by Ryan (1979) in the title of her paper: “Why do low-pres
tige varieties persist?”
At a practical level, attempts to abandon nonprestigious varieties are fraught with dan
ger. If we (p. 415) remember that speakers of these varieties are often more materially dis
advantaged, too, it may be that linguistic replacement alone will have little if any effect
on desired social or occupational mobility. The most obvious examples here have to do
with group differences—like skin color—that are indelible. In fact, research has shown
that black speakers with “white” speech may be perceived more negatively. Unfavorable
reactions here sometimes come from within the black community itself—see Ogbu (1999)
and Smitherman (2006) on “talking white”—and sometimes from without, as Giles and
Bourhis (1976) first reported; see also Edwards (2009; 2010). Readers may like to consid
er just why a perceived mismatch between color and speech patterns should elicit ad
verse reactions.
In addition to those speakers whose fixed characteristics may make language alteration
seem a rather pointless exercise, others may also have qualms about shifting. Moving in
to new social roles means a departure from the known features of group identity, and lin
guistic mobility may therefore prove risky. There are two chief dangers here. The first is
that an imperfect or inadequate assumption of a new linguistic persona may lead to social
marginalization, a sense of not being a full member of any social group. The second is
that even a relatively seamless movement may lead to deeper divergences than are
wished for. The Mexican American who abandons Spanish for the socioeconomic rewards
of English may be seen by friends as a vendido (a “sell-out”), and his or her French Cana
dian counterpart as a vendu.
Considering all this, and bringing the discussion back to the level of the classroom, we
can see that a logical course of action might involve educational policies of linguistic
repertoire expansion, rather than replacement. This would reflect the research bearing
on the validity of all languages and dialects and the importance of not severing important
linkages between maternal varieties and social identities. At the same time, it would re
flect the realities of life in stratified societies and aim to equip children to better succeed
in them; it would aim in effect to produce bidialectal or bilingual children. It can immedi
ately be seen that a policy of addition or expansion, while both reasonable and plausible,
does nonetheless run the risk of actually promoting replacement. It is a matter of some
sensitivity, after all, to suggest to children that the language they bring to school is per
fectly valid while in the same breath indicating that it will not serve them well in desir
able social settings. They could be forgiven for thinking this to be a hypocritical stance—
they would be wrong, but considerable damage might be done before they came to real
ize the subtlety of the points at issue.
In such sensitive contexts, indirect example is more appropriate than direct and formal
instruction; the relevant findings are summarized in Edwards (2010). Beyond what I have
already touched on here, this is also supported by research from a number of contexts
showing that children who speak nonstandard dialects generally comprehend more stan
Page 13 of 22

dard ones quite well; given the most common social circumstances, they are reasonably
bidialectal—at least at a passive or receptive level—from an early age. This is a basis from
which a more active use of standard variants can emerge, according to the wishes and
needs of the speakers themselves. Lest this seem rather too laissez-faire, we should bear
in mind the difficulties already cited regarding direct or formal teacher intervention, how
ever well-intended. Teachers should, of course, be sympathetic and accepting where non
standard varieties are concerned but present themselves as the standard-speaking mod
els they most typically are—without, of course, calling direct attention to that presenta
tion. Where reading and writing are concerned, it is worth remembering that all children
have to master this formal encoding of oral language, and that—other things (like appro
priate classroom sensibilities) being equal—there is no reason to think that nonstandard-
dialect speakers will have more difficulty than others. (Indeed, studies dating to the time
of Basil Bernstein and his “restricted” and “elaborated” codes—see Edwards [1989]—
showed that, in formal writing exercises, the grammatical usage of standard- and non
standard-speaking schoolchildren was not substantially different.)
The single most important factor here is the creation by teachers of classroom atmos
pheres in which children are not made self-conscious about their language. From a basis
of acceptance, the expansion of linguistic capabilities may be expected—and achieved. At
tempts at instruction that, however subtly, suggest that the maternal varieties first
brought to school are less than adequate are likely to foster linguistic insecurity: not a
good position from which to proceed.
Classrooms and Beyond: Language Power and

Preservation
The same sensitivities that are desirable in classrooms containing dialectal heterogeneity
—or, (p. 416) indeed, where there is nonstandard-dialect homogeneity among the children
—are equally to be encouraged where differences exist at the level of language. Even
though language variations are obviously much greater than those at the dialect level, the
issues they raise are rather more straightforward. Dialects, after all, are mutually intelli
gible varieties of the same language, and speakers of nonstandard forms—even where
these are socially denigrated—can therefore expect to be understood by others. For
speakers of mutually unintelligible varieties (i.e., different languages), however, difficul
ties of classroom communication are immediate and obvious.5
Just as it has often been thought right to work to eradicate “incorrect” dialects and re
place them with “proper” standard ones, so schools have often considered—implicitly or
directly—that the sooner foreign-language-speaking pupils engage with language shift,
the better. At the same time, however, many jurisdictions have come to understand that
the expansion of linguistic repertoires is an important facet of the educational process.
Many obvious tensions arise here. For example, the same school that values and teaches
French or Spanish may do little or nothing to recognize, adapt to, or build on the Hausa,
Turkish, and Bulgarian that come in the door with new immigrant pupils. Mackey’s obser
Page 14 of 22

vation notwithstanding, some languages are more equal than others: one implication is
that it will be easier to teach German in Nijmegen than in Nebraska—easier, and more
immediately recognized as a useful and rational thing to do. Social realities and attitudes,
rather than any intrinsic linguistic features, dictate the teaching of Spanish and the ignor
ing of Bulgarian. In a more enlightened climate, however, we can imagine that—given a
sufficiency of student numbers—it might be a good thing to teach the latter as part of a
bilingual education initiative, for instance. Acceptance of current realities, of course,
might also suggest a limited utility for such a move. Another sort of argument has some
times emerged here, and it rests on the notion that improved recognition and treatment
of hitherto unregarded (at least locally) “lesser-used languages” will—apart from all else
—somehow “empower” their speakers.
Would making room in the classroom for Hausa or Bulgarian materially affect immigrant
children from West Africa and eastern Europe? The obvious answer is yes—if it were
practicable, it would at the least ease their passage into and through school. Contrary to
the popular equation, I do feel obliged to point out, however, that language is not power—
not inevitably, at any rate—and that exercises in linguistic “empowerment” are essentially
compensatory in nature. This is not to deny that there may be further positive benefits (in
terms of children’s “self-esteem,” perhaps) to be gained by attending more closely to ma
ternal varieties, but it is generally a mistake—and sometimes an unrealistic elevation of
expectations—to think of schools as agents of power through language per se (in the
sense under discussion here).6
Teaching about or through foreign languages has always had an honored and important
place in the classroom, and it is unfortunate that their presence has diminished in many
settings. One of the most lamentable aspects of the rise and scope of “big” languages is
that their penetration and their power push other varieties into the shadows. With global
English being the main contemporary case in point, we observe—in the context of this
chapter but not, obviously enough, in broader global context—that the main victims are
anglophones themselves. In a world made increasingly safe for their language, the incen
tive and (one regrets to add) the need for them to expand their repertoire lessen.
A related point, however, arises when we turn our particular attention to the matter of mi
nority or “at-risk” languages at school. There is a large and growing literature on endan
gered varieties and on the use of schools to revitalize them; the collections edited by Flo
res Farfán and Ramallo (2010), and Hornberger (2011) provide excellent recent
overviews. Schools have always been deeply embedded in society at large, of course, and
educational decisions about what (and what not) to teach or encourage both reflect and
help sustain cultural norms and attitudes. Difficulties, however, have arisen when schools
are asked to act in ways that are not strongly reflected in the social currents outside their
gates. When schools have been charged with the maintenance or revival of minority lan
guages, the results have been rather tepid—at least in terms of any broad revernacular
ization.
Page 15 of 22

With independence, Ireland enshrined Irish as the national language and the first of the
two official varieties. The schools were given the lion’s share of the effort to maintain and
revive Irish, and, although they have succeeded in giving everyone a thin wash of Irish,
they have not brought the language back into general use—results that are unsurprising
to any but the most myopic nationalist. “Heritage-language” schools in Canada and bilin
gual schools in the United States have also been unable to significantly affect the public
face of (p. 417) language. (The ever-increasing visibility of Spanish in America—which is
dramatically altering longstanding linguistic patterns—owes little or nothing to educa
tional postures.) Where some modest success can be attributed to school programs, we
typically find strong extra-academic support—with some aboriginal-language innovations
in Canada, for example, or with the Māori “language nests” in New Zealand—without
which the educational interventions would count for very little.
Still, such language intervention duties continue to be assigned to schools. In a way, this
is entirely understandable, for reasons already touched on here, reasons having to do
with the powerful and enduring influences that classrooms exert on their students. In an
other way, however, asking schools to swim against strong social and linguistic tides is
clearly asking too much. This is particularly so when we remember that language dynam
ics are often symptoms of broader cultural contact. Thus, for example, the social forces
that placed Irish in a subordinate position to English in Ireland are not reversible by edu
cational intervention; indeed, it is doubtful that any practical, democratic, and generally
desirable set of circumstances could effect the return of Irish to its former position.7 The
same may be said for most contact settings involving minority languages, either immi
grant or indigenous.
Fuller and more honest recognition of the social embeddedness of language would surely
suggest that radical social steps are required to halt or reverse unwanted linguistic shift.
And it is noteworthy that those who are most animated and activist in this regard do not
endorse those radical steps. Rather, the desire is for only some aspects of cultural and lin
guistic life to be altered, rectified, or brought out from under oppressive thumbs. One
hardly needs to point out that such selectivity is very difficult to accomplish and that the
hope for it—whether articulated or not—is precisely what hamstrings closely focused ap
proaches, exactly what has condemned many well-meaning school-based efforts to either
outright failure or, at best, to triumphs that rarely leave the school precincts. Further
comments here may be found in Edwards (2012), a piece in which I concluded by pointing
out that “educational programmes of language revitalisation are the lamp-posts in whose
light we hope to recover things that were lost elsewhere” (p. 203).
We can end here by returning very briefly to bilingual education. Not only has it had the
function of better accommodating children who arrive at school speaking a “foreign” lan
guage, but it has also been pressed into service in the revitalization capacity just dis
cussed. It is in this latter and more politically sensitive regard that, as noted, it has come
under pressure. The manifestations of that pressure in the United States provide clear ev
idence of the degree to which education is deeply socially embedded. Fears of the in
creasing strength of Spanish, in a country where Hispanic Americans now outnumber
Page 16 of 22

black Americans, have naturally if very regrettably put pressure on schools to become
stronger agents of assimilation—this despite a great deal of evidence showing that pro
viding education through Spanish does not impede Hispanic movement into the cultural
mainstream. It may, in fact, expedite it. The point is quite lost, of course, on simpler
minds obsessed by some looming linguistic and culture balkanization—or, worse (al
though even more unlikely), an actual replacement of English in that cultural main
stream.
Conclusion
Although schools cannot work linguistic and social miracles, and although their efforts
are severely compromised if they are not reinforced in the larger society, a simple and
highly relevant fact remains: entering that earliest classroom represents children’s first
substantial break from the familiarity of home life, and it ushers in a period that—for
most pupils—will extend over at least a dozen years or so. It is inconceivable that lan
guage matters, whether directly or indirectly touched on over that span of time, will not
prove important and formative. Sensitive and well-informed educational practice may not
be able to vanquish powerful social bigotry and intolerance nor bring about full linguistic
enlightenment, but it can certainly be useful—if only by not exacerbating prejudice and
ignorance. As regards favorable alterations in social assessments of linguistic variation,
we can agree with Trudgill (1975, p. 70): “school is a good place to start.”
References
Abedi, J. (2004). The No Child Left Behind Act and English language learners: Assessment
and accountability issues. Educational Researcher, 33, 4–14.
American Dialect Society (ADS). (1943). Needed research in American English. Chicago:
A.D.S.
American Dialect Society (ADS). (1964). Needed research in American English (1963).
Tuscaloosa: University of Alabama Press.
American Dialect Society (ADS). (1984). Needed research in American English (1983).
Tuscaloosa: University of Alabama Press.
Andersson, L., & Trudgill, P. (1990). Bad language. Oxford: Blackwell.
Brouwer, N., & Korthagen, F. (2005). Can teacher education make a difference? American
Educational Research Journal, 42, 153–224.
Bullivant, B. (1981). The pluralist dilemma in education. Sydney: Allen & Unwin.
Burtonwood, N., & Bruce, R. (1999). Cultural sensitivity: American and British teacher
trainees compared—a response to Thomas Deering. Educational Research, 41, 94–99.
Page 17 of 22

Cochran-Smith, M. (1995). Color blindness and basket making are not the answers: Con
fronting the dilemmas of race, culture, and language diversity in teacher education.
American Educational Research Journal, 32, 493–522.
Corson, D. (2001). Language diversity and education. Mahwah, NJ: Lawrence Erlbaum.
Crawford, J. (2004). Educating English learners: Language diversity in the classroom. Los
Angeles: Bilingual Education Services. (The 5th edition of Bilingual education: history,
politics, theory and practice)
Dewar, D. (1812). Observations on the character, customs and superstitions of the Irish.
London: Gale & Curtis.
Edwards, J. (1981). Disadvantage: Guilt by association. Educational Psychology, 1, 101–

103.
Edwards, J. (1985). Language, society and identity. Oxford: Blackwell (in association with
André Deutsch).
Edwards, J. (1989). Language and disadvantage (2nd revised edition). London: Cole &
Whurr. (Originally published by Edward Arnold, 1979).
Edwards, J. (1995). Multilingualism. London: Penguin.
Edwards, J. (2008). Obituary: Peter Hans Nelde (1942–2007). Journal of Multilingual and
Multicultural Development, 29, 447.
Edwards, J. (2009). Language and identity. Cambridge: Cambridge University Press.
Edwards, J. (2010). Language diversity in the classroom. Bristol: Multilingual Matters.
Edwards, J. (2012). Review of Can schools save indigenous languages? (Nancy Hornberg
er, ed.). Language Policy, 11, 201–203.
Edwards, J., & McKinnon, M. (1976). The continuing appeal of disadvantage as deficit: A
Canadian study in a rural context. Canadian Journal of Education, 12, 330–349.
Evans, B., & Hornberger, N. (2005). No Child Left Behind: Repealing and unpeeling feder
al language education policy in the United States. Language Policy, 4, 87–106.
Ferguson, G. (2006). Language planning and education. Edinburgh: Edinburgh University

Press.
Fine, M., Jaffe-Walter, R., Pedraza, P., Futch, V., & Stoudt, B. (2007). Swimming: On oxy
gen, resistance, and possibility for immigrant youth under siege. Anthropology and Edu
cation Quarterly, 38, 76–96.
Flores Farfán, J. A., & Ramallo, F. (2010). New perspectives on endangered languages.
Amsterdam: John Benjamins.
Page 18 of 22

Giles, H., & Bourhis, R. Y. (1976). Black speakers with white speech: A real problem? In
Gerhard Nickel (Ed.), Proceedings of the Fourth International Congress of Applied Lin
guistics. Stuttgart: Hochschulverlag.
Godley, A., Carpenter, B., & Werner, C. (2007). “I’ll speak in proper slang”: Language ide
ologies in a daily editing activity. Reading Research Quarterly, 42, 100–131.
Grenoble, L., & Whaley, L. (2006). Saving languages: An introduction to language revital
ization. Cambridge: Cambridge University Press.
Grinberg, J., & Saavedra, E. (2000). The constitution of bilingual/ESL education as a disci
plinary practice: Genealogical explorations. Review of Educational Research, 70, 419–
441.
Halliday, M. A. K. (1968). The users and uses of language. In Joshua Fishman (Ed.), Read
ings in the sociology of language. The Hague: Mouton.
Hélot, C., & Young, A. (2005). The notion of diversity in language education: Poli
(p. 419)
cy and practice at primary level in France. Language, Culture and Curriculum, 18, 242–
257.
Hornberger, N. (2011). Can schools save indigenous languages? Basingstoke: Palgrave

Macmillan.
Kautzsch, A. (2006). Review of African American English (Lisa Green). Language in Soci
ety, 35, 149–152.
Labov, W. (1976). Language in the inner city. Philadelphia: University of Pennsylvania

Press.
Labov, W. (1994). Principles of linguistic change. Two volumes. Oxford: Blackwell.
Lambert, W. E., Hodgson, R., Gardner, R., & Fillenbaum, S. (1960). Evaluational reactions
to spoken languages. Journal of Abnormal and Social Psychology, 60, 44–51.
Lippi-Green, R. (1997). English with an accent: Language, ideology and discrimination in

the United States. London: Routledge.
Mackey, W. (1978). The importation of bilingual education models. In James Alatis (Ed.),
Georgetown University round table on languages and linguistics. Washington: George
town University Press.
McDiarmid, G. W. (1992). What to do about differences? A study of multicultural educa

tion for teacher trainees in the Los Angeles Unified School District. Journal of Teacher Ed
ucation, 43(2), 83–93.
McDiarmid, G. W., & Price, J. (1990). Prospective teachers’ views of diverse learners. East
Lansing, Michigan: National Center for Research on Teacher Education, Michigan State
University.
Page 19 of 22

Ogbu, J. (1999). Beyond language: Ebonics, proper English and identity in a Black-Ameri
can speech community. American Educational Research Journal, 36, 147–184.
Phillips, M. (1997). All must have prizes. New York: Little, Brown.
Preston, D. (Ed.). (2003). Needed research in American dialects. Durham, NC: Duke Uni
versity Press.
Quiocho, A., & Rios, F. (2000). The power of their presence: Minority group teachers and
schooling. Review of Educational Research, 70, 485–528.
Rokkan, S., & Urwin, D. (1983). Economy, territory, identity. London: Sage.
Ryan, E. B. (1979). Why do low-prestige varieties persist? In H. Giles & R. St. Clair (Eds.),
Language and social psychology. Oxford: Blackwell.
Siegel, J. (2007). Creoles and minority dialects in education: An update. Language and Ed
ucation, 21, 66–86.
Smith, A. (1981). The ethnic revival. Cambridge: Cambridge University Press.
Smitherman, G. (2006). Word from the mother: Language and African Americans. London:
Routledge.
Steiner, G., & Ladjali, C. (2003). Éloge de la transmission: le maître et l’élève. Paris: Albin
Michel.
Swann, M. (1985). Education for all: The Report of the Committee of Inquiry into the Edu
cation of Children from Ethnic Minority Groups. London: HMSO, Department of Educa
tion and Science.
Tenenbaum, H., & Ruck, M. (2007). Are teachers’ expectations different for racial minori
ty than for European American students? A meta-analysis. Journal of Educational Psychol
ogy, 99, 253–273.
Tomlinson, S. (1984). Home, school and community. In M. Craft (Ed.), Education and cul
tural pluralism. London: Falmer.
Trudgill, P. (1975). Accent, dialect and the school. London: Edward Arnold.
Valenzuela, A., Prieto, L., & Hamilton, M. (2007). No Child Left Behind and minority
youth: What the qualitative evidence suggests. Berkeley: University of California Press.
Walsh, C. (1991). Pedagogy and the struggle for voice. New York: Bergin & Garvey.
Zéphir, F. (1997). Haitian Creole language and bilingual education in the United States:
Problem, right or resource? Journal of Multilingual and Multicultural Development, 18,
223–237.
Page 20 of 22

Zéphir, F. (1999). Challenges for multicultural education: Sociolinguistic parallels be

tween African American English and Haitian Creole. Journal of Multilingual and Multicul
tural Development, 20, 134–154.
Zientek, L. (2007). Preparing high-quality teachers: Views from the classroom. American
Educational Research Journal, 44, 959–1001.
Notes:
(1) . I should point out here that, hoping to produce a readable chapter that is both fluent
and succinct, I have kept intrusive references to a minimum. The reader is directed to
easily available existing work—particularly Edwards (1989, 2009, and 2010)—for more
specific details of all the main themes that are only touched on here. I am grateful to the
St. Francis Xavier University Council for Research for their support.
(2) . I use words like “think,” “recognize,” and “stated” here for the obvious reason that
concern with language variation in the classroom— where it involves a repudiation of ear
lier and more brutal language practices—still often exists only at a theoretical level, or,
indeed, is only given lip service.
(3) . At the same time as Bullivant was writing, the dynamic was also being considered by
others, using different but analogous dichotomies: see, for example, Rokkan and Urwin’s
(1983) treatment of “roots and options” or Smith’s (1981) distinction between “state” and
“community.” All such considerations of tensions between minorities and majorities, be
tween “big” and “small” societies, between traditional and modern values and practices—
and, hence, between linguistic “mainstreams” and “peripheries”—pay direct or indirect
homage to Tönnies’s famous distinction between Gemeinschaft and Gesellschaft and to
Saussure’s conflicting tendencies (la force d’intercourse and l’esprit de clocher) (see Ed
wards, 1985, 1995).
(4) . The sentiment expressed here is also very interesting in its suggestion that multicul
tural programs would be better aimed at majority group children. Who, after all, requires
heightened cross-cultural awareness—the socially dominant or the socially dominated?
(5) . Of course, while mutual intelligibility is a defining characteristic of dialect variants,

it can sometimes exist only at a theoretical level. We have all heard speakers of our own
language whose variety is so far removed from our own that they are extremely difficult
to understand; see Edwards (1995) for further discussion of the matter.
(6) . Self-esteem is not a negligible matter, of course, but recent discussion and debate
has suggested that too much attention to it may be counterproductive. If children are
never to fail at school, for instance, if they are to progress through the grades on the ba
sis of matters of “self-esteem” rather than on those of academic accomplishment, are we
not simply delaying decisions that will prove more and more difficult, and more and more
important, as time moves on? When an entire culture operates and rewards in terms of
competitions and rankings of all sorts, what are the implications of bringing up young
Page 21 of 22

children in “none-must-fail” and “all-are-to-have-prizes” atmospheres? A very polemical

treatment by Phillips (1997) may be consulted here.
(7) . I prescind entirely, of course, from any matters of ethics that may bear on the rela
tionship between Irish and English. Some would claim that history has placed the former
on the higher moral ground but, even if this were universally admitted, it need not have
the slightest bearing on revival success.
John Edwards
John Edwards, Department of Psychology, St. Francis Xavier University
Page 22 of 22

Pragmatic Processes in Survey Interviewing

Frederick G. Conrad , Michael F. Schober , and Norbert Schwarz

Subject: Psychology, Social Psychology, Psychological Methods and Measurement
Online Publication Date: Dec 2013 DOI: 10.1093/oxfordhb/9780199838639.013.005
In interpreting survey questions, respondents rely as much on the pragmatics of everyday

communication as they do on interpreting syntax or literal meaning. Respondents make
inferences about the intended meaning of questions based on a large array of contextual
cues, from judgments about the researcher to their sense of how common ground builds
up in ordinary conversation and their perception of the actions of interviewers (or auto
mated interviewing systems). How respondents answer questions—the speed and fluency
of their answers or requests for clarification—is informative about the accuracy and relia
bility of their answers. Although survey respondents try to make sense of questions using
whatever conversational resources they can bring to bear, researchers are often insensi
tive to these processes and surprised by their impact.
Keywords: pragmatics, survey, questions, dialogue, conversation, inferences, accuracy, clarification, interviewer,
automated system
When people fill out online customer satisfaction questionnaires, respond to political polls
on the telephone, talk to census enumerators, or participate in laboratory research that
involves questionnaires, they are engaged in a complex communicative act that involves
more processes than are immediately apparent. Of course, they must comprehend the lit
eral meaning of the words in the questions, retrieve relevant facts and opinions, and for
mulate their answers. But they also make inferences about the intentions behind the
questions based on a large array of contextual cues—what is ordinarily considered to be
long to the domain of linguistic pragmatics, as opposed to semantics or syntax. Some of
these inferences are made on the basis of prior knowledge that respondents apply to the
survey situation, and may make in real time while hearing or reading questions. For ex
ample, respondents may use their knowledge about the questioner’s affiliation, and there
fore likely intentions, or knowledge about what speakers usually mean when they ask
more than one question on the same topic. Other inferences are made on the basis of sub
sequent turns of dialogue, whether this “dialogue” is with a human interviewer or an au
tomated interviewing system. For example, a respondent may alter his or her interpreta
Page 1 of 30

tion of what a word or phrase in a question means after receiving clarifying information
or even a verbatim repetition of the question from an interviewer.
In this essay, we review evidence about these kinds of pragmatic processes in survey in
terviews. We argue that they occur because survey respondents bring their ordinary con
versational practices to survey interviews, even though surveys are a specialized form of
interaction (see, e.g., Houtkoop-Steenstra, 2000; Schaeffer, 1991; Schober, 1999; Schober
& Conrad, 2002; Schwarz, 1994). In everyday conversations, people have an array of
tools available to bridge the gaps between what they mean and what listeners take them
to mean. In general, they assume (p. 421) that they are engaged in a cooperative endeavor
with their interlocutor—that they are following what Grice (1975) called the cooperative
principle. So listeners try to make sense of what has been said, even if the literal content
doesn’t make sense; because they assume that a speaker is trying to communicate some
thing, they infer an intended meaning, even if that meaning is indirect, ironic, or hostile.
As we will see, in surveys, this can lead respondents to make unwarranted inferences in
assuming that the formal features of questionnaires have been designed following that
cooperative principle, even though they often have not (Schwarz, 1994, 1996). In typical
conversations, people also assume that they have the ability to ground their understand
ing with their partner (Clark, 1996; Clark & Schaefer, 1989; Clark & Wilkes-Gibbs, 1986)
—to engage in additional turns of clarification dialogue to make sure they have under
stood each other sufficiently for their current purposes. When survey respondents are
confronted with interviewing situations that explicitly prevent grounding dialogue, they
can end up interpreting survey questions differently than researchers intend and ulti
mately provide inaccurate answers without knowing it.
Surveys as an Unusual Conversational Situa

tion
Survey interactions are unusual on several fronts. First, whether the survey is adminis
tered by an interviewer or self-administered (for example a paper-and-pencil question
naire), the primary interlocutors are the respondent and an absent researcher or re
searching agency who have designed the questions (or commissioned someone else to de
sign them). In an interviewer-administered survey, there is an intermediary, the interview
er, who presents the questions on behalf of the researchers and facilitates the data collec
tion by keeping the interview on track, motivating the respondent, and recording the an
swers. In self-administered surveys, there may be no extra party involved in the “dia
logue” between the researchers and the respondent, although modern technologies are
blurring the distinction between interviewer- and self-administered surveys as they em
bed features of human interviewing, such as interactive dialogue, into their systems.1
Second, even when an interviewer is present in a survey interview, the goal of standard
ized survey interviewing is to minimize any potentially biasing influence of the interview
er on the responses. A set of standardized practices has evolved to reduce the likelihood
that an individual interviewer will bias answers (see Fowler & Mangione, 1990, for a re
Page 2 of 30

view of the state of the art and the underlying rationale; see also Beatty, 1995). The key
features are that interviewers read questions exactly as worded, probe neutrally, and nev
er allow their own ideas to influence the respondents’ answers. Observance of these
methodological rules is assumed to reduce or eliminate what survey researchers refer to
as interviewer-related error— any systematic effect of a particular interviewer on survey
responses.
Third, the requirement of standardized interviewers to administer only “neutral” or

nondirective probes creates an unusual dynamic in the interaction (see Schaeffer, 1991;
Schober & Conrad, 2002; Suchman & Jordan, 1990, 1991). Consider an interview in
which the interviewer presents the question “How many hours per week do you usually
work?” If a respondent asks the interviewer for clarification—“Should I consider business
lunches to be work?”— she is unlikely to get a satisfying answer under standardized in
terviewing conditions. To prevent the interviewer’s providing additional substantive infor
mation to this respondent that would not be available to a respondent who didn’t inquire,
an interviewer following strictly standardized procedures is required to reply with a neu
tral probe such as rereading the question (“Let me repeat the question: How many hours
per week do you usually work?”) or indicating that the interpretation of the question is
the sole obligation of the respondent (e.g., “Whatever it means to you”). This kind of
move would be extremely strange in an everyday conversation, and, even if respondents
tend to be aware that they are in an unusual interactive situation, there is evidence that
this kind of interaction can be frustrating (Suchman & Jordan, 1990, 1991). Aside from
frustration, this leads to the unusual feature that, in survey interviews, interviewers relin
quish control over what the question means. In most conversations, speakers are respon
sible for the meaning of what they say (Clark, 1996; Clark & Schober, 1992), and they
usually have the ability to engage in additional dialogue to clarify and verify—to ground—
what was intended.
Nonetheless, survey interactions do share many other important features with everyday
conversations, most notably the back-and-forth structure of question-asking and answer
ing and the reliance on additional contextual information to make sense of what is being
said. We argue that it is exactly because people do not (perhaps cannot) turn off their or
dinary reasoning and meaning-making when participating in a survey interview that the
pragmatics of (p. 422) survey interviewing are particularly important to understand.
Inferring Meaning from Features of the Ques

tionnaire
The constraints that standardized communication impose on the mutual negotiation of
meaning force participants to rely on other information to make sense of the researcher’s
questions. To do so, they make extensive use of contextual information, going beyond the
literal meaning of the question asked to infer what the question designers intended, as li
censed by Grice’s (1975, 1978) cooperative principle. Listeners make sense of utterances
on the assumption that speakers are following (at least) four of what Grice called maxims.
Page 3 of 30

Listeners assume that speakers make their contribution such that it can be understood by
their audience, not only avoiding ambiguity and wordiness, but also taking the character
istics of their audience into account, designing their utterance in a way that the audience
can figure out what they mean (Grice’s maxim of manner). Central to this process are
speakers’ assumptions about the information that they share with recipients—that is, the
common ground (Schiffer, 1972; Stalnaker, 1978); listeners interpret the speaker’s utter
ance against what they assume to constitute the common ground (e. g., Clark, Schreuder,
& Buttrick, 1983). Moreover, listeners assume that each successful contribution to the
conversation extends the common ground of the participants, in line with their assump
tion that speakers are trying to make all contributions relevant to the aims of the ongoing
conversation (Grice’s maxim of relation). Listeners thus use the context of an utterance to
disambiguate its meaning by making bridging inferences (Clark, 1977); they assume
speakers will not make a contribution to a conversation that is irrelevant to its goal, un
less it is marked as such. Listeners assume that speakers make their contribution as in
formative as is required, but not more informative than is required (Grice’s maxim of
quantity), without reiterating information that the recipient already has or may take for
granted anyway (Clark & Haviland, 1977). Finally, listeners assume that speakers do not
say things they believe to be false or lack adequate evidence for (Grice’s maxim of
quality).
As Grice (1975) noted, these maxims apply most directly to situations in which people at
tempt to exchange information to get things done. Of course, many conversations are
characterized by other goals, such as entertaining one another (see Higgins, Fondacaro,
& McCann, 1982), and such additional goals can be at play in survey interviews, as inter
viewers work to build rapport and motivate respondents. Nonetheless, the assumptions
underlying the cooperative principle remain at play even when additional goals are
present, in a number of ways.
Researcher’s Affiliation
In natural conversations, listeners take their knowledge about the speaker into account
when they interpret his or her utterances. So do research participants. For example,
Norenzayan and Schwarz (1999) asked respondents to explain a case of workplace homi
cide, described in a newspaper clipping. The otherwise identical questionnaire was either
printed on the letterhead of an “Institute for Personality Research” or on the letterhead of
an “Institute for Social Research.” As expected, respondents’ open-ended explanations
entailed a greater emphasis on personality variables or on social-contextual variables, de
pending on whether they thought the researcher was a personality psychologist or a so
cial scientist. Similarly, Galesic and Tourangeau (2007) asked respondents otherwise iden
tical questions about workplace behavior as part of an alleged “Sexual Harassment Sur
vey” conducted for Women Against Sexual Harassment or as part of an alleged “Work At
mosphere Survey” conducted for a Work Environment Institute. Respondents perceived
the same behaviors as more likely to represent sexual harassment when they were pre
Page 4 of 30

sented as part of a sexual harassment survey—after all, why else would they have been
included?
Such findings highlight that participants take the researchers’ likely epistemic interest in
to account to provide an answer that is relevant to the goals of the ongoing conversation,
much as they would be expected to in daily life. Beyond simply providing the kind of an
swer that they believe the researcher wants to hear, respondents can perceive the do
main of inquiry differently depending on the questioner’s identity and thus his or her pre
sumed research goals. Unfortunately, researchers themselves are often insensitive to the
information that their institutional affiliation or the title of their study may convey.
Preceding Questions
In natural conversations, listeners draw on the already established common ground when
they interpret a speaker’s utterances. In stark contrast, researchers would often prefer
respondents to (p. 423) answer each question in isolation, thus avoiding the emergence of
any order effects. Respondents, however, may not always be aware that the usual rules of
conversational conduct do not apply. Hence, they do their best to make sense of the ques
tions asked by drawing on the context in which they are presented, including the content
of preceding questions.
Suppose, for example, that respondents are asked questions about highly obscure or even
completely fictitious issues, such as the “Agricultural Trade Act of 1978” (e.g., Bishop,
Oldendick, & Tuchfarber, 1986; Schuman & Presser, 1981). In a typical study, some 30
percent of the respondents of a representative sample are likely to report an opinion on
these issues, despite the fact that they can hardly know anything about it. Some re
searchers attributed this to respondents’ “fear of appearing uninformed,” which may in
duce them “to conjure up opinions even when they had not given the particular issue any
thought prior to the interview” (Erikson, Luttberg, & Tedin, 1988, p. 44), casting serious
doubt on the meaningfulness of survey responses. From a conversational perspective,
however, respondents merely do what they would be expected to do in any other conver
sation: they make sense of an ambiguous utterance by drawing on its context. Testing the
possibility that context created by prior questions can affect interpretation of subsequent
questions, Strack, Schwarz, and Wänke (1991, experiment 1) asked German college stu
dents about their attitude toward the introduction of an “educational contribution” al
legedly discussed in state parliament. For some participants, this question was preceded
by a question about the average tuition fees that students have to pay at US universities
(in contrast to Germany, where university education is free), whereas others had to esti
mate how much money Swedish students receive from their government as financial sup
port. At the end of the study, participants were asked what the “educational contribution”
implied. As expected, they inferred that students would receive money when the fictitious
issue was preceded by the Swedish support question, but that students would have to pay
money when it was preceded by the American tuition question. Not surprisingly, they fa
Page 5 of 30

vored the introduction of an “educational contribution” in the former case, but opposed it
in the latter case.
Open Versus Closed Question Formats
Respondents also rely on contextual information when the question is less obviously am
biguous than in the just presented case. Suppose, for example, that you are asked to re
port what you have done today. Most likely, you would not include in your report that you
took a shower, that you dressed, and so on. However, if these activities were included in a
list of response alternatives, you would probably endorse them. This thought experiment
reflects a robust set of findings (see Schwarz & Hippler, 1991, for a review): any given
opinion or behavior is less likely to be volunteered in an open response format (the re
spondent answers in his or her own words) than to be endorsed in a closed response for
mat (multiple choice options), if presented. On the other hand, opinions or behaviors that
are omitted from the set of response alternatives in a closed format are unlikely to be re
ported at all, even if an “other” category is explicitly offered, which respondents in gener
al rarely use. Several processes contribute to this reliable pattern.
Most importantly, respondents are unlikely to spontaneously report, in an open answer

format, information that seems self evident or irrelevant, as in the shower example just
given. In refraining from these responses, they follow the conversational maxim that an
utterance should be informative and should provide the information that the recipient is
interested in. This results in an underreporting of presumably self evident information
that is eliminated by closed response formats, in which the explicit presentation of the
proper response alternative indicates the investigator’s interest in this information. Simi
larly, a generic “other” response provides little information and would be considered in
adequate as an answer in most conversations; hence, it is rarely checked. In addition, the
response alternatives may remind respondents of options that they may otherwise not
have considered.
The Numeric Values of Rating Scales
When constructing a rating scale, researchers typically pay attention to the verbal end
anchors used and to the number of scale points presented (see Dawes & Smith, 1985, for
a review). Having settled for a 7-point rating scale, for example, they are less likely to
worry whether those seven points should be represented by unnumbered boxes, by num
bers ranging from 1 to 7, or by numbers ranging from –3 to +3. However, the specific nu
merical values used may again change the inferred meaning of the question, strongly af
fecting the obtained responses.
For example, Schwarz, Knäuper, Hippler, Noelle-Neumann, and Clark (1991a, experiment
1) asked a representative sample of German (p. 424) adults, “How successful would you
say you have been in life?” This question was accompanied by an 11-point rating scale,
with the endpoints labeled “not at all successful” and “extremely successful.” In one con
dition, the numeric values of the rating scale ranged from 0 (“not at all successful”) to 10
Page 6 of 30

(“extremely successful”), whereas in the other condition they ranged from –5 (“not at al
successful”) to +5 (“extremely successful”). The results showed a pronounced impact of
the numeric values used. Whereas 34 percent of the respondents endorsed a value be
tween 0 and 5 on the 0–10 scale, only 13 percent endorsed one of the formally equivalent
values between –5 and 0 on the –5 to +5 scale.
As subsequent experiments showed (Schwarz et al., 1991a), respondents draw on the nu
meric values of the rating scale to interpret the verbal anchor “not at all successful”—
does this term refer to the absence of noteworthy success or to the presence of explicit
failure? When the numeric values of the scale run from 0 to 10, respondents assume that
the researcher has a unipolar dimension in mind, in which different values reflect differ
ent degrees of the same attribute. Conversely, when the numeric values run from –5 to
+5, they infer that the researcher has a bipolar dimension in mind, in which one endpoint
(–5) refers to the opposite of the other (+5). Accordingly, “not at all successful” refers to
the mere absence of success when combined with the numeric value “0” but to the pres
ence of failure when combined with the numeric value “–5.”
Respondents’ use of numeric values in making sense of verbal labels is not restricted to
variations that include or omit negative numbers. For example, Schwarz, Grayson, and
Knäuper (1998) asked undergraduates how often they engaged in a variety of low-fre
quency activities. In all conditions, the 11-point rating scale ranged from “rarely” to “of
ten.” However, “rarely” was combined with the numeric value 0 or 1. As expected, re
spondents interpreted “rarely” to mean “never” when combined with 0, but to mean “a
low frequency” when combined with 1. As a result, they provided higher mean frequency
ratings along the 0–10 (M = 2.8) than the 1–11 scale (M = 1.9; scale recoded to 0–10).
Frequency Scales
Research participants are often asked to report the frequency with which they engage in
a behavior by checking the appropriate value from a set of frequency response alterna
tives provided to them (see Table 26.1 for an example). Again, the range of response alter
natives may serve as a source of information for respondents (see Schwarz, 1999b, for a
review). In general, respondents assume that researchers construct a meaningful scale
that reflects appropriate knowledge about the distribution of the behavior. Accordingly,
values in the middle range of the scale are assumed to reflect the “average” or “typical”
behavior, whereas the extremes of the scale are assumed to correspond to the extremes
of the distribution. These assumptions influence respondents’ interpretation of the ques
tion, their behavioral reports, and related judgments.
Page 7 of 30

Table 26.1 Reported daily TV consumption as a function of response

alternatives
Low Frequency High Frequency
Alternatives Alternatives
Up to 1/2 h 7.4% Up to 2 1/2h 62.5%
1/2 h to 1h 17.7% 2 1/2h to 3h 23.4%
1h to 1 1/2h 26.5% 3h to 3 1/2h 7.8%
1 1/2h to 2h 14.7% 3 1/2h to 4h 4.7%
2h to 2 1/2h 17.7% 4h to 4 1/2h 1.6%
More than 2 1/2h 16.2% More than 4 1/2h 0.0%
Question Interpretation
Suppose, for example, that respondents are asked to indicate how frequently they were
“really irritated” recently. To give an informative answer, they must decide what the re
searcher means by “really irritated”: does the term refer to major or minor irritations?
When the scale presents high-frequency response alternatives, participants may infer that
the researcher is interested in relatively frequent events, whereas low-frequency re
sponse alternatives presumably indicate an interest in rare events. Given that major irri
tations are less frequent than minor ones, the range of frequency values may therefore in
directly specify the kind of irritations the researcher has in mind.
Empirically, this is the case. For example, Schwarz, Strack, Müller, and Chassein (-1988)
asked participants to report how frequently they were “really irritated” along a high-fre
quency scale ranging from “several times daily” to “less than once a week” or along a
low-frequency scale ranging from “several times a year” to “less than once every three
months.” Subsequently, participants reported a typical example of irritation. As expected,
their examples were more extreme when the preceding question presented a low- rather
than high-frequency scale, indicating that they drew on the frequency scale to determine
the intended meaning of the question. Accordingly, identical question stems in combina
tion with different frequency scales result in different interpretations and hence assess
different experiences (see Schwarz, 1999, for a review).
Frequency Estimates
Even if the behavior under investigation is reasonably well defined, however, the range of
response alternatives may strongly affect respondents’ frequency estimates. Mundane be
Page 8 of 30

haviors of a high frequency, such as watching TV, for example, are not represented in
memory as distinct episodes; instead, the individual episodes blend into generic (p. 425)
representations that lack time and space markers (see Belli, 1998, for a discussion). Ac
cordingly, respondents cannot recall and count relevant episodes but have to resort to
various estimation and inference strategies to arrive at a plausible answer (for more de
tailed treatments, see Brown, 2002; Schwarz & Oyserman, 2001). One of these strategies
draws on the range of the frequency scale as a frame of reference, resulting in higher fre
quency estimates along scales that present high rather than low frequency response al
ternatives.
For example, only 16.2 percent of a sample of German consumers reported watching TV
for more than 2.5 hours a day when presented with the low-frequency scale shown in Ta
ble 26.1, whereas 37.5 percent did so when presented with the high-frequency scale (Sch
warz, Hippler, Deutsch, & Strack, 1985). Similar results have been obtained for a wide
range of different behaviors, including health behaviors (e.g., Schwarz & Scheuring,
1992), sexual behaviors (e.g., Tourangeau & Smith, 1996), and consumer behaviors (e.g.,
Menon, Raghubir, & Schwarz, 1995). More demanding estimation tasks increase the de
gree to which scale values affect participants’ personal frequency estimates (e.g., Bless,
Bohner, Hild, & Schwarz, 1992), and, the less concrete information people have in memo
ry, the more they rely on the scale when making their judgments (e.g., Schwarz & Bi
enias, 1990). In contrast, the impact of response alternatives is weak or absent when the
question pertains to highly regular behaviors (e.g., brushing teeth), for which respon
dents can draw on rate-of-occurrence information (e.g., “twice a day”; Menon, 1994;
Menon et al., 1995).
Comparative Judgments
In addition, the frequency range of the response alternatives has been found to affect
subsequent comparative judgments. Given the assumption that the scale reflects the dis
tribution of the behavior, checking a response alternative is the same as locating one’s
own position in the distribution. Accordingly, respondents extract comparison information
from their own location on the response scale and use this information in making subse
quent comparative judgments.
For example, checking 2h on the low-frequency scale shown in Table 26.1 implies that a
respondent’s TV consumption is above average, whereas checking the same value on the
high-frequency scale implies that his or her TV consumption is below average. As a re
sult, respondents in the Schwarz et al. (1985) studies reported that TV plays a more im
portant role in their leisure time (experiment 1) and described themselves as less satis
fied with the variety of things they do in their leisure time (experiment 2) when they had
to report their TV consumption on the low- rather than the high-frequency scale. More
over, these frame of reference effects are not limited to respondents themselves, but in
fluence the users of their reports as well. For example, in a study by Schwarz, Bless,
Bohner, Harlacher, and Kellenbenz (1991, experiment 2), experienced medical doctors
considered having the same physical symptom twice a week to reflect a more severe med
Page 9 of 30

ical condition when “twice a week” was a high rather than a low response alternative on
the symptoms checklist presented to them.
Reference Periods
Similarly, Winkielman, Knäuper, and Schwarz (1998) observed that the length of the refer
ence period can profoundly affect question interpretation. In their studies, respondents
were either asked how frequently they had been angry “last week” or “last year.” Again,
they inferred that the researcher is (p. 426) interested in more frequent and less severe
episodes of anger when the question pertained to one week rather than one year, and
their examples reflected this differential question interpretation.
On theoretical grounds, one may further expect that formal features of questionnaires,
like the values of a frequency scale or the length of a reference period, seem more rele
vant when they are unique to the question asked rather than shared by many heteroge
neous questions. In the latter case, respondents may conclude that this is the format used
for all questions, rendering it less informative for the intended meaning of any given one.
Empirically, this is the case. For example, using the same reference period for several
substantively unrelated questions attenuates its influence on question interpretation rela
tive to conditions in which each question is associated with a unique reference period
(Igou, Bless, & Schwarz, 2002).
Summary
As the reviewed examples illustrate, minor variations in question wording or format can
profoundly affect respondents’ answers (Schwarz, 1999a). They do so because respon
dents draw on the context in which a question is asked to infer its intended meaning. In
most surveys, respondents have no other choice because there is either nobody to ask for
clarification (as in the case of self-administered questionnaires, mail, and most web sur
veys) or a well-trained standardized interviewer will refuse to provide additional informa
tion. This reliance on contextual information is licensed by the cooperative principle
(Grice, 1975); respondents merely do what they would be expected to do in everyday in
teractions. Unfortunately, their efforts are rarely appreciated by most researchers, who
consider formal characteristics of questionnaires and the content of adjacent questions
substantively irrelevant and treat their influence as an undesirable artifact. In most stud
ies, researchers are unlikely to even note such influences, given that most surveys only
ask questions in one way, without experimentally manipulating question format.
Giving an Informative Answer: Question Repe

tition and Meaning Change
The preceding examples illustrated how research participants draw on conversational
maxims in interpreting the researcher’s contributions (i.e., the questions). Not surprising
ly, participants also heed the maxims that underlie their conduct of conversation in daily
life when they design their own answer to the researcher’s questions. Consistent with the
Page 10 of 30

maxim of quantity, they observe the common ground and try to provide information that
is new to the researcher, rather than information they have already given earlier—or in
formation the researcher may take for granted anyway.
As an example, consider a study in which respondents were asked to report their marital
satisfaction, as well as their general life-satisfaction in different orders (Schwarz, Strack,
& Mai, 1991b). As shown in the first column of Table 26.2, the answers to both questions
correlated r = .32 when the general life-satisfaction question preceded the more specific
marital satisfaction question. Reversing the question order increased this correlation to r
= .67. This reflects that answering the marital satisfaction question brought marriage-re
lated information to mind that respondents might otherwise not have considered. This in
terpretation is supported by a correlation of r = .61 when a reworded version of the gen
eral question explicitly asked respondents to take their marriage into account.
Page 11 of 30

Table 26.2 Correlation of relationship satisfaction and life-satisfaction

as a function of question order and conversational context
Number of Specific Questions
One Three
Condition
General-specific .32* .32*
Specific-general .67* .46*
Specific-general,
with joint lead-in .18 .48*
Specific-general,
explicit inclusion .61* .53*
Specific-general,
explicit exclusion .20 .11
N = 50 per cell, except for “Specific-general with joint lead-in,” N =

56. Correlations marked by an asterisk differ from chance, p <.05.
Adapted from Schwarz, N., Strack, F., & Mai, H. P. (1991). Assimila
tion and contrast effects in part-whole question sequences: A conver
sational logic analysis. Public Opinion Quarterly, 55, 3–23. Reprinted
by permission.
In a third condition, Schwarz and colleagues deliberately evoked the conversational norm
of nonredundancy (maxim of quantity). To do so, they introduced both questions with a
joint lead-in that (p. 427) read, “We now have two questions about your life. The first per
tains to your marital satisfaction and the second to your general life-satisfaction.” Under
this condition, the same question order that resulted in r = .67 without a joint lead-in now
produced a low and nonsignificant correlation of r = .18. This suggests that respondents
deliberately ignored information that they had already provided in response to a specific
question when making a subsequent general judgment. Apparently, they interpreted the
general question as if it were worded, “Aside from your marriage, which you already told
us about, how satisfied are you with other aspects of your life?” Consistent with this inter
Page 12 of 30

pretation, a correlation of r = .20 was obtained when the general question was reworded
in this way.
Note, however, that the pragmatic implications of the norm of nonredundancy may
change when several specific questions precede a more general one. Suppose, for exam
ple, that respondents are asked to report on their job satisfaction, their leisure time satis
faction, and their marital satisfaction before a general life-satisfaction question is pre
sented. In this case, they may interpret the general question in two different ways. On the
one hand, they may assume that it is a request to consider still other aspects of their life,
much as if it were worded, “Aside from what you already told us,…” On the other hand,
they may interpret the general question as a request to integrate the previously reported
aspects into an overall judgment, much as if it were worded, “Taking these aspects to
gether,…” This interpretational ambiguity does not arise if only one specific question is
asked, as in this example. In that case, an interpretation of the general question in the
sense of “taking all aspects together” would make little sense because only one aspect
was addressed, thus rendering this interpretation of the general question completely re
dundant with the specific one. If several specific questions are asked, however, both inter
pretations are viable. In this case, an integrative judgment is informative because it does
provide “new” information about the relative importance of the respective domains.
Moreover, “summing up” at the end of a series of related thoughts is acceptable conver
sational practice, whereas there is little to sum up if only one thought was offered.
To explore these possibilities, other respondents were asked three specific questions per
taining to their leisure time satisfaction, their job satisfaction, and, finally, their marital
satisfaction. As shown in the second column of Table 26.2, the correlation between mari
tal satisfaction and life satisfaction increased from r = .32 to r = .46 when answering the
specific questions first brought information about one’s marriage to mind. However, this
increase was less pronounced than when the marital satisfaction question was the only
specific question that preceded the general one (r = .67), reflecting that the three specif
ic questions brought a more varied set of information to mind. More importantly, intro
ducing the three specific questions and the general one by a joint lead-in did not reduce
the resulting correlation, r = .48. This indicates that respondents adopted a “Taking-all-
aspects-together” interpretation of the general question if it was preceded by three,
rather than one, specific question(s). This interpretation is further supported by a correla
tion of r = .53 when the general question was reworded to request an integrative judg
ment and a correlation of r = .11 when the reworded question required the consideration
of other aspects of one’s life.
In combination, these findings further illustrate that the interpretation of an identically

worded question may change as a function of conversational variables, resulting in
markedly different responses. Moreover, the resulting differences are not restricted to
the means or margins of the response distribution, as social scientists have frequently as
sumed. Rather, context variables may result in different correlational patterns, thus vio
lating the assumption that context effects would be restricted to differences in the means,
whereas the relationship between variables would be “form resistant” (Schuman & Dun
Page 13 of 30

can, 1974). Obviously, we would draw very different substantive conclusions about the
contribution of marital satisfaction to overall life satisfaction, depending on the order in
which these questions are presented and the perceived conversational context evoked by
their introduction.
Adjusting Interpretation in Survey Dialogue
Pragmatic processes in survey interviewing are at play not only in respondents’ initial in
terpretation of questions, but also in the interaction—the conversational turns or subse
quent respondent behaviors with an automated system—that follows the presentation of
the questions. In those rare interviews that allow clarification dialogue, respondents can
adjust their interpretation of questions from definitions that interviewers (or automated
interviewing systems) give or instructions about the response task (e.g., the degree of
precision required in an answer); clarification can be given either as a result of the
respondent’s direct request (p. 428) for help (“what do you mean by ‘work for pay’?”) or
because the interviewer has diagnosed the respondent’s need for additional help from the
respondent’s tentative-sounding answer (“uh…yes?”). In a strictly standardized interview,
respondents cannot rely on explicit clarification or help, but they still can adjust their in
terpretations on the basis of interviewer actions and what those actions might mean in an
everyday conversation.
It is worth highlighting that more goes on in survey dialogue than one would think from
simply looking at an interviewer’s script. Beyond interviewers asking question and re
spondents immediately providing answers followed immediately by the next question, in
terviewers often accept answers (“got it,” “okay”) in ways that can inform respondents
about the suitability of their answers and potentially about the interviewer’s evaluation of
the content of the answers, although interviewers are trained to avoid giving evaluative
feedback on answers. Interviewers can repeat a respondent’s answer to verify that they
heard it correctly (“three times a week?”)—which has the potential to sound like a chal
lenge or surprise at the answer. Just as almost any feature of a questionnaire can provide
information about what a researcher might have intended, there is virtually nothing an in
terviewer can say or do that doesn’t have the potential to influence a respondent’s inter
pretation—nodding or not nodding, writing down an answer or not writing down an an
swer, smiling or not smiling, all have the potential to be treated as meaningful by respon
dents and to be used to infer researchers’ intentions.
Inferring Interpretation in Strictly Standardized Interviews
As we noted earlier, strictly standardized interviewers are trained to behave in ways that
would be quite unusual in an everyday conversation. This can limit the kinds of conversa
tional moves they can make in ways that have the potential to seriously affect survey data
quality.
In particular, respondents in strictly standardized interviews cannot engage in the kind of

clarification dialogue that would help align their interpretation of individual words with
Page 14 of 30

the intentions of the question authors. And the likelihood of misalignment can be much
greater than one might think; respondents’ interpretations of ordinary terms in even well-
pretested questions can be surprisingly variable. For example, Belson (1981) reported
that respondents’ interpretation of “reading a magazine” covered a wide range of differ
ent behaviors, from having seen the magazine at a newsstand to having read it cover-to-
cover or having subscribed to it. Similarly, Suessbrick, Schober, and Conrad (2000) asked
respondents, “Have you smoked at least 100 cigarettes in your entire life?” Whereas 54
percent of the respondents interpreted this as including “cigarettes that you took a puff
or two from,” 23 percent thought it referred to “cigarettes you partially smoked” (more
than a puff or two but not finished), and 23 percent interpreted it as including “only ciga
rettes that you finished.” There was a similar range in whether respondents counted only
tobacco cigarettes, herbal and marijuana cigarettes, pipes, cigars, or chewing tobacco;
whether they counted cigarettes that they smoked without inhaling; and whether they
counted only cigarettes they themselves had bought.
Respondents’ interpretations were variable enough that when they were reasked the
same question, asking them to include any cigarettes that they took even one puff of but
only tobacco cigarettes, 10 percent of the respondents changed their answers, indicating
that their initial interpretation of the question was not shared with the researchers or
with each other. Their initial interpretation also had the problematic effect of leading
them down different paths of the survey, and thus answering quite different subsequent
questions, than they would have with a different starting interpretation. This kind of con
ceptual variability in interpretation can thus harm survey estimates for more than just a
single question.
As this example makes clear, the nature of ambiguity in surveys goes beyond simple lexi
cal ambiguity, which usually refers to words having different senses (e.g., “bank” as a fi
nancial institution vs. edge of a river). The ambiguity in the smoking example has to do
with the range of activities that a respondent includes in the term—which may not be sta
ble across all different moments of use of “smoking” in a respondent’s life or day because
people may have different judgments about what is relevant to include in different con
versations. In surveys that ask respondents about their own lives, there is an additional
layer of ambiguity having to do with the mapping between the researcher’s concept, the
respondent’s (chronic or momentary) concept, and the respondent’s actual life circum
stances. For example, a respondent who has never touched a cigarette or one who
smokes tobacco cigarettes all the time may conceive of smoking rather differently than
the researcher, but their answer to the smoking question will be unaffected by the differ
ence. It is (p. 429) only for respondents who have smoked occasionally and whose inter
pretation of smoking differs from the researcher’s that the misalignment can lead to an
swers that are inconsistent with researcher’s concepts and thus problematic survey esti
mates.
As another example, consider these questions from major surveys of US government

agencies. “During the past year, have you purchased or had expenses for household furni
ture?” If you bought an end table, you shouldn’t have much trouble answering “Yes.” But
Page 15 of 30

what if you bought a floor lamp—does that qualify as “furniture”? Or consider the ques
tion, “Last week, how many hours did you work?” If you have a 9-to-5 job, the meaning
may not be ambiguous, but what if you do business over lunch or solve work-related prob
lems while jogging? How about an academic who keeps up with developments in her field
by reading at night? As the examples illustrate, many questions have a high degree of
what we have called mapping ambiguity; that is, the mapping between the question term
and the respondent’s circumstances remains unclear.
As we have noted, the ideal of strictly standardized interviewing is to leave the interpreta
tion of survey terms entirely up to respondents and to deal with any requests for clarifica
tion with neutral probes that return the responsibility for interpretation to the respon
dent. Nonetheless, interviewers’ neutral probes can (intentionally or not) direct respon
dents to particular interpretations. Consider this example from a telephone interview
(Schober & Conrad, 1997), in which the interviewer repeats the question with contrastive
stress on the critical word (“COLLEGE”) in a way that clearly informs the respondent
about the researcher’s interpretation of college tuition:
INTERVIEWER (I): Has Alexander purchased or had expenses for college tuition
or fixed fees?
RESPONDENT (R): Um, he’s got…um tuition for secretarial school.
I: Pardon me, I …I didn’t…
R: He has tuition for secretarial school.
I: Has Alexander purchased or had expenses for COLLEGE tuition?
R: No.
I: Or fixed fees?
R: No.
Here, the interviewer has followed the letter of the law for standardized interviewing but
has nonetheless provided additional information to this respondent (and probably not to
others), at odds with the goals of neutral probing. The respondent has used his knowl
edge of the pragmatics of dialogue to determine the researcher’s interpretation in a way
that would be invisible to the researcher without close examination of the interview dia
logue.
Clarification Dialogue in Surveys

Although strictly standardized interviews try to prevent the use of everyday techniques
for grounding understanding, one could imagine administering surveys in ways that en
courage interviewers to use nonstandardized wording to help respondents interpret sur
vey questions more uniformly, in keeping with respondents’ (and interviewers’) everyday
Page 16 of 30

conversational practices. This approach has been explored in a series of studies by Con
rad, Schober, and colleagues that compare response accuracy and interaction patterns in
strictly standardized and more collaborative “conversational” interviews, as well as for a
range of interviewing techniques in between. These studies have produced consistent
findings across a range of survey modes involving human interviewers (telephone and
face-to-face) and automated interviewing systems (textual web surveys, virtual animated
interviewers, speech interactive voice response [IVR]) that allow clarification in different
ways.
In one set of studies, professional interviewers telephoned naïve respondents either in the
laboratory (Schober & Conrad, 1997; Schober, Conrad, & Fricker, 2004) or at home (Con
rad & Schober, 2000) and asked questions from large US government surveys. In the lab
oratory studies, respondents answered the questions on the basis of scenarios that de
scribed the work, housing, and purchases of fictional people. Because the researchers
created the scenarios, they knew the correct answers—according to official definitions—
for each question-scenario combination, allowing the determination of response accuracy.
In the household study, respondents reported about their own lives, and response accura
cy was evaluated by less direct reinterview methods.
In all interviews, questions were first posed exactly as worded, essentially standardizing
the “starting point” under both standardized and conversational interviewing approach
es; the difference between approaches concerns whether it is ultimately wording or un
derlying meaning that is standardized. In the more collaborative interviews, interviewers
were (p. 430) then encouraged to ground understanding of question meaning. Across the
experimental conditions in the different studies, they were trained to do this either by
providing clarification only when respondents explicitly requested it, as in this example
(from Schober & Conrad, 1997):
I: Has Kelly purchased or had expenses for household furniture?
R: Um…is a lamp furniture?
I: No sir, we do not include lamps and lighting fixtures.
R: Okay, no.
or by also providing clarification when, in their judgment, respondents seemed to need it

(based on how they had answered or not answered the question), as in this example
(Schober, Conrad, & Fricker, 2004):
I: How many hours per week does Mindy usually work at her job?
R: (long pause)
I: And by usually I mean fifty percent of the time or more, or the most frequent
schedule during the past four or five months.
R: Uh, fifty hours.

Page 17 of 30

I: Okay.
In some conditions they were trained to provide full verbatim definitions when they pro
vided clarification, and, in others, they could paraphrase what they deemed to be the
most relevant parts of definitions. In all studies, all interviewers learned the official defin
itions of key concepts in the questions, whether or not they were to provide definitions in
the interview, so that differential interviewer knowledge could not account for any effects
on response accuracy.
Fig. 26.1 Response accuracy for straightforward and

complicated scenarios across five interviewing condi
tions: (1) no clarification (standardized), (2) respon
dent-initiated clarification provided verbatim, (3) re
spondent-initiated clarification that interviewers
could paraphrase, (4) mixed initiative (respondent- or
interviewer-initiated) provided verbatim, and (5)
mixed initiative (respondent- or interviewer-initiated)
clarification that interviewers could paraphrase.
Across all studies with well-specified fictional scenarios (Schober & Conrad, 1997;
Schober, Conrad, & Fricker, 2004), a reliable pattern emerged (see Figure 26.1). When
the mappings between question concepts and people’s circumstances (scenarios) were
straightforward, all interviewing techniques led to nearly perfect accuracy; virtually all
respondents interpreted question concepts in the ways that survey designers intended.
But when mappings between question concepts and people’s circumstances were compli
cated, response accuracy was lowest (28 percent) under strictly standardized interview
ing conditions (Schober & Conrad, 1997). Response accuracy improved dramatically
when interviewers provided clarification on request (54 and 59 percent in Schober, Con
rad, & Fricker, 2004, experiment 1), but not as much as when they could also give clarifi
cation when they thought (p. 431) it might help (87 percent in Schober & Conrad, 1997).
Whether interviewers’ clarifications were scripted (i.e., they provided definitions verba
tim) or unscripted (i.e., they could paraphrase definitions) (Schober, Conrad, & Fricker,
2004, experiment 1), did not really matter. Even when interviewers were merely instruct
ed to do what they ordinarily do, they sometimes provided clarification (Schober, Conrad,
& Fricker, 2004, experiment 2), and, in these cases, response accuracy was much higher
Page 18 of 30

(79 percent) than when they left the interpretation of questions entirely up to respon
dents (23 percent).
These results were mirrored in a household study (Conrad & Schober, 2000) in which par
ticipants reported on their own lives. In this study, respondents’ actual circumstances
were unknown, and response accuracy was assessed through a reinterview; the initial in
terview was standardized and the second interview was conversational for half the re
spondents and standardized for the other half. Respondents changed their answers more
when their second interview was conversational (22 percent) than when it was strictly
standardized (11 percent), a change that must be attributed to the fact that the clarifica
tion available in the conversational reinterview led to a reinterpretation of the questions
consistent with the researchers’ intended interpretations. Consistent with this finding,
when asked about their purchases, respondents were more likely to answer on the basis
of “legal” purchases (i.e., those that fit the definition of the purchase category) when the
second interview was conversational (95 percent) than when it was standardized (57 per
cent). For example, in conversational interviews, respondents who answered “yes” to a
question about moving expenses were more likely to do so because they were considering
expenses included in the definition (e.g., they had hired commercial movers) than in stan
dardized interviews, where they were more likely to respond “yes” because of excluded
expenses (e.g., do-it-yourself moving). This could clearly be attributed to interviewers’ ex
plaining what to count and what not to count (i.e., grounding the meaning of the term).
The improved response accuracy in conversational interviews suggests that respondents’
actual circumstances (as opposed to the fictional scenarios presented in the lab studies)
are complicated sufficiently often—at least for these questions—to justify a collaborative
approach to survey interviewing.
Despite the improved accuracy that comes with conversational interviewing and the feasi
bility of the technique for use by ordinary telephone interviewers, the benefits come at an
operational cost. Conversational interviews took substantially longer than standardized
interviews (from 80 to 300 percent longer) depending on the number of complicated map
ping situations. It is difficult to determine how much of a cost issue this is for surveys be
cause the frequency of complicated mappings across a survey population and for a partic
ular topic domain is usually unknown, and conversational partners (here, survey inter
viewers and respondents) often don’t recognize that they might have misaligned interpre
tations (Schober, 2005).
The findings from these studies replicate in studies of surveys using automated interview
ing agents, as opposed to human interviewers. Using the same fictional scenarios from
the Schober and Conrad (1997) study, Conrad, Schober, and Coiner (2007) demonstrated
that laboratory respondents in a self-administered web (browser-based) survey that al
lowed respondents to click for clarification of survey terms provided more accurate re
sponses when answering about complicated mapping scenarios than did respondents who
were unable to click for clarification. Respondents answered even more accurately for
complicated mapping scenarios when the automated system could also volunteer clarifi
cation even if the respondent hadn’t explicitly requested it, based on long delays before
Page 19 of 30

responding. Similarly, Ehlen et al. (2007) showed that respondents in an automated

speech dialogue survey system who could request clarification or who were provided with
clarification when they took too long to answer provided more accurate responses for
complicated mapping scenarios than did respondents who did not have access to clarifi
cation dialogue.
Conversational interviewing seems to improve comprehension in much the same way, not
only in surveys with nonsensitive fact-based questions but also when respondents answer
survey questions about their opinions, where it is less immediately evident what an accu
rate answer would be. As Hubbard, Antoun, and Conrad (2012) demonstrated, conversa
tional interviewing improved response accuracy for attitude questions about drowsy dri
ving (e.g., “In your opinion, how risky is drowsy driving? Would you say that it is extreme
ly risky, very risky, somewhat risky, not too risky, or not at all risky?”) and the economy
(e.g., “We are interested in how people are getting along financially these days. Would
you say that you [and your family living there] are better off or worse off financially than
you were a year ago?”). The experimental (p. 432) questions were administered twice, ini
tially with standardized techniques and subsequently with either standardized or conver
sational techniques that allowed clarification about the labels on the response scale (e.g.,
“By ‘extremely risky’ we mean it causes an accident every time you do it. By ‘not at all
risky’ we mean it never causes an accident”) or about the key concept in the question
(e.g., “By ‘better off financially’ we mean that you and your family have more left over at
the end of each month than you did a year ago. Or, if you’re spending more than you’re
earning, that you’re adding less debt each month than you did a year ago”). The critical
variable was response change between initial and subsequent administration, mirroring
the logic used in Conrad and Schober (2000). When the questions were readministered
with conversational interviewing, responses changed 18.1 percent of the time versus 12.7
percent under standardized readministration.
The evidence is thus that across both interviewer- administered and automated self-ad
ministered surveys, involving both behavioral and opinion questions, the ability to ground
meaning through additional dialogue measurably improves respondents’ comprehension
and ultimately response accuracy. The inability to ground meaning in strictly standard
ized surveys has corresponding harmful effects on comprehension and thus response ac
curacy.
Evidence About Comprehension in Survey Dia

logue
As these studies show, when interviewers actively diagnose respondents’ need for clarifi
cation, they can improve response accuracy even more than when they rely only on re
spondents to explicitly request clarification, which in turn improves response accuracy
relative to when they cannot provide clarification at all. This mirrors what happens in
everyday conversations, in which people not only wait to be asked for clarification but al
so actively monitor each other closely for evidence of whether they are being understood
Page 20 of 30

and adjust what they say accordingly. The evidence people rely on is not only verbal but
paralinguistic and visual, and it is used systematically enough that, as we will see, inter
viewers and automated interviewing systems can potentially make use of it. But people al
so vary in how sensitive they are in everyday conversations to evidence of each other’s
comprehension, and they vary in how they indicate their own state of comprehension:
whether they are prone to explicitly asking for clarification as opposed to relying on oth
ers to detect their need for clarification and when they produce which kinds of cues. This
variability is also present in survey interviews.
In survey interviews and interactions with automated interviewing systems, many respon
dents do not explicitly request clarification even when other evidence suggests that they
need it (Conrad & Schober, 2000; Schober & Conrad, 1997). They probably are relying on
what Clark and Schober (1992) called the “presumption of interpretability”: they assume
that researchers have designed questions with them in mind and that if the researchers
meant something different by a question than what it appears to the respondent to mean,
then the researcher would have qualified the question. Nonetheless, respondents have
been shown to display paralinguistic and visual evidence that they need clarification even
when they don’t explicitly request it.
For example, in an analysis of the laboratory telephone interviews from Schober and Con
rad (1997), in which respondents answered questions about fictional scenarios so that the
researchers could know when respondents were encountering a complicated mapping sit
uation that was likely to require clarification, respondents’ answers were more likely to
include a filler (an um or uh) in answers to questions about complicated scenarios than in
answers about straightforward scenarios (Schober & Bloom, 2004). Schober, Conrad,
Dijkstra, and Ongena (2012) replicated this finding and also extended it to examine visual
evidence of need for clarification in a corpus of Dutch face-to-face and telephone inter
views in which students answered questions about their own lives rather than fictional
scenarios. Answers that were disfluent (that included fillers and other disfluencies) and
answers during which respondents looked away from the interviewer (what has been
called “gaze aversion”) were significantly more likely to be unreliable—that is, to change
within a question-answer sequence in the interview or to change when respondents were
provided with definitions for survey terms post interview. In both studies, the interview
ers’ ability to provide clarification (i.e., in administering conversational vs. standardized
interviews) affected respondents’ likelihood of providing disfluent answers and answers
with gaze aversion. That is, while answering in conversational interviews, respondents
were more likely to be disfluent and avert their gaze from the interviewer when their an
swers turned out to be unreliable. This demonstrates that the cues of need (p. 433) for
clarification that a respondent produces can be affected by the interviewer’s ability to re
spond to those cues, highlighting the depth of interactivity at play in survey dialogue.
Similar kinds of interactivity are also at play in dialogue with automated survey systems.
In the Conrad, Schober, and Coiner (2007) study, in which laboratory respondents inter
acted with a browser-based (i.e., self-administered) questionnaire, respondents’ delays in
answering were informative about the extent to which they needed clarification. Conrad
Page 21 of 30

et al. hypothesized that respondent inactivity (no typing, no clicking) in this situation is
likely to reflect comprehension difficulty, and they tested this by programming the survey
interface to provide unsolicited clarification when respondents’ inactivity exceeded a
question-specific threshold (based on median response times for complicated scenarios in
control conditions). In experiment 2, the thresholds were set differently for younger and
older respondents, based on evidence from the cognitive aging literature strongly linking
aging with slowing of behavior and mental processing. The evidence in both experiments
was that response accuracy improved when respondents whose inactivity exceeded the
thresholds were provided with unsolicited clarification; group-based modeling with differ
ent thresholds for different respondents improved accuracy more than a single threshold
for all respondents.
Fig. 26.2 Response accuracy for straightforward and

complicated scenarios across five interviewing condi
tions: (1) no clarification (standardized), (2) respon
dent-initiated clarification, (3) system-initiated clari
fication delivered after generic pause duration (re
spondents can also request clarification), (4) system-
initiated clarification delivered after stereotyped
pause duration, i.e., different durations for younger
and older groups (respondents can also request clari
fication), and (5) required clarification, delivered au
tomatically after every question for all respondents.
These findings were mirrored in the Ehlen et al. (2007) speech dialogue system study, in
which, in addition to pauses (the spoken equivalent of web browser inactivity), the re
searchers modeled other potential speech indicators of need for clarification, including
fillers, restarts, hedges, and the like. Respondent pauses emerged as the best predictor of
need for clarification—but with a twist: not only response times that were particularly
slow but also response times that were particularly fast predicted inaccurate answers.
This led them to identify an ideal range of response times for each question, within which
answers were reliably more likely to be accurate. They demonstrated (see Figure 26.2)
that when the speech dialogue interviewing system embodied these “just right” ranges
(with different ranges for younger and older respondents), then providing unsolicited
clarification when answers were too fast or too slow reliably improved response accuracy
—in fact, just as much as providing (p. 434) clarification all the time (which has the down
Page 22 of 30

side of being irritating and increasing interview duration unnecessarily for all respon
dents).
Together, these findings strongly support the view that respondents’ paralinguistic dis
plays provide systematic evidence of comprehension difficulty, in both speech and text di
alogue systems, and in much the same way that they do in human interviews. Further
more, the very same display can mean different things about different respondents: the
same delay in responding by one respondent (a younger one in both these studies) can be
informative about the need for clarification in a way that the same delay by another re
spondent (an older one in these studies) is not. The same is likely to be true of any other
paralinguistic or visual displays: a filler produced during an answer by a chronically dis
fluent respondent should be less informative than the same filler produced by a highly flu
ent respondent, and gaze aversion during an answer by a respondent who never looks at
an interviewer should mean less than gaze aversion by a respondent who makes continu
al eye contact (see Conrad, Schober, & Dijkstra, 2007). Consistent with a view that takes
the base rates of respondents’ displays into account, Coiner, Schober, and Conrad (2011)
demonstrated in a browser-based lab study that an individual respondent’s general speed
of processing (measured through the Wechsler Adult Intelligence scale [WAIS] subtests at
the individual level, rather than based on broad age groups) can strongly predict respon
dents’ likelihood of requesting clarification and eventual response accuracy.
These studies demonstrate that an automated interviewing system’s ability to detect and
make use of a respondent’s paralinguistic displays can affect eventual response accuracy.
Are the same dynamics at work in human interviews, where interviewers rely on their
judgment (rather than preprogrammed algorithms) about whether and when to intervene
with clarification? The evidence (Hubbard et al., 2012) is that interviewers can vary in
their nonverbal sensitivity, as measured by the Profile of Non-Verbal Sensitivity, or PONS
test (Hall, Murphy, & Schmid Mast, 2006; Rosenthal, Hall, Archer, DiMatteo, & Rogers,
1979) and that this affects their interviewing behavior and ultimately respondents’ accu
racy. Interviewers who scored higher on the PONS were more efficient at communicating
the intended meaning of questions in conversational interviews, taking less time and us
ing fewer words. This is probably because they were less likely to engage in “preemptive
strikes” (i.e., providing definitions before a respondent had a chance to indicate confu
sion or misinterpretation) than were conversational interviewers who scored lower on
this attribute. Presumably, interviewers who are more sensitive to a speaker’s (in this
case a respondent’s) nonverbal cues administer help more judiciously and efficiently,
whereas less sensitive interviewers do not discriminate as well between responses that
do and do not indicate the need for clarification.
Conclusion
Our argument is that pragmatic processes are at play in survey interviews to a far greater
extent than most survey practitioners or researchers (let alone journalists or members of
the general public) recognize. These processes have substantial effects on how respon
Page 23 of 30

dents interpret questions, how survey dialogue unfolds, and what researchers should
make of survey findings. In particular, we have reviewed evidence that:
1. Survey respondents—as they do in everyday conversation—make use of contextual

and pragmatic information to infer a question’s intended meaning.
When respondents answer, they rely on apparently formal features of the questionnaire,
the content of preceding questions, and what they have answered already to interpret the
questions. Unlike researchers, they may not treat the questions as independent, and they
may take into account what they judge to be the researchers’ intentions (e.g., based on
the researcher’s affiliation). In other words, they bring to the survey interaction their full
interpretive capacities from everyday conversation, even when the inferences this leads
them to make are (from the researcher’s perspective) unwarranted. Also, as in everyday
conversation, survey respondents are likely to assume that their interpretation of a sur
vey question is the same as the researcher’s, even though this is not necessarily the case
(in everyday conversation or surveys). The question that a respondent answers may not
be the question that the researcher wanted to ask, nor do the answers provided by differ
ent respondents necessarily pertain to the same behavior.
2. Survey interviews involve more than simply question asking and answering.
Pragmatic processes are involved not only in respondents’ immediate interpretation of

the question, but in subsequent dialogue, whether that dialogue is strictly standardized
or more collaborative. (p. 435) How respondents answer questions—how long they take to
answer, how fluent their answers are, whether they request clarification or not—is infor
mative about the accuracy and reliability of their answers. This all stems from the fact
that the back-and-forth of survey dialogues is related to the back-and-forth of everyday
conversation, in which people assume that they can work with each other collaboratively
to ground their understanding. They use these skills to interpret what happens even in
strictly standardized interviews that are not intended to allow grounding and clarifica
tion; survey respondents try to make sense of the questions they are asked using whatev
er conversational resources they can bring to bear.
It is certainly not our intention to suggest that all survey research is flawed because of
the kinds of pragmatic processes in survey interviews that we are reviewing. But we pro
pose that understanding what happens in survey responding requires more careful atten
tion to how the questions are presented and how the answers are produced in dialogue.
The implication for researchers is that they would be wise to assume that respondents
are engaged in all the pragmatic processes we have reviewed and to design question
naires and interviewing procedures that build on them, rather than ignoring them or
wishing they were not present.
References
Beatty, P. (1995). Understanding the standardized/non-standardized interviewing contro
versy. Journal of Official Statistics, 11, 147–160.
Page 24 of 30

Belli, R. F. (1998). The structure of autobiographical memory and the event history calen
dar: Potential improvements in the quality of retrospective reports in surveys. Memory, 6,
383–406.
Belson, W.A. (1981). The Design and Understanding of Survey Questions. Aldershot: Gow
er.
Bishop, G. F., Oldendick, R. W., & Tuchfarber, R. J. (1986). Opinions on fictitious issues:
The pressure to answer survey questions. Public Opinion Quarterly, 50, 240–250.
Bless, H., Bohner, G., Hild, T., & Schwarz, N. (1992). Asking difficult questions: Task com
plexity increases the impact of response alternatives. European Journal of Social Psychol
ogy, 22, 309–312.
Brown, N. R. (2002). Encoding, representing, and estimating event frequencies: Multiple

strategy perspective. In P. Sedlmeier & T. Betsch (Eds.), Etc.: Frequency processing and
cognition (pp. 37–54). New York: Oxford University Press.
Clark, H. H. (1977). Inferences in comprehension. In D. La Berge & S. Samuels (Eds.), Ba

sic processes in reading: Perception and comprehension (pp. 243–263). Hillsdale, NJ: Erl
baum.
Clark, H. H. (1996). Using language. Cambridge MA: Cambridge University Press.
Clark, H. H., & Haviland, S. E. (1977). Comprehension and the given-new contract. In R.
O. Freedle (Ed.), Discourse production and comprehension (pp. 1–40). Hillsdale, NJ: Erl
baum.
259–294.
Clark, H. H., & Schober, M. F. (1992). Asking questions and influencing answers. In J. M.
Tanur (Ed.), Questions about questions: Inquiries into the cognitive bases of surveys (pp.
15–48). New York: Russell Sage.
Clark, H. H., Schreuder, R., & Buttrick, S. (1983). Common ground and the understanding
of demonstrative reference. Journal of Verbal Learning and Verbal Behavior, 22, 245–258.

22, 1–39.
Coiner, T., Schober, M. F., & Conrad, F. G. (2011). Which web survey respondents are
most likely to click for clarification? Paper presented at the 66th annual conference of the
American Association for Public Opinion Research, Phoenix, AZ.
Conrad, F. G., & Schober, M. F. (2000). Clarifying question meaning in a household tele
phone survey. Public Opinion Quarterly, 64, 1–28.
Page 25 of 30

Conrad, F. G., Schober, M. F., & Coiner, T. (2007). Bringing features of dialogue to web
surveys. Applied Cognitive Psychology, 21, 165–187.
Conrad, F. G., Schober, M. F., & Dijkstra, W. (2007). Cues of communication difficulty in
telephone interviews. In J. M. Lepkowski, C. Tucker, M. Brick, E. de Leeuw, L. Japec, P.
Lavrakas, M. Link, & R. Sangster (Eds.), Advances in telephone survey methodology (pp.
212–230). New York: Wiley.
Dawes, R. M., & T. Smith (1985). “Attitude and Opinion Measurement.” In G. Lindzey & E.
Aronson (Eds.), Handbook of social psychology (vol. 2, pp. 509–566). New York: Random
House.
Ehlen, P., Schober, M. F., & Conrad, F. G. (2007). Modeling speech disfluency to predict
conceptual misalignment in speech survey interfaces. Discourse Processes, 44(3), 245–
265.
Erikson, R. S., Luttberg, N. R., & Tedin, K. T. (1988). American public opinion (3rd ed.).
New York: Macmillan.
Fowler, F. J., & Mangione, T. W. (1990). Standardized survey interviewing: Minimizing in
terviewer-related error. Newbury Park, CA: Sage.
Galesic, M., & Tourangeau, R. (2007). What is sexual harassment? It depends on who
asks! Applied Cognitive Psychology, 21, 189–202.
semantics, 3: Speech acts (pp. 41–58). New York: Academic Press.
Grice, H. P. (1978). Further notes on logic and conversation. In Cole, P. (Ed.), Syntax and
semantics, 9: Pragmatics (pp. 113–128). New York: Academic Press.
Hall, J. A., Murphy, N. A., & Schmid Mast, M. (2006). Recall of nonverbal cues: Exploring
a new definition of interpersonal sensitivity. Journal of Nonverbal Behavior. 30, 141–155.
Higgins, E. T., Fondacaro, R. A., & McCann, C. D. (1982). Rules and roles: The “communi
cation game” and speaker-listener (p. 436) processes. In W. P. Dickson (Ed.), Children’s
oral communication skills (pp. 289–312). New York: Academic Press.
Houtkoop-Steenstra, H. (2000). Interaction and the standardized survey interview: The

living questionnaire. Cambridge, UK: Cambridge University Press.
Hubbard, F., Antoun, C., & Conrad, F. (2012). Conversational interviewing, the compre
hension of opinion questions and nonverbal sensitivity. Paper presented at the annual
meeting of the American Association for Public Opinion Research.
Igou, E. R., Bless, H., & Schwarz, N. (2002). Making sense of standardized survey ques
tions: The influence of reference periods and their repetition. Communication Mono
graphs, 69, 179–187.
Page 26 of 30

Menon, G. (1994). Judgments of behavioral frequencies: Memory search and retrieval

strategies. In N. Schwarz & S. Sudman (Eds.), Autobiographical memory and the validity
of retrospective reports (pp. 161–172). New York: Springer Verlag.
Menon, G., Raghubir, P., & Schwarz, N. (1995). Behavioral frequency judgments: An ac
cessibility-diagnosticity framework. Journal of Consumer Research, 22, 212–228.
Norenzayan, A., & Schwarz, N. (1999). Telling what they want to know: Participants tailor
causal attributions to researchers' interests. European Journal of Social Psychology, 29,
1011-1020.
Rosenthal, R., Hall, J. A., Archer, D., DiMatteo, M. R., & Rogers, P. L. (1979). The PONS
test: Measuring sensitivity to nonverbal cues. In S. Weitz (Ed.), Nonverbal communica
tion: Readings with commentary, 2nd ed. New York: Oxford University Press.
Schaeffer, N. C. (1991). Conversation with a purpose—or conversation? Interaction in the

standardized interview. In P. P. Biemer, R. M. Groves, L. E. Lyberg, N. A. Mathiowetz, & S.
Sudman (Eds.), Measurement errors in surveys (pp. 367–391). New York: Wiley.
Schiffer, S. (1972). Meaning. Oxford, UK: Clarendon Press.
Schober, M. F. (1999). Making sense of questions: An interactional approach. In M. G.

Sirken, D. J. Hermann, S. Schechter, N. Schwarz, J. M. Tanur, & R. Tourangeau (Eds.),
Cognition and survey research (pp. 77–93). New York: John Wiley & Sons.
Schober, M. F. (2005). Conceptual alignment in conversation. In B. F. Malle & S. D.

Hodges (Eds.), Other minds: How humans bridge the divide between self and others (pp.
239–252). New York: Guilford Press.
Schober, M. F., & Bloom, J. E. (2004). Discourse cues that respondents have misunder
stood survey questions. Discourse Processes, 38, 287–308.
Schober, M. F., & Conrad, F. G. (1997). Does conversational interviewing reduce survey
measurement error? Public Opinion Quarterly, 61, 576–602.
Schober, M. F., & Conrad, F. G. (2002). A collaborative view of standardized survey inter
views. In D. W. Maynard, H. Houtkoop-Steenstra, N. C. Schaeffer, & J. van der Zouwen
(Eds.), Standardization and tacit knowledge: Interaction and practice in the survey inter
view (pp. 67–94). New York: Wiley.
Schober, M. F., Conrad, F. G., & Bloom, J. E. (2000). Clarifying word meanings in comput
er-administered survey interviews. In L. R. Gleitman & A. K. Joshi (Eds.), Proceedings of
the 22nd Annual Conference of the Cognitive Science Society (pp. 447–452). Mahwah, NJ:
Lawrence Erlbaum Associates.
Schober, M. F., Conrad, F. G., Dijkstra, W., & Ongena, Y. P. (2012). Disfluencies and gaze
aversion in unreliable responses to survey questions. Journal of Official Statistics, 28(4),
555–582.
Page 27 of 30

Schober, M. F., Conrad, F. G., & Fricker, S. S. (2004). Misunderstanding standardized lan
guage in research interviews. Applied Cognitive Psychology, 18, 169–188.
Schuman, H., & Duncan, O. D. (1974). Questions about attitude survey questions. In H. L.
Costner (Ed.), Sociological methodology. San Francisco: Jossey-Bass.
Schuman, H., & Presser, S. (1981). Questions and answers in attitude surveys. New York:
Academic Press.
Schwarz, N. (1994). Judgment in a social context: Biases, shortcomings, and the logic of
conversation. Advances in Experimental Social Psychology, 26, 123–162.
Schwarz, N. (1996). Cognition and communication: Judgmental biases, research methods

and the logic of conversation. Hillsdale, NJ: Erlbaum.
Schwarz, N. (1999a). Self-reports: How the questions shape the answers. American Psy
chologist, 54, 93–105.
Schwarz, N. (1999b). Frequency reports of physical symptoms and health behaviors: How
the questionnaire determines the results. In Park, D. C., Morrell, R. W., & Shifren, K.
(Eds.), Processing medical information in aging patients: Cognitive and human factors
perspectives (pp. 93–108). Mahaw, NJ: Erlbaum.
Schwarz, N., & Bienias, J. (1990). What mediates the impact of response alternatives on
frequency reports of mundane behaviors? Applied Cognitive Psychology, 4, 61–72.
Schwarz, N., Bless, H., Bohner, G., Harlacher, U., & Kellenbenz, M. (1991). Response
scales as frames of reference: The impact of frequency range on diagnostic judgment. Ap
plied Cognitive Psychology, 5, 37–50.
Schwarz, N., Grayson, C. E., & Knäuper, B. (1998). Formal features of rating scales and
the interpretation of question meaning. International Journal of Public Opinion Research,
10, 177–183.
Schwarz, N., & Hippler, H. J. (1991). Response alternatives: The impact of their choice
and ordering. In P. Biemer, R. Groves, N. Mathiowetz, & S. Sudman (Eds.), Measurement
error in surveys (pp. 41–56). Chichester: Wiley.
Schwarz, N., Hippler, H. J., Deutsch, B., & Strack, F. (1985). Response scales: Effects of
category range on reported behavior and subsequent judgments. Public Opinion Quarter
ly, 49, 388–395.
Schwarz, N., Knäuper, B., Hippler, H. J., Noelle-Neumann, E., & Clark, F. (1991a). Rating
scales: Numeric values may change the meaning of scale labels. Public Opinion Quarterly,
55, 570–582.
Schwarz, N., & Oyserman, D. (2001). Asking questions about behavior: Cognition, com
munication and questionnaire construction. American Journal of Evaluation, 22, 127–160.
Page 28 of 30

Schwarz, N., & Scheuring, B. (1992). Selbstberichtete Verhaltens- und Symptomhäu

figkeiten: Was Befragte aus Anwortvorgaben des Fragebogens lernen. [Frequency-reports
of psychosomatic symptoms: What respondents learn from response alternatives.]
Zeitschrift für Klinische Psychologie, 22, 197–208.
Schwarz, N., Strack, F., & Mai, H. P. (1991b). Assimilation and contrast effects in part-
whole question sequences: A conversational logic analysis. Public Opinion Quarterly, 55,
3–23.
Schwarz, N., Strack, F., Müller, G., & Chassein, B. (1988). The range of response alterna
tives may determine the meaning of the question: Further evidence on informative func
tions of response alternatives. Social Cognition, 6, 107–117.
Stalnaker, R. C. (1978). Assertion. In P. Cole (Ed.), Syntax and semantics, Vol. 9: Pragmat
ics (pp. 315–332). New York: Academic Press.
Strack, F., Schwarz, N., & Wänke, M. (1991). Semantic and pragmatic aspects of
(p. 437)
context effects in social and psychological research. Social Cognition, 9, 111–125.
Suchman, L., & Jordan, B. (1990.) Interactional troubles in face-to-face survey interviews.
Journal of the American Statistical Association, 85, 232–253.
Suchman, L., & Jordan, B. (1991). Validity and the collaborative construction of meaning
in face-to-face surveys. In J. M. Tanur (Ed.), Questions about questions: Inquiries into the
cognitive bases of surveys (pp. 241–267). New York: Russell Sage Foundation.
Suessbrick, A. L., Schober, M. F., & Conrad, F. G. (2000). Different respondents interpret
ordinary questions quite differently. Proceedings of the American Statistical Association,
Section on Survey Research Methods. Alexandria, VA: American Statistical Association.
Tourangeau, R., & Smith, T.W. (1996). Asking sensitive questions. The impact of data col
lection, mode, question format, and question context. Public Opinion Quarterly, 60,
275-304.
Winkielman, P., Knäuper, B., & Schwarz, N. (1998). Looking back at anger: Reference pe
riods change the interpretation of emotion frequency questions. Journal of Personality
Notes:
(1) . For example, web surveys can allow respondents to request clarification with a
mouse (e.g., Conrad, Schober, & Coiner, 2007), and speech dialogue interviewing systems
(speech-IVR, interactive voice response) can, in principle, recognize spoken requests for
clarification or even diagnose paralinguistic evidence of confusion or misunderstanding
(see Schober, Conrad, & Bloom, 2000; Ehlen, Schober, & Conrad, 2007).
Page 29 of 30

Frederick G. Conrad
Frederick G. Conrad is Research Professor and Director, Michigan Program in Sur

vey Methodology, University of Michigan and Joint Program in Survey Methodology,
University of Maryland.
Michael F. Schober
Michael F. Schober is Professor of Psychology at the New School for Social Research.
Norbert Schwarz
Norbert Schwarz is Provost Professor in the Department of Psychology and Marshall

School of Business at the U of Southern California.
Page 30 of 30

Language and the Law: Illustrations from Cases of Disputed Sexual Consent
Language and the Law: Illustrations from Cases of Dis

puted Sexual Consent
Deborah Davis and J. Guillermo Villalobos
Print Publication Date: Sep 2014 Subject: Psychology, Social Psychology, Forensic Psychology
Online Publication Date: Jun 2014 DOI: 10.1093/oxfordhb/9780199838639.013.014
This chapter aims to provide a general review of issues psychology has addressed regard
ing language and the law. It focuses on three primary areas of application in legal set
tings: (1) encoding and memory of language and conversation; (2) language, trickery, and
persuasion; and (3) language as it affects classification and judgment. The authors illus
trate these issues largely with the example of disputed sexual consent and further point
the reader to resources on language and law addressed more fully by other disciplines
(e.g., the Oxford Handbook of Language and the Law: Tiersma & Solan, 2012a) or in oth
er chapters of this volume. The chapter begins with a review of how statements are en
coded and remembered because these issues provide the basis for understanding the use
of language for trickery and persuasion, as well as for classification and judgment.
Keywords: Language, law, encoding, memory, conversation, persuasion, classification, judgment, sexual consent
Encoding and Memory of Language and Con

versation
Social psychologists have devoted considerable attention to issues of comprehension and
memory of language in legal settings. Comprehension of literal meaning has been investi
gated primarily as it applies to understanding of legal rights to silence among criminal
suspects, of jury instructions among jurors, and failures of successful communication in
such areas as contract disputes or professional communication with clients. Implied
meaning has been investigated in a range of contexts, such as manipulative use of prag
matic implication among law enforcement; commission of language crimes, such as
bribery or solicitation, through indirect speech; negotiation of sexual encounters as ap
plied to claims of sexual harassment or sexual assault; broader inferences regarding
speaker and hearer (such as competence or culpability); and others.
Page 1 of 35

Although the language at issue is sometimes written, for many disputes, the communica
tions occurred in conversations between interested parties. Although memory for precise
wording is often crucial in legal contexts, research on memory for conversational state
ments has shown that such memories tend to be for the gist of the main points of what
was said rather than for exact wording. Moreover, they are subject to the same sources of
failures and distortion as memories for people, objects, or events (see Davis & Friedman,
2007; Davis, Kemmelmeier, & Follette, 2005, for reviews). Yet, in some respects, sources
of such failures for conversational statements can be more complicated. This is particu
larly true with respect to how meaning is encoded.
As context for our discussion of legal issues regarding comprehension and memory of
both written language and conversation, we first review the complications of accurate en
coding of written (p. 439) and oral statements. We then illustrate these issues of compre
hension and memory with the example of memory reports of sexual interactions and the
issue of consent.
Levels of Encoding
Statements are encoded on at least seven levels, each subject to failure or inaccuracy
(see Davis & Friedman, 2007; Holtgraves, 2002; Turnbull, 2003). First, sounds must be
correctly perceived and parsed into phonemes and words that, second, must be under
stood as words, syntax, and sentence structure. Successful encoding at this level can be
compromised by personal failures of perceptual abilities, by difficulties in understanding
unfamiliar accents or languages, or because of environmental interference with hearing
or vision (that might otherwise clarify nonverbal cues).
Third, one must understand the propositions of the utterance (e.g., “It is 12 o’clock.”).
Tracking simple assertions can sometimes be difficult in that speakers may be unclear,
lengthy, and disorganized and hearers (particularly where speech is in a second language
for them) may understand some words and not others. Moreover, statements tend to be
encoded in terms of perceived meaning (“It is time to go”) rather than exact wording
(“It’s 12 o’clock.”), which can be the first level at which incorrect memory (vs. memory
failure) can become a problem. One interpretation may reflect criminal action or inten
tion, but a related one not. A man may say to a woman “You’re clearly asking for it! I’m
gonna teach you a lesson!” (a clear threat or intention). But a similar statement, such as
“You’re clearly asking for it! Someone needs to teach you a lesson,” which may be simply
a statement of opinion, may be remembered equivalently and reported as the speaker
threatening the woman or stating a clear intention to teach her a lesson.
Fourth, one may understand words and sentences correctly, but not understand what is
being talked about. Bransford and Johnson (1972) famously demonstrated complete fail
ure to understand the meaning of a paragraph describing the process of doing laundry
unless participants were first told the paragraph was about “doing laundry.” All words
and sentences are comprehensible on their own, but one still has no sense of understand
Page 2 of 35

ing what the referents are (e.g., “First, you arrange things into different groups.”) with
out the defining context. What things?
The hearer may fail to understand simple referents such as “he” or “she”—and therefore
be unclear about who is being talked about—but may also misunderstand or misinterpret
more complicated referents. This can occur because the speaker fails to provide needed
contextual information to the hearer or because the hearer doesn’t have sufficient contex
tual information or knowledge or fails to consider the perspective of the speaker ade
quately when interpreting the remarks. These are a few of the many forms of failure of
the “common ground” (shared knowledge and context) necessary for hearers to under
stand remarks the way speakers intend.
Among the more spectacular failures of common ground applicable to the legal system
are those posed by discrepancies between the meaning of legal language as intended by
those writing contracts or formulating instructions to trial juries versus lay understand
ing of the words, concepts, sentences, and intended use in practice (see section on “Com
prehension and the Jury”). Among lay persons and jurors, much legal language is either
not understood at all or is misunderstood in ways that may promote severe negative con
sequences (sometimes the opposite of what is intended); for example, the imposition of
the death penalty. Such misunderstandings have prompted a “plain language” movement
among legal professionals and language scholars (Adler, 2012).
Failures of common ground can be more profound for those who overhear conversations
and lack much of the history between the original participants, their shared knowledge,
the conversational context, what they are talking about, who is being spoken to, and oth
er crucial elements of common ground. Judges and juries, for example, must often deter
mine the meaning of recorded conversations for which they lack such significant features
of common ground between the conversational participants. Speakers may also use delib
erate strategies to cause failures of common ground in either the person they are talking
to or those who may witness or listen to recordings of the conversation later. Undercover
agents, for example, sometimes strive to elicit seemingly incriminating admissions or in
formation from suspects who are misled to assume they are talking about something dif
ferent from the appearance the agent intends to create among those who will later listen
to and judge the exchange (see section in this chapter on law enforcement).
Fifth, the hearer may understand all previous levels, but fail to understand what the
speaker actually intends to communicate or what “speech act” (Austin, 1962; Grice, 1975;
Searle, 1969) the speaker intends to accomplish. Speech act theory (p. 440) identified the
manner in which intended meaning can differ from literal propositional content. The in
tended meaning is communicated “indirectly” or by “pragmatic implication,” whereby the
listener interprets the intended (implied) meaning based on contextual cues (e.g., tone of
voice or facial expression) and understanding of normal “rules” of conversation (e.g.,
Grice, 1975). Irony, for example, is often accomplished by stating the opposite of what is
meant (violating the conversational maxim to tell the truth), and the speaker’s meaning
can be clarified by ironic tone or facial expression.
Page 3 of 35

Indirect communications are inherently ambiguous in that they may be interpreted liter
ally or according to one or more plausible intended meanings. For example, if a wife ar
rives home from work and says in front of her husband and her complaining mother “I
have a splitting headache,” she may be stating a simple fact, commenting on her day at
work, signaling her husband not to expect sex, telling her mother to shut up, or making
an indirect request for her husband to get her an aspirin, among other possibilities. Such
multiple possible meanings raise the risk that indirect communications may not be under
stood (or remembered) as intended.
Such ambiguities give rise to disputes in legal settings over such issues as whether cru
cial information was conveyed (Did the doctor’s statement “If you take this drug with oth
ers such as X it can cause a fatal overdose” constitute adequate warning or instruction
not to take the drugs together?), whether an order was given (Is “We need to make sure
that X is done by the end of the day” an order or a simple statement of fact or opinion?),
whether a message was or was not communicated (Is the indirect statement “I can’t af
ford to risk getting pregnant” a refusal of sex or an advisement of how to have sex?),
whether incentives were offered (by those accused of bribery), whether an oral contract
was/was not made, and many others.
Additional problems occur when such indirect communications become targets of legal
dispute, particularly when no recording exists to clarify them. Since the hearer’s interpre
tation (rather than literal content) of indirect statements tends to be remembered, judges
and juries will typically either never hear the actual content of the speaker’s statements
or be faced with divergent accounts of what was said. As noted in the section in this chap
ter on police interrogation, this is often the case when unrecorded content of police inter
rogations is litigated. Suspects are likely to “remember” having been explicitly threat
ened or having been made promises contingent on confession, whereas officers who con
veyed such inducements indirectly will deny having made them at all.
If there is a recording, judges and juries hear the indirect speech act(s) in the context of
specific claims regarding intended meaning, with significant potential to bias interpreta
tions to differ from the speaker’s intended meaning. This is particularly problematic in
that the full context in which statements are made is not typically available to those who
must judge after the fact (see section on the language of trickery and deceit). Participants
are not always visible, for example, thus depriving those who judge of important nonver
bal cues to meaning. The recordings may be audio only, and, for police interrogations,
video recordings often offer a view of the suspect only while the interrogator is not visible
(Lassiter, Ware, Lindberg, & Ratcliff, 2010).
The sixth level of interpretation for conversational statements is that of relational commu
nication, which concerns intended and unintended messages about the relationship be
tween participants, including how speakers see themselves and the hearers, their under
standing of their relationships within and outside the immediate context, how they want
the relationships to progress (or not) in the future, and so on. Central to relational mes
sages are status and solidarity (liking, closeness). Misunderstandings of relational mean
Page 4 of 35

ing are often at the heart of both criminal and civil cases (Davis & Friedman, 2007). Mis
interpreted implications for sexual or romantic intentions are central to claims of ac
quaintance rape or sexual harassment (see the section “A Microcosm of Miscommunica
tion”). Similarly, misinterpreted relational messages can be at the heart of student or em
ployee complaints. A very problem-focused criticism (e.g., “This argument isn’t very
clear”) can feel personal to the student and therefore later be remembered as a more ad
hominem criticism (e.g., “You just don’t know how to make a good argument”) or worse
(“You’re just stupid!”). Such memory failures can fuel hostility and lead to formal com
plaints against a professor or employer.
The final level at which conversational statements (or conversations as a whole) are en
coded concerns broader impressions or inferences based on their content, including at
tributes of the participants (such as likeability, status, competence and others), their atti
tudes toward one another (“He didn’t like her (me)”), the overall emotional tone of the
conversation (“They (We) were fighting!”), the likely hearer (“He seemed to be talking to
a child”), whether the statements are truthful, and many (p. 441) other inferences (see the
section on “Classification and Judgment”).
Among the important legal issues raised by such broad inferences is that of the impact of
suggestive questioning on witness reports. The massive social science literature address
ing suggestive questions has shown that to the extent statements or questions presume
information, the hearer is likely to assume the speaker believes the information is the
case, to therefore believe the same themselves (or to believe the speaker wants them to
believe and report it), and to offer memory reports consistent with the suggestions. Eliza
beth Loftus’s classic research on the misinformation effect (e.g., Loftus & Palmer, 1974;
see Loftus, 2005, for review) showed just such phenomena. Participants asked whether
they saw “the” stop sign were more likely to answer “yes” (although there was none) than
were those asked about “a” stop sign. When asked how fast the car was going when it
“smashed” into another, participants reported greater speed than when the words “hit” or
“contacted” were used. Since that time, suggestive questioning has been implicated in
the development of more dramatic false memory reports—such as for child sexual abuse,
alien abduction, satanic ritual sexual abuse, and many others (see Loftus & Davis, 2006).
Together, the many subjective judgments necessary to process conversation—such as

searching common ground to determine what is being talked about, selecting direct ver
sus indirect meanings, understanding relational meanings, and forming broader impres
sions of participants and their intended messages—offer substantial potential for error
and bias in encoding and memory for conversation and for interpretation of its broader
meaning. Indeed, adding to contextual influences on such judgments, individual differ
ences in interpretation of language exist in contexts ranging from gender or personality
differences in understanding of sexual intentions (Abbey, 2011; Koukounas & Letch,
2001) to relationships of attitudes and personality to comprehension of jury instructions
(see section “Comprehension and the Jury”).
Page 5 of 35

For conversational participants as well as those who judge them, multiple contextual fea
tures are necessary to accurately interpret conversational meaning at all levels (from
speech perception through relational and other inferences; Davis & Friedman, 2007).
These include the ability to see both participants clearly (in part to see what the speaker
is responding to); absence of personal impairments, distractions, or interference with per
ception; access to nonverbal signs that clarify or alter interpretation (such as tone of
voice, gestures, facial expressions, and others); knowledge of the general context for
statements, including what and who are present, when the conversation is taking place,
nature of the environment, preceding behaviors and statements of others, knowledge of
the recent and historical context of current remarks, personal acute states and chronic
characteristics of those present (and what each believes of others), the general purpose
and goals of the conversation and parties, and much more. Absent any one of these, accu
racy of encoding, memory, impressions, and inferences can be impaired. But, even if all
relevant context is known, perceivers must also know how meaning should be adjusted
based on it. A tall order indeed, particularly for those who did not participate in the origi
nal conversation but must nevertheless decide important legal outcomes based on inter
preted meaning. In sections to come, we illustrate many of these problems of encoding
and memory with the example of disputed claims of date rape, which we believe incorpo
rate all or most of the problems just discussed.
A Microcosm of Miscommunication: Communi

cating, Interpreting, and Remembering Sexual
Intentions and Consent
The areas of legal application best illustrating the multiple levels of failure of communica
tion are, arguably, those involving disputed sexual/romantic intentions and consent. The
primary issue is often that of whether and how inappropriate invitations or sexually coer
cive behaviors were performed by the accused and/or whether and how the victim clearly
communicated lack of consent to the accused. Relevant communications often unfold over
time; involve multiple (often intentionally) indirect communications entailing significant
underlying relational meanings; include important nonverbal behaviors as primary mes
sages and as contextual clarifiers of verbal messages; take place in the context of commu
nicative and relational histories relevant to interpretation of current communications; oc
cur in the context of intoxication of accused and/or accuser; and become subject to
schematic, hindsightful, and motivational distortions and other errors in perceived mean
ing by speaker, hearer, attorneys, judges, and jurors as they become the centerpiece of le
gal disputes. These (p. 442) characteristics can first contribute to events such as date rape
or sexual harassment and later cause problems of memory and reporting among witness
es and errors of judgment in those who decide legal outcomes.
Page 6 of 35

Indirect Meaning and the Communication of Desire and Refusal
In their theory of indirect speech and the “strategic speaker,” Lee and Pinker (2010)
specified conditions under which the use of “off-record” indirect speech acts are most
likely. “Off-record” indirect speech refers to statements offering plausible deniability of
the intended meaning, as opposed to “on record” indirect statements for which meaning
is conventionalized or otherwise perfectly clear. The authors suggested that deniability
tends to motivate the use of “off-record” indirect speech acts under two primary condi
tions: (1) when there is ambiguity concerning how the hearer will respond to a direct
statement and (2) when potentially negative responses are severe, such as rejection, em
barrassment, or criminal charges. Indirect speech becomes more likely when the commu
nication can pose a threat to the face of either speaker or hearer, and particularly when
the communication would suggest a change in social identity (as when a bribe suggests
an identity of “dishonest cop”) or in the nature of relationship between speaker and hear
er (such as a change from nonsexual to sexual).
The authors provided examples of strategic use of indirect speech to avoid definitive com
munication of illegal intent (such as offering bribes, solicitation of another to commit a
crime, blackmail, veiled threats, perjury, and others) or potentially embarrassing interper
sonal suggestions (such as sexual come-ons). Although such indirect communications typ
ically can be easily understood in terms of their intended meaning, it is not the case that
they must be so interpreted, thereby offering some protection from negative legal or so
cial consequences.
Moreover, even if the intended meaning will be perfectly clear to speaker and hearer, the
speaker may use indirect speech to maintain “higher order” plausible deniability beyond
the immediate hearer (e.g., if content might be judged negatively or might provide the ba
sis of legal sanctions; Lee & Pinker, 2010; see also Tiersma & Solan, 2012b). Lee and
Pinker (2010) further demonstrated that receptive hearers operate at lower thresholds
for accurate perception of the intended bribe, sexual invitation, and the like, such that
the risk that indirect meaning will not be understood is low. However, a higher standard
of clarity and directness is required for the original target or for those judging alleged
criminality of the statements to respond with highly negative personal reactions or legal
sanctions. In this way, indirect speech can clearly elicit the cooperation of a receptive
hearer while minimizing the risk of negative consequences from an unreceptive one (or
by prosecutors, judges, or juries). Hence, potentially unwelcome or illegal suggestions
are often stated indirectly, thereby giving rise to a host of legal cases hinging on their dis
puted interpretation.
Prominent among them are cases involving disputed sexual intentions or consent (see al
so section on law enforcement). Attempts to initiate romantic or sexual activities are of
ten communicated indirectly, via hints and other means that provide plausible deniability
in the event of rejection (or legal claims of sexual harassment). Indeed, Lee and Pinker
(2010) found that indirect sexual advances are viewed as more likely to allow a relation
ship to continue as friends in the event the target rejects the advance. Similarly, attempts
Page 7 of 35

to reject such advances are often communicated indirectly in order to avoid unambiguous
and potentially hurtful personal rejection of the initiator (Woodhams, Hollin, Bull, &
Cooke, 2012). The rejector may change the subject, offer reasons why he or she can’t
accept, pretend not to notice an indirect invitation or hint, use nonverbal distancing
rather than explicit verbal refusals of physical advances, and otherwise use indirect
means of communicating refusal. In this way, an accused sexual harasser may be con
vinced no unwelcome advances occurred (and indeed may not have intended sexual
meaning), whereas the complaining “victim” clearly perceived repeated unwelcome re
quests. A man may have sex with a woman who feels she clearly indicated that he should
stop, even though he feels she consented. Such miscommunications can be fueled by mo
tivated biases in perceptions of the relational implications of communications. Men, for
example, have been shown to perceive more sexual interest in the communications of
women than the women themselves intend (e.g., Abbey, Jacques-Tiura, & LeBreton,
2011). Moreover, each may remember and report the communications involved in the
contested events in a manner distorted toward their own motivations and perspective.
Remembering Communications of Interest and Consent
Memory for precise language is of particular relevance to cases involving allegations of

rape, sexual (p. 443) assault, or sexual harassment between acquainted parties. A number
of memory issues are raised by the questions of whether advances were or were not made
and whether consent was or was not present, many unique to memory for conversation
and specific wording. We illustrate this here with respect to allegations of acquaintance
rape.
There is a great deal of information to be remembered relevant to establishing lack of

consent. This may include contextual information regarding the time period preceding
the alleged assault; many details of specific statements and behaviors of each party pre
ceding and during the alleged assault; emotions; reactions to and interpretations of the
other’s behaviors and statements; their own intentions; and other subjective thoughts (as
they occurred at the time). Many details become important only in hindsight, after the ac
cuser has reported the incident, the legal implications of the issues have become high
lighted, and both are asked to recall the many relevant details. What is not attended to at
the time it occurs is unlikely to be remembered later, and what is not considered impor
tant at the time is less likely to be attended to.
Second, whereas memory for exactly how (and how effectively) consent or nonconsent
was conveyed (such that a reasonable person would have understood) is crucial for litiga
tion of claims of sexual assault or harassment, memory tends to be for gist rather than for
exact detail—particularly so for conversation, exact wording, or nonverbal behaviors.
Thus, memory reports regarding indicators of consent based on the “gist” of what was
done to convey consent can appear more definitive than what was actually done at the
time.
Page 8 of 35

Victim reports of what was said or done to convey nonconsent may be conveyed in gener
al terms—such as “I made it clear that I wanted him to stop.” Such reports state an al
leged outcome (“I clearly conveyed that I wanted him to stop”) of unspecified statements
or behaviors rather than exact statements or behaviors that presumably led to the alleged
outcome (“I kicked him in the nuts and screamed at him to leave me alone”). The alleged
perpetrator may likewise remember in general that the alleged victim “acted like she
wanted it” or “didn’t do anything to indicate she wanted me to stop.” The former charac
terization can reflect incorrect relational attributions drawn from behaviors the accuser
meant as “friendly,” “polite,” or “nice”—such as smiling, expressing interest in what the
accused was saying, complimenting him.
In such cases, the witness may not be remembering specific behaviors at all. The accuser
may accurately remember that she didn’t like the sexual interaction and assume that the
defendant understood her feelings based on “memories” that can range from the simple
assumption that he would “just know” or that how she felt would have to show overtly, to
memories of nonverbal behaviors of varying subtlety, to those of very clear verbal com
mands to stop. Unfortunately, such general descriptions can be significantly misleading to
prosecutors, judges, and juries who determine important outcomes for the accuser and
accused.
An additional problem of memory arises in that false memories can occur for having actu
ally said what one actually only thought about or had wanted or intended to say, but
didn’t (see Davis & Friedman, 2007, for review). Generally, things actively thought about
or imagined can later be falsely remembered as having actually happened—as shown in a
very large body of research on “imagination inflation.” Such false memories can develop
for the very mundane (such as falsely remembering saying something one only intended
to say) to the seemingly impossible (such as falsely remembering being the victim of sa
tanic ritual abuse spanning multiple years). False memories such as the latter typically
develop through a combination of activities entailing actively picturing and imagining the
events coupled with suggestion from one or more apparently credible sources that the
events did occur (see Loftus & Davis, 2006). Thus, it is very possible for the accuser to
falsely remember saying or doing more to indicate nonconsent than she actually overtly
did, including falsely remembering saying things that she felt. Similarly, the accused may
falsely remember having done more to seek or verify consent than he overtly did. Each
can have false memories based on actively reliving, imagining, and recounting the con
tested events repeatedly in the course of the investigation and prosecution of the case.
Fourth, in cases involving allegations of rape or sexual harassment, issues of shame can
be crucial. Memory is subject to distortion toward consistency with one’s needs and de
sires, including maintenance of self-esteem. To the extent that voluntary sexual activity
with the defendant would be regarded as shameful for any reason, the accuser may mis
remember the encounter as nonconsensual, and the nonconsent as more explicitly com
municated, to avert such feelings. Shame is likely to be an issue for the accused as well.
Emotions and self-justification (p. 444) motives that arise once the alleged rape is report
Page 9 of 35

ed can drive reinterpretation of previous behaviors and reconstructive influences on

memory for both accused and accuser (Tavris & Aronson, 2007).
Fifth, memories of consent can be directly affected by others who apply the linguistic la
bel “rape” to the defendant’s actions. In a contested case of rape within an ongoing rela
tionship, for example, the accuser had initially viewed a sexual interaction as acceptable
and consistent with sexual activities within their relationship, even though it had been
more forceful or aggressive than what might seem normal to others. She began to report
it as “rape” after the relationship had terminated and after having discussions with a
friend who told her the actions she described were clearly “rape.” The application of such
a label in hindsight can lead an accuser to mentally review what happened in this new
light and come to remember it differently, including potentially more aggressive behav
iors by the defendant and more forceful or explicit protests by the accuser than occurred
at the time. The general principle that memory for previous events can be reconstructed
after the fact by the application of a new label or other information or interpretations en
countered after the event has been demonstrated across a large number of memory tar
gets (see Davis & Loftus, 2007, for review).
Sixth, negative feelings toward a particular person, arising either at the time of the
events in question or after the fact, can distort interpretation of that person’s statements
and behaviors. Our original interpretation of events, as well as later memory for those
events, tends to be distorted toward consistency with other relevant beliefs (Brainerd &
Reyna, 2005). If one possesses a generally negative impression of a particular person, his
or her behaviors and statements are generally more likely to be interpreted negatively
(particularly if at all ambiguous). If a negative impression of the person develops later,
the same actions tend to be reinterpreted in hindsight and can appear and be reported in
more negative terms than they were experienced at the time they occurred. The possibili
ty of such reconstructive processes is often raised in cases of sexual assault, in which
women who had relationships or encounters with the defendant in the past testify that he
had previously “raped” them (although they never reported rape at the time or discussed
it as such with others). Although motivations such as hatred, jealousy, revenge, or hurt
feelings may lead to deliberate false reports, the same motivations can also directly affect
interpretation and lead to honest errors of memory for communication of consent.
Seventh, when memory for the original event is unclear—for example, due to intoxication
during the event—the person may unknowingly add details to the “memory” based largely
on what would make sense given everything the person does believe, know, and remem
ber. In cases involving issues of sexual consent, the accuser or accused may unknowingly
misremember what happened based in part on beliefs about what (s)he believes (s)he
would have done in such a situation—including, for example, beliefs about whether (s)he
would have had voluntary sex (or coerced the accuser) with the person in the circum
stances at hand, and how.
This can be particularly problematic when an interviewer wants very specific answers
and the witness has only “gist” memories. Witnesses generally try to answer the ques
Page 10 of 35

tions asked of them, and the pressure to do so becomes greater the more often the inter
viewer asks for specific information or the more he or she tries to get answers from the
witness. If the actual memory is a gist-based account such as “I was very clear that I
didn’t want to have sex,” an accuser asked to say exactly how she made this clear may
well answer based on fuzzy thoughts about how she would have said it (based in part on
vague memories of what happened combined with guesses or assumptions about how it
would have happened). These processes can also be exaggerated through suggestive
questioning. An uncertain interviewee’s answers are more likely to be shaped by cues
from the interviewer regarding what might have or probably did happen.
Finally, alcohol use by one or both parties is a factor in the majority of alleged sexual as
saults. Alcohol can affect the clarity and timing of communication of sexual interest and
(non)consent. Intoxicated women tend to offer less assertive denials of consent and do so
later in the sexual encounter than do those who are not intoxicated, and intoxicated men
are less able to interpret communications of nonconsent accurately due to alcohol’s ef
fects on cognition (see Abbey, 2011; Davis & Loftus, 2004, for reviews). Generally, alcohol
tends to impair encoding processes, rendering memory for the event and its details poor
er and more subject to distortion or influence from others (Soraci et al., 2007).
We (Davis & Loftus, 2004) have suggested that these impairments of memory, combined
with the tendency to guess what happened based on assumptions about what one would
have done in the circumstances or with a particular person provide a (p. 445) potential
pathway to false allegations of rape—one we have seen suggestive evidence of in several
cases of alleged rape. In one such case, an accuser was drunk during the encounter and
suffered a severe head injury from a fall down stairs when she was leaving. Not surpris
ingly, she admitted having no memory of the sexual encounter (which both agreed took
place), but nevertheless alleged rape, stating that she “would not have” voluntarily had
sex with the accused. Notably, although her initial statements were more positive toward
the accused and less suggestive of “rape,” she arrived at this firm claim after the police
interviewer had recounted the accused’s negative reputation and asked in a tone drip
ping disgust “You wouldn’t have voluntarily had sex with him, would you?”
Investigating the Allegations: The “Pretext” Call
Once an allegation of rape (or many other crimes) has been made, police may ask the ac
cuser or other witnesses to call the accused to try to elicit an admission of guilt. Un
known to the accused, police have instructed the accuser concerning exactly what admis
sions to elicit from the accused to support the charges to be filed. The call is recorded
and admissible as evidence against the accused. The caller is to elicit the admissions
while not giving any hint of the real purpose of the call or the fact that it is being record
ed. The accuser will bring up the alleged rape and may be asked to mention her own be
haviors indicating protest and lack of consent (in the hope the accused will acknowledge
they occurred), ask the accuser why he did it (assuming any explanation is a presumptive
admission of guilt), talk about how he hurt her, ask for an apology, and so on. These are
intended at minimum to elicit admissions that sex did occur, that the accuser didn’t con
Page 11 of 35

sent, and that this was clear to the accused. An interesting challenge for callers is to
make clear for later hearers what is being talked about and what it means. Simultaneous
ly, she must not insert context that may be needed by the later hearers but that would
seem suspicious to the accused and cause him to become uncooperative.
Such pretext calls are often subject to battles of interpretation, in part due to missing
contextual information needed to understand what is and is not being admitted to. Non
verbal reactions, for example, are missing in audio recordings. If the accused is silent in
response to an accuser’s assertion, but hearers can’t see his facial expressions of confu
sion, they may view his silence as an “adoptive admission” of the truth of the accusation
(see the section “Language, Classification, and Judgment”). Generally, hearers may miss
nonverbals that negate the verbal responses of the accused, that clearly convey lack of
understanding or recognition of what the accuser is talking about, that indicate what is
agreed to is compliance in the absence of actual belief in its truth, and other important in
dicators of the actual significance of his statements. An apology is considered further evi
dence of nonconsent and rape, although, from the point of view of an accused, it can be
an expression of sympathy and regret for the distress of the accuser (see Shuy, 2005, for
a case example of disputed meaning for such an apology). Moreover, missing context for
understanding what is being discussed can promote misunderstanding. Ambiguities pro
duced by lack of nonverbal and other context leave an open door for expectations and
confirmation biases to influence interpretation among prosecutors, judges, and juries—of
ten to the detriment of the accused. Instead of asking “Must I interpret these statements
as admissions,” they often seem to ask only “Can I interpret them as admissions.” The lat
ter, of course, is a much less stringent criterion for imprisonment.
Comprehension and Assertion of Rights to

Avoid Self-Incrimination
Criminal suspects in the United States, Canada, and jurisdictions in Europe have legal
rights to avoid self-incrimination in interactions with law-enforcement. Here, we focus on
understanding and exertion of rights to avoid questioning and search as they apply in the
United States (see Rock, 2012, for similar information regarding Canadian and European
jurisdictions).
Criminal suspects in the United States have no obligation to speak to police who wish to
question them—nor do those who are not, or do not believe themselves to be, suspects at
the time of the request (see http://www.youtube.com/watch?v=6wXkI4t7nuc for an excel
lent discussion of risks of submitting to police questioning for the innocent). Scholars
have addressed the comprehension of such rights primarily with regard to custodial sus
pects who, prior to questioning or to interrogation while under arrest, are actively ad
vised of their rights under Miranda v. Arizona (1966) to remain silent, to access to an at
torney free of charge (for those who cannot afford to pay), of the intended use of anything
said against the suspect in court, and of the fact that these rights can be claimed at any
time, and, to a lesser extent, have examined (p. 446) comprehension with regard to the
Page 12 of 35

right to refuse consent to search. The later entails no formal reading of rights. Rather, an
officer requests to search a person or his property and may “seize” either (detain the per
son or take property as evidence). A suspect may later contest the voluntariness of a
waiver of Miranda rights or consent to search, which, if successful, will prevent the use of
the results of questioning or search at trial.
For waivers of Miranda rights to be considered valid, they must be made knowingly, intel
ligently, and voluntarily. “Knowing and intelligent” refers to awareness of both the nature
of rights to be waived and the consequences of doing so, whereas “voluntariness” refers
to the extent to which the suspect is personally vulnerable to coercion and the extent to
which police behaved coercively. Trial courts may consider the wording or comprehensi
bility of the warning and characteristics of the suspect or police behavior to assess the
“totality of circumstances” surrounding the waiver of rights and the extent to which it
was or was not voluntary.
The voluntariness of consent to search, crucial under Fourth Amendment rights to avoid
“unreasonable” search and seizure, concerns two issues. First, was the person detained
and thereby “seized” such that a reasonable person would not feel free to leave. If the
person was seized inappropriately, even if consent to search is granted, anything found
following this unlawful seizure is inadmissible as evidence against the suspect. Second,
the consent to search must be “voluntary”—a test that again rests on a “totality of cir
cumstance” standard. That is, under the circumstances at hand, would a reasonable per
son have felt free to refuse consent. Courts may consider the nature of police requests for
consent, including politeness, asking versus demanding permission to search, physical in
timidation, and others.
Researchers have addressed two broad issues regarding waivers of rights in these two
contexts: (1) the extent to which the nature of rights against self-incrimination are under
stood and (2) the pragmatics of police communications surrounding these rights and the
extent to which reasonable persons indeed feel free to refuse to talk to police, to leave
when detained, or to refuse consent to search.
Comprehension of the Asserted Nature of Rights
The bulk of social science research on comprehension of rights against self-incrimination

has addressed comprehension of Miranda warnings in the United States or the “caution”
in England, Canada, and other countries. This research has generally contradicted wide
spread assumptions among criminal defendants, college students, the general public, law
enforcement, and the judiciary that they and others know their rights and understand the
warnings (see Rogers et al., 2010).
The U.S. Supreme Court affirmed that the rights to be conveyed by Miranda warnings can
vary in how they are stated, thus opening the door to hundreds of variations in specific
wording across jurisdictions (Rogers, Hazelwook, Harrison, Sewell, & Shuman, 2008),
varying in length from 49 to 527 words, with reading levels necessary for comprehension
varying from third grade to post-college. Average sentence complexity exceeded that of
Page 13 of 35

IRS 1040 EZ instructions. Although most words could be understood at lower grade lev
els (67–84 percent of persons at that grade level can recognize the definition among three
alternatives), many crucial terms required higher grade levels including “admission (10),
alleged (13), appointed (13), coercion (13), counsel (12), executed (10), indigent (16), in
voke (12), leniency (12), retain (12), revoked (12), terminate (12), waive (13), and waiver
(16)” (Rogers et al., 2008, p. 130).
These numbers are particularly worrisome given that more than two-thirds of criminal de
fendants have literacy levels at or below sixth-grade level, and many suffer retardation or
mental illness, speak English as a second language, and/or are juveniles—all populations
showing impaired Miranda comprehension (see Follette, Davis, & Leo, 2007; Goldstein,
Kelley, Romaine, & Zelle, 2012, for reviews). Moreover, we have recently demonstrated
(Davis, 2013) that expecting to be questioned about past criminal (vs. thrill-seeking) be
haviors impairs memory for Miranda warnings, particularly among minority students who
suffer increased cognitive load posed by “stereotype threat” (awareness of stereotypes
linking one’s ethnic group to crime). Comprehension can be expected to also decline
among those experiencing unique stressors affecting cognition during police encounters,
including the very fact of being arrested and accused of a crime (e.g., Rogers, Gillard,
Wooley, & Fiduccia, 2011).
Moreover, failures of comprehension are only partially reflected in failures of literal un
derstanding. The criteria for voluntariness of Miranda waivers include the requirement
that the person must fully comprehend the nature of the rights he is entitled to and the
consequences of his waiver. This level of comprehension is even more rare among crimi
nal suspects. For example, the right to silence (p. 447) is interpreted by some as the right
not to start a conversation, while the person believes he is nevertheless obligated to an
swer questions if asked. Few Miranda warnings specify that one does not have to answer
questions. Rogers and colleagues (Rogers et al, 2010) examined a number of beliefs about
the meaning of various components of Miranda rights, revealing serious consequential er
rors such as that exertion of the right to silence will be used as evidence against the sus
pect in trial (31 percent) or result in “piling on” of charges (9.4 percent); that if the sus
pect says something “off the record,” it will not be used against him (52 percent); that if
the suspect doesn’t sign the waiver (vs. agree verbally) his statements will not be used
against him (26 percent); or even that if rights are asserted, questioning may continue
until the attorney arrives (30 percent). All such issues of comprehension, whether of the
literal meaning or of the implications of the rights, can be worsened by the way they are
actually read, paraphrased, or explained or by other police behaviors conveying inconsis
tent messages. Similar issues of comprehension occur regarding the “caution” against
self-incrimination in England and other nations (see Rock, 2012).
Little evidence exists concerning the more complicated and crucial question of the extent
to which suspects understand what interrogation practices they might be subject to if
they waive their rights, as well as the fact that most subjected to interrogation do confess
(Kassin et al., 2010). This level of understanding is central to whether suspects recognize
implications of waiver. But, in our experience, even most defense attorneys have little to
Page 14 of 35

no awareness of interrogation practices and the way they work to promote confession. A
few studies have addressed this issue. Rogers et al., for example, did find significant mis
conceptions concerning permissibility of the common practice among police of lying to
suspects about evidence or charges against them (55–64 percent incorrectly believed ly
ing of this sort is not allowed), which is considered by interrogation scholars to be a pri
mary cause of false confession (see Kassin et al., 2010, for review). This and other mis
conceptions concerning the nature and outcomes of interrogation have been further doc
umented in surveys of beliefs concerning interrogation and confession (e.g., Chojnacki,
Cicchini, & White, 2008; Henkel, Coffman, & Dailey, 2008). These and other tests of
knowledge of Miranda rights and their implications have provided overwhelming evi
dence of failures of comprehension. But perhaps the most compelling evidence comes
from data indicating that most suspects waive their Miranda rights, that innocent sus
pects are particularly likely to do so, and that many believe that exertion of rights will
have negative consequences, such as drawing suspicion or eliciting harsher treatment
from police (see Davis, Hernandez, Follette, & Leo, 2010a; Kassin et al., 2010).
Pragmatic Implication and the Perception of Freedom to Refuse
An important consideration regarding understanding of rights to silence or to refuse

search and seizure is the nature of contextual forces that imply lack of choice even where
it legally does exist. For both situations, police have expressed a desire to talk with and/or
search the suspect, which places considerable subjective pressure on suspects to comply,
and undermines the perception of choice. Prior to the administration of Miranda
warnings, for example, police often say something like: “We’d like to talk to you to get
this thing straightened out. But before we do that we need to get you to ‘sign this paper’
or ‘take care of this formality’” (or similar statements implying that the officer is going to
talk to the suspect and the reading and signing of the waiver just needs to happen first).
Additionally, the location may imply lack of freedom (e.g., the police station), as can the
uniforms and authority of police, the presence of their weapons, the manner in which
they talk to suspects, or statements they make implying lack of choice. If suspects never
theless refuse to speak or ask for an attorney, the officer may say something like, “Well,
we can get you an attorney, and that is your right. But I have ta tell ya that we’ll have to
wonder why you think you need one.”
As Nadler and Trout (2012) commented in their review of pragmatics of consent searches:
People know that they should be courteous to the police, that police carry guns
and handcuffs, that they make mistakes that can cause you harm, and that addi
tional police are just a radio call away. They know that the police can handcuff you
and take you to the station for processing and that it can take hours or days to
sort out a misunderstanding. So, if a police officer asks to check my backpack or
luggage—even if they inform me that I have the right to refuse—I would naturally
worry that a refusal would be viewed as grounds for suspicion. (p. 333)
Page 15 of 35

Indeed, survey responses to scenarios depicting the most benign of police requests (to an
swer a few questions) indicated that most respondents would not feel free to refuse.
Moreover, among motorists (p. 448) stopped for traffic violations and then asked for per
mission to search, 76 percent reported negative reactions such as fear, humiliation,
shame, or a sense of violation (although the vast majority had no illegal objects to find)
and 91 percent agreed to the search. Of these, 96 percent reported doing so because of
fears of the consequences of refusal, ranging from the relatively benign (such as undue
delay) to serious (being searched anyway, beaten, or arrested; damage to their property)
to catastrophic (being killed). The assumptions underlying these reactions appear imper
vious to explicit police statements that the suspect has the right to refuse search (similar
to Miranda warnings of rights to silence) because studies of refusal rates among mo
torists who were or were not advised of this right were comparable. Thus, as is true for
Miranda rights and police interrogation, pragmatic implications of compulsion and au
thority in consent searches appear to override messages or knowledge concerning the na
ture of one’s rights (see Nadler & Trout, 2012, for review).
As has been the case for other issues for which pragmatics surrounding police communi
cations dramatically affect suspect perceptions (see the section “Language and Persua
sion: Law Enforcement and the Language of Trickery and Deceit”), judicial reactions to
disputed consent communications have often (though not uniformly) unrealistically ig
nored the pragmatics of consent communications. The U.S. Supreme Court, for example,
in its decision in U.S. v. Drayton (2000), denied the contextual importance of the authority
of police in consent searches, stating that “the presence of a holstered firearm…is unlike
ly to contribute to the coerciveness of the encounter absent active brandishing of the
weapon.” Moreover, it clearly indicated that, regardless of their nature, pragmatic consid
erations are not relevant to the voluntariness of a consent search. Instead, the Court indi
cated that if the request is in the form of a question and politely stated, it can be freely
denied, regardless of contextual influences suggesting otherwise. That is, the verbally po
lite question is assumed to override inferences based on police behaviors consistent with
compulsion (such as blocking exits, removal of means to leave, overbearing physical de
meanor, the compliance of others preceding the request to a specific person). Since this
ruling, lower court rulings have frequently rested primarily, if not solely, on the cordiality
or politeness of the officer’s request to search. Ironically, a suspect’s behaviors can be
viewed as implied consent (such as raising one’s hands upon being ordered to exit a vehi
cle) in the absence of verbal consent (see Nadler & Trout, 2012, for review and for excep
tions among some state courts).
Comprehension and the Jury: The Case of Judi

cial Instructions
At the end of civil and criminal trials by jury, the judge must instruct the jury on the law
applicable to the case so that the jury can apply it in arriving at their verdicts. Specific
verdicts require specific findings of fact, such as when a civil verdict of medical malprac
Page 16 of 35

tice might require a finding of negligence (departure from professional standards of care)
on the part of the doctor being sued and findings of harm to the plaintiff resulting from
the doctor’s negligence. In many instances, the elements of proof are more numerous
and/or more complicated. For example, a verdict of “fraud” requires proof of five ele
ments: defendant (1) knowingly, (2) made a misrepresentation to plaintiff, (3) that was
relevant to a decision made by plaintiff, (4) plaintiff relied on that information, and (5)
was harmed by doing so. Many concepts involved in such elements of proof are more diffi
cult to comprehend and to apply, such as “insanity,” “mitigation,” “preponderance of evi
dence,” “proximately caused,” “malice and aforethought,” and others.
Jury instructions may also concern the proper use of evidence, indicating what things
may and may not be considered to determine whether the legal elements for the verdict
have been met. Additionally, the jury may be instructed concerning how to evaluate a spe
cific piece of evidence. For example, judges commonly instruct the jury concerning fac
tors to take into account when evaluating the accuracy of an eyewitness. Unfortunately,
such instructions may themselves be based on ignorance or erroneous assumptions about
the evidence. Instructions concerning eyewitness accuracy, for example, are often inaccu
rate regarding how some factors are related to accuracy, are accurate regarding others,
or may simply instruct the jury to “consider” a factor with no instructions as to how it af
fects accuracy (Devenport, Kimbrough, & Cutler, 2009).
Jury instructions are, in effect, generally designed to fail. They are designed to facilitate
legal accuracy rather than juror comprehension and are therefore full of jargon, difficult
and complex sentence structure, and abstract rather than concrete language, and, conse
quently are often not understood at all or as intended. They are typically read at the end
of the trial, often by a judge who speaks too rapidly, in monotone, and/or without empha
sis on important (p. 449) concepts, and jurors are typically neither given written copies
nor provided further explanation if they ask questions of clarification. Instructions are of
ten so lengthy (sometimes more than an hour) that jurors cannot possibly remember all of
them even had they been worded comprehensibly. And finally, some concepts are so diffi
cult to convey that even with the best efforts of social scientists and linguists to render
them more understandable, they are still widely misunderstood.
Given such problems, psychologists have devoted considerable attention to such issues as
the extent of failure to understand specific types of instructions, how specific misunder
standings are related to verdicts, how jurors tend to decide the cases when they don’t un
derstand, individual differences in understanding, extralegal influences on understanding
and use of instructions (e.g., defendant race, victim impact, emotion, and other contextu
al variables), how to redesign instructions to facilitate comprehension, and others (see
Haney, 2005; Lieberman & Sales, 1997; Marder, 2012).
Because arguably the most consequential impact of juror misunderstanding is when the
jury must decide whether a defendant lives or dies, it is perhaps not surprising that much
of this work has concerned the issue of death penalty instructions. Capital sentencing in
structions were designed to guide jurors’ penalty deliberations to be less arbitrary or sub
Page 17 of 35

ject to biases, emotions, and other extralegal factors. Most instructions offer guidelines
regarding what factors should weigh toward death (“aggravators”) versus toward life
(“mitigators”)—although some states list them without specifying whether they are aggra
vators or mitigators. They instruct jurors how to balance and combine aggravators and
mitigators to decide the verdict. Importantly, nothing in sentencing instructions specifies
conditions mandating a sentence of death, but they do specify conditions that allow (or
are necessary for) a verdict of death and that actively mandate life.
Tragically, although juror comprehension can mean the difference between life and death,
it is discouragingly poor. Many instructions fail to clearly explain, and jurors fail to under
stand, central concepts of “aggravating” versus “mitigating” or “extenuating.” Even
worse, studies have almost universally shown that capital instructions tend to be misun
derstood in a way promoting imposition of the death penalty. The concept of mitigation is
much more poorly understood than aggravation, with many jurors failing completely to
understand the concept. Among those who do understand what mitigation means, many
fail to understand the breadth of its applicability, restricting consideration to narrow fac
tors relating to the crime and failing to consider the broader context of the defendant’s
life history and other non–crime related factors favoring life sentencing over death. More
over, many jurors essentially disagree with the judiciary concerning which factors should
be considered aggravating versus mitigating. For most factors, more than 20 percent
place them in the opposite category from that intended (see Haney, 2005, for review).
Jurors also tend to misunderstand instructions for balancing aggravating versus mitigat
ing factors in a manner favoring the death penalty. Although no instruction requires that
any imbalance favoring death means that death must be imposed, many jurors interpret
the instruction in just this way, sometimes assuming that they are obligated to impose
death if any aggravators are present or if there is any imbalance (however small) favoring
aggravators over mitigators. As well, some instructions concern circumstances under
which death must not be imposed. For example, the law in California mandates a life ver
dict when the aggravators and mitigators are equally balanced or when mitigators carry
more weight. Nevertheless, as many as half of California samples fail to understand that
life is mandated when mitigators predominate and roughly only one-sixth understand life
is mandated when aggravators and mitigators were equal.
Perhaps more disturbingly, jurors in real-life death penalty cases have reported they in
terpreted the instructions to mean that they were, in effect, following legal “orders” re
garding the imposition of the death penalty, thereby removing moral responsibility from
them for the decision. Some reported that they relied heavily on the instructions because
it allowed them to feel more comfortable with the task and less personally responsible for
their death verdicts. Finally, these attitudes and misapprehensions are more common
among those favoring the death penalty or those who are “death qualified” and therefore
able to serve on capital juries. Thus, those on trial for their lives are most likely to be
tried by jurors predisposed to misunderstand and misuse sentencing guidelines (see
Haney, 2005; Smith & Haney, 2011, for reviews).
Page 18 of 35

Researchers have used psycholinguistic techniques such as simplified vocabulary; avoid

ing double negatives, difficult grammatical structure, or use of compound sentences;
clearer organization; and others to attempt to improve comprehension. These efforts have
met with mixed success, and (p. 450) substantial misunderstanding of capital sentencing
instruction persists in studies of comprehension. Moreover, many techniques have been
tested in the laboratory but not deployed in trials (see English & Sales, 1997; Haney, 2005;
Smith & Haney, 2011; Wiener et al., 2004, for reviews). Although social science research
offers compelling evidence of problems with capital sentencing instruction (and others),
efforts to affect design of jury instructions have been frustrating, to say the least, despite
some judicial indications of recognition of the problem. The Court in Boyde v. California.
(1990, at 380) stated that if there is a “reasonable likelihood” that capital sentencing in
structions have misled a jury into finding for death, the instructions are constitutionally
defective. It went on to establish a “what a reasonable juror could understand” standard
for comprehension of capital instructions (Francis v. Franklin, 1985). Nevertheless, in for
mulating pattern instructions, the judiciary shows inadequate understanding of what rea
sonable (or typical) jurors do understand and appears to presume much greater common
ground in understanding among jurors than is warranted by the body of social science re
search.
Haney (2005) provided an excellent review of judicial reactions to the issue of juror com
prehension of death penalty instructions. Generally, his review of relevant case law indi
cated that judges often have and/or express difficulty understanding what is confusing to
jurors. He noted the following reactions and assumptions in published decisions: (1) the
content of capital instructions can be understood with ease by jurors of ordinary intelli
gence; (2) jurors will “just know” to reach beyond what instructions explicitly state to un
derstand and use appropriate factors in sentencing; (3) instructions are clearly stated and
conveyed, any misunderstanding requires strained parsing of their language and reason
able jurors would not have misunderstood; (4) even if there is some doubt as to clarity of
the instructions alone, jurors will manage to divine proper meaning and application of the
concepts in the instructions from the entire trial context in which they were given; (5)
even if jurors express confusion in notes and questions concerning instructions (and the
judge refuses to answer or simply restates the instructions as originally read), there is on
ly a small possibility that this confusion will persist and cause a miscarriage of justice; (6)
there is a presumption of understanding among jurors and that, absent that presumption,
a reversal would be required each time a jury had inquired about a matter of constitution
al import; (7) the terms “aggravating” and “mitigating” are commonly used, and the jury
needs no definition of them—nor does it need to be told whether specific given factors
should be considered mitigating versus aggravating because their nature is self-evident;
and (8) individual juror misunderstandings will be corrected through attorney arguments
or deliberation. The American Bar Association maintains a “Death Penalty Moratorium”
website where the bases of their advocacy of a death penalty moratorium, including jury
instructions, are reviewed state by state (http://www.americanbar.org/groups/
individual_rights/projects/death_penalty_moratorium_implementation_project.html). The
nature of and assumptions underlying death penalty instructions across the country sug
Page 19 of 35

gest the judiciary has failed in perspective-taking and provision of the common ground
necessary to facilitate juror understanding and proper use of the instructions.
Language and Persuasion: Law Enforcement

and the Language of Trickery and Deceit
Perhaps the most widely investigated issues of language and persuasion in legal settings
have concerned law enforcement personnel and manipulative, deceptive interactions with
suspects. Here, we focus on two such forms of interaction: police interrogation of crimi
nal suspects and undercover police operations targeting criminal suspects.
Police Interrogation and the Manipulative Use of Pragmatic Implication

A large and growing body of documented interrogation-induced false confessions in the
United States and around the globe has fueled research among psychologists regarding
their causes (Kassin et al., 2010). Two general causes have been identified as primary
contributors to the majority of false confessions: (1) distress intolerance and confession to
escape the interrogation and (2) the mistaken conviction that confession will be benign or
in one’s best long-term legal interests. Indeed, the formal instruction of interrogators fo
cuses on strategies designed explicitly to convince the suspect that confession will best
serve those interests (e.g., Inbau, Reid, Buckley & Jayne, 2011; Zulawski & Wicklander,
2002).
Because the law prohibits use of explicit threats regarding failure to confess or promises
regarding benefits of confession, and because violation of these prohibitions can result in
the confession becoming inadmissible as evidence at trial, interrogation (p. 451) technique
has evolved to convey the incentives through implication. In this way, threats and promis
es can be clearly stated by the interrogator, and understood as such by the suspect, while
not violating the letter of the law. Interrogation technique, from the level of individual
words or sentences to strategies that unfold over long stretches of interrogator argu
ments, is carefully scripted to convey four central messages: (1) sufficient evidence exists
against the suspect such that there is no hope of exoneration; (2) nevertheless, the conse
quences are flexible and there may be a way to minimize or avoid them altogether; (3) the
interrogator has the authority and desire to achieve (or at least facilitate) the best legal
outcomes for the suspect; and (4) this can only be accomplished if the suspect confesses
and explains his actions during the interrogation. Each of these messages is conveyed in
whole or part through pragmatic implication (see Davis, 2010; Davis & O’Donohue, 2004;
Davis & Leo, 2012).
In what we have dubbed the “Borg Maneuver,” the interrogator first attempts to establish
a sense of hopelessness in the suspect (Resistance is futile! We have the evidence to con
vict.). The interrogator states that the results of “our” investigation clearly establish that
the suspect committed the crime (implying that multiple parties have come to the conclu
sion of the suspect’s guilt). He then presents real, fabricated, or implied evidence to the
suspect to facilitate the sense of hopelessness. He may imply the existence of evidence by
Page 20 of 35

such statements as “Is there any reason why we would have found your fingerprints at the
scene of the crime?” The interrogator is instructed to use such implied evidence when the
real evidence is not available or yet collected. This indirect implication offers plausible
deniability in case the suspect knows for sure the fingerprints are not there and prevents
the suspect from perceiving the detective as dishonest and generally discounting the
interrogator’s messages as a result.
The messages that the consequences of the suspect’s guilt are flexible and that the inter
rogator has the authority and desire to facilitate the best legal outcomes for the suspect
are communicated almost entirely through implication. Interrogators are trained to ask
the suspect a question such as the following at the beginning of the interrogation: “Tell
me Jack, what do you think should happen to the person who did this. Should they just go
to jail, or are there any circumstances where he should maybe get a second chance—not
go to jail, but maybe get counseling instead?” After accusing the suspect of the crime, the
interrogator offers a “pretext” (Inbau et al., 2011) for continuing the interrogation. He
tells the suspect that he is not there to determine whether the suspect committed the
crime, but rather why he did it and what kind of person he is: that these things are impor
tant to know. These messages clearly imply that the consequences are flexible and that
the suspect’s explanation of himself and his actions will affect the consequences (explana
tions that entail confession, of course). Our research has shown that these two messages
affect the degree to which participants believe the interrogator can choose between
charges or let the suspect go without charges versus has no choice in whether or what
charges to file (e.g., Davis et al., 2010a) and, furthermore, that these perceptions predict
the perceived wisdom of false confession. These messages are supported by other refer
ences to what is being done as “clearing this up” “getting this straightened out,” each im
plying that the problem will be over once the suspect explains himself.
The interrogator further conveys his apparent authority to affect suspects’ outcomes with
a tactic we have dubbed the “sympathetic detective with a time-limited offer.” The inter
rogator flatters the suspect, says he doesn’t believe the suspect is a bad guy (rather a
good guy who got caught up in a bad situation), and states that he would like to help the
suspect but that he can only help him if the suspect “explains” himself (i.e., confesses
what he did, how and why) while in the interrogation. After the suspect leaves the inter
rogation, the detective may say, the DA and others to come will no longer be interested in
listening and will just file charges based on the information they have. The sympathetic
detective tactic increases perceptions of the detective’s authority to affect what, if any,
charges are filed, his liking of the suspect, and his desire to see the suspect get the best
legal outcomes. Furthermore, these perceptions (detective beneficence and authority)
predict the perceived wisdom of false confession (Davis, Leo, & Follette, 2010b).
Having set the agenda for the interrogation as allowing the suspect to explain his actions
to achieve the detective’s help and allegedly the suspect’s best outcomes, the detective
turns to communicating relative benefits and costs of confession versus denial. The inter
rogator attempts to lower the perceived costs of confession through a process known as
“theme development” (Inbau et al., 2011), whereby he offers various scenarios for how
Page 21 of 35

and why the crime was committed that appear to minimize its seriousness (“minimiza
tion”). Interrogation (p. 452) manuals offer long lists of possible themes tailored for use in
each of many specific kinds of crimes.
Essentially, the interrogator tells the suspect a story about why he believes the crime
might have been committed that appears relatively (or completely) excusable. He may
suggest it was unintentional, accidental, self-defense, something provoked by the alleged
victim, something other people would have done in the same circumstances, or similarly
seemingly excusable versions of the crime. Having suggested such a scenario, he may
then say something such as “If that’s what happened, I can understand that. That’s no big
deal. We can work with that.” All such statements imply less serious consequences—or
even no consequences (“That’s no big deal”). The relatively minimal consequences of the
scenario the detective suggests is eventually contrasted with a more serious version: “But
if instead, you planned this in advance and took a weapon with you and looked for these
guys to start a fight, that’s completely different. I don’t even want to talk to you anymore.
Which was it Jack, is this something you didn’t plan and they started, or did you go out
planning to get into a fight?” This “alternative question” technique (Inbau et al., 2011) is
often successful in eliciting a first admission. Indeed, research testing the effectiveness of
“minimization” and the implied promises the technique conveys has shown it to be equal
ly effective in eliciting true and false confessions as an explicit promise of leniency (e.g.,
Russanno, Meiissner, Narchet, & Kassin, 2005).
The implied promises of benefits of confession are supplemented by implied threats of

consequences. “How do you think a judge and a jury is going to feel about someone who
won’t stand up like a man and take responsibility for what he’s done, someone who lies
about what he’s done. Would you feel better about that guy, or one who says ‘Yeah, I did it
and I’m sorry.’—someone who takes responsibility like a man.” “Others are talking and
telling their side of the story. And guess where their fingers are pointing? They’re saying
you’re the one who planned this and the one who had the gun. If you don’t take this op
portunity to tell your side of the story, we’re gonna have to go with the one they’re telling
us, and the DA is gonna have to file charges based on their story.” Some such messages
walk a thin line between explicit and implicit threats: for example, “You’re at a crossroads
today and you’ve got to decide which path to take. One way is life, and the other is death.
I’d like to see you choose life.”
These tactics are extremely effective, eliciting confessions from approximately 42–76 per
cent of suspects in American and English field studies involving actual criminal acts (Gud
jonsson, 2003; Thomas, 1996) and in as many as 100 percent in laboratory studies involv
ing noncriminal acts (e.g., Kassin & Kiechel, 1996). The implied threats and promises
have proven very effective in convincing suspects that confession is in their best inter
ests, even that there will be no consequences at all. Roughly 40 percent of self-reported
false confessors explain that they did so to be let go or avoid detention (e.g., Sigurdsson
& Gudjonsson, 1996). Although it is more than clear that the implied threats and promis
es central to modern police interrogation strategies are clearly perceived as such by sus
pects, the law views them as acceptable, in contrast to the same messages conveyed ex
Page 22 of 35

plicitly. Thus, confessions elicited through interrogations full of implied threats and
promises contingent on confession are routinely viewed as voluntary and admitted as evi
dence against the defendant at trial (see Davis & Leo, 2010, for review).
Linguistic Strategies for Undercover Deception

In his book Creating Language Crimes: How Law Enforcement Uses (and Misuses) Lan
guage, Roger Shuy offers a detailed account of how undercover law enforcement agents
manipulate the perceptions of their targets, as well as those of the prosecutors, judges,
and juries who will later interpret the meaning of taped conversations. A number of con
versational strategies are utilized by undercover agents to portray the illusion that sus
pects are indeed involved in criminal activities, even when the available evidence is weak
at best. Many entail the manipulative misuse of context and common ground.
A fundamental problem in such undercover interactions is that agents are often indirect
in their own statements about the illegal actions in question. They may be asked to draw
a suspected criminal into talking about past criminal actions and/or planning future crimi
nal actions and may suggest incentives for involvement that could be viewed as entrap
ment. Not wishing to cause suspicion in their targets, and wanting the target to be the
one to appear to suggest the illegal acts, they may hint and speak indirectly. But, as with
all indirect statements, they can be subject to substantial disputes over meaning, particu
larly when clarifying contextual cues are absent in recordings. Unfortunately, although
failures of context may simply obscure meaning, context can also be manipulated to cre
ate the illusion of culpability.
A number of such strategies are designed to mislead those who listen to the
(p. 453)
recorded conversation regarding who is being spoken to or about. A target can leave the
room or walk out of earshot (unknown to listeners), whereupon the agent will continue to
talk (as if to the target) about illegal past or future actions. Silence can then be argued to
be an “adoptive admission” of guilt by an allegedly still-present target. Similarly, the
agent may mislead regarding the referent. While both look at TV, the agent could say
something such as “Look at her! Good thing she’s about to get whacked!”—thereby giving
the impression that the target was party to plans to murder a victim. Necessary context
may also be obscured by selectively turning the tape on and off, strategically causing stat
ic in the microphone, talking over the target, and other strategies to mislead listeners re
garding what is being talked about or agreed to.
Some strategies entail use of failures of knowledge and common ground with the target,
which can lead him to plan or agree to illegal actions on tape while thinking the conversa
tion is about something else. The agent may say to the target “Let’s go to Jack’s and pick
up some ‘speakers’ tomorrow.” “Speakers” is street slang for drugs. If the slang use is un
known to the target, he may agree to a plan he has no inkling is illegal. Shuy described a
case of alleged solicitation of murder in which the undercover officer referred to “help
ing” the suspect with his “wee problem,” “getting it sorted out,” and so on, rather than to
murdering the person posing the “wee problem.” He showed that the defendant thought
Page 23 of 35

they were talking about help with finding and getting his daughter (who had eloped)
back, rather than killing the man she had eloped with. In these separate contexts, the
words “help,” “get him out of the picture,” “get it done,” “he’s got to go,” “I’ll do the job
for you,” and others had very different meanings (one criminal and the other not). Such
problems remind one of the “Who’s on first?” routine by Abbott and Costello!
Undercover “cooperating witnesses” may also be asked to wear a microphone or make

recorded phone calls to elicit incriminating recorded statements from suspects, as dis
cussed in the earlier section on “pretext” calls. However, as with attempts to elicit admis
sions of rape, other attempts by cooperating witnesses can result in ambiguities subject
to disputed meaning. For example, the witness may talk about “doing” someone: which
can be interpreted as having sex, killing, doing their hair, and countless other actions.
Moreover, the intentions may be stated indirectly. Thus, what might be an actual instance
of planning murder functions as less compelling proof than intended. The addition of
video can make clear that two speakers who seem on the audio recording to be engaging
in one conversation on one topic are actually talking about two different things (as made
clear by their physical actions). Therefore, a speaker may appear to agree to something
entirely different from what he thinks he is talking about.
The preceding examples illustrate only a few of the many problems of context and com
mon ground that can compromise interpretation of recorded statements or actions. Such
problems are often unwitting, often result from attempts to be indirect and “off the
record,” and may also be deliberate attempts to create the illusion of culpability when at
tempts to get the necessary statements without trickery are failing. Regardless, under
cover recordings offer great potential to mislead, particularly in the absence of careful
dissection by a skilled linguist.
Language, Classification, and Judgment

As noted earlier, statements are likely to be interpreted in terms of implications for the
relationship between conversational participants and for broader impressions of the
speaker and listener. Social psychologists have investigated several such attributions with
potential importance in legal settings: such as causality and attributions of blame, per
ceptions of individual character, detection of deception, evaluations of argument credibili
ty, and others. In legal contexts, such language-based attributions can shape outcomes
ranging from assessment or charging of criminal and civil plaintiffs and defendants (e.g.,
Black & Downie, 2010) to juror reactions to the parties or to attorney presentations dur
ing trial (e.g., Schmid & Fiedler, 1998).
Attribution of Responsibility/Blame
Several linguistic indicators have been found to exert an influence on attributions of

causality and responsibility—judgments crucial to jury verdicts and sentencing. One of
the most common of these is the use of passive versus active voice. For example, “Jane
was raped by John” is a passive voice statement that tends to increase ambiguity, conceal
Page 24 of 35

agency, and focus attention on the recipient of the action (Jane) while diverting attention
from its performer (John). In contrast, “John raped Jane” (the active voice counterpart)
draws attention to the agent of the action while presenting the victim as a secondary ele
ment of the sentence.
This process of enhancing ambiguity and shifting attention to the target of an ac
(p. 454)
tion can be even more effective if no agent is referenced at all (e.g., Penelope, 1990).
Bohner (2001) illustrated the use of the verb get to imply victim responsibility when no
agent is specified (e.g., “Jane got raped”), suggesting that such sentences obscure the
agent while effectively suggesting the recipient (Jane) actively participated in what was
done to her (getting raped).
Fausey and Boroditsky (2010) showed the importance of active versus passive voice by
analyzing trials held at London’s central criminal court between 1674 and 1913. Cases
presenting predominantly nonagentive phrases (it broke, it burned, and died) resulted in
a lower rate of guilty verdicts, than those predominantly using their agentive counter
parts (broke it, burned it, and killed). The authors followed with a series of experiments in
which participants were exposed to accident descriptions depicted in agentive or nona
gentive language. Across studies, participants attributed more blame to perpetrators and
assigned higher monetary penalties when the event was described using agentic versus
nonagentic language. Interestingly, evaluations were still affected by the linguistic fram
ing used to describe the event, even after a fictitious liability evaluation made by an inde
pendent panel (which attributed the agent low, medium, or high blame) was presented.
Moreover, such differences remained after participants read and thought about a well-
known event and even after they were exposed to a video of the event.
The authors further explored the role of agentivity on blame attribution by examining the
differences in agentive language use between English and Spanish speakers in describing
accidental events (Fausey & Boroditsky, 2011). Spanish speakers described accidental
events using more nonagentive descriptions than those of English speakers, and English
speakers were more likely to remember the agent of accidental events than are Spanish
speakers. The authors concluded that frequency of agentic language use might influence
how information is encoded and retrieved, suggesting that cross-linguistic differences in
agent identification for accidental events might be attributed to English speakers encod
ing more information about the agent, compared to speakers of languages in which
agency is less common to describe such events.
Research further suggests that people tend to use passive or nonagentive descriptions of
questionable behaviors when the agent is not perceived to be fully responsible for the ac
tions involved or when attempting to portray the agent—perhaps the speaker—in a more
favorable light. For instance, Bohner (2001) asked participants to view a video of a rape
scene and then describe their perceptions of responsibility for the perpetrator and victim.
The video either incorporated circumstances matching rape myths (the victim flirted with
the aggressor and was wearing scanty clothing) or not (e.g., the victim was unacquainted
with the aggressor and dressed modestly). Participants exposed to the myth-consistent
Page 25 of 35

video were more likely to use passive voice to describe the perpetrator’s actions and to
assign lower responsibility to the perpetrator than were those exposed to myth-inconsis
tent video. Likewise, in an archival study of scientific articles involving male-to-female do
mestic violence, Lamb (1991) found a high prevalence of passive forms (among other lin
guistic indicators) that presumably served the purpose of concealing a perpetrator’s re
sponsibility for an incident. Furthermore, although not necessarily framed in passive
form, language used by sexual offenders to justify their actions has been found to regular
ly focus on victim actions that led to the assault, in a presumed attempt to justify the
perpetrator’s actions and shift the blame onto the victim him- or herself (e.g., Muchoki,
2011).
Luchjenbroers and Aldridge (2007) described the use of metaphors and frames to influ
ence perceptions, claiming that words have inherent meanings and associations that are
evoked and encoded into memory regardless of the context in which the words are uti
lized. Allusions to the future, for example, tend to be interpreted in terms of moving for
ward, whereas words related to the past tend to be associated with events that have al
ready occurred and are now left behind. The authors claimed that these past–future
metaphors are regularly employed by police and in courtroom settings to implicitly ma
nipulate listeners’ perceptions. For example, statements like “she backed away from a full
confession” (p. 343) describing the presumed victim of sexual assault in a court hearing
would suggest cowardice and that the victim is not willing to “move forward,” thus cast
ing her in a negative light. Likewise, a question like “did you say it was clear where the
night was headed?” (p. 343) directed to the presumed victim would imply that she was
fully aware of the sequence of events before they occurred and that she should have tak
en the necessary steps to take herself out of that path, an implication that aims to trans
fer responsibility to the victim. Other metaphorical statements and framings are likewise
used to manipulate judgments, including dehumanizing labels (calling the suspect
(p. 455)
an animal), attributing human qualities to ideas or inert objects (stating that evidence in
criminates the suspect), and using euphemistic framings (referring to domestic violence
as marital discord). Such “metaphors” and “frames” can be viewed as “schemas,” “stereo
types,” or anchors (Wagenaar, 1995), serving as context and guiding interpretation and
evaluation of testimony and evidence.
Consistent with such notions, a study examining a newspaper’s reporting of violent sexu
al crimes (Clark, 1992) found a correlation between labels used to refer to alleged victims
and perpetrators. Victims were often called by neutral or sympathetic labels (e.g., moth
er, schoolgirl) when the perpetrator was referred to by a subhuman label (e.g., fiend,
monster), whereas the victim was often assigned a label suggesting blame or “sexual
availability” (e.g., blonde divorcee, Lolita) when the perpetrator was referred to in neu
tral terms (e.g., his name or occupation). Moreover, Clark found that many headlines at
tempted to provide a cause for the crime in the victim or her actions (e.g., “Mower hit me
for saying no to sex,” p. 221), even if evidence was insufficient to establish a cause. Final
ly, in an analysis of scientific articles dealing with male-to-female physical aggression,
Lamb (1991) unveiled several components of linguistic avoidance—that is, text structures
designed to minimize the perpetrator’s responsibility—including passive voice usage,
Page 26 of 35

nominalization, and an effort to diffuse responsibility by making agency ambiguous and

using language that minimizes the violence. Taken together, these studies support the no
tion that people employ a wide array of linguistic techniques to manipulate perceptions of
responsibility in legal contexts, with varying degrees of effectiveness.
Attribution and the “Adoptive Admission”
Although these examples illustrate the relationship between language and attributions
among observers, the law sometimes takes a formal position on the meaning of linguistic
behavior—unfortunately, without scientific basis. A particularly disturbing example is law
related to “adoptive admission,” in which a person is assumed to have admitted to what
he or she is accused of by virtue of being silent in response to an accusation (see
Ainsworth, 2012, for review). Ainsworth provided an excellent discussion of linguistic re
search identifying multiple interpretations of the meaning of silence (e.g., disapproval,
confusion, fear) and evidence that such meaning is highly dependent on contextual, cul
tural, and personal circumstances. Nevertheless, the law of adoptive admissions is based
on the simplistic assumption that people confronted with false accusatory statements
would naturally respond to them with objection and that responding with silence can be
interpreted as a tacit admission of the statement’s truth.
Such “confessions” are not only extracted and used against suspects at court proceed
ings, but also during police questioning, even when suspects are simply trying to exercise
their Miranda rights. The true meaning of silence following an accusation can be purpose
fully muddled by tactics aimed at obtaining such silent responses at all costs, including
making indirect accusations (i.e., made in presence of the accused but not directed to
her) and purposely using ambiguous or confusing language when making an accusation
to promote a confused silence that can be misinterpreted as assent (see section in this
chapter on undercover police work).
Language and the Assessment of Deception
Among the most heavily investigated issues of language and assessment is the associa
tion between language and deception. Law enforcement and those who train agents offer
systems for diagnosis of deception based on analysis of wording, almost none of which
has been supported by scientific research. The most common of these is offered by John
Reid Associates, the most prominent interrogation training organization in America. Simi
lar methods are taught by the U.S. Federal Bureau of Investigation (FBI), police depart
ment trainers, and independent training organizations such as Zulawski & Wicklander,
among others. The specificity of claims regarding differences between truthful and decep
tive statements can be rather startling. I (Davis) attended a Reid Associates seminar in
which the trainer (as also stated in their written manuals) made such claims as: A truthful
person faced with an accusation will say “I didn’t kill her,” whereas a deceptive person
will say “I did not kill her.” Although the Reid method of detecting deception relies on
both verbal and nonverbal indicators, other methods rely on linguistic indicators alone.
Aldert Vrij (2008) has recently provided a very detailed and thorough review of all law en
Page 27 of 35

forcement systems for detecting deception, showing that neither the Reid method nor the
many others (with few exceptions) have adequate (or any) scientific foundation, nor have
most been supported by existing empirical tests. Some, such as Reid, worsen perfor
mance compared to untrained controls.
(p. 456) Attributions of Character
A number of studies have addressed the use of linguistic strategies to convey positive or
negative character and competence. Use of stigmatizing language and labeling in crimi
nal justice settings has been a pervasive issue, particularly in regards to the diagnosis of
mental health ailments (Black & Downie, 2010). Mellsop, Fraser, Tapsell, and Menkes
(2011) claimed that psychiatric labeling is often misused to affect legal assessments such
as competence to stand trial or criminal culpability. One study, for example, showed that
when a juvenile criminal suspect was described as meeting criteria for diagnosis of psy
chopathy, he was deemed less dangerous and less deserving of harsh punishment than if
he was simply labeled as a “psychopath” (Boccaccini, Murrie, Clark, & Cornell, 2008),
suggesting that labeling (regardless of accuracy) might play an important role in individ
ual assessments in criminal justice settings.
Conclusion
We have barely scratched the surface of the many applications of language to law or of
the many findings within the areas we did address. We focused primarily on those most
widely investigated (particularly by social psychologists) and left out several areas essen
tially covered by other chapters in this volume (such as language and jury persuasion: see
Dillard, this volume). Therefore, we end by pointing out several important areas of appli
cation, reviews of which can be found in the recent Oxford Handbook of Language and
Law (Tiersma & Solan, 2012a).
Among the most important of these are issues surrounding the rights and involvement of
linguistic minorities in the legal system. Some issues concern the extent to which linguis
tic minorities may be disadvantaged in interactions with police or others who fail to pro
vide interpreters or who translate inadequately. Moreover, what are the effects of using
interpreters? How do they alter meaning? What about use of interpreters in court to en
able use of jurors for whom English is a second language? Research on these and other
issues for linguistic minorities is in its infancy but is likely to grow substantially as multi
lingualism increases across the globe.
Other interesting areas include disputed trademark or copyright infringement; language

as “match” evidence (identifying authorship or detecting plagiarism or country of origin
of the writer), and others. Social psychologists may find themselves studying issues of
confirmation and other biases affecting such judgments, much as they have with finger
prints and other forensic evidence. The future of language and law research is full of pos
sibility and excitement!
Page 28 of 35

References
Abbey, A. (2011). Alcohol’s role in sexual violence perpetration: Theoretical explanations,
existing evidence and future directions. Drug and Alcohol Review, 30(5), 481–489. doi:
10.1111/j.1465-3362.2011.00296.x
Abbey, A., Jacques-Tiura, A. J., & LeBreton, J. M. (2011). Risk factors for sexual aggres
sion in young men: An expansion of the confluence model. Aggressive Behavior, 37(5),
450–464. doi: 10.1002/ab.20399
Adler, M. (2012). The plain language movement. In L. Solan & P. Tiersma (Eds.), The Ox
ford handbook on language and law (pp. 67–86). Cambridge: Oxford University Press.
Ainsworth, J. (2012). The meaning of silence in the right to remain silent. In L. Solan & P.
Tiersma (Eds.), The Oxford handbook on language and law (pp. 287–298). Cambridge: Ox
Austin, J. L. (1962). How to do things with words. Oxford: Clarendon Press.
Black, M., & Downie, J. (2010). Watch your language: A review of the use of stigmatizing
language by Canadian judges. Journal of Ethics in Mental Health, 5(1), 1–8.
Boccaccini, M. T., Murrie, D. C., Clark, J. W., & Cornell, D. G. (2008). Research report: De
scribing, diagnosing, and naming psychopathy: How do youth psychopathy labels influ
ence jurors? Behavioral Sciences & the Law, 26(4), 487–510. doi: 10.1002/bsl.821
Bohner, G. (2001). Writing about rape: Use of the passive voice and other distancing text
features as an expression of perceived responsibility of the victim. British Journal of So
cial Psychology, 40(4), 515–529. doi: 10.1348/014466601164957
Boyde v. California, 494 U.S. 370 (1990)
Brainerd, C. J., & Reyna, V. F. (2005). The science of false memory. New York: Oxford Uni
versity Press.
Bransford, J. D., & Johnson, M. K. (1972). Contextual prerequisites for understanding:

Some investigations of comprehension and recall. Journal of Verbal Learning & Verbal Be
havior, 11(6), 717–726. doi:10.1016/S0022-5371(72)80006–9
Chojnacki, D. E., Cicchini, M. D., & White, L. T. (2008). An empirical basis for the admis
sion of expert testimony on false confessions. Arizona State Law Journal, 40, 1–45.
Clark, K. (1992). The linguistics of blame: Representations of women in The Sun’s report
ing of crimes of sexual violence. In M. Toolan (Ed.), Language, text and context: Essays in
stylistics (pp. 208–224). London: Routledge.
Page 29 of 35

Davis, D. (2010). Lies, damned lies, and the path from police interrogation to wrongful
conviction. In M. H. Gonzales, C. Tavris, & J. Aronson (Eds.), The scientist and the human
ist: A festschrift in honor of Elliot Aronson (pp. 211–247). New York: Psychology Press.
Davis, D. (2013). Interrogation-related cognitive decline: The roles of failures of self-regu

lation and stereotype threat. Paper presented at the American Psychology-Law Society,
Portland, Oregon.
Davis, D., & Friedman, R. D. (2007). Memory for conversation: The orphan child of wit
ness memory researchers. In M. P. Toglia, J. D. Read, D. F. Ross, & R. C. L. Lindsay
(p. 457) (Eds.), The handbook of eyewitness psychology, Vol. I: Memory for events. (pp. 3–
52). Mahwah, NJ: Lawrence Erlbaum Associates Publishers.
Davis, D., Hernandez, J., Follette, W. C., & Leo, R. A. (2010a, January). Interrogation
through pragmatic implication: Communicating beneficence and promises of leniency. Pa
per presented at the Society for Personality and Social Psychology, Las Vegas, Nevada.
Davis, D., Kemmelmeier, M., & Follette, W. C. (2005). Conversational memory on trial. In
Y. I. Noy & W. Karwowski (Eds.), Handbook of human factors in litigation (pp. 12-1 to
12-29). New York: CRC Press.
Davis, D., & Leo, R. A. (2010). Commentary: Overcoming judicial preferences for person-
versus situation-based analyses of interrogation-induced confessions. Journal of the Amer
ican Academy of Psychiatry & the Law, 38(2), 187–194.
Davis, D., & Leo, R. A. (2012). Interrogation through pragmatic implication: Sticking to
the letter of the law while violating its intent. In L. Solan & P. Tiersma (Eds.), The Oxford
handbook on language and law (pp. 354–366). Cambridge: Oxford University Press.
Davis, D., Leo, R. A., & Follette, W. C. (2010b). Selling confession: Setting the stage with
the “sympathetic detective with a time-limited offer”. Journal of Contemporary Criminal
Justice, 26(4), 441–457. doi: 10.1177/10439862103772
Davis, D., & Loftus, E. F. (2004). What’s good for the goose cooks the gander: Inconsisten
cies between the law and psychology of voluntary intoxication and sexual assault. In W. T.
O’Donohue & E. R. Levensky (Eds.), Handbook of forensic psychology: Resource for men
tal health and legal professionals. (pp. 997–1032). New York: Elsevier Science.
Davis, D., & Loftus, E. F. (2007). Internal and external sources of distortion in adult wit
ness memory. In M. P. Toglia, J. D. Read, D. R. Ross, & R. C. L. Lindsay (Eds.), Handbook
of eyewitness memory (Vol. 1): Memory for events (Vol. 1, pp. 195–237). Mahwah, NJ: Erl
baum.
Davis, D., & O’Donohue, W. T. (2004). The road to perdition: Extreme influence tactics in
the interrogation room. In W. T. O’Donohue & E. Levensky (Eds.), Handbook of forensic
psychology (pp. 897–996). New York: Elsevier, Academic Press.
Page 30 of 35

Devenport, J. L., Kimbrough, C. D., & Cutler, B. L. (2009). Effectiveness of traditional safe
guards against erroneous conviction arising from mistaken eyewitness identification. In
B. L. Cutler (Ed.), Expert testimony on the psychology of eyewitness identification. (pp.
51–68). New York: Oxford University Press.
English, P. W., & Sales, B. D. (1997). A ceiling or consistency effect for the comprehension
of jury instructions. Psychology, Public Policy, and Law, 3(2–3), 381–401. doi:
10.1037/1076-8971.3.2-3.381
Fausey, C. M., & Boroditsky, L. (2010). Subtle linguistic cues influence perceived blame
and financial liability. Psychonomic Bulletin & Review, 17(5), 644–650. doi: 10.3758/pbr.
17.5.644
Fausey, C. M., & Boroditsky, L. (2011). Who dunnit? Cross-linguistic differences in eye-
witness memory. Psychonomic Bulletin & Review, 18(1), 150–157. doi: 10.3758/
s13423-010-0021-5
Follette, W. C., Davis, D., & Leo, R., A. (2007). Mental health status and vulnerability to in
terrogative influence. Criminal Justice, 22(3), 42–49.
Francis v. Franklin, 471 U.S. 307, 332 (1985).
Goldstein, N. E. S., Kelley, S. M., Romaine, C. L. R., & Zelle, H. (2012). Potential impact of
juvenile suspects’ linguistic abilities on Miranda understanding and application. In L.
Solan & P. Tiersma (Eds.), The Oxford handbook on language and law (pp. 299–311). Cam
bridge: Oxford University Press.
Gudjonsson, G. H. (2003). The psychology of interrogations and confessions: A handbook.

New York: John Wiley & Sons Ltd.
Haney, C. (2005). Death by design: Capitol punishment as a social psychological system.

New York: Oxford University Press.
Henkel, L. A., Coffman, K. J., & Dailey, E. M. (2008). A survey of people’s attitudes and be
liefs about false confessions. Behavioral Sciences & the Law, 26(5), 555–584. doi:10.1002/
bsl.826
Inbau, F., E., Reid, J. E., Buckley, J. P., & Jayne, B. C. (2011). Criminal interrogation and
confessions (5th ed.). Burlinton, MA: Jones & Bartlett Learning.
Kassin, S. M., Drizin, S. A., Grisso, T., Gudjonsson, G. H., Leo, R. A., & Redlich, A. D.
(2010). Police-induced confessions: Risk factors and recommendations. Law and Human
Behavior, 34(1), 3–38. doi: 10.1007/s10979-009-9188-6
Page 31 of 35

Kassin, S. M., & Kiechel, K. L. (1996). The social psychology of false confessions: Compli
ance, internalization, and confabulation. Psychological Science, 7(3), 125–128. doi:
10.1111/j.1467–9280.1996.tb00344.x
Kassin, S. M., & McNall, K. (1991). Police interrogations and confessions: Communicating
promises and threats by pragmatic implication. Law and Human Behavior, 15(3), 233–
251. doi: 10.1007/bf01061711
Koukounas, E., & Letch, N. M. (2001). Psychological correlates of perception of sexual in
tent in women. The Journal of Social Psychology, 141(4), 443–456. doi:
10.1080/00224540109600564
Lamb, S. (1991). Acts without agents: An analysis of linguistic avoidance in journal arti
cles on men who batter women. American Journal of Orthopsychiatry, 61(2), 250–257. doi:
10.1037/h0079243
Lassiter, G., Ware, L. J., Lindberg, M. J., & Ratcliff, J. J. (2010). Videotaping custodial inter
rogations: Toward a scientifically based policy. In G. Lassiter, C. A. Meissner (Eds.), Police
interrogations and false confessions: Current research, practice, and policy recommenda
tions (pp. 143–160). Washington, DC: American Psychological Association. doi:
10.1037/12085-009
Lee, J. J., & Pinker, S. (2010). Rationales for indirect speech: The theory of the strategic
speaker. Psychological Review, 117(3), 785–807. doi: 10.1037/a0019688
Lieberman, J. D., & Sales, B. D. (1997). What social science teaches us about the jury in
struction process. Psychology, Public Policy, and Law, 3(4), 589–644. doi:
10.1037/1076-8971.3.4.589
Loftus, E. F. (2005). Planting misinformation in the human mind: A 30-year investigation

of the malleability of memory. Learning & Memory, 12(4), 361–366. doi: 10.1101/lm.94705
Loftus, E. F., & Davis, D. (2006). Recovered memories. Annual Review of Clinical Psychol
ogy, 2, 469–498. doi: 10.1146/annurev.clinpsy.2.022305.095315
Loftus, E. F., & Palmer, J. C. (1974). Reconstruction of automobile destruction: An exam

ple of the interaction between (p. 458) language and memory. Journal of Verbal Learning
and Verbal Behavior, 13, 585–589.
Luchjenbroers, J., & Aldridge, M. (2007). Conceptual manipulation by metaphors and

frames: Dealing with rape victims in legal discourse. Text & Talk, 27(3), 339–359. doi:
10.1515/text.2007.014
Marder, N. S. (2012). Interrogation through pragmatic implication: Sticking to the letter

of the law while violating its intent. In L. Solan & P. Tiersma (Eds.), The Oxford handbook
on language and law (pp. 435–445). Cambridge: Oxford University Press.
Page 32 of 35

Mellsop, G. W., Fraser, D., Tapsell, R., & Menkes, D. B. (2011). Courts’ misplaced confi
dence in psychiatric diagnoses. International Journal of Law and Psychiatry, 34(5), 331–
335. doi: 10.1016/j.ijlp.2011.08.005
Miranda v. Arizona, 384 U.S. 336 (1966).
Muchoki, S. (2011). Vocabulary used by sexual offenders: Meaning and implications. Cul
ture, Health & Sexuality, 13(1), 101–113. doi: 10.1080/13691058.2010.503971
Nadler, J., & Trout, J. D. (2012). The language of consent in police encounters. In L. Solan
& P. Tiersma (Eds.), The Oxford handbook on language and law (pp. 326–339). Cam
bridge: Oxford University Press.
Penelope, J. (1990). Speaking freely: Unlearning the lies of the fathers’ tongues. New
York: Pergamon Press.
Rock, F. (2012). The caution in England and Wales. In L. Solan & P. Tiersma (Eds.), The
Oxford handbook on language and law (pp. 312–325). Cambridge: Oxford University
Press.
Rogers, R., Gillard, N. D., Wooley, C. N., & Fiduccia, C. E. (2011). Decrements in Miranda
abilities: An investigation of situational effects via a mock-crime paradigm. Law and Hu
man Behavior, 35(5), 392–401. doi: 10.1007/s10979-010-9248-y
Rogers, R., Hazelwood, L. L., Sewell, K. W., Harrison, K. S., & Shuman, D. W. (2008). The
language of Miranda warnings in American jurisdictions: A replication and vocabulary
analysis. Law and Human Behavior, 32(2), 124–136. doi: 10.1007/s10979-007-9091-y
Rogers, R., Rogstad, J. E., Gillard, N. D., Drogin, E. Y., Blackwood, H. L., & Shuman, D. W.
(2010). Everyone knows their Miranda rights: Implicit assumptions and countervailing ev
idence. Psychology, Public Policy, and Law, 16(3), 300–318. doi: 10.1037/a0019316
Russano, M. B., Meissner, C. A., Narchet, F. M., & Kassin, S. M. (2005). Investigating true
and false confessions within a novel experimental paradigm. Psychological Science, 16(6),
481–486.
Schmid, J., & Fiedler, K. (1998). The backbone of closing speeches: The impact of prose
cution versus defense language on judicial attributions. Journal of Applied Social Psychol
ogy, 28(13), 1140–1172. doi: 10.1111/j.1559–1816.1998.tb01672.x
Searle, J. R. (1969). Speech acts: An essay in the philosophy of language. Cambridge:

Shuy, R. (2005). Creating language crimes: How law enforcement uses (and misuses) lan
guage. Oxford: Oxford University Press.
Page 33 of 35

Sigurdsson, J. F., & Gudjonsson, G. H. (1996). The psychological characteristics of “false

confessors”. A study among Icelandic prison inmates and juvenile offenders. Personality
and Individual Differences, 20(3), 321–329. doi: 10.1016/0191-8869(95)00184-0
Smith, A., & Haney, C. (2011). Getting to the point: Attempting to improve juror compre
hension of capital penalty phase instructions. Law and Human Behavior, 35(5), 339–350.
doi: 10.1007/s10979-010-9246-0
Soraci, S. A., Carlin, M. T., Read, J. D., Pogoda, T. K., Wakeford, Y., Cavanagh, S., & Shin,
L. (2007). Psychological impairment, eyewitness testimony, and false memories: Individ
ual differences. In M. P. Toglia, J. D. Read, D. F. Ross, & R. C. L. Lindsay (Eds.), The hand
book of eyewitness psychology, Vol. I: Memory for events (pp. 261–297). Mahwah, NJ:
Lawrence Erlbaum Associates Publishers.
Tavris, C., & Aronson, E. (2007). Mistakes were made (but not by me): Why we justify
foolish beliefs, bad decisions, and hurtful acts. Orlando, FL: Harcourt.
Thomas, G. C. I. (1996). Plain talk about the Miranda empirical debate: A “steady-state”
theory of confessions. U.C.L.A. Law Review, 43, 933–959.
Tiersma, P. M., & Solan, L. M. (Eds.). (2012a). The Oxford handbook of language and law.
Oxford, UK: Oxford University Press.
Tiersma, P. M., & Solan, L. M. (2012b). The language of crime. In L. Solan & P. Tiersma
(Eds.), The Oxford handbook on language and law (pp. 340–353). Cambridge: Oxford Uni
versity Press.
Turnbull, W. (2003). Language in action: Psychological models of conversation. New York:

Routledge.
United States v. Drayton, 536 U.S. (2002).
Vrij, A. (2008). Detecting lies and deceit: Pitfalls and opportunities (2nd ed.). Chichester,
UK: John Wiley.
Wagenaar, W. A. (1995). Anchored narratives: A theory of judicial reasoning, and its con
sequences. In G. Davies, S. Lloyd-Bostock, M. McMurran & C. Wilson (Eds.), Psychology,
law, and criminal justice: International developments in research and practice (pp. 267–
285). Oxford, UK: Walter De Gruyter.
Wiener, R. L., Rogers, M. A., Winter, R., Hurt, L., Hackney, A., Kadela, K.,…Morasco, B.
(2004). Guided jury discretion in capital murder cases: The role of declarative and proce
dural knowledge. Psychology, Public Policy, and Law, 10(4), 516–576. doi:
10.1037/1076-8971.10.4.516
Woodhams, J., Hollin, C. R., Bull, R., & Cooke, C. (2012). Behavior displayed by female vic
tims during rapes committed by lone and multiple perpetrators. Psychology, Public Policy,
and Law, 18(3), 415–452. doi: 10.1037/a0026134
Page 34 of 35

Zulawski, D., & Wicklander, D. (2002). Practical aspects of interview and interrogation.
Boca Raton, FL: CRC Press.
Deborah Davis
Deborah Davis, PhD is Professor of Psychology at the University of Nevada, Reno,

and President of Sierra Trial and Opinion Consultants. She has worked as a jury con
sultant and as an expert witness on witness memory, coerced confessions, and sexual
consent. She publishes frequently in these areas, including a 100-page review article
on interrogations and a chapter on jury selection in the recent Handbook of Forensic
Psychology, and chapters on sources of distortion in eyewitness memory and on
memory for conversation in the recent Handbook of Eyewitness Psychology. Website:
http://www.unr.edu/psych/faculty/davis.html.
J. Guillermo Villalobos
J. Guillermo Villalobos, University of Nevada, Reno
Page 35 of 35

The Role of Language in Conflict and Conflict Resolution
The Role of Language in Conflict and Conflict Resolu

tion
Paul J. Taylor

Conflicts are a ubiquitous part of social life. This chapter explores the role of language in
shaping the way conflicts unfold and resolve. The first section examines the functions of
language in conflict and how different communicative acts relate to speakers’ motivation
al goals and conflict outcome. The second section examines these acts as part of an un
folding interaction. At the microlevel, it examines different forms of cue-response se
quences and their role in managing information exchange and structuring relationships in
conflict. At the macro-level, it examines how episodes of language produce phases and cy
cles that escalate conflict or move it toward a resolution. The final section examines the
link between thought and talk, showing that basic language choices have a profound ef
fect on perceptions and cooperation, which in turn shape language. The chapter ends by
posing unanswered research questions that address the social psychological theory of
language in conflict and the practical issue of how to better use language in conflict reso
lution.
Keywords: Interpersonal processes, cooperation, motivations, goals, outcome, cue-response sequence, phases, ne
gotiation
The US government recently adopted the term “the good stranger” to refer to military
personnel who are adept at gaining cooperation from civilians who might otherwise be
antagonistic or distrustful (DARPA, 2011). The term reflects what has no doubt been ob
served through extended military campaigns in Afghanistan and Iraq (Goodwin, 2005):
adopting a certain demeanor and communicative style is critical in conflicts where one
must work with another whose priorities and beliefs are very different from one’s own. Of
course, the value of understanding what makes a “good stranger” is not confined to mili
tary peacekeeping. It instantiates a question about the nature of what escalates and
deescalates conflicts in contexts as diverse as marital conflict (Gottman, Markman, & No
tarius, 1977), teacher–school board disputes (Putnam, Wilson, & Turner, 1990), and nego
tiations with hostage takers (Giebels & Taylor, 2009). This chapter explores one aspect of
Page 1 of 21

what makes a good stranger; namely, what is known about the role of language in educ
ing cooperation and resolving conflicts.
Language of Conflict and Conflict Resolution

Competition and Cooperation
The basis of much of the social psychological research on language and conflict is a dis
tinction between competitive and cooperative communication. Competitive language is
characterized by behaviors such as justifications, irrelevant arguments, personal attacks,
and excessive demands and threats (Giebels & Noelanders, 2004; Olekalns & Smith,
2003). By contrast, cooperative language is associated with behaviors such as proposals
and counterproposals, agreements, expressions of confidence (p. 460) in the other’s abili
ty, and humor (Donohue & Roberto, 1996; Putnam & Jones, 1982). A person’s use of these
two forms of language depends on their orientation to the conflict. The use of competitive
language is associated with a focus on self and a motivation to maximize personal out
come even at the expense of the other party (often referred to as a distributive or pro-self
orientation). The use of cooperative language is associated with a focus on fairness and a
desire to find ways to satisfy the needs of all parties (often referred to as a integrative or
pro-social orientation). The extent to which these orientations to communication are used
during a conflict has been found to depend on a range of factors. These include person
factors such as culture (e.g., Adair & Brett, 2005) and individual differences (e.g., Park &
Antonioni, 2007) and situational factors such as power differences (e.g., Giebels, De Dreu,
& Van de Vliert, 2000) and emotions (e.g., van Kleef, De Dreu, & Manstead, 2004).
As one might expect, the use of competitive and cooperative language impacts the out
come of conflict. Cooperative language tends to promote conflict resolution and increases
efforts to identify solutions that benefit both parties (Taylor, 2002a). By contrast, competi
tive language is associated with conflict spiraling and a failure to identify areas of com
mon ground and “win-win” solutions (Weingart, Prietula, Hyder, & Genovesse, 1999).
However, the associations between language and outcome are not clear-cut. On some oc
casions, the use of cooperative language has negative consequences, such as when a
shrewd counterpart takes advantage of a cooperator’s goodwill (Murnigham, Babcock,
Thompson, & Pillutla, 1999) or when cooperative messages are overshadowed by the
messages of a single “hawk” (Steinel, De Dreu, Ouwehand, & Ramirez-Marin, 2008). Sim
ilarly, in certain low-stakes scenarios, use of competitive language, such as expressions of
anger, may lead to preferential outcomes, at least for self (Van Kleef, van Dijk, Steinel,
Harinck, & van Beest, 2008). However, in high-stakes scenarios, the use of aggressive
language often leads to reciprocal aggression and conflict spiraling (Donohue & Taylor,
2003; Giebels & Taylor, 2009). Thus, depending on context, the use of cooperative and
competitive language may play an important role for the “good stranger” who seeks to re
solve a conflict. Identifying when and how both types of language work to resolve conflict
remains a dominant topic of research in the field.
Page 2 of 21

Motivational Goals
The distinction between competitive and cooperative language is a broad one that cap
tures a person’s overall orientation to conflict. As such, it does not speak to the variety of
motives or goals that underlie a communicative act (Taylor, 2002b; Wilson & Putnam,
1990). For example, consider a hostage siege in which the police enquire about the
hostages’ welfare and offer to deliver food, while the perpetrator complains about the po
lice snipers and the fact his threats are not taken seriously. These messages serve differ
ent purposes: the police are trying to establish information and initiate a substantive ex
change over food; the perpetrator is trying to assert his “identity” and relay his concerns
about personal safety. Thus, communication may address different aspects of a conflict,
as determined by a person’s strategic choice (De Dreu & Carnevale, 2003), their percep
tions of the other’s intentions (Sillars, Coletti, Parry, & Rogers, 1982), and the way in
which an individual “frames” the interaction (Drake & Donohue, 1996).
Of the possible communicative frames, by far the most studied is how language allows
people to resolve substantive or “instrumental” issues. This is understandable, given the
strong grounding of early conflict resolution research in game theory and social exchange
paradigms (Roloff, 1981; Schelling, 1960). Instrumental goals are broadly concerned with
material “transactions,” which speakers resolve through positional arguments, offers and
counteroffers, rationale persuasion, and so on. This kind of behavior most easily aligns
with a view that actors are “motivated information processors” (De Dreu & Carnevale,
2003) who seek to make sense of the other person’s communication in order to under
stand their goals and beliefs. The language of the motivated information processor is thus
a direct response to his or her inferences about how best to achieve a desired outcome,
and, if he or she is orientated cooperatively, about how best to achieve the other person’s
desired outcome at the same time. When interactants fail to act in ways that maximize
their material reward from the interaction, this is explained by a set of sociocognitive bar
riers. As De Dreu, Beersma, Steinel, and van Kleef (2007) identify, three salient barriers
are heuristics-led judgments, whereby the listener attends to only the salient features of
the other’s message; a naïve realism, whereby the listener interprets the other’s actions
as being motivated by an equivalent set of goals; and ego defensiveness, whereby conflict
triggers a “contend” reaction that leads to competitive (p. 461) behavior against the other.
Support for the importance of these barriers comes from research showing that such bi
ases can be reduced when individuals are motivated to process the other party’s mes
sages in detail (De Dreu & Van Kleef, 2004; Galinsky & Mussweiler, 2001).
A second, quite different kind of language use concerns the “relational” dynamic between
the parties. These messages are less concerned with resolving the substantive disagree
ments of the conflict and more concerned with shaping the affiliation and interdepen
dence between the parties (Donohue, 2001). Speakers manage relational goals by ex
pressing affiliation and liking (e.g., humor), asserting rights and obligations (e.g., justifi
cations and appeals), establishing trust and rapport (e.g., reassurances and promises),
and so on. These behaviors are particularly important outside of the laboratory, where in
teractants are mindful of their reputations, are keen to build long-term relationships, and
Page 3 of 21

are more likely to be under the spotlight of external perceptions (Donohue & Taylor,
2007). Thus, relational language is evident, for example, in messages promoting violent
extremism, which tend to highlight ingroup norms, outgroup immorality, and peer accli
mations over instrumental for-and-against arguments (Prentice, Taylor, Rayson, Hoskins,
& O’Loughlin, 2011). It becomes the prevalent focus of communication in contexts where
the value of relationships is salient because of factors such as cultural norms (Cross, Mor
ris, & Gore, 2002), “power moves” by one party (Donohue & Hoobler, 2002), or a percep
tion that the interaction has consequences beyond the current conflict (Greenhalgh &
Gilkey, 1993). It is also linked to the success of conflict resolution, serving particularly to
establish the interactional roles and trust necessary for information sharing and problem
solving at later stages (Wilson & Putnam, 1990).
A third motivational focus for language concerns the manipulation of interactants’ identi
ty or “face.” Face may be conceived broadly as an “individual’s claimed sense of positive
image in the context of social interaction” (p. 398, Oetzel, Ting-Toomey, Yokochi, Masumo
to, & Takai, 2000; see also Ting-Toomey, chapter 28). A person’s face can be undermined
by competitive “attacking” message that are direct, such as insults or criticisms, or indi
rect, such as negative emotions and commands. Alternatively, face may be addressed co
operatively by messages that “give” or “restore” face, such as compliments and apologies
(Ting-Tooney, 1994). These messages can have a profound impact on the resolution of
conflict because people respond to defend not only their material interests but also their
self-image and social honor. Rogan and Hammer (1994) present clear evidence of this in
their study of crisis negotiations in which the perpetrator was contemplating suicide. In
these interactions, the language of the police negotiators focused on restoring the perpe
trators’ face, whereas the language of the perpetrator focused on restoring self-face and,
critically for the incident that ended in suicide, attacking self-face. However, it is impor
tant to realize that identity-focused messages also impact on conflicts not necessarily dri
ven by issues related to personal crises. For example, disputes between eBay buyers and
sellers are perceived as having a higher likelihood of settlement following interaction in
which sellers offer apologies and confessions and a lower likelihood of settlement follow
ing interactions dominated by commands or negative emotions (Brett et al., 2007). Simi
larly, in cross-cultural conflicts, having one’s honor harmed or insulted can provoke be
havioral reactions of retaliation against the transgressor (Nisbett & Cohen, 1996). Inter
actants with cultural backgrounds in which face is important respond particularly poorly
to implied threats and attempts at intimidation (Adair & Brett, 2005; Giebels & Taylor,
2009).
Language Intensity
A final facet to consider when examining the nature of language in conflict is the role of
language intensity. This aspect of language is often overlooked in studies that examine ut
terance content alone (e.g., from a transcript) rather than content alongside the form of
delivery. High-intensity dialogue includes anger and threats, profanity, obscure
metaphors, dramatic changes in intonation, unqualified compliance, and so on (Bowers,
1963). It reflects “deviat[ion] from neutrality” (Bowers, 1963) in terms of emotional stress
Page 4 of 21

and relational affect (Bradac, Bowers, & Courtright, 1979; Donohue, 2001), as well as a
rigidity of commitment on substantive factors such as persuasion or threat conviction
(Hamilton & Stewart, 1993). For example, relentlessly threatening action if a demand is
not met signifies a high degree of concern for a substantive issue because it does not al
low communication to deviate to alternative possibilities (Taylor, 2002b). Similarly, Mat
sumoto, Hwang, and Frank (2012) demonstrated that leaders of ideologically motivated
groups tend to increase their expressions of anger, contempt, and disgust in public
speeches immediately before acts of violence. Interestingly, (p. 462) although strategic use
of intensity can direct the other party’s attention to an issue of personal concern and can
enhance message clarity (Hamilton, Hunter, & Burgoon, 1990), overuse of language in
tensifiers can reduce the credibility of an argument, particularly when this use contra
venes the receiver’s expectations (Burgoon & Stewart, 1974).
The Structure of Language in Conflict
Fig. 28.1 Schematic representation of language use

during conflict.
Recently, the various facets of language use just described have been shown to be part of
a single communication “structure” in which the various distinctions relate to one anoth
er in distinct ways. Across a number of studies (Bilsky, Tebrugge, & Webel-Therhorn,
2010; Taylor, 2002b; Taylor & Donald, 2004, 2007), individuals have been shown to adopt
an avoidant, competitive, or cooperative orientation to interaction and, at any one time,
pursue either identity, instrumental, or relational goals with varying degrees of intensity.
As individuals progress through a conflict, so do they move through these different re
gions of discussion, ideally moving toward a cooperative interaction. This coming togeth
er of the different facets of language can be depicted graphically as a cylinder, as shown
in Figure 28.1. Over time within a conflict, dialogue may be characterized as being fo
cused on the different areas of the cylinder. On some occasions, for example, a couple
may yell abuse and insults as they thrash out competitive identity-related issues. When
Page 5 of 21

they discuss their child, however, they may revert to more cooperative, instrumental be
havior as they seek to do what’s best for their child.
The cylinder structure highlights other important aspects of the nature of language in
conflict. First, the model encompasses an “avoidant” orientation. People in conflict often
engage in withdrawal behaviors, remaining inactive and avoiding their counterparts as a
result of either fear or strategic intention (Wang, Fink, & Cai, 2012). When used strategi
cally, avoidance allows an interactant to work around an issue that is likely to disrupt
wider progress (Roloff & Ifert, 2000) or stonewall or devalue the other person and the re
lationship through a lack of engagement (Loving, Le, & Crockett, 2009). Second, the
model’s structure suggests that there is a linear relationship from avoidance to competi
tive to cooperative behavior, such that a move from avoidance to cooperation requires a
move through competition. This reasserts the importance of competitive and cooperative
language to successful conflict resolution, and it is consistent with research showing that
a “differentiation” of issues (i.e., in which parties mark their interests) typically precedes
an “integration” of possibilities (i.e., in which parties find mutually beneficial solutions;
De Dreu et al., 2007; Olekalns, Brett, & Weingart, 2003). Third, the identity, relational,
and instrumental goals that interactants may pursue are orthogonal to the possible orien
tations that they may adopt while pursuing their goal. Interactants may take an avoid
ance, competitive, or cooperative approach to each of their different goals, demonstrating
the importance of not conflating a particular orientation to interaction (e.g., cooperation)
with a particular communicative goal. This is sometimes forgotten within the literature,
with some authors conceptualizing relational language as a friendly, cooperative act and
instrumental language as primarily competitive (Pinkley, 1990). Fourth, the model brings
together the distinct sociocognitive systems cited as explanations for each of the various
forms of language use. An individual’s overall orientation to the conflict will be governed
in part by his or her social motivation (Liu & Wilson, 2011), instrumental behavior gov
erned by motivated information processing (De Dreu et al., 2007), relational behavior by
evaluations of emotion as social information (van Kleef, 2009), and so on.
So how do “good strangers” use the various forms of language mapped out in Figure
28.1? An early answer to this question is provided in Ormerod, Barrett, and Taylor’s
(2008) analysis of interactions during hostage negotiations. They showed that interac
tants match one another’s communicative goals and orientations for sustained periods,
with (p. 463) this matching increasing over time for conflicts that ended peacefully but de
creasing over time for those that ended violently. Part of this divergence in matching ap
peared to be due to the fact that, in successful incidents, police negotiators were more
likely to switch their language style to that of the perpetrator. The behavior of these nego
tiators is consistent with “active listening,” which is a communication technique that in
volves the listener beginning his or her response with a reiteration of what he or she has
just heard, in order to confirm a common understanding of what has been said (Royce,
2005). In successful resolutions, police negotiators reduced the amount they spoke by
more than 40 percent during transitional periods in which the motivation behind a
Page 6 of 21

perpetrator’s communication was unclear, presumably in an attempt to reengage their

sense of the perpetrator’s concerns.
The Process of Conflict Resolution

Building Blocks of Conflict Resolution
The different forms of language described in the last section are, of course, the result of
an unfolding exchange of cues and responses. Understanding how conflict occurs at this
micro-level is important to our understanding of how language builds into the larger pat
terns of observed cooperation and conflict spiraling. This is illustrated in an early study
by Gottman, Markman, and Notarius (1977), which showed that marital discussion could
be decomposed into two-behavior loops and three-behavior chains (e.g., probe feeling,
mind reading). These sequences served a variety of functions, such as “ascertaining the
other party’s concerns,” “forming a ‘contract’ regarding future behavior,” and “summariz
ing self versus other’s feelings.” The relative use of these communicative “building
blocks” by couples was sufficient to differentiate those in distressed and nondistressed
relationships.
A body of research has examined the nature of this “lawfulness and inter-connected
ness” (Auld & White, 1959, p. 100) of utterances. The goal of this research is to derive
generalizations about how language organizes over time with a view to understanding the
types of building blocks that differentiate successful and unsuccessful interactions (Wat
zlawick, Beavin, & Jackson, 1968). Kelley (1997), in particular, described the importance
of a sequence of four behaviors comprising an initial cue, a response by the second speak
er, a subsequent adjustment by the initial speaker, and a final closure message that moves
the interaction in a particular direction (Cappella & Planalp, 1981; Taylor & Donald,
2003; Watzlawick, Beavin, & Jackson, 1968). This “triple interact” may be viewed as a lo
cal context or “move” in which interactants modify one another’s understanding and tran
sition to a new position. The importance of this triple interact is borne out in studies of
the interaction sequence. When modeling interactions using Markov chain analysis or
similar techniques, researchers are typically only able to predict a person’s behavior
when two (Krain, 1973; Mark, 1971; Weick, 1969) or three (Cappella & Planalp, 1981;
Mishler, 1975; Watzlawick et al., 1968) previous behaviors are taken into account. Thus, a
person’s behavior in conflict is not a mere reaction to the previous cue of his or her coun
terpart, but a response dependent on at least both parties’ prior behaviors.
The cue–response sequences that tend to dominate conflict interactions are those that
reciprocate or “match” the other party’s actions (Smith, Pruitt, & Carnevale, 1982). Reci
procity allows interactants to confirm their common perspectives and understanding of
the issues at hand while preventing exploitation by parties who strategically compete to
gain advantage (Putnam, 1990; Putnam & Jones, 1982). Although reciprocation is usually
associated with the positive impact it has when parties develop a common solution, the
reciprocation of competitive language is equally difficult to break away from (O’Connor &
Page 7 of 21

Arnold, 2001). Even warning parties about the dangers of conflict spiraling from recipro
cal competition does not reduce the tendency for it to occur (Weingart et al., 1999). Yet,
breaking a chain of reciprocal competition is possible. Brett, Shapiro, and Lytle (1998)
show how negotiators can do this by adopting a form of language that combines a com
petitive position with a more cooperative possibility and by changing their orientation to
one of avoidance (e.g., by labeling the process as ineffectual). These two acts move dia
logue in opposing directions on the orientation axes of the cylinder model in Figure 28.1,
nicely reflecting the different ways in which behavioral chains shape interaction.
As the Brett et al. study suggests, a “good stranger” must ideally do more than recipro
cate cooperation and avoid competition if she or he is to resolve a conflict. Indeed, dyads
that follow a tight, predictable pattern of reciprocating the other sides’ behavior are more
likely to reach impasse (Putnam & Jones, 1982). Examining how interactants achieve a
more dynamic pattern of behavior has led researchers to identify two other kinds of lan
guage sequence. One of these is the complementary sequence, which is formed when
speakers (p. 464) match one another’s language orientation but not their strategic goal.
For example, a complementary chain would be a speaker who focuses on the partner’s
feelings in response to an utterance about a potential custodial solution over the child.
This kind of sequence is often associated with impasse, even when it involves complemen
tary cooperative behaviors (Olekalns & Smith, 2000). Oleklans and Smith argue that the
reason for this failure is that complementarity reflects speakers having different interests
at any one time and a failure to blend cooperative and competitive behavior to address ei
ther party’s issues. A second kind of nonreciprocal sequence is a mismatching of orienta
tions. Although mixing competitive and cooperative behaviors often leads to a win-lose
outcome, it is not always a detrimental sequence. Mismatching in early phases of a con
flict can be beneficial because it allows one side to gain a better understanding of the
partner’s bargaining strengths and limits (Pruitt & Carnevale, 1993). Similarly, as a dead
line looms, it may be more beneficial to mismatch in order to reach an agreement and
avoid impasse (Pruitt & Carnevale, 1993).
Consistent with the research on general language use, some evidence suggests that inter
actants’ preferences for different cue–response sequences is affected by individual differ
ence and context variables. For example, Donohue, Diez, and Hamilton (1984) found
strong role differences in the kinds of messages communicated during labor–manage
ment negotiations. In general, the labor representatives were more likely to reciprocate
the actions of the management representatives, whereas management were more likely
to respond in a complementary way to the labor negotiators (Donohue et al., 1984). Dono
hue et al. also compared these interactions to a set of simulated conflict interactions and,
in doing so, demonstrated that conflict resolution in natural settings involves more se
quences of attacking behavior than found in simulated interactions. This is a critical find
ing because it suggests that the composition of the language building blocks that charac
terize “real-world” conflicts is qualitatively different from that of experimental interac
tions, even though both may yield successful and unsuccessful outcomes.
Page 8 of 21

The influence of external variables also extends to sequences involving relational and
identity concerns (Beune, Giebels, & Taylor, 2010). This is particularly true of cross-cul
tural research, in which these goals can take on particular importance because they
shape how people engage with strategic sequences. For example, Adair and Brett (2005)
compared a group of negotiators from high-context cultures (e.g., Japan, Russia) whose
communication is often implicit and reliant on social expectations with a group of nego
tiators from low-context cultures (e.g., Germany, United States) whose communication is
explicit, with meaning transmitted through the message itself. They found that comple
mentary sequences play a more central role for high-context communicators because the
increased diversity of communication provides the flexibility needed to manage the indi
rect, relational nature of the expected interaction. Giebels and Taylor (2009) also exam
ined the differences across high- and low-context communicators but in the high-stakes
context of hostage negotiation. Using a more flexible measure of the interrelationships
among behaviors known as proximity coefficients, they found that high-context hostage
takers were less likely to engage in persuasive arguments or to respond to them positive
ly and that high-context perpetrators were more likely to reciprocate threats, particularly
when they were made about self. This, they argue, is because high-context communica
tors expect strategic sequences to emphasize relational and identity issues over the ex
change of rational arguments, such that they fail to engage in the persuasive sequences
and respond negatively when the identity dynamic is challenged. This is consistent with
subsequent findings of Beune, Giebels, and Taylor (2010) who showed that high-context
rather than low-context suspects responded negatively to police efforts at being kind, but
positively to intimidating behavior when it was directed toward the social group rather
than toward self.
Building Blocks Make Phases
Although studies of interaction sequences reveal how different kinds of utterances struc
ture the local cue–response exchange, they tell us less about how such dialogue comes to
gether to produce a complete conflict interaction. To address this question, other re
search has sought to understand how dialogue comprises episodes or phases of talk that
serve distinct purposes. These efforts are well illustrated by case study research on his
toric conflicts. For example, in his analysis of negotiations between Spain and the United
States over military base rights, Druckman (1986) showed the cyclical nature of behavior
over time. This multistage negotiation moved through periods of negotiators using com
petitive tactics, such as retractions and threats, and cooperative tactics, such as initia
tions and promises. As the weaker party, the Spanish government (p. 465) typically de
fended its position by using three times more competitive tactics than the US team. How
ever, there were occasions when the opposite was true; US negotiators “ramped up” their
competitive behavior, with the result that the Spanish responded with more cooperative
behavior. This observed mirroring in behavior instantiates what is often referred to as the
“negotiation dance,” with interactants engaging in a “dynamic processes of mutual ac
commodation” (Schelling, 1960, p. 102).
Page 9 of 21

The idea that conflicts move through a series of stages is encapsulated in phase models of
interaction. A phase model presents either a descriptive or prescriptive account of how a
set of coherent stages of communication change as a conflict unfolds (Holmes, 1992). For
example, in her field observations of private disputes, Douglas (1957) found that interac
tions move through three phases that cannot be skipped or ordered differently. These
were establishing the conflict parameters through a competitive exchange, a cooperative
reconnoitering to identify possible solutions, and a final period of competition to “sort out
the final details.” Others have since refined Douglas’s proposition by suggesting addition
al phases. For example, in his analysis of conflict interactions across several contexts,
Gulliver (1979) identifies eight phases: search for an arena, agenda formation, explo
ration of the issue range, narrowing of the differences, preliminaries, final bargaining, rit
ual confirmation, and execution of the settlement. As the Douglas and Gulliver examples
illustrate, the number of phases identified by researchers depends on the granularity of
the analysis undertaken. Indeed, Gulliver’s eight phases are often paired to produce the
four phases of establishing contact, relationship building, problem solving, and resolution
(e.g., Donohue, Kaufmann, Smith, & Ramesh, 1991; Holmes, 1992).
The majority of studies of conflict utilize phase models in a prescriptive sense by dividing
the interaction under examination into the requisite number of periods. If a four-phase
model is considered an appropriate framework for understanding the organization of the
interaction, then the dialogue itself is broken into four, equally-sized segments and
changes in language are examined across these segments. Although this approach ig
nores the possibility that phases may take different amounts of time to complete (Taylor,
2002b), it nevertheless reveals some important patterns in language use. For example,
proposals increase in a linear fashion over time, with the use of instrumental language
peaking during the second and particularly third quarter of the negotiation, in line with
the idea of a problem-solving phase (Lytle, Brett, & Shapiro, 1999). By contrast, use of af
fect during conflict tends to peak during the early establishing contact stage of interac
tion, only then to reemerge during the problem-solving stage as parties compete over po
sition (Adair & Brett, 2005; Rogan & Hammer, 1995). This second peak in language affect
appears particularly prominent in high-stakes scenarios in which impasse and threats of
withdrawal may dominate prior to the “endgame” (Abbott, 1986).
Although a significant number of studies use phase models as frameworks for examining
interaction over time, only a few have tested whether moving through the prescribed
phases promotes successful conflict resolution. If prescribed phase models do not pro
mote conflict resolution, then there is little point in the “good stranger” trying to adhere
to the event order that the models prescribe. In an early study, Hirokawa (1983)
compared the phases of discussion produced by groups who developed successful and un
successful solutions to a traffic-flow problem. He found no clear distinction in the set of
phases associated with success or failure, but a clear difference in the timing with which
successful and unsuccessful groups engaged in problem solving; successful groups began
a problem-solving phase of interaction far sooner than did their unsuccessful counter
parts. In a more detailed study, Sambamurthy and Poole (1992) examined the relationship
of conflict management phases to the degree of consensus change, perceptions of deci
Page 10 of 21

sion quality, and satisfaction with the resolution. Groups who showed the most consensus
change and the highest decision quality moved through phases of interaction that map
onto the prescribed phase sequences for negotiations. This is consistent with research
showing that negotiation supports systems that guide users through prescribed phases of
negotiation, increase the joint outcome from a conflict, and produce more balanced con
tracts (Foroughi, Perkins, & Jelassi, 1995).
One of the key questions for phase models of dialogue is what prompts changes in the fo
cus of an interaction. In conflict research, these points are often referred to as turning
points, defined as the “events or processes that mark passage from one stage to the next,
signaling progress from earlier to later phases” (Druckman, 1997, p. 92). Druckman
(2001) identifies three causes of turning points: procedural (e.g., a change from public to
private venue), substantive (e.g., new concepts introduced), or external (e.g., introduction
of a (p. 466) third party). One precursor for turning points appears to be a crisis that jeop
ardizes the interactions (Druckman, 1986). For example, Harinck and De Dreu (2004)
examined what occurred after a period of temporary impasses, when two parties dead
locked on a particular issue. They found early competitive language was related to im
passes within a negotiation, but that such impasses were valuable in the long run be
cause they prompted problem-solving dialogue later in the interaction. It appears that im
passe allows negotiators to step back from the interaction in a way that facilitates a
switch away from escalating conflict.
The Link Between Thought and Talk

In examining the language of conflict resolution, it is easy to focus on the content and
outcome of communication while giving little consideration to how the message is being
consumed (Olekalns & Smith, 2005). For many purposes, it is quite reasonable to assume
that speakers process the messages of the other party in a way consistent with their own
orientations and goals. However, to derive a full account of how cooperation emerges, it
is useful to model the factors that influence the way in which individuals makes sense of a
counterpart’s communication. The importance of addressing the issue of interpretation is
made clear in Sillars, Roberts, Leonard, and Dun’s (2000) analysis of partners’ in vivo
thoughts during marital conflict. Sillars et al. video taped couples discussing a current is
sue on which they were in conflict and then presented this recording to each partner, ask
ing him or her to describe his or her thoughts and feelings as experienced across the in
teraction. Overall, spouses viewed their own communication in more favorable terms than
their partner’s communication. In particular, the extent to which speakers believed that
they were acting in a confrontational manner and the extent to which their partners be
lieved that they were acting in a confrontational manner was negatively correlated, sug
gesting that partners in conflict build opposing interpretative frameworks of what is be
ing discussed. Husbands, in particular, perceived their messages as being far more coop
erative than independent judges who rated their language use.
Page 11 of 21

So, what is the relationship between goals, interpretations, and talk? According to com
munication accommodation theory (Giles & Coupland, 1991), individuals’ use of language
is important to the creation and maintenance of the social dynamic between themselves
and the other party. The interaction alignment model (Pickering, this volume) takes this
one step further by arguing that successful dialogue and shared understanding is
achieved through a mechanistic process of coordination across multiple levels of lan
guage, from lexical choice, to syntactic and semantic preferences, to message framing. In
this model, a speaker’s lexical, syntactic, and semantic choices are part of his cognitive
representation of the conflict. Their utterance leads to the activation of a matching repre
sentation in the other speaker, which has the behavioral consequence of shaping that
person’s response. This exchange reciprocates until speakers come to view the world in a
similar way. The key implication of this model is that progress toward cooperation comes
not only from high-level alignment of orientation and goal, as explored in the previous
section, but also from a more basic coordination of language.
The notion that conflict is played out at the level of word use, as well as at the level of
strategic choice is supported by research. Indirect support comes from evidence showing
that the persuasive effectiveness of a message is positively related to perceived communi
cator–recipient similarity in lexical diversity (Bradac et al., 1979), language intensity
(Aune & Kikuchi, 1993), and speech rate (Street, Brady, & Putnam, 1983). Since persua
sion is an important facet of language use in conflict (Giebels & Taylor, 2009), these cor
relates suggest that language similarity may impact conflict outcome. More direct evi
dence comes from studies showing that the use of assents and positive emotional lan
guage has a positive impact on negotiation success, whereas the use of negative emotion
al language may promote competition (Curhan & Pentland, 2007; Van Beest, Van Kleef, &
Van Dijk, 2008). This pattern of language is also evidence in coalition formation in multi
party negotiations, in which group members show greater convergence in language use
as they move toward agreement and divergence when they move apart (Huffaker, Swaab,
& Diermeier, 2011).
What is perhaps particularly interesting about the link between language use and conflict
outcome is that cooperation appears to be associated with the coordination of language
style. A person’s language style is constructed by his or her choice of function words and
reflects not the content of his or her message but the way in which it is being conveyed.
These function words include articles, auxiliary verbs, and pronouns and are typically un
derstood to be outside of conscious control (Ireland & Pennebaker, 2010). Using a specif
ic measure of function word coordination known as language style matching (LSM; see
Pennebaker, this volume), a number of (p. 467) studies have shown a positive relationship
between LSM and interaction outcome. Gonzales, Hancock, and Pennebaker (2010) found
that cooperation in small-group activities goes hand in hand with greater matching of cat
egories of word use. Scissors, Gill, and Gergle (2009) linked increased language mimicry
with less likelihood of defecting in a trader task. Taylor and Thomas (2008) showed that
successful hostage negotiations were associated with higher aggregate levels of LSM
than were unsuccessful negotiations due to dramatic fluctuations of LSM during unsuc
cessful negotiations. Negotiators in these situations appeared unable to maintain the con
Page 12 of 21

stant levels of rapport and coordination that occurred in successful negotiations. More
over, by examining LSM at the local turn-by-turn level, Taylor and Thomas further re
vealed complex but organized variations in behavior across outcome. In comparison to
unsuccessful negotiations, the dialogue of successful negotiations involved greater coor
dination of turn taking, a focus on the present rather than the past, reciprocation of posi
tive affect, and a focus on alternatives rather than on competition.
The impact of language on thought can be quick, with people’s opening words in a con
flict impacting the subsequent outcome of their interaction. In their examination of the
first 5 minutes of a simulated employment negotiation, Curhan and Pentland (2007) found
that conversational engagement, prosodic emphasis, and vocal mirroring predicted 30
percent of the variance in individual outcomes. The conversational dynamics associated
with success among high-status parties were different from those associated with success
among low-status parties. Examining whether the timing of language mimicry matters to
outcome, Swaab, Maddux, and Sinaceur (2011) found that high levels of language mimic
ry in the first 10 minutes of an interaction was conducive of better negotiation outcomes
than high levels of language mimicry in the last 10 minutes of interaction. This was be
cause early mimicry elicited more trust from the speaker’s counterpart, which in turned
allowed individuals to maximize their outcomes. By contrast, late mimickers tended to
mimic language that was more accommodating to the other negotiator. Such mimicking of
accommodating language in the final phase of interaction seemed to impair individuals’
ability to protect their own interests.
Being the Good Stranger

This chapter has set out some of the major themes in research on the social psychology of
language in conflict resolution. In doing so, it has shown how educing cooperation from
those who may have alternative motivations and goals requires a comprehensive under
standing of interaction processes. Conflict resolution cannot be distilled into an examina
tion of positions or to the use of particular influence tactics. It is a process in which dif
ferent levels and different kinds of language may shape progression, be that in the first
10 seconds or in the 10th hour of negotiations. As a process, it is also not isolated to lan
guage. Critical to many conflict resolution scenarios is also a set of nonverbal dynamics,
contextual constraints such as expectations from social groups, and preconceptions and
emotional history that can radically shape expectations and behavior. However, what is
clear is that language is the vehicle for much of what occurs across a conflict. Thus,
learning how language structures an unfolding conflict and its move toward resolution
will play a key role in identifying the “good stranger.”
Unanswered Research Questions

• What role, if any, do interpersonal skills in impression management and interperson
al sensitivity play in efforts to engender cooperation?
Page 13 of 21

• How do we reconcile the top-down cognitive model of conflict behavior, as encapsu

lated by theories such as motivated information processor, with bottom-up models of
interaction processes?
• To what extent do findings from experimental studies of conflict resolution translate
to real world conflicts?
• How can third parities be effective?
• How does language interact with nonverbal behavior to shape cooperation?
References
Abbott, A. (1986). Time-phase model of hostage negotiation. The Police Chief, pp. 34–35.
Alexandria, VA: International Asssociation of Chiefs of Police. ISSN 0032-2571.
Adair, W. L., & Brett, J. M. (2005). The negotiation dance: Time, culture, and behavioral
sequences in negotiation. Organization Science, 16, 33–51.
Auld, F., & White, A. M. (1959). Sequential dependencies in psychotherapy. Journal of Ab
normal and Social Psychology, 58, 100–104.
Aune, R. K., & Kikuchi, T. (1993). Effects of language intensity similarity on perceptions of
credibility, relational attributions, and persuasion. Journal of Language and Social Psy
chology, 12, 224–238.
Beune, K., Giebels, E., & Taylor, P. J. (2010). Patterns of interaction in police interviews:
The role of cultural dependency. Criminal Justice and Behavior, 37, 904–925.
Bilsky, W., Tebrugge, B., & Webel-Therhorn, D. (2010). Escalation and deescalation in
hostage-negotiation. In R. G. Rogan & F. J. Lanceley (Eds.), Contemporary theory, re
search (p. 468) and practice of crisis and hostage negotiation (pp. 119–140). Cresskill, NJ:
Hampton Press.
Bowers, J. W. (1963). Language intensity, social introversion, and attitude change. Speech
Monographs, 30, 415–420.
Bradac, J. C., Bowers, J. W., & Courtwright, J. A. (1979). Three language variables in com
munication research: Intensity, immediacy and diversity. Human Communication Re
search, 5, 257–269.
Brett, J. M., Olekalns, M., Friedman, R., Goates, N., Anderson, C., & Lisco, C. C. (2007).
Sticks and stones: Language, face, and online dispute resolution. Academy of Manage
ment Journal, 50, 85–99.
Brett, J. M., Shapiro, D. L., & Lytle, A. L. (1998). Breaking the bonds of reciprocity in ne
gotiations. Academy of Management Journal, 41, 410–424.
Page 14 of 21

Burgoon, M., & Stewart, D. L. (1974). Toward a message-centered theory of persuasion:

Empirical investigations of language intensity: The effects of sex of source, receiver, and
language intensity on attitude change. Human Communication Research, 1, 241–248.
Cappella, J. N., & Planalp, S. (1981). Talk and silence sequences in informal conversations
III: Interspeaker influence. Human Communication Research, 7, 117–132.
Cross, S. E., Morris, M. L., & Gore, J. S. (2002). Thinking about oneself and others: The
relational-interdependent self-construal and social cognition. Journal of Personality and
Curhan, J. R., & Pentland, A. (2007). Thin slices of negotiation: Predicting outcomes from
conversational dynamics within the first 5 minutes. Journal of Applied Psychology, 92,
802–811.
DARPA. (2011). Strategic social interaction modules (DARPA-BAA-11–32). Retrieved No

vember 30, 2012 from https://www.fbo.gov/download/ddb/
ddb2a40c0a44d3f611a91d6992c05e73/DARPA-BAA-11-32_SSIM.pdf
De Dreu, C. K. W., Beersma, B., Steinel, W., & Van Kleef, G. A. (2007). The psychology of
negotiation: Principals and basic processes. In A. W. Kruglanski, & E. T. Higgins (Eds.),
Handbook of basic principles in social psychology (2nd ed., pp. 608–629). New York: Guil
ford.
De Dreu, C. K. W., & Carnevale, P. J. (2003). Motivational bases of information processing

and strategy in conflict and negotiation. Advances in Experimental Social Psychology, 35,
235–291.
De Dreu, C. K. W., & van Kleef, G. A. (2004). The influence of power on the information
search, impression formation, and demands on negotiation. Journal of Experimental So
Donohue, W. A. (2001). Resolving relational paradox: The language of conflict in relation

ships. In W. F. Eadie, & P. E. Nelson (Eds.), The language of conflict resolution (pp. 21–46).
Thousand Oaks, CA: Sage.
Donohue, W. A., Diez, M. E., & Hamilton, M. (1984). Coding naturalistic negotiation inter
action. Human Communication Research, 10, 403–425.
Donohue, W. A., & Hoobler, G. D. (2002). Relational frames and their ethical implications
in international negotiation: An analysis based on the Oslo II negotiation. International
Negotiation, 7, 143–167.
Donohue, W. A., Kaufmann, G., Smith, R., & Ramesh, C. (1991). Crisis bargaining in in
tense conflict situations. International Journal of Group Tensions, 21, 133–153.
Page 15 of 21

Donohue, W. A., & Roberto, A. J. (1996). An empirical examination of three models of inte
grative and distributive bargaining. International Journal of Conflict Management, 7, 209–
229.
Donohue, W. A., & Taylor, P. J. (2003). Testing the role effect in terrorist negotiations. In
ternational Negotiation, 8, 527–547.
Donohue, W. A., & Taylor, P. J. (2007). Role effects in negotiation: The one-down phenome
non. Negotiation Journal, 23, 307–331.
Douglas, A. (1957). The peaceful settlement of industrial and intergroup disputes. Journal
of Conflict Resolution, 1, 69–81.
Drake, L. E., & Donohue, W. A. (1996). Communicative framing theory in conflict resolu
tion. Communication Research, 23, 297–322.
Druckman, D. (1986). Stages, turning points, and crises: Negotiating military base rights,
Spain and the United States. Journal of Conflict Resolution, 30, 327–360.
Druckman, D. (1997). Negotiating in the international context. In W. Zartman, and J. L.

Rasmussen (Eds.), Peacemaking in international conflict: Methods and techniques, I (pp.
81–124). Washington DC: United States Institute of Peace Press.
Druckman, D. (2001). Turning points in international negotiation: A comparative analysis.

Journal of Conflict Resolution 45, 519–544.
Foroughi, A., Perkins, W. C., & Jelassi, M. T. (1995). An empirical study of an interactive,
session-oriented computerized negotiation support system (NSS). Group Decision and Ne
gotiation, 4, 485–512.
Galinsky, A. D., & Mussweiler, T. (2001). First offers as anchors: The role of perspective-
taking and negotiator focus. Journal of Personality and Social Psychology, 81, 657–669.
Giebels, E., De Dreu, C. K. W., & van de Vliert, E. (2000). Interdependence in negotiation:
Effects of exit options and social motive on distributive and integrative negotiation. Euro
pean Journal of Social Psychology, 30, 255–272.
Giebels, E., & Noelanders, S. (2004). Crisis negotiations: A multiparty perspective. Vee
nendaal, The Netherlands: Universal Press.
Giebels, E., & Taylor, P. J. (2009). Interaction patterns in crisis negotiations: Persuasive
arguments and cultural differences. Journal of Applied Psychology, 94, 5–19.
Giles, H., & Coupland, N. (1991). Language: Contexts and consequences. Pacific Grove,
CA: Brooks Cole.
Gonzales, A. L., Hancock, J. T., & Pennebaker, J. W. (2010). Language style matching as a
predictor of social dynamics in small groups. Communication Research. 37, 3–19.
Page 16 of 21

Goodwin, D. (2005). The military and negotiation: The role of the solider-diplomat. Abing
don, Oxford: Frank Cass.
Gottman, J. M., Markman, H., & Notarius, C. (1977). The topography of marital conflict: A
sequential analysis of verbal and non-verbal behavior. Journal of Marriage and the Family,
39, 461–477.
Greenhalgh, L., & Gilkey, R. W. (1993). The effect of relationship orientation on negotia
tors’ cognitions and tactics. Group Decision and Negotiation, 7, 465–489.
Gulliver, P. H. (1979). Disputes and negotiations: A cross-cultural perspective. New York:

Academic Press.
Hamilton, M. A., Hunter, J. E., & Burgoon, M. (1990). An empirical test of an axiomatic
model of the relationship between language intensity and persuasion. Journal of Lan
guage and Social Psychology, 9, 235–56.
Hamilton, N. A., & Stewart, B. L. (1993). Extending an information processing model of

language intensity effects. Communication Quarterly, 41, 231–246.
(p. 469) Harinck, F., & De Dreu, C. K. W. (2004). Negotiating values or resources: The
moderating impact of time pressure. European Journal of Social Psychology, 34, 595–612.
Hirokawa, R. Y. (1983). Group communication and problem-solving effectiveness: An in

vestigation of group phases. Human Communication Research, 4, 312–321.
Holmes, M. E. (1992). Phase structures in negotiation. In L. L. Putnam, & M. E. Roloff

(Eds.), Communication and negotiation (pp. 83–105). Newbury Park, CA: Sage.
Huffaker, D., Swaab, R. I., & Diermeier, D. (2011). The language of coalition formation in
online multiparty negotiations. Journal of Language and Social Psychology, 30, 66–81.
Ireland, M. E., & Pennebaker, J. W. (2010). Language style matching in writing: Syn
chrony in essays, correspondence, and poetry. Journal of Personality and Social Psycholo
gy, 99, 549–571.
Kelley, H. H. (1997). Expanding the analysis of social orientations by reference to the se
quential-temporal structure of situations. European Journal of Social Psychology, 27, 373–
404.
Krain, M. (1973). Communication as a process of dyadic organization and development.

Journal of Communication, 23, 392–408.
Liu, M., & Wilson, S. R. (2011). Negotiation tactics and outcomes: A dyad-level analysis
across two cultures. Communication Research, 38, 248–277.
Loving, T. J., Le., B., & Crockett, E. E. (2009). The physiology of feeling hurt. In A. L. Van
gelisti (Ed.), Feeling hurt in close relationships (pp. 359–375). New York: Cambridge Uni
versity Press.
Page 17 of 21

Lytle, A. L., Brett, J. M., & Shapiro, D. L. (1999). The strategic use of interests, rights, and
power to resolve disputes. Negotiation Journal, 15, 31–51.
Mark, R. A. (1971). Coding communication at the relationship level. Journal of Communi

cation, 21, 221–232.
Matsumoto, D., Hwang, H. C., & Frank, M. G. (2012). Emotions expressed in speeches by
leaders of ideologically motivated groups predict aggression. Behavioral Sciences of Ter
rorism and Political Aggression, iFirst Article, 1–18. doi: 10.1080/19434472.2012.716449
Mishler, E. G. (1975). Studies in dialogue and discourse: Types of discourse initiated by

and sustained through questionings. Journal of Psycholinguistic Research, 4, 99–121.
Murnigham, J. K., Babcock, L., Thompson, L., & Pillutla, M. (1999). The information
dilemma in negotiations: Effects of experience, incentives and integrative potential. Inter
national Journal of Conflict Management, 10, 313–339.
Nisbett, R. E., & Cohen, D. (1996). Culture of honor: The psychology of violence in the
South. Boulder, CO: Westview.
O’Connor, K. M., & Arnold, J. A. (2001). Distributive spirals: Negotiation impasses and the
moderating role of disputant self-efficacy. Organizational Behavior and Decision Process
es, 84, 148–176.
Oetzel, J. G., Ting-Toomey, S., Yokochi, Y., Masumoto, T., & Takai, J. (2000). A typology of
facework behaviors in conflicts with best friends and relative strangers. Communication
Quarterly, 48, 397–419.
Olekalns, M., Brett, J. M., & Weingart, L. R. (2003). Phases, transitions and interruptions:
The processes that shape agreement in multi-party negotiations. International Journal of
Conflict Management, 14, 191–211.
Olekalns, M., & Smith, P. L. (2000). Understanding optimal outcomes: The role of strategy
sequences in competitive negotiations. Human Communication Research, 26, 527–557.
Olekalns, M., & Smith, P. L. (2003). Simple frequency effects: Motivational orientation,
strategy selection and sequencing, and outcome optimality in negotiations. Journal of Ex
perimental Social Psychology, 39, 101–117.
Olekalns, M., & Smith, P. L. (2005). Cognitive representations of negotiation. Australian

Journal of Management, 39, 57–76.
Ormerod, T., Barrett, E., & Taylor, P. J. (2008). Investigative sensemaking in criminal con
texts. In J. M. Schraagen, L. G. Militello, T. Ormerod, & R. Lipshitz (Eds.), Naturalistic de
cision making and macrocognition (pp. 81–102). Aldershot, UK: Ashgate.
Park, H., & Antonioni, D. (2007). Personality, reciprocity, and strength of conflict resolu
tion strategy. Journal of Research in Personality, 41, 110–125.
Page 18 of 21

Pinkley, R. L. (1990). Dimensions of conflict frame: Disputant interpretations of conflict.

Journal of Applied Psychology, 75, 117–126.
Prentice, S., Taylor, P. J., Rayson, P., Hoskins, A., & O’Loughlin, B. (2011). Analyzing the
semantic content and persuasive composition of extremist media: A case study of texts
produced during the Gaza conflict. Information Systems Frontiers, 13, 61–73.
Pruitt, D. G., & Carnevale, P. J. (1993). Negotiation in social conflict. Pacific Grove, CA:
Brooks-Cole.
Putnam, L. L. (1990). Reframing integrative and distributive bargaining: A process per

spective. In B. H. Sheppard, M. H. Bazerman, & R. J. Lewicki (Eds.), Research on negotia
tion in organizations (pp. 3–30). Greenwich, CN: JAI Press.
Putnam, L. L., & Jones, T. S. (1982). Reciprocity in negotiations: An analysis of bargaining

interaction. Communication Monographs, 49, 171–191.
Putnam, L. L., Wilson, S. R., & Turner, D. B. (1990). The evolution of policy arguments in
teachers’ negotiations. Argumentation, 4, 129–152.
Rogan, R., & Hammer, M. (1994). Crisis negotiations: A preliminary investigation of face
work in naturalistic conflict. Journal of Applied Communication Research, 22, 216–231.
Rogan, R., & Hammer, M. (1995). Assessing message affect in crisis negotiations: An ex
ploratory study. Human Communication Research, 21, 553–574.
Roloff, M. E. (1981). Interpersonal communication: The social exchange approach. Bever

ly Hills, CA: Sage.
Roloff, M. E., & Ifert, D. E. (2000). Conflict management through avoidance: Withholding
complaints, suppressing arguments, and declaring topics taboo. In S. Petronio (Ed.), Bal
ancing the secrets of private disclosure (pp. 151–163). Mahwah, NJ: Erbaum.
Royce, T. (2005). The negotiator and the bomber: Analyzing the critical role of active lis
tening in crisis negotiation. Negotiation Journal, 21, 5–27.
Sambamurthy, V., & Poole, M. S. (1992). The effects of variations in capabilities of GDSS
designs on management of cognitive conflicts in groups. Information Systems Research,
3, 224–251.
Schelling, T. C. (1960). The strategy of conflict. Boston: Harvard University Press.
Scissors, L. E., Gill, A. G., Geraghty, K., & Gergle, D. (2009). In CMC we trust: The role of
similarity. In Proceedings of the ACM conference on computerhuman interaction (CHI)
2009 (pp. 527–536). New York: ACM Press.
Sillars, A., Coletti, S. F., Parry, D., & Rogers M. A. (1982). Coding verbal conflict tactics:
Nonverbal and perceptual correlates of the “avoidance-distributive-integrative” distinc
tion. Human Communication Research, 9, 83–95.
Page 19 of 21

Sillars, A., Roberts, L. J., Leonard, K. E., & Dun, T. (2000). Cognition during mari
(p. 470)
tal conflict: The relationship of thought and talk. Journal of Social and Personal Relation
ships, 17, 479–502.
Smith, D, L., Pruitt, D. G., & Carnevale, P. J. (1982). Matching and mismatching: The ef
fect of own limit, other’s toughness, and time pressure on concession rate in negotiation.
Steinel, W., De Dreu, C. K. W., Ouwehand, E., & Ramirez-Marin, J. Y. (2008). When con
stituencies speak in multiple tongues: The relative persuasiveness of hawkish minorities
in representative negotiation. Organizational Behavior and Human Decision Processes,
109, 67–78.
Street, R. L., Jr., Brady, R., & Putnam, W. (1983). The influence of speech rate stereotypes
and rate similarity on listener’s evaluations of speakers. Journal of Language and Social
Swaab, R. I., Maddux, W. W., & Sinaceur, M. (2011). Early words that work: When and
how virtual linguistic mimicry facilitates negotiation outcomes. Journal of Experimental
Taylor, P. J. (2002a). A partial order scalogram analysis of communication behavior in cri

sis negotiation with the prediction of outcome. International Journal of Conflict Manage
ment, 13, 4–37.
Taylor, P. J. (2002b). A cylindrical model of communication behavior in crisis negotiations.

Human Communication Research, 28, 7–48.
Taylor, P. J., & Donald, I. J. (2003). Foundations and evidence for an interaction based ap
proach to conflict. International Journal of Conflict Management, 14, 213–232.
Taylor, P. J., & Donald, I. J. (2004). The structure of communication behavior in simulated
and actual crisis negotiations. Human Communication Research, 30, 443–478.
Taylor, P. J., & Donald, I. J. (2007). Testing the relationship between local cue-response
patterns and global dimensions of communication behavior. British Journal of Social Psy
chology, 46, 273–298.
Taylor, P. J., & Thomas, S. (2008). Linguistic style matching and negotiation outcome. Ne
gotiation and Conflict Management Research, 1, 263–281.
Ting-Toomey, S. (1994). Managing intercultural conflicts effectively. In L. Samovar & R.

Porter (Eds.), Intercultural communication: A reader (7th ed., pp. 360–372). Belmont, CA:
Wadswort.
Van Beest, I., & Van Kleef, G. A., & Van Dijk, E. (2008). Get angry, get out: The interper
sonal effects of anger communication in multiparty negotiation. Journal of Experimental
Page 20 of 21

Van Kleef, G. A. (2009). How emotions regulate social life: The emotions as social informa
tion (EASI) model. Current Directions in Psychological Science, 18, 184–188.
Van Kleef, G. A., De Dreu, C. K. W., & Manstead, A. S. R. (2004). The interpersonal effects
of anger and happiness in negotiations. Journal of Personality and Social Psychology, 86,
57–76.
Van Kleef, G. A., van Dijk, E., Steinel, W., Harinck, F., & van Beest, I. (2008). Anger in so
cial conflict: Cross-situational comparisons and suggestions for the future. Group Deci
sion and Negotiation, 17, 13–30.
Wang, Q., Fink, E. L., & Cai, D. A. (2012). The effect of conflict goals on avoidance strate
gies: What does not communicating communicate? Human Communication Research, 38,
222–252.
Watzlawick, P., Beavin, J. H., & Jackson, D. D. (1968). Pragmatics of human communica
tion. London: Faber and Faber.
Weick, K. (1969). The social psychology of organizing. Reading, MA: Addison-Wesley.
Weingart, L. R., Prietula, M. J., Hyder, E. B., & Genovese, C. R. (1999). Knowledge and the
sequential processes of negotiation; A Markov chain analysis of response-in-kind. Journal
of Experimental Social Psychology 35, 366–393.
Wilson, S. R., & Putnam, L. L. (1990). Interaction goals in negotiation. In J. Anderson

(Ed.), Communication yearbook 13 (pp. 374–406). Newbury Park, CA: Sage.
Paul J. Taylor
Paul Taylor, Fylde College, Lancaster University
Page 21 of 21

Computer-Mediated Communication
Susan R. Fussell and Leslie D. Setlock

Over the past 30 years, an increasing proportion of communication has occurred over the
Internet, using a wide array of new technologies including video conferencing, audio con
ferencing, text-based messaging, and social media, and prompting a new field of research
on the interaction between media and communication. This chapter explores the ways
that computer-mediated communication (CMC) scholars have conceptualized differences
among media, with a focus on Clark and Brennan’s (1991) influential theory of media af
fordances, and it reviews empirical research that has examined the role of these affor
dances in CMC. The role of individual characteristics (e.g., expertise, cultural back
ground, native language) and group dynamics (e.g., group size, status) as mediating fac
tors of the effects of media affordances on conversational grounding and task perfor
mance are discussed.
Keywords: Computer-mediated communication, social media, affordances, communication, empirical, group com
munication, collaborative work
Over the past 30 years, an increasing proportion of people’s interactions with others have
occurred over the Internet, using a wide array of new technologies including video con
ferencing, audio conferencing, text-based messaging, and social media. These new tech
nologies each support some aspect of face-to-face communication while failing to support
others. For example, video conferencing allows people to see each other’s faces and hear
one another, but other aspects of communication such as body posture and hand gestures
are frequently not visible. Since at least Chapanis’s early studies (Chapanis, 1975), re
searchers in the area of computer-mediated communication (CMC) have been interested
in how the features and constraints of different media affect fundamental communicative
processes such as message construction, message comprehension, conversational
grounding, and relationships between communicators.
Computer-mediated communication is a broad field, encompassing psychological, socio

logical, organizational science, communication, computer science, and information sci
ence perspectives. Research has focused on the dyadic level (e.g., use of different entry
methods for text messaging; Hancock & Dunham, 2001), small-group or team level (e.g.,
Page 1 of 40

how well media affect group decision-making; McGuire, Kiesler, & Siegel, 1987), organi
zational level (e.g., to what extent employees use different media; Rice, 1993), and soci
etal level (e.g., how the adoption of text messaging and other communication media is
changing society; Turkle, 2011). In this chapter, we focus on the dyadic and small-group
level interactions. We specifically present research that has its theoretical roots in the
psychology literature and common psychology methods such as laboratory studies.
We first discuss ways that CMC scholars have conceptualized differences among media,
with a (p. 472) focus on Clark and Brennan’s (1991) influential theory of media affor
dances. We then review empirical research that has examined the role of these affor
dances in CMC. Next, we consider how some individual characteristics (e.g., expertise,
cultural background, native language) and the number of participants in the conversation
shape the effects of media affordances on conversational grounding and task perfor
mance. Finally, we discuss several current limitations of the framework with respect to
the changing landscape of CMC tools and the tendency to focus on task-related communi
cation in lieu of other dimensions of an interaction.
Conceptualizing Differences Among Media

People have a wide array of tools available to communicate with partners in other loca
tions, some of the most popular of which include the phone, text media such as instant
messaging (IM), Twitter or short message service (SMS), and video conferencing. A fun
damental assumption of much if not all research on CMC is that the specific features of
individual tools shape the ways that people use them for conversation and with what suc
cess.
Early CMC theories attempted to catalogue and distinguish media features along one or
two dimensions, such as Daft and Lengel’s (1984) notion of media richness. In their view,
“[m]edia can be characterized as high or low in ‘richness’ based on their capacity to facil
itate shared meaning. A rich medium facilitates insight and rapid understanding” (Daft,
Lengel, & Trevino, 1987). E-mailed memos to all employees of a company would be con
sidered very low in richness because they are impersonal and lack the nonverbal cues of
face-to-face conversation. Video conferencing would be higher in richness (although not
as high as face-to-face conversation) because conversations are personal and some non
verbal cues are available. According to Daft and Lengel, the successful use of CMC re
quired selecting the right level of richness for the task. For example, an e-mailed memo
might be appropriate for broadcasting general news, but it would be inappropriate for,
say, firing an employee. Short, Williams, and Christie (1976) suggested a similar dimen
sion of social presence, as the awareness of others in an interaction combined with an ap
preciation of the interpersonal aspects of that interaction. Therefore, social presence
would provide the speaker with awareness of who was hearing (seeing) the memo, includ
ing but not limited to those most directly affected by it, and the social impact of the news
the memo would deliver.
Page 2 of 40

It quickly became apparent, however, that single-dimension theories could not differenti
ate clearly enough among media. A more sophisticated framework for thinking about dif
ferences among media was provided by Clark and Brennan (1991), who used a finer-
grained analysis of the features or affordances media provide and how these affect the
nature of communication using those media (see Table 29.1). For example, telephone calls
and video conferencing provide audibility and thus afford the use of vocal speech produc
tion, whereas IM does not. Video conferencing provides visibility so facial expressions can
be seen, whereas telephone calls do not. In this framework, communication over different
media will entail different costs for producing messages, receiving and understanding
messages, changing speakers, and repairing misunderstandings.
Page 3 of 40

Table 29.1 Some affordances of communication media and their typical presence (Y), partial presence (P), or ab
sence (N) in face-to-face communication (FtF), video conferencing, telephone conversations, and instant messaging
(IM).
Affordance Definition FTF Video Phone Text
Audibility Participants + + + -
hear other peo
ple and sounds
in the environ
ment.
Visibility Participants see + + - -

other people
and objects in
the environ
ment.
Co-presence Participants are + ? - -

mutually aware
that they share
a physical envi
ronment.
Page 4 of 40

Cotemporality Messages are + + + -

received close
to the time that
they are pro
duced, permit
ting fine-
grained interac
tivity.
Simultaneity Multiple partici + + + -

pants can send/
receive mes
sages at the
same time, al
lowing
backchannel
communication.
Sequentiality Participants + + + -
take turns in an
orderly fashion
in a single con
versation.
Reviewability Messages do - - - +
not fade over
time.
Page 5 of 40

Revisability Messages can - - - +

be revised be
fore being sent.
Adapted from Clark & Brennan, 1991.
Page 6 of 40

Since its publication, Clark and Brennan’s framework has inspired hundreds of studies
comparing different media, including those they covered in their chapter and more recent
tools like Twitter. In part, this is because the clarity of the framework makes it easy to ap
ply to both the tools available in 1991 and later developments such as Twitter. In addi
tion, the notion of a range of affordances—as opposed to a single dimension of, say, “rich
ness”—made it easier to see how the most appropriate medium would likely depend sig
nificantly on the specific task the communicators were trying to perform. Some types of
work (e.g., negotiations) might benefit from seeing a partner’s facial expressions, where
as other types of work (e.g., brainstorming) might not. Finally, Clark and Brennan’s dis
cussion of tradeoffs made apparent some less intuitive aspects of media, such as the time
required to start up the CMC system.
Effects of Media on Task Performance and Con

versational Grounding
A long line of CMC research is motivated by Clark and colleagues’ theory of conversation
al grounding (Clark & Henetz, this volume, Clark & Marshall, 1981; Clark & Wilkes-
Gibbs, 1986). This approach sees communication as a process of building up mutual
knowledge or common ground. Speakers create messages based on their expectations of
what their addressee is likely to know, based on community co-membership (membership
in a common group), linguistic co-presence (the history of prior messages exchanged),
and physical co-presence (objects, people, and activities in the shared environment).
Speakers and addressees then (p. 473) work back and forth to ensure that the message
has been “grounded” or understood as intended (Clark & Krych, 2004; Clark & Wilkes-
Gibbs, 1986). Collaborators working across distances continue to face disadvantages over
collocated teams in terms of their ability to achieve and maintain common ground, and
this impacts their collaborations even with richer media (Olson & Olson, 2000).
Computer-mediated communication tools such as the telephone, text messaging, or video

conferencing, because of their affordances or lack thereof, can make it harder for commu
nicators to both identify existing common ground and ground new messages. Early Inter
net relay chat (IRC) chatrooms, for example, masked information about people’s charac
teristics and therefore eliminated most cues to community co-membership beyond mutual
participation in the specific chatroom in question. Text CMC rules out the use of gesture
for pointing, which plays an important role in many types of collaboration (e.g., Bekker,
Olson, & Olson, 1995; Clark & Krych, 2004).
Features of CMC tools can also shift the strategies communicators use to achieve ground
ing. When they can’t use gestures, for instance, they may use lengthier verbal substi
tutes; for example, “the long black piece over in the corner,” in place of “that one” plus a
deictic gesture (e.g., Brennan, 1998; Doherty-Sneddon et al., 1997; Kraut, Fussell, &
Siegel, 2003). To capture differences in grounding strategies, researchers have developed
a number of different conversational coding schemes. Possibly the most widely used of
these is conversational games analysis (Kowtko, Isard, & Doherty-Sneddon, 1991), which
Page 7 of 40

codes messages in terms of dialogue moves such as instructions, queries, and checks on
comprehension (see Table 29.2). In this coding scheme, instructions are the messages
that must be grounded if the appropriate task action is to be taken; queries, alignments,
and checks on comprehension are used to ensure grounding has occurred. Hence, the
numbers of these “moves” can indicate the ease of grounding and specific grounding
strategies that occur under specific media conditions. As we show later, conversational
games analysis and other types of conversational coding have often shown that grounding
strategies differ among media even when no performance effects can be identified.
Page 8 of 40

Table 29.2 Message categories in conversational games analysis
MOVE Definition
INSTRUCT Communicates a direct or indirect request for action

or instruction.
CHECK Listener checks self-understanding of a previous mes

sage or instruction from conversational partner by re
questing confirmation that the interpretation is cor
rect.
ALIGN Speaker confirms the listener’s understanding of a

message or accomplishment of some task. ALIGNs al
so check attention, agreement, or readiness. (Note
that in certain respects CHECK and ALIGN carry simi
lar functions but differ in terms of who initiates the
game. CHECKs are initiated by the listener of a previ
ous message, whereas ALIGNs are initiated by the
speakers of a message. Both serve to establish that
mutual understanding is maintained.)
QUERY-YN Yes–no question. A request for affirmation or negation

regarding new or unmentioned information about
some part of the task (not checking an interpretation
of a previous message).
QUERY-W Open-answer question. Requests more than affirma

tion or negation regarding new or unmentioned infor
mation about some part of the task (not checking an
interpretation of a previous message).
EXPLAIN Freely offered information regarding the task not

elicited by coparticipant.
Adapted from Doherty-Sneddon et al., 1997; Kowtko et al., 1991.
In the remainder of this section, we review studies examining how several key affor
dances of media suggested by Clark and Brennan (1991) influence the grounding process:
visibility, co-presence, and reviewability/revisibility. We also discuss research on the role
of tools to support gesture in remote collaboration, which draws on both visibility and co-
presence.
Page 9 of 40

(p. 474) Visibility
According to Clark and Brennan (1991), visibility is defined as, “A and B are visible to
each other. In face-to-face conversation, the participants can see each other, and in other
media they cannot. They may also be able to see each other, as in video teleconferencing,
without being able to see what each other is doing or looking at” (p. 141). Much early
CMC research focused on the presence or absence of video conferencing capabilities and
posited that visibility would be critical for coordinating collaborative work (e.g., Daft &
Lengel, 1986; Short et al., 1976). Underlying this hypothesis was the notion that the clos
er a medium is to face-to-face interaction, the better it would be for remote communica
tion and collaboration.
Despite the intuitive plausibility of this hypothesis, many empirical studies, over many
decades and improvements to video quality, failed to show a critical role for visibility in
terms of task completion times or the quality of work (e.g., Chapanis, 1975; Chapanis,
Parrish, Ochsman, & Weeks; 1977; Ochsman & Chapanis, 1974; Sellen, 1995; see Nardi &
Whittaker, 2002; Whittaker & O’Connail, 1997; Williams, 1977, for reviews). For problem-
solving tasks and those involving manipulating objects in the physical world, studies have
found minimal benefit to seeing a partner’s face (either face-to-face or by video confer
encing; Daly-Jone, Monk, & Watts, 1998; Reid, 1977; Short et al., 1976; Williams, 1977). A
possible exception is negotiation tasks, where the task is, itself, a highly social one and
may make more direct use of nuanced facial expressions and body position.
Although visibility may not affect task performance in many situations, there is consider
able evidence that it shapes grounding processes. Comparisons of audio telephony to
face-to-face conversations suggest that when visibility is absent, communicators use more
words and turns (e.g., Boyle, Anderson, & Newlands, 1994; Reid, 1977; Rutter, 1987),
suggesting that the absence of visual cues makes grounding more difficult. Doherty-Sned
don et al. (1997, study 1) found that instruction-givers used more ALIGN moves and in
struction-followers used more CHECK moves in the audio-only condition, and these
ALIGN and CHECK moves tended to occur at the same points at which speakers in a face-
to-face setting looked at their partner’s face. This suggests that verbal moves had re
placed visual cues in the grounding process.
Substituting verbal cues for visual cues may also impact turning-taking processes. Turn-
taking in video conferencing also tends to be more formal than in face-to-face conversa
tions (e.g., O’Conaill et al. 1993; Sellen, 1995), closer to telephone calls than to face-to-
face conversations. It appears that the quality of visual cues in these video conferencing
systems did not provide the same value for establishing mutual understanding as do visu
al cues in face-to-face interactions. This may suggest that visibility is a dimension rather
than a property that media do or do not have.
Depending on the image size, resolution, and bandwidth, video conferencing systems may
or may not allow for the use of visual cues in the grounding (p. 475) process to the same
extent as in face-to-face settings. For example, O’Conaill, Whittaker, and Wilbur (1993)
found that when a video system transmitted data quickly, had better audio, and had bet
Page 10 of 40

ter resolution, the use of backchannels and interruptions was higher and more similar to
face-to-face conversation than when the system transmitted data more slowly and had
poorer audio and video quality. O’Malley et al. (1996) found longer dialogues with more
interruptions in video conferencing versus face-to-face conversations. They manipulated
bandwidth and found more errors in the low bandwidth condition. They argue that with
higher quality video participants can use gaze to ensure that messages have been under
stood as intended, whereas with lower quality video this is not possible. Interestingly,
they also found more gaze and speech in video conferencing than in face-to-face conver
sations, which they suggest may be an “overcompensation” effect for uncertainty as to
whether messages have been appropriately grounded.
Most video conferencing systems do not allow for direct eye contact because of misalign
ments of the cameras relative to the displayed image of a remote partner. In higher end
video systems, this problem is often corrected. One question that arises is whether the
potential for mutual gaze affects conversation. Doherty-Sneddon et al. (1997, study 2)
compared audio conferencing, video conferencing with eye contact, and video conferenc
ing without eye contact. As in previous studies of visibility, there was no effect of medium
on task performance, but the availability of gaze and mutual gaze affected several as
pects of the conversation, including total words exchanged, turn-taking, and interrup
tions. Gaze analysis showed that partners looked at each others’ faces more often when
the video system allowed for eye contact than when it did not. In both video conditions,
participants looked at each others’ faces more often than in the face-to-face condition in
study 1, which they call “over-gazing” and liken to television watching.
Several studies have examined where people look in greater detail, particularly in collab
orations in which there are multiple potentially useful sources of visual information. For
example, Fussell, Setlock, and Parker (2003) found that in a collaborative assembly task,
partners looked at each other’s faces very rarely but instead focused gaze on task objects
and actions. Setlock, Quinones, and Fussell (2007) found that participants in a decision-
making and negotiation task looked at one another less than 20 percent of the time and
only rarely established mutual gaze. The results from studies like these suggest that gaze
may be important for other purposes than visibility of a partner’s face and gestures.
Specifically, gaze at activities and actions in the environment may play a critical role in
communication and coordination. We turn to this use of gaze in the next section.
Co-Presence
Clark and Brennan define co-presence as, “A and B share the same physical environment.
In face-to face conversation, the participants are usually in the same surroundings and
can readily see and hear what each other is doing and looking at. In other media there is
no such possibility” (p. 141). Co-presence is especially important when people are en
gaged in collaborative physical tasks, tasks in which two or more individuals work togeth
er to perform actions on objects in the three-dimensional (3D) or two-dimensional (2D)
world. For example, a remote expert might guide a worker’s performance in emergency
repairs to an aircraft, a group of students might collaborate to build a Lego robot, or a
Page 11 of 40

medical team might work together to save a patient’s life. Because of the importance of
such tasks in many domains, including education, design, industry, and medicine, consid
erable CMC research has focused on how best to support them.
In collaborative physical tasks, people’s speech and actions are intricately related to the
position and dynamics of objects, other people, and ongoing activities (e.g., Flor, 1998;
Ford, 1999; Goodwin, 1996; Kuzuoka & Shoji, 1994; Tang, 1991). Conversations during
collaborative physical tasks typically focus on the identification of target objects and loca
tions, descriptions of actions to be performed on those targets or to reach those locations,
and confirmation that the actions have been performed successfully.
When people can see one another’s actions in the environment, much of the conversation
al grounding in collaborative physical tasks can be done nonverbally. If a doctor in an op
erating room asks a nurse to “pass the forceps,” the doctor can see immediately whether
the nurse has picked up the right implement. If the nurse happened to pick up a scalpel
instead, the doctor can intervene immediately with more a more detailed description of
the desired implement. But if the doctor had to relay instructions via audio only, he or she
would need explicit confirmation from the nurse (e.g., “I found them”) before assuming
the message was grounded.
By the principle of least collaborative effort (Clark & Wilkes-Gibbs, 1986), we would ex
pect (p. 476) that reliance on verbal grounding would be reduced when visual cues of un
derstanding are present in the shared environment, and a number of studies provide evi
dence that this is the case. Clark and Krych (2004), for example, manipulated whether
participants could see each other’s faces and whether instructors could see worker’s
work areas. A view of partners’ faces had no effect, but a view of the workspace facilitat
ed task performance. Both instructors and workers adjusted their grounding strategies to
the channels of communication available to them, using more deictic expressions when
the workspace was visible.
How to use CMC tools to create virtual as opposed to physical co-presence has been a fo
cus of much research in the past few decades. There are two somewhat separate types of
tasks that rely on electronically mediated co-presence. In one scenario, remote
assistance, one person works with objects in the 3D world and a partner assists with the
task by giving instructions or other advice. In this case, there is a clear asymmetry in the
relationships between the team members on the objects to be manipulated. In a slightly
different scenario, two or more people work with objects in a 2D space, such as a shared
map, puzzle, or digital design task. In this case, partners may or may not have symmetric
access to the objects and locations relevant to the task. For both types of tasks, CMC re
searchers have assumed that performance and communication will be facilitated if the
CMC environment can provide some of the key elements of physical co-presence found in
face-to-face settings.
Determining how to create co-presence in CMC is a complex issue. When people are
physically co-present, the environment provides a number of distinct sources of informa
tion that can be used for conversational grounding, including mutual knowledge of the
Page 12 of 40

entities in the scene, awareness of others’ focus of attention, feedback from facial expres
sions, and so forth. It is rarely possible for virtual co-presence to provide all these aspects
of physical co-presence. To address this issue, Kraut and colleagues proposed a decompo
sitional model of co-presence (Kraut, Fussell, Brennan, & Siegel, 2002; Kraut et al., 2003)
that attempts to identify the pertinent dimensions of physical co-presence and to deter
mine the effects of each dimension on communication (see Table 29.3). For example, a
tool for visual co-presence may show differing amounts of the environment or have differ
ent visual fidelity. An underlying assumption of this model is that the extent to which re
mote video-based collaborations match the success of comparable face-to-face interac
tions is in large part a function of the extent to which the virtually shared environment
contains the key visual cues of the corresponding face-to-face setting.
Table 29.3 Some dimensions of physical co-presence
Dimensions of Physi Definition

cal Co-presence
Field of view People can see what entities (participant,

task, other) each person is oriented to
ward.
Spatial perspective People can see their partners’ perspec

tives toward the environment.
Display symmetry Both parties can see equivalent aspects of

one another’s environments.
Dimensionality People have a 3D view of the environment.
Spatial resolution People have high-quality views of the envi

ronment.
Temporal delay Images as received relative when they

were sent.
Adapted from Kraut, Fussell, Brennan, & Siegel, 2002.
Page 13 of 40

Fig. 29.1 Examples of collaborative physical tasks

used in computer-mediated communication (CMC)
studies.
Clockwise from upper left: Collaborative bicycle re

pair (Kraut, Fussell, & Siegel, 2003), large toy robot
(Fussell et al., 2004), stylized online puzzle task (Ger
gle et al., 2004a, 2004b), HCRC map task (Doherty-
Sneddon et al., 1997).
The effects of the dimensions in Table 29.3 on CMC, alone or in combination, have been
studied in a variety of laboratory experiments. Studies often use a within-subjects design
in which participants perform a number of similar collaborative physical tasks, each us
ing a different form of CMC, to control for general spatial abilities. Tasks are generally
structured such that one person (the “helper” or “instructor”) provides guidance to an
other person (the “worker” or “follower”). Examples include repairing a bicycle (Fussell,
Kraut, & Siegel, 2000; Kraut et al., 2003), building a toy robot (Fussell, Setlock, & Kraut,
2003; Fussell et al., 2004), solving a stylized puzzle task (Gergle, Kraut, & Fussell, 2004a;
2013; Ou, Oh, Yang, & Fussell, 2005), giving map directions (Anderson et al., 1991;
(p. 477) Doherty-Sneddon et al., 1997), and constructing a Lego design (Clark & Krych,
2004) (see Figure 29.1. Dependent measures include task performance (time, errors), the
amount of words exchanged, and message content.
Across numerous studies, there is clear evidence that dimensions of visual co-presence,
unlike visibility, affect task performance. For example, Fussell and colleagues (Fussell et
al., 2003) found that a scene-oriented camera that provided a wide-angle view of the
workspace led to better instruction in a robot construction task than a head-mounted
camera that showed a narrow view of what the partner was focused on. Asymmetries in
displays (in which one person can see an environment but the other cannot) can lead to
poorer performance on tasks within that environment than when both parties can see the
environment (Brennan, 1990; Gergle et al., 2004a; Gergle et al., 2013; Kraut et al., 2003).
Decreased task performance has also been linked to delays in transmitting the video im
age (Gergle et al., 2004a; 2013), reduced size of the shared view with respect to the work
environment (Fussell, Setlock, & Kraut, 2003; Gergle et al., 2013), and when communica
tors have different perspectives on the work environment (Gergle et al., 2013).
Page 14 of 40

The benefits of visual co-presence on performance may be directly attributed to the cues
provided for grounding. When conversationalists can see the same environment, they use
deictic expressions (e.g., “put that piece there”) to refer efficiently to task objects and lo
cations (e.g., Brennan, 1990). Detailed examination of verbal behavior (Gergle, Kraut, &
Fussell, 2004b) during an online puzzle task shows that actions substitute for verbal mes
sages in the grounding process when a shared visual space is present. When they know
their actions are visible to the help-giver, workers say little, letting their actions serve as
confirmation of understanding. When they know their actions aren’t visible to the help-
giver, they provide explicit verbal feedback. At the same time, helpers used the worker’s
actions as intended—as indicators of comprehension—and moved on to the next step in
the instructions. Gergle and colleagues suggest that visual co-presence shifts responsibili
ty for ensuring that grounding has occurred: when helpers can see the workspace, they
decide when to move on to the next step in the instructions; when helpers cannot see the
workspace, the workers must take responsibility for indicating that the task is complete.
For instance, in Fussell et al., workers who lacked visual co-presence needed to signal
readiness to proceed with verbal feedback (e.g., “ok,” “got it”), whereas those with visual
support could move their bodies or the task to allow the helper to see.
The dimensions of co-presence also interact with features of the task in terms of their ef
fects on communication and performance (Whittaker et al., 1993). When objects are hard
to describe (e.g., plaids vs. solid colors, shifting colors), visual co-presence is more impor
tant than when the (p. 478) objects are easier to describe (Gergle et al., 2004a, Kraut, Ger
gle, & Fussell, 2002).
Most of the studies on visual co-presence have examined co-presence alone, without con
sidering how co-presence and visibility might interact to shape communication and per
formance. Monk and Gale (2002) manipulated the presence or absence of full gaze aware
ness: participants can see each other’s faces, establish mutual gaze, and see where each
person is looking in the environment. In face-to-face settings, people are quite accurate at
assessing where partners are looking (Gale & Monk, 2000). Pairs conversed using video
conferencing with full gaze awareness, video conferencing with mutual gaze but no full
gaze awareness, video conferencing without mutual gaze or full gaze awareness, and au
dio only. They found that those with full gaze awareness used fewer words and turns to
complete the task, suggesting that even when people have the option of looking at each
other’s faces, they find looking at the visual environment more useful for grounding.
These results are consistent with those of Fussell, Setlock, and Parker (2003), who found
few glances at a partner’s face but many and longer glances at the robot pieces and the
worker’s hands.
Gesture
Studies of face-to-face collaboration on physical tasks suggests that hand gestures play
an important role in conversational grounding (see Bavelas, Gerwing, & Healing, this vol
ume). People use pointing gestures to refer to task objects and locations (e.g., “that piece
goes over there”), and they use representational gestures such as hand shapes and move
Page 15 of 40

ments to represent the form of task objects and the nature of actions to be used with
those objects (Bekker et al., 1995; Clark & Krych, 2004; Kuzuoka & Shoji, 1994; McNeill,
1992). For example, a speaker might say, “turn the screw,” while using his or her hands to
indicate the direction to turn it.
In face-to-face settings, people can readily use both pointing and representational ges
tures because the setting affords visibility (of a partner’s hands) and physical co-presence
(to see where he or she is pointing in the environment). Providing remote support for ges
ture is complicated, however, because the visual requirements for pointing and represen
tational gestures differ: pointing gestures require co-presence, whereas representational
gestures require visibility. Furthermore, many technologies that support remote pointing
(e.g., Bauer, Kortuem, & Segall, 1999; Kuzuoka, Kosuge, & Tanaka, 1994) require users to
control the pointer using devices such as mouses and joysticks, which restrict their hands
from performing other movements, including representational gestures.
In most of the remote collaboration scenarios used to study visual co-presence (e.g.,
Fussell, Setlock & Kraut, 2003; Kraut et al., 2003), there is an asymmetry in the ability to
gesture between the two conversational partners in the video conditions. The workers re
pairing the bike or building the robot could use their hands to point directly at task ob
jects, and the remote instructor could see these gestures in addition to the rest of the
workspace. But the remote instructors had no mechanism to point in the workspace and
often expressed frustration about this limitation of the technology. To address this issue,
Fussell and colleagues (Fussell et al., 2004, study 1) first explored the value of a simple
cursor pointer that remote instructors could use to point at a specific spot on their view
of the worker’s workspace. The combined view of the workspace plus the cursor overlay
was then projected on a monitor in front of the worker. Although instructors reported that
the cursor pointer made it easier for them to refer to entities in the workspace, there
were no objective effects on task performance.
Fig. 29.2 Close-up of the DOVE drawing tool on the

Helper’s tablet PC (left front insert) and on the
worker’s monitor (right).
From Fussell et al., 2004.
Page 16 of 40

One reason why the cursor pointer might not have improved communication and perfor
mance is that a cursor allows for only one of several types of gestures important for col
laborative physical tasks (see Table 29.4). Other key gestures include iconic representa
tional gestures that can be used to indicate the shape of a tool or task component, spatial/
distance gestures that can be used to show how far apart pieces should be, and kinetic/
motion gestures that show what action should be taken on a task object (Bekker et al.,
1995). To allow instructors to use representational gestures even though their hands
were not visible to the worker, Fussell et al. (2004, study 2) created the Drawing Over
Video Environment (DOVE) system that allowed remote helpers to draw as well as point
in the shared view of the workspace (see Figure 29.2. In comparison to visual co-presence
without gesture support, participants using DOVE communicated more efficiently and
performed better on the robot construction task. Conversation and performance was also
better when DOVE was configured such that the drawn gestures were erased after 3 sec
onds (making them fleeting, like hand gestures) than when the instructors had to manual
ly erase the gestures.
Page 17 of 40

Table 29.4 Definitions and possible functions of gestures used in col

laborative physical tasks
Type of Definition Possible Functions

Gesture
Deictic Orienting a finger or hand Reference to objects

(Pointing) toward a point in the envi and locations
ronment
Concrete
represen
tational
Iconic rep Forming hands to show what Reference to objects,

resenta a piece looks like, or to show procedural instruc
tions how two pieces should be tions (particularly ori
positioned relative to one an entation), descriptions
other of task status
Spatial/ Indicating through use of Procedural instruc

Distance one or both hands how far tions, descriptions of
apart two objects should be task status
or how far to move a given
object
Kinetic/ Demonstrating through use Procedural instruc

Motion of hands what action should tions
be performed on a task ob
ject
Adapted from Fussell et al., 2004.
Other investigators have compared DOVE-like systems (that rely on video overlay) with
alternative (p. 479) gesture tools that allow the remote helper to make natural hand ges
tures. For example, Kirk and Stanton Fraser (2006) compared sketching versus hand ges
tures, presented either on a monitor in front of the workspace (as in the DOVE system) or
projected directly onto the workspace. They found that hand gestures led to better per
formance than sketching, but it did not matter whether the gestures were presented on a
monitor or projected directly onto the workspace. Alem and Li (2011) also compared
drawing versus direct hand gesturing and found no significant differences in performance
(time or errors), but helpers rated the quality of the collaboration higher when they could
use their hands to gesture.
Page 18 of 40

Subsequent work has tried to elucidate the manner in which gesture tools can support
grounding and task performance. Kirk, Rodden, and Danton Fraser (2007) showed that
hand gestures projected on a work space facilitated conversational (p. 480) grounding ear
ly in a task, as the pair was establishing task-related common ground (e.g., establishing
names for Lego pieces). When gestures were not available, workers had to put more ef
fort into the initial grounding process. However, after this initial grounding period was
over, both gesture and no gesture groups performed equally well. When using a hands
overlay style of gesturing, participants develop strategies for grounding such as “waver
ing hands” to indicate what pieces should be selected and “mimicking hands” (illustrating
actions) or “inhabited hands” (overlaid on the worker’s hands to show fine-grained as
sembly movements) to show what to do with those pieces (Kirk, Crabtree, & Rodden,
2005). Interestingly, the benefits of gesturing for the worker have also been shown to
benefit the instructor when later completing the same task (Kirk & Stanton Fraser, 2005).
Other recent research has tried to improve the process of gesturing for both the remote
instructor and the local worker. For example, Yamashita and colleagues (Yamashita, Kaji,
Kuzuoka, & Hirata, 2011) explored gesture in the context of tabletop displays and found
that creating gestural persistence made it easier for remote partners to notice the ges
tures. Lanir and colleagues (Lanir, Stone, Cohen, & Gurevich, 2013) examined the effects
of giving the helper versus the worker control of camera positioning in a DOVE-like sys
tem. For one of the two tasks, building a Lego construction, helpers gestured more when
they had control of the camera. For the second task, however, there was no difference be
tween conditions.
Reviewability and Revisibility in Text-Based CMC
The discussion in the preceding sections suggests that richer media can facilitate ground
ing, especially during collaborative physical tasks. Text-based CMC, which lacks most of
the affordances of face-to-face interaction outlined in Table 1, would seem poorly suited
for conversation at a distance. Yet, substantial work in organizations is conducted using
text-based media such as e-mail, IM, and group chat (e.g., Birnholtz, Finholt, Horn, &
Bae, 2005; Ducheneaut & Bellotti, 2001; Handel & Herbsleb, 2002; Isaacs, Walendowski,
Whittaker, Schiano, & Kamm, 2002; Nardi, Whittaker, & Bradner, 2000). In part, the suc
cess of text media may be due to the types of work being conducted at a distance, which
often may not need to refer to objects and locations in the physical environment.
The widespread use of text CMC likely also stems from two affordances that text media
possess that others do not: revisibility of messages before they are sent and reviewability
afterward, perhaps indefinitely in e-mail archives. Revisibility allows communicators to
put extra conscious effort into the audience design process, which should lead to mes
sages that are easier to ground (Epley, Keysar, VanBoven, & Gilovich, 2004), although
there is some evidence that even when they have time to consider their partners’ per
spectives, people overestimate the extent to which they share common ground with recip
ients of their e-mail (Kruger, Epley, Parker, & Ng, 2005).
Page 19 of 40

Early research on text CMC found that tasks using text took longer and people exchange
fewer messages, presumably due to the greater difficulty of typing rather than speaking
(e.g., Chapanis et al., 1977; Galegher & Kraut, 1990; Lea and Spears, 1991; McGuire et
al., 1987; Ochsmans & Chapanis, 1974; Siegel, Dubrovsky, Kiesler, & McGuire, 1986). At
the time of these early studies, many of the participants had little experience with text-
based communication, so their generalizability to text communication today may be mini
mal. Subsequent work showed that people can adapt to text-based CMC over time (e.g.,
Walther, 1995). Newlands et al. (2003), for example, examined changes in people’s com
munication strategies when they conducted map tasks via IM over 3 days. Although initial
performance was much worse in the chat versus face-to-face settings, this difference be
came minimal after repeated trials with different maps. However, conversational coding
of the dialogues suggested that participants did not use interactive mechanisms for
grounding (e.g., questions) to the same extent in IM as in face-to-face; instead, they ap
pear to have made use of message revisibility and reviewability.
As with other modes of CMC, however, the precise features of the text tool matter. For ex
ample, having an explicit end-of-turn marker leads to better turn-taking and grounding
than when no explicit end of turn marker is present (Hancock & Dunham, 2001). Gergle
and colleagues (Gergle, Millen, Kraut, & Fussell, 2004) found that providing a history of
preceding conversational turns improved grounding over seeing just one turn at a time.
However, there may be limits to the amount of history that is useful because McCarthy
and Monk (1994) found few differences between a large chat space (30 lines of 10 words
each) and a smaller chat space (6 lines of 10 words each). Whether a (p. 481) text CMC
tool provides a mechanism for sharing artifacts (diagrams, notes, maps, etc.) also makes
a difference because these shared artifacts can help facilitate conversational grounding
in a manner similar to sharing visual co-presence. In support of this view, Gergle, Millen,
et al. (2004) found that differences between one- and six-line chat boxes were significant
only when partners didn’t have access to a shared visual space (a view of the worker’s
puzzle).
Effects of Individual and Group Characteristics

Newer CMC research has focused on understanding how characteristics of individuals or
groups influence the complex relationships between media affordances and conversation
al grounding, asking such questions as, “does the value of video differ depending upon
the expertise of the helper”? Four areas that have received some attention in recent years
are the expertise of the communicators, their cultural backgrounds, their native lan
guages, and the size of the group. We consider each of these areas briefly here.
Expertise
Many of the studies we have reviewed describe the dialogue participants as “experts” and
“novices” but in fact the expert is not trained at the task but rather is someone who has
required information such as an instruction manual for assembling a robot. Earlier work
Page 20 of 40

(Fussell & Krauss, 1992; Isaacs & Clark, 1987) showed that expertise, and the match of
expertise between helper and worker, had a large effect on the grounding process.
Specifically, if both parties to a conversation are experts in a subject area, reference is
significantly more efficient than when one or both are novices. When participants in, say,
our robot construction task (Fussell et al., 2004) are given an instruction manual, they
have greater knowledge than their partners about the task, but they are not in fact ex
perts because they haven’t themselves built the robot.
There is growing evidence that expertise may affect how features of media affect ground
ing. For example, support for gesture appears to be more important for conversational
grounding when communicators are novices rather than experts in the subject matter of
the discussion (Anderson, Sanford, Thomson, & Ion, 2007). Gaze cues from trained ex
perts can be valuable to less expert partners (Stein & Brennan, 2004), which suggests
that differences between video conferencing tools that support mutual gaze versus full
gaze awareness (e.g., Monk & Gale, 2002) may be even stronger when the remote party is
an expert in the subject area.
In some cases, features of media can hinder communication between experts and novices.
When the interface provides technical information (e.g., specialized medical terminology),
experts are less able to tailor appropriate messages for lay people, instead using too
many expert terms that novices would not know (Bromme, Jucks, & Runde, 2005; Jucks,
Bromme, & Runde, 2005, 2007). Similarly, if laypersons use more technical terminology
(perhaps because they have been coached as to how to ask the questions), experts re
spond with messages that are too technical for a layperson (Jucks & Bromme, 2011).
Bromme, Jucks, and colleagues conducted their studies on text-based CMC, and we know
of no systematic examination of how features of CMC tools might affect expert–novice
communication. However, we have anecdotal evidence of the phenomena they observe in
our bike repair studies, in which experts—who did have some degree of training in the
task—may use language (e.g., “derailleur”) that was unfamiliar to the worker, the expla
nation of which required additional verbal grounding work (Kraut, Fussell, & Siegel,
2003).
Culture and Intercultural Interaction
Most of the research on media, grounding, and task performance was conducted in West
ern societies, raising the question of how well the findings will generalize to other cul
tures or to intercultural interaction. There are some reasons to believe that this general
ization might not be straightforward. Theorists, for example, have distinguished between
individualistic cultures, in which people tend to identify themselves as individuals and fo
cus on their own personal needs and goals (including those of their immediate family),
and collectivistic cultures, in which people identify themselves as a member of a collec
tive and focus on the betterment of that collective (e.g., Hofstede, 1983, 2001; Triandis,
1989; see Oyserman et al., 2002, for a review). Nisbett (2003) describes a wide range of
cognitive processes affected by membership in individualistic versus collectivistic cul
tures, including reasoning styles and memory processes, that could potentially influence
Page 21 of 40

how affordances of media shape CMC both within different cultures and in cross-cultural
settings.
Much of the fairly limited research on culture and CMC has drawn on one or both of two
frameworks proposed by cultural theorists: high- versus low-context communication
(Hall, 1976) or task versus relationship orientation (Triandis, 1995). (p. 482) Hall
(1976/1981) argued that communication in low-context, typically Western, cultures tends
to be verbally explicit, to the point, with relatively little attempt to mask one’s feelings. In
contrast, communication in high-context, typically Eastern cultures tends to be more indi
rect, often ambiguous, and sensitive to the context in which it occurs (e.g., the relation
ship between speaker and addressee, nuances of facial expressions or tone of voice). Al
though people in all cultures use both communication styles, research suggests that low-
context communication is preferred in individualistic societies and high-context communi
cation is preferred in collectivistic societies (Gudykunst & Ting-Toomey, 1988; Gudykunst
et al., 1996). Triandis (1995) distinguishes between task-oriented cultures, which tend to
focus on getting work done, versus relationship-oriented cultures, which focus on estab
lishing rapport with one’s partners.
Of particular interest for building theories of culture and CMC is the idea that cultures
vary in their strategies for grounding conversations (e.g., Li, 1999a, 1999b). Specifically,
it has been suggested that audibility and visibility may be more important for grounding
in high-context cultures than in low-context cultures because awareness of how others
are reacting to one’s messages is an important aspect of high-context communication
(Hall, 1976/1981). This notion is supported indirectly by Veinott et al. (1999), who found
that non-native English speakers benefited from video over audio conferencing, whereas
native English speakers did not. They infer that the richer cues to mutual understanding
provided by visibility (e.g., quizzical looks, raised eyebrows) were especially valuable for
non-native speakers. However, this study confounded native language with intercultural
communication, so we don’t know which factor accounts for the results.
Research by Setlock and colleagues (Setlock, Fussell, & Neuwirth, 2004; Setlock et al.,
2007; Stewart, Setlock, & Fussell, 2007) suggests a less clear-cut relationship between
the availability of auditory and visual cues and conversational grounding in low- versus
high-context cultures. In one study (Setlock et al., 2004), American, Chinese, and mixed
American–Chinese dyads performed a negotiation task face-to-face or via IM. We hypothe
sized that the lack of social/contextual cues in IM would make it poorly suited for commu
nication among members of high-context cultures but not affect communication among
members of low-context cultures. Consistent with this hypothesis, we found no difference
between media in terms of how much grounding American pairs required to complete the
task, but a large culture-by-medium interaction, such that Chinese pairs spoke much
more face-to-face. At the same time, and contrary to the argument that visibility is central
to high-context communication, Setlock et al. (2007) found no main effects of culture or
medium when comparing American dyads, Chinese dyads, and American–Chinese dyads
using audio versus video conferencing for the same task. Thus, audibility seems to be im
portant for Chinese dyads, but adding visual cues via video conferencing did not provide
Page 22 of 40

additional benefit. These results conflict with those of Veinott et al., reported earlier, and
suggest the need to a more detailed examination of factors that differed across the two
studies (e.g., tasks, specific cultural backgrounds of participants).
Native Language
Some inconsistencies in the cross-cultural CMC research may stem from a confounding of
culture with native language. In a few cases, investigators have compared participants
speaking in their own native language as opposed to English as a second language. Set
lock and Fussell (2011), for example, showed that the interactions between culture and
media they identified in their earlier studies (Setlock et al., 2004, 2007) were also found
when Chinese participants conversed in Chinese. In other cases, investigators have com
pared members of different cultures each speaking in their own native language (e.g.,
Zhang & Marksbury, 2011).
Studies that require participants to use English as a second language can be problematic
for identifying the effects of culture on grounding. Speaking a non-native language can be
cognitively taxing and decrease the mental resources available for a task (Takano & No
da, 1993, 1995). Non-native speakers may also fear appearing incompetent if they misun
derstand conversational partners and therefore may not request clarification when it is
needed (Harzing & Feely, 2008; Park et al., 1996). Because second-language speakers
tend to have more limited linguistic resources (e.g., constrained vocabulary size), they of
ten have to use more complex communication strategies, such as rephrasing or repeating
previous utterances, to bridge the gap between their second-language proficiency and
their communicative intentions (Dornyei & Scott, 1997). These considerations illustrate
the difficulty of distinguishing effects of culture versus native language on the grounding
process. They have also (p. 483) led to new research on the use of machine translation in
CMC (e.g., Gao et al., 2013; Hautassari, 2010; Yamashita & Ishida, 2006; Yamashita, Ina
ba, Kuzuoka, & Ishida, 2009).
Group Size
The majority of experimental studies of CMC have used dyads, and it is an open question
to what extent the results will generalize to larger groups. Anderson (2006) and Branni
gan (2006) suggest that grounding processes differ between dyads and larger groups, in
part because common ground may need to be established between each pair of partici
pants independently. There is a smattering of evidence that this generalization may not
be straightforward. For example, Jackson and colleagues (Jackson, Anderson, McEwan, &
Mullin, 2000) found that the quality of a video conferencing system affected the number
of words exchanged in dyads but not in four-party groups. Anderson and colleagues (An
derson, Mullin et al., 1999) found that three-party conversations required more words (ef
fort) than two-party conversations about the same task.
Other affordances of media also vary as group size increases. In larger groups, communi
cators appear to spend a larger percentage of time looking at the person speaking than
Page 23 of 40

they do in dyadic settings (Vertegaal, Slagter, van der Veer, & Nijholt, 2001), possibly be
cause gaze direction in a multiparty setting can serve as a visible signal of attention that
is seen by all members of the group. Gesture likewise becomes more complicated to sup
port when group sizes are larger (Tang, Pahud, Inkpen, Benko, Tang, & Buxton, 2010) be
cause there are more potential overlaps of gesture outputs (e.g., crossing of virtual
“arms”). Finally, studies examining grounding in multilingual or multicultural settings
(e.g., Wang & Fussell, 2010; Yamashita et al., 2009) show how problems can multiply
when grounding needs to be established across multiple pairs of communicators with dif
ferent degrees of common ground.
Limitations
The conversational grounding approach to understanding CMC has been very fruitful in
terms of identifying how specific features of media affect specific conversational prac
tices, as well as in terms of helping understand how task features affect the relationship
among media, communication, and performance. Yet, this general research framework
suffers from some limitations that suggest it will need to be expanded to apply to the
modern media landscape.
Changing Media Landscape
First, Clark and Brennan’s general framework will need some updating and expansion to
address today’s use of technology. People’s facility with media, and therefore the costs
and benefits of using these tools, can change dramatically over time. Early research on
text CMC, for example, suggested that people were very poor typists and thus took nearly
three times as long to do a task than would be required in a face-to-face setting (Siegel et
al. 1986). This is no longer the case (e.g., Newlands et al., 2003), so the relative cost of
typing versus oral communication has shifted over time. Video conferencing systems such
as Skype are also very easy to set up relative to their predecessors.
In addition, new CMC tools provide the original set of affordances in new ways that may
improve remote collaboration, such as the Kinect (e.g., Genest, Gutwin, Tang, Kalyn, &
Ivkovic (2013) or tabletop conferencing systems (e.g., Yamashita et al., 2011). There have
also been developments in the area of text-based tools, with increased file sharing
through cloud services, shared bookmarking and tagging sites, and the like. To date,
there has been only a smattering of studies on how grounding occurs with respect to
practices such as naming of shared files (Rader, 2010) or the creation of socially shared
tags (e.g., Mamykina et al., 2011), although earlier work (e.g., Carroll, 1985; Krauss,
Vivekananthan, & Weinheimer, 1968; Fussell & Krauss, 1989a, 1989b) suggests that tags
or file names created for oneself will be problematic for others to understand.
Page 24 of 40

Changes in How People Use CMC
Second, much early CMC research implicitly assumed that people would be using only
one medium at a time (e.g., text, chat, or video). This assumption no longer holds because
of the multifunctionality of new CMC tools and because of increased multitasking. With
respect to multifunctionality, a new “media,” such as a social network site, can include
many different affordances depending on which subsets of functions one uses. Facebook
can be used for text chat, picture exchange, posting videos, and so forth. Understanding
when and why people choose one subset of features over others has been the focus of
growing social science research (e.g., Lampe, Wash, Velasquez, & Ozkaya, 2010).
The assumption of “one medium at a time” also doesn’t hold today because of multitask
ing. While engaged in a video conference or face-to-face meeting, a person might simulta
neously be text (p. 484) chatting with one or more friends (e.g., Rennecker, Dennis, &
Munson, 2010). How features of video and text CMC influence communication might be
expected to shift when they are not an undivided focus of attention. Similarly, Isaacs and
colleagues (Isaacs, Szymanski, Yamauchi, Glasnapp, & Iwamoto, 2012) have shown how
people blend conversations across face-to-face settings and various types of CMC, creat
ing yet more novel communication phenomena. Herring (2013) proposed the notion of
convergent media computer-mediated communication (CMCMC) to capture new uses of
CMC tools, such as text plus television (Zelenkauskaite & Herring, 2008). Such consider
ations suggest that the benefits and limitations to various tools identified in CMC lab
studies may not readily generalize to their use in the real world. For example, the bene
fits of visual co-presence may be reduced if a remote helper is simultaneously engaged in
additional conversations via text because his or her visual attention may not be on the
workspace.
Mobility is another change in how people use technology that may have consequences for
grounding in CMC. Work in this area is in its infancy. For example, Giusti et al. (2013)
performed a qualitative analysis of how people configure mobile phones for remote col
laboration on physical tasks, and several investigators have looked at conversational dy
namics while using cell phones in the car (e.g., Iqbal, Horvitz, Ju, & Mathews, 2011; Sch
neider & Kiesler, 2005)
Relationship-Oriented Versus Task-Oriented Goals
Most of the research inspired by Clark and Brennan’s framework has focused on task-ori
ented conversation, using measures such as performance time, errors, efficiency of com
munication, and coding of task-oriented conversational content. Less attention has been
paid to relational aspects of communication, which are concerned not with what informa
tion is conveyed but with how that information is conveyed and what this indicates about
the relationship between speaker and addressee(s). In face-to-face communication, non
verbal cues such as eye gaze (e.g., Argyle & Cook, 1976; Burgoon, Buller, Hale, & Turck,
1984; Exline & Winters, 1965; Kleinke, 1986), facial expressions (e.g., Ekman, 1982), and
posture (e.g., Mehrabian, 1967) can be used to indicate intimacy, trust, and attraction. In
Page 25 of 40

addition, messages themselves can be formulated in different ways to establish, maintain,

and/or build closeness with a partner. For example, indirect requests such as, “Could you
close the door?” are more polite than directives such as “Close the door!” (Holtgraves,
1997). Similarly, hedging an opinion (e.g., “I think you might be wrong”) is more polite
than directly stating that opinion (e.g., “You are wrong”). As a rule, the polite phrases are
also longer, so choices relating to politeness may interact with those related to features of
the media in which efficiency or brevity are prioritized.
Features of media have been shown to affect relational communication (e.g., Herring,
1994; Hiltz, Johnson, & Turoff, 1986; Kiesler et al., 1984; see Whittaker, 2003, for a re
view). In some cases, greater negative emotion and “flaming” has been found in text com
munication (e.g., Kiesler et al., 1984; Sproull & Kiesler, 1986), a finding that Kiesler and
colleagues have attributed to the lack of social context cues in text communication but
which others have attributed to the difficulty of producing politeness markers in typed
discourse (e.g., Brennan & Ohaeri, 1999; Walther, 1992, 1995). Other studies have found
more relational communication in face-to-face settings than in text communication (e.g.,
Hiltz & Turoff, 1978). Early studies also suggested that relational aspects of communica
tion are reduced when conversations take place over the phone versus face-to-face (Rut
ter, 1987; Stephenson, Ayling, & Rutter, 1976). Less research has compared relational
communication in audio versus video conferencing, although many media theories (e.g.,
Daft & Lengel, 1984; Short et al., 1976) suggest that video will better support relational
communication.
The lack of clear-cut evidence of how media affordances influence relational communica
tion likely stems from two causes. First, many of the CMC researchers interested in con
versational grounding and task performance fail to include any analysis of the relational
components of their work. Conversational games analysis, for example, is presented in
terms of task “moves,” with no categories for coding messages, such as “we’re doing
great,” that are relational in nature (Kowtko et al., 1991). At the same time, much of the
CMC research on relational communication has focused on text-based media and eluci
dated how intimacy (e.g., Walther, 1992, 1995), conversational involvement (e.g., Nguyen
& Fussell, 2012), emotion (e.g., Hancock, Gee, Ciaccio, & Lin, 2008; Hancock, Landrigan,
& Silver, 2007), trust (e.g., Scissors, Gill, & Gergle, 2008), and deception (e.g., Hancock
et al., 2009) can occur when none of the customary nonverbal cues is available. Only a
handful of studies (p. 485) (e.g., Bietz, 2008; Hancock, Currya, Goorhaa, & Woodworth,
2007; Zheng, Veinott, Bos, Olson, & Olson, 2002) have compared relational communica
tion across multiple media.
Truthful Versus Deceptive Messages
Finally, a growing number of studies have examined how media affordances facilitate or
hinder deceptive communication (e.g., Hancock et al., 2009; Hancock, Currya, Goorhaa,
& Woodworth, 2007; Hancock, Thom-Santelli, & Ritchie, 2004; see Hancock & Gonzalez,
2012, for a review). Hancock (2007) distinguishes between identity-based deception (pre
tending to be someone other than who one is) and message-based deception (making
Page 26 of 40

false statements in a message) and argues that affordances of media can facilitate or hin
der both types of deception. For example, a man can pretend to be a woman or vice versa
much more easily when no visual or auditory cues are present. Similarly, it is easier to lie
about one’s whereabouts (e.g., in the library as opposed to in the bar) when no visual
cues are present. At the same time, the persistence of text-based media present chal
lenges to deception since the message recipient can go back and check what was said
(e.g., identifying when the same excuse has been used repeatedly, or noting mismatches
in what was said on different occasions).
Over the past decade, a growing number of studies have supported the idea that media
affordances shape deceptive behavior. In a diary study, Hancock, Thom-Santelli, and
Ritchie (2004) found that people reported lying more often face-to-face and on the phone
than in text-based media, a finding attributed in part to the synchronous and fleeting na
ture of face-to-face and phone conversations as opposed to the recordability of text me
dia. In subsequent work, Hancock and colleagues (Hancock, Birnholtz, et al., 2009)
demonstrated how the lack of visibility or visual co-presence in IM allows users to pro
duce what they call “butler lies,” or minor untruths, for managing social interactional
processes. For example, a person might politely end a conversation with a false to-do item
(e.g., “I have to grab my laundry from the dryer”). This interesting area of research is
continuing to grow as investigators examine deception across a range of media, including
online dating sites (e.g., Hancock, Toma, & Ellison, 2007), text messaging (Birnholtz, Guil
lory, Hancock, & Bazarova, 2010; Reynolds, Smith, Birnholtz, & Hancock, 2013), and
avatar-mediated communication (e.g., Steptoe, Steed, Rovira, & Rae, 2010).
Conclusion
Clark and Brennan’s (1991) theory of media affordances has served as a guiding frame
work for many decades of research on CMC. Their general strategy of characterizing me
dia in terms of their costs and benefits along multiple dimensions has proved more useful
for generating research hypotheses and building theory than earlier views that consid
ered media solely in terms of richness (Daft & Lengel, 1984) or presence (Short et al.,
1976). Research on some affordances, such as visibility, has suggested that they are less
important than might initially have been thought, but, at the same time, it has served to
clarify when and why visibility is desirable in CMC. Research on co-presence, conversely,
has consistently shown its value for tasks in which people are interacting with 2D or 3D
objects, and much subsequent theorizing has taken a similar “decomposing” strategy to
understand the different elements of co-presence (e.g., Kraut et al., 2002). Although addi
tional research will be needed to flesh out components of Clark and Brennan’s framework
in light of new technologies, multitasking, and individual characteristics, there can be no
doubt that this framework will continue to motivate CMC research for many years to
come.
Page 27 of 40

References
Alem, L., & Li, J. (2011). A study of gestures in a video-mediated collaborative assembly
task, Advances in Human-Computer Interaction, 2011, 1–7.
Anderson, A. (2006). Achieving understanding in face-to-face and video-mediated multi

party conversations. Discourse Processes, 41, 251–287.
Anderson, A., Bader, M., Bard, E. G., Boyle, E. A., Doherty, G. M., Garrod, S.,…Weinert, R.
(1991). The HCRC Map Task Corpus. Language and Speech, 34, 351–366.
Anderson, A. H., Mullin, J., Katsavras, E., Brundell, P., McEwan, R., Grattan, E., et al.
(1999). Multimediating multiparty interactions. Proceedings of INTERACT 99, 313–320.
Anderson, T., Sanford, A. J., Thomson, A., & Ion, W. (2007). Computer-supported and face-
to-face collaboration on design tasks. Discourse Processes, 43(3), 201–228.
Argyle, M., & Cook, M. (1976). Gaze and mutual gaze. London: Cambridge University
Press.
Bauer, M., Kortuem, G., & Segall, Z. (1999). “Where are you pointing at?” A study of re
mote collaboration in a wearable video conference system. Proceedings of the Third Inter
national Symposium on Wearable Computers, 3, 151–158.
Bekker, M. M., Olson, J. S., & Olson, G. M. (1995). Analysis of gestures in face-to-face de
sign teams provides guidance for how to use groupware in design. Proceedings of the 1st
Conference/Designing Interactive Systems: Processes, Practices, Methods, & Techniques,
1, 157–166.
Bietz, M. J. (2008). Effects of communication media on the interpretation of critical feed

back. Proceedings of the Conference on Computer-Supported Cooperative Work, 12, 467–
476.
Birnholtz, J. P., Finholt, T. A., Horn, D. B., & Bae, S. J. (2005). Grounding needs:
(p. 486)
Achieving common ground via lightweight chat in large, distributed, ad-hoc groups. Pro
ceedings of the SIGCHI Conference on Human Factors in Computing Systems, 24 (1), 21–
30.
Birnholtz, J. P., Guillory, G., Hancock, J. T., & Bazarova, N. (2010). “On my way”: Deceptive
texting and interpersonal awareness narratives. Proceedings of the 2010 ACM Confer
ence on Computer Supported Cooperative Work, 13, 1–4.
Boyle, E. A., Anderson, A. H., & Newlands, A. (1994). The effects of visibility on dialogue
and performance in a cooperative problem solving task. Language and Speech, 37, 1–20.
Brannigan, H. (2006). Perspectives on multi-party dialogue. Research on Language and

Computation, 4, 153–177.
Page 28 of 40

Brennan, S. E. (1990). Seeking and providing evidence for mutual understanding. Unpub
lished doctoral thesis. Stanford, CA: Stanford University.
Brennan, S. E. (1998). The Grounding Problem in Conversations With and Through Com
puters. Social and cognitive psychological approaches to interpersonal communication,
201–225.
Brennan, S. E., & Ohaeri, J. O. (1999). Why do electronic conversations seem less polite?
The costs and benefits of hedging. Proceedings from International Joint Conference on
Work Activities, Coordination, and Collaboration, 227–235.
Bromme, R., Jucks, R., & Runde, A. (2005). Barriers and biases in computer-mediated ex
pert–layperson communication: An overview and insights into the field of medical advice.
In R. Bromme, F. W. Hesse, & H. Spada (Eds.), Barriers and biases in computer-mediated
knowledge communication—and how they may be overcome (pp. 89–119). New York:
Springer.
Burgoon, J. K., Buller, D. B., Hale, J. L., & Turck, M. A. (1984). Relational messages associ
ated with nonverbal behaviors. Human Communication Research, 10, 351–378.
Carroll, J. M. (1985). What’s in a name? New York: Freeman.
Chapanis, A. (1975). Interactive human communication. Scientific American, 232, 36–42.
Chapanis, A., Parrish, R. N., Ochsman, R. B., & Weeks, G. D. (1977). Studies in interactive
communication: II. The effects of four communication modes on the linguistic perfor
mance of teams during cooperative problem solving. Human Factors, 19, 101–126.
Clark, H. H., & Brennan, S. E. (1991). Grounding in communication. In L. B. Resnick, R.

M. Levine, & S. D. Teasley (Eds.), Perspectives on socially shared cognition (pp. 127–149).
Washington, DC: APA Press.
Clark, H. H., & Krych, M. A. (2004). Speaking while monitoring addressees for under
standing. Journal of Memory & Language, 50, 62–81.
Clark, H. H., & Marshall, C. R. (1981). Definite reference and mutual knowledge. In A. K.
Joshi, B. L. Webber, & I. A. Sag (Eds.), Elements of discourse understanding. Cambridge:

22, 1–39.
Daft, R. L., & Lengel, R. H. (1984). Information richness: A new approach to managerial
behavior and organizational design. In L. L. Cummings & B. M. Staw (Eds.), Research in
organizational behavior 6 (pp. 191–233). Homewood, IL: JAI Press.
Daft, R. L., & Lengel, R. H. (1986). Organizational Information Requirements, Media

Richness and Structural Design. Management Science, 32, 5, 554-571.
Page 29 of 40

Daft, R. L., Lengel, R. H., & Trevino, L. K. (1987). Message equivocality, media selection,
and manager performance: Implications for information systems. MIS Quarterly, 11, 355–
366.
Daly-Jones, O., Monk, A. F., & Watts, L. (1998). Some advantages of video conferencing
over high-quality audio conferencing: Fluency and awareness of attentional focus. Inter
national Journal of Human–Computer Studies, 49, 21–58.
Doherty-Sneddon, G., Anderson, A. H., O’Malley, C., Langton, S., Garrod, S., & Bruce, V.
(1997). Face-to-face and video mediated communication: A comparison of dialogue struc
ture and task performance. Journal of Experimental Psychology: Applied, 3, 105–125.
Dornyei, Z., & Scott, M. (1997). Communication strategies in a second language: Defini
tions and taxonomies. Language Learning, 47, 173-210.
Ducheneaut, N., & Belloti, V. (2001). Email as habitat: An exploration of embedded per
sonal information management. Interactions of the ACM, 8 (5), 30–38.
Ekman, P. (1982). Emotions in the human face. Cambridge: Cambridge University Press.
Epley, N., Keysar, B., VanBoven, L., & Gilovich, T. (2004). Perspective taking as egocentric
anchoring and adjustment. Journal of Personality and Social Psychology, 87, 327–339.
Exline, R. V., & Winters, L. C. (1965). Affective relations and mutual glances in dyads. In
S. S. Tomkins and C. W. Izard (Eds.), Affect, cognition and personality. New York:
Springer-Verlag.
Flor, N. (1998). Side-by-side collaboration: A case study. International Journal of Human-

Computer Studies, 49, 201–222.
Ford, C. E. (1999). Collaborative construction of task activity: Coordinating multiple re

sources in a high school physics lab. Research on Language and Social Interaction, 32,
369–408.
Fussell, S. R., & Krauss, R. M. (1989a). The effects of intended audience on message pro
duction and comprehension: Reference in a common ground framework. Journal of Exper
imental Social Psychology, 25, 203–219.
Fussell, S. R., & Krauss, R. M. (1989b). Understanding friends and strangers: The effects
of audience design on message comprehension. European Journal of Social Psychology,
19, 509–525.
Fussell, S. R., & Krauss, R. M. (1992). Coordination of knowledge in communication: Ef

fects of speakers’ assumptions about what others know. Journal of Personality and Social
Psychology (JPSP), 62, 378–391.
Page 30 of 40

Fussell, S. R., Kraut, R. E., & Siegel, J. (2000). Coordination of communication: Effects of
shared visual context on collaborative work. Proceedings of the Conference on Computer-
Supported Cooperative Work, 8, 21–30.
Fussell, S. R., Setlock, L. D., & Kraut, R. E. (2003). Effects of head-mounted and scene-
oriented video systems on remote collaboration on physical tasks. Proceedings of the
SIGCHI Conference on Human Factors in Computing Systems, 22 (1), 513–520.
Fussell, S. R., Setlock, L. D., & Parker, E. M. (2003). Where do helpers look? Gaze targets
during collaborative physical tasks. Companion Proceedings of the SIGCHI Conference on
Human Factors in Computing Systems, 22 (2), 768–769.
Fussell, S. R., Setlock, L. D., Yang, J., Ou, J., Mauer, E. M., & Kramer, A. (2004). Gestures
over video streams to support remote collaboration on physical tasks. Human-Computer
Interaction, 19, 273–309.
Gale, C., & Monk, A. F. (2000). Where am I looking? The accuracy of video-mediated gaze
awareness. Perception & Psychophysics, 62, 586–595.
(p. 487) Galegher, J., & Kraut, R. E. (1990). Computer-mediated communication for intel
lectual teamwork: A field experiment in group writing. Proceedings of the Conference on
Computer-Supported Cooperative Work, 3, 65–78.
Gao, G., Wang, H-C., Cosley, D., & Fussell, S. R. (2013). Same translation but different ex
perience: The effects of highlighting on machine-translated conversations. Proceedings of
the SIGCHI Conference on Human Factors in Computing Systems, 32 (1), 449–458.
Genest, A. M., Gutwin, C., Tang, A. Kalyn, M., & Ivkovic, Z. (2013). KinectArms: A toolkit
for capturing and displaying arm embodiments in distributed tabletop groupware. Pro
ceedings of the Conference on Computer-Supported Cooperative Work, 16 (1), 157–166.
Gergle, D., Kraut, R. E., & Fussell, S. R. (2004a). Language efficiency and visual technolo
gy: Minimizing collaborative effort with visual information. Journal of Language and So
Gergle, D., Kraut, R. E., & Fussell, S. R. (2004b). Action as language in a shared visual
space. Proceedings of the Conference on Computer-Supported Cooperative Work, 10,
487–496.
Gergle, D., Kraut, R. E., & Fussell, S. R. (2013). Using visual information for grounding
and awareness in collaborative tasks. Human Computer Interaction, 28, 1–39.
Gergle, D., Millen, D. R., Kraut, R. E., & Fussell, S. R. (2004). Persistence matters: Mak
ing the most of chat in tightly-coupled work. Proceedings of the SIGCHI Conference on
Goodwin, C. (1996). Professional vision. American Anthropologist, 96, 606–633.
Page 31 of 40

Giusti, L., Xerxes, K., Schladow, A., Wallen, N., Zane, F., & Casalegno, F. (2013). Spatial
setups: A study of mobile configurations in remote collaborative tasks. Proceedings of the
1st Symposium on Spatial User Interaction, 1, 49–52.
Gudykunst, W. B., Matsumoto, Y., Ting-Toomey, S., Nishida, T., Kim, K. S., & Heyman, S.
(1996). The influence of cultural individualism-collectivism, self construals, and individual
values on communication styles across cultures. Human Communication Research, 22,
510–543.
Gudykunst, W. B., & Ting-Toomey, S. (1988). Culture and interpersonal communication.

Newbury Pk, CA: Sage.
Hall, E. (1976/1981). Beyond culture. New York: Doubleday/Anchor Books.
Hancock, J. T. (2007). Digital deception: Why, when and how people lie online. Oxford
handbook of internet psychology (pp. 287–299). Oxford: Oxford University Press.
Hancock, J. T., Birnholtz, J. P., Bazarova, N., Guillory, J., Perlin, J., & Amos, B. (2009). But
ler lies: Awareness, deception and design. Proceedings of the SIGCHI Conference on Hu
man Factors in Computing Systems, 28, 517–526.
Hancock, J. T., Currya, J., Goorhaa, S., & Woodworth, M. (2007). On lying and being lied
to: A linguistic analysis of deception in computer-mediated communication. Discourse
Processes, 45, 1, 1–23.
Hancock, J. T., & Dunham, P. J. (2001). Language use in computer-mediated communica

tion: The role of coordination devices. Discourse Processes, 31, 91–110.
Hancock, J. T., Gee, K., Ciaccio, K., & Lin, J. M. -H. (2008). I’m sad you’re sad: Emotional
contagion in CMC. Proceedings of the Conference on Computer-Supported Cooperative
Work, 12, 295–298.
Hancock, J. T., & Gonzalez, A. (2012). Deception in CMC. In S. Herring, D. Stein, & T. Vir
tanen (Eds.), Pragmatics of computer-mediated communication (p. 736). Berlin: Mouton
de Gruyter.
Hancock, J. T., Landrigan, C., & Silver, C. (2007). Expressing emotion in text-based com
munication. Proceedings of the SIGCHI Conference on Human Factors in Computing Sys
tems, 26 (1), 929–932.
Hancock, J. T., Thom-Santelli, J., & Ritchie, T. (2004). Deception and design: The impact of
communication technologies on lying behavior. Proceedings of the SIGCHI Conference on
Human Factors in Computing, 23 (1), 130–136.
Hancock, J. T., Toma, C., & Ellison, N. (2007). The truth about lying in online dating pro
files. Proceedings of the SIGCHI Conference on Human Factors in Computing Systems, 26
(1), 449–452.
Page 32 of 40

Handel, M., & Herbsleb, J. (2002). What is chat doing in the workplace? Proceedings of
the Conference on Computer-Supported Cooperative Work, 9 (1), 1–10.
Harzing, A. W. K., & Feely, A. J. (2008). The Language barrier and its Implications for HQ-
Subsidiary Relationships. Cross Cultural management: An International Journal, 15 (1):
49–60.
Hautassari, A. (2010). Machine translation effects on group interaction: An intercultural

collaboration experiment. Proceedings of the International Conference on Intercultural
Collaboration, 69–78.
Herring, S. C. (1994). Politeness in computer culture: Why women thank and men flame.
In M. Bucholtz, A. Liang, L. Sutton, & C. Hines (Eds.), Cultural performances: Proceed
ings of the Third Berkeley Women and Language Conference (pp. 278–294). Berkeley, CA:
Berkeley Women and Language Group.
Herring, S. C. (2013). Discourse in Web 2.0: Familiar, reconfigured, and emergent. In D.

Tannen & A. M. Tester (Eds.), Georgetown University Round Table on Languages and Lin
guistics 2011: Discourse 2.0: Language and new media (pp. 1–25). Washington, DC:
Georgetown University Press.
Hiltz, S. R., Johnson, K., & Turoff, M. (1986). Experiments in group decision making: Com
munication process and outcome in face-to-face versus computerized conferences. Hu
man Communications Research, 13, 225–252.
Hiltz, S. R., & Turoff, M. (1978). The network nation: Human communication via comput
er. Reading, MA: Addison-Wesley.
Hofstede, G. (1983). National cultures revisited. Cross-cultural Research, 18, 4, 285–305.
Hofstede, G. (2001). Culture’s consequences: Comparing values, behaviors, institutions,

and organizations across nations. Thousand Oaks CA: Sage.
Holtgraves, T. (1997). Styles of language use: Individual and cultural variability in conver
sational indirectness. Journal of Personality and Social Psychology (JPSP), 73, 624–637.
Iqbal, S. T., Horvitz, E., Ju, Y-C., & Mathews, E. (2011). Hang on a sec!: Effects of proac
tive mediation of phone conversations while driving. Proceedings of the SIGCHI Confer
ence on Human Factors in Computing Systems, 30 (1), 463–472.
Isaacs, E., & Clark, H. H. (1987). References in conversation between experts and
novices. Journal of Experimental Psychology: General, 116, 26–37.
Isaacs, E., Szymanski, S., Yamauchi, Y., Glasnapp, J., & Iwamoto, K. (2012). Integrating lo
cal and remote worlds through channel blending. Proceedings of the Conference on Com
puter-Supported Cooperative Work, 15 (1), 617–626.
Page 33 of 40

Isaacs, E., Walendowski, A., Whittaker, S., Schiano, D. J., & Kamm, C. (2002). The charac
ter, functions, and styles of instant messaging in the workplace. Proceedings of the
(p. 488) Conference on Computer-Supported Cooperative Work, 9 (1), 11–20.
Jackson, M., Anderson, A. H., McEwan, R., & Mullin, J. (2000). Impact of video frame rate
on communicative behaviour in two and four party groups. Proceedings of the Conference
on Computer-Supported Cooperative Work, 8, 11–20.
Jucks, R., & Bromme, R. (2011). Perspective taking in computer-mediated instructional

communication. Journal of Media Psychology: Theories, Methods, and Applications, 23(4),
192–199.
Jucks, R., Bromme, R., & Runde, A. (2005). The impact of external representations on pro
viding online instructions. Proceedings of the 27th Annual Conference of the Cognitive
Science Society, 27, 1072–1077.
Jucks, R., Bromme, R., & Runde, A. (2007). Explaining with non-shared illustrations: How
they constrain explanations. Learning and Instruction, 17, 204–218.
Kiesler, S., Siegel, J., & McGuire, T. W. (1984). Social psychological aspects of computer-
mediated communication. American Psychologist, 39, 1123–1134.
Reprinted in
Greif (Ed.), Computer-supported cooperative work: A book of readings (pp. 657–682). San
Mateo, CA: Morgan Kaufmann.
Kirk, D., Crabtree, A., & Rodden, T. (2005). Ways of the hands. Proceedings of the Ninth
Conference on European Conference on Computer Supported Cooperative Work, 9, 1–21.
Kirk, D., Rodden, T., & Stanton Fraser, D. (2007). Turn it this way: Grounding collabora
tive action with remote gestures. Proceedings of the SIGCHI Conference on Human Fac
tors in Computing Systems, 26 (1), 1039–1048.
Kirk, D., & Stanton Fraser, D. (2005). The effects of remote gesturing on distance instruc
tion. Proceedings of the 2005 Conference on Computer Support for Collaborative Learn
ing: Learning 2005: The Next 10 Years! 301–310.
Kirk, D., & Stanton Fraser, D. (2006). Comparing remote gesture technologies for sup
porting collaborative physical tasks. Proceedings of the SIGCHI Conference on Human
Factors in Computing Systems, 25 (1), 1191–1200.
Kleinke, C. (1986). Gaze and eye contact: A psychological review. Psychological Bulletin,
100, 78–100.
Kowtko, J. C., Isard, S., & Doherty-Sneddon, G. (1991). Conversational games in dialogue.
In A. Lescardes (Ed.), Proceedings of the ESPRIT/ DAWDI Workshop on Discourse Coher
ence. Edinburgh, Scotland: University of Edinburgh.
Page 34 of 40

Krauss, R. M., Vivekananthan, P. S., & Weinheimer, S. (1968). “Inner speech” and “exter
nal speech”: Characteristics and communication effectiveness of socially and nonsocially
encoded messages. Journal of Personality and Social Psychology, 9, 295–300.
Kraut, R. E., Fussell, S. R., Brennan, S. E., & Siegel, J. (2002). Understanding effects of
proximity on collaboration: Implications for technologies to support remote collaborative
work. In P. Hinds & S. Kiesler (Eds.), Distributed work (pp. 137–162). Cambridge, MA:
MIT Press.
Kraut, R. E., Fussell, S. R., & Siegel, J. (2003). Visual information as a conversational re
source in collaborative physical tasks. Human Computer Interaction, 18, 13–49.
Kraut, R. E., Gergle, D., & Fussell, S. R. (2002). The use of visual information in shared vi
sual spaces: Informing the development of virtual co-presence. Proceedings of the Confer
ence on Computer-Supported Cooperative Work, 9 (1), 31–40.
Kruger, J., Epley, N., Parker, J., & Ng, Z. –W. (2005). Egocentrism over e-mail: Can we com
municate as well as we think? Journal of Personality and Social Psychology, 89, 925–936.
Kuzuoka, H., Kosuge, T., & Tanaka, K. (1994). GestureCam: A video communication sys
tem for sympathetic remote collaboration. Proceedings of the Conference on Computer-
Supported Cooperative Work, 5, 35–43.
Kuzuoka, H., & Shoji, H. (1994). Results of observational studies of spatial workspace col
laboration. Electronics and Communications in Japan, 77, 58–68.
Lampe, C., Wash, R., Velasquez, A., & Ozkaya, E. (2010). Motivations to participate in on
line communities. Proceedings of the ACM Conference on Human Factors in Computing
Systems, 29 (1), 1927–1936.
Lanir, J., Stone, R., Cohen, B., & Gurevich, P. (2013). Ownership and control of point of
view in remote assistance. Proceedings of the SIGCHI Conference on Human Factors in
Computing Systems, 32 (1), 2243–2252.
Lea, M., & Spears, R. (1991). Computer-mediated communication, de-individuation and

group decision-making. International Journal of Man-Machine Studies, 34, 283–301.
Li, H. Z. (1999a). Grounding and information communication in intercultural and intracul

tural dyadic discourse. Discourse Processes, 28, 195.
Li, H. Z. (1999b). Communicating information in conversations: A cross-cultural compari

son. International Journal of Intercultural Relations, 23, 387.
Mamykina, L. Miller, A. D., Grevet, C., Medynskiy, Y., Terry, M. A., Mynatt, E. D., & David
son, P. R. (2011). Examining the impact of collaborative tagging on sensemaking in nutri
tion management. Proceedings of the SIGCHI Conference on Human Factors in Comput
ing Systems, 30 (1), 657–666.
Page 35 of 40

McCarthy, J., & Monk, A. F. (1994). Measuring the quality of computer-mediated commu
nication. Behavior & Information Technology, 13(5), 311–319.
McGuire, T. W., Kiesler, S., & Siegel, J. (1987). Group and computer-mediated discussion
effects in risk decision making. Journal of Personality and Social Psychology, 52, 917–930.
McNeill, D. (1992). Hand and mind: What gestures reveal about thought. Chicago: Uni
versity of Chicago Press.
Mehrabian, A. (1967). Orientation behaviors and nonverbal attitude communication. Jour

nal of Communication, 16, 324–332.
Monk, A. F., & Gale, C. (2002). A look is worth a thousand words: Full gaze awareness in
video-mediated conversation. Discourse Processes, 33, 257–278.
Nardi, B. A., & Whittaker, S. (2002). The place of face-to-face communication in distrib
uted work. In P. Hinds, S. Kiesler (Eds.), Distributed work (pp. 83-110). Cambridge, MA
US: MIT Press.
Nardi, B. A., Whittaker, S., & Bradner E. (2000). Interaction and outeraction: Instant mes
saging in action. Proceedings of the Conference on Computer-Supported Cooperative
Work, 8, 79–88.
Newlands, A., Anderson, A. H., & Mullin, J. (2003). Adapting communicative strategies to
computer-mediated communication: An analysis of task performance and dialogue struc
ture. Applied Cognitive Psychology, 17, 325–348.
Nguyen, T. D., & Fussell, S. R. (2012). How did you feel during our conversation? Retro
spective analysis of intercultural and same-culture Instant Messaging conversations. Pro
ceedings of the Conference on Computer-Supported Cooperative Work, 15 (1), 117–126.
Nisbett, R. E. (2003). The geography of thought: How Asians and Westerners think differ
ently, and why. Free Press, New York.
Ochsman, R. B., & Chapanis, A. (1974). The effects of ten communication modes on the
behaviour of teams during (p. 489) cooperative problem solving. International Journal of
Man–Machine Systems, 6, 579–619.
O’Conaill, B., Whittaker, S., & Wilbur, S. (1993). Conversations over video conferences: An
evaluation of the spoken aspects of video-mediated communication. Human–Computer In
teraction, 8, 389–428.
Olson, G. M. & Olson, J. S. (2000). Distance Matters. Human-Computer Interaction, 15,

139–178.
O’Malley, C., Langton, S., Anderson, A. H., Doherty-Sneddon, G., & Bruce, V. (1996). Com
parison of face-to-face and video-mediated interaction. Interacting with Computers, 8,
177–192.
Page 36 of 40

Ou, J., Oh, K., Yang, J., & Fussell, S. R. (2005). Effects of task properties, partner actions,
and message content on eye gaze patterns in a collaborative physical task. Proceedings of
Oyserman, D., Coon, H. M. & Kemmelmeier, M. (2002). Rethinking individualism and col
lectivism: Evaluation of theoretical assumptions and meta-analyses. Psychological Bul
letin, 128, 3–72.
Park, H., Hwang, S. D. & Harrison, J. K. (1996). Sources and consequences of communica
tion problems in foreign subsidiaries: The case of United States firms in South Korea. In
ternational Business Review, 5, 79–98.
Rader, E. (2010). The effect of audience design on labeling, organizing, and finding
shared files. Proceedings of the SIGCHI Conference on Human Factors in Computing Sys
tems, 29 (1), 777–786.
Reid, A. (1977). Comparing telephone with face-to-face contact. In I. de Sola Pool (Ed.),
The social impact of the telephone (pp. 386-415). Cambridge: MIT Press.
Rennecker, J. A., Dennis, A. R., & Munson, S. (2010). “Invisible whispering”: Restructur
ing meeting processes with instant messaging. In D. M. Kilgour & C. Eden (Eds.), Hand
book of group decision and negotiation. Advances in group decision and negotiation vol. 4
(pp. 25–45). New York: Springer.
Reynolds, L., Smith, M. E., Birnholtz, J. P., & Hancock, J. T. (2013). Butler lies from both
sides: Actions and perceptions of unavailability management in texting. Proceedings of
the 2013 Conference on Computer Supported Cooperative Work, 16 (1), 769–778.
Rice, R. (1993). Media appropriateness: Using Social Presence Theory to compare tradi
tional and new organizational media, Human Communication Research, 19, 451–484.
Rutter, M. (1987). Communicating by telephone. Oxford: Pergamon.
Schneider, M., & Kiesler, S. (2005). Calling while driving: Effects of providing remote traf
fic context. Proceedings of the SIGCHI Conference on Human Factors in Computing Sys
tems, 24 (1), 561–569.
Scissors, L. E., Gill, A. J., & Gergle, D. (2008). Linguistic mimicry and trust in text-based
CMC. Proceedings of the Conference on Computer-Supported Cooperative Work, 12, 277–
280.
Sellen, A. J. (1995). Remote conversations: The effects of mediating talk with technology.
Human Computer Interaction, 10, 401–444.
Setlock, L. D., & Fussell, S. R. (2011). Culture or fluency? Unpacking interactions be
tween culture and communication medium. Proceedings of the SIGCHI Conference on
Page 37 of 40

Setlock, L. D., Fussell, S. R., & Neuwirth, C. (2004). Taking it out of context: Collaborat
ing within and across cultures in face-to-face settings and via instant messaging. Proceed
ings of the Conference on Computer-Supported Cooperative Work, 10, 604–613.
Setlock, L. D., Quinones, P. A., & Fussell, S. R. (2007). Does culture interact with media
richness? The effects of audio vs. video conferencing on Chinese and American dyads.
Hawaii International Conference on System Sciences, 40, 13–23.
Short, J., Williams, E., & Christie, B. (1976). The social psychology of telecommunication.
London: John Wiley.
Siegel, J., Dubrovsky, V., Kiesler, S., & McGuire, T. W. (1986). Group Processes in Comput
er-Mediated Communication. Organizational Behavior and Human Decision Processes,
37, 157–186.
Sproull, L., & Kiesler, S. (1986). Reducing social context clues: Electronic mail in organi
zational communication. Management Science, 32, 1492–1512.
Stein, R., & Brennan, S. E. (2004). Another person’s eye gaze as a cue in solving program
ming problems. Proceedings of the 6th International Conference on Multimodal Inter
faces, 6, 9–15.
Stephenson, G., Ayling, K., & Rutter, D. (1976). The role of visual communication in social
exchange. British Journal of Social and Clinical Psychology, 15, 113–120.
Steptoe, W., Steed, A., Rovira, A., & Rae, J. (2010). Lie tracking: Social presence, truth
and deception in avatar-mediated telecommunication, Proceedings of the SIGCHI Confer
ence on Human Factors in Computing Systems, 29 (1), 1039–1048.
Stewart, C. O., Setlock, L. D., & Fussell, S. R. (2007). Conversational argumentation in de
cision-making: Differences across cultures and communication media. Discourse Process
es, 44, 113–139.
Takano, Y., & Noda, A. (1993). A temporary decline of thinking ability during foreign lan
guage processing. Journal of Cross-Cultural Psychology, 24, 445–462.
Tang, A., Pahud, M., Inkpen, K., Benko, H. Tang, J. C., & Buxton, B. (2010). Three’s com
pany: Understanding communication channels in three-way distributed collaboration. Pro
ceedings of the Conference on Computer-Supported Cooperative Work, 13, 271–280.
Tang, J. C. (1991). Findings from observational studies of collaborative work. Internation

al Journal of Man-Machine Studies, 34, 143–160.
Triandis, H. (1989). Cross-cultural studies of individualism and collectivism. Nebraska

Symposium on Motivation. Nebraska Symposium on Motivation, 37, 41–133.
Triandis, H. (1995). Individualism and Collectivism. Boulder, CO: Westview Press.
Page 38 of 40

Turkle, S. (2011). Alone together: Why we expect more from technology and less from
each other. New York: Basic Books.
Veinott, E., Olson, J. S., Olson, G. M. & Fu, X. (1999) Video helps remote work: Speakers
who need to negotiate common ground benefit from seeing each other. Proceedings of the
SIGCHI Conference on Human Factors in Computer Systems, 18 (1), 302–309.
Vertegaal, R., Slagter, R., van der Veer, G., & Nijholt, A. (2001). Eye gaze patterns in con
versations: There is more to conversational agents than meets the eyes. Proceedings of
Walther, J. B. (1992). Interpersonal effects in computer-mediated interaction: A relational

perspective. Communication Research, 19, 52–90.
Walther, J. B. (1995). Relational aspects of computer-mediated communication: Experi

mental observations over time. Organization Science, 6, 186–203.
Wang, H-C., & Fussell, S. R. (2010). Groups in groups: Conversational similarity in

(p. 490)
online multicultural multiparty brainstorming. Proceedings of the Conference on Comput

er-Supported Cooperative Work, 13, 351–360.
Whittaker, S. (2003). Theories and methods in mediated communication. In A. C. Graess

er, M. Gernsbacher, & S. Goldman (Eds.), The handbook of discourse processes (pp. 243–
286). Mahwah, NJ: Lawrence Erlbaum Associates.
Whittaker, S., Geelhoed, E., & Robinson, E. (1993). Shared workspaces: how do they work
and when are they useful?. International Journal of Man-Machine Studies, 39, 5, 813–842.
Whittaker, S., & O’Conaill, B. (1997). The role of vision in face-to-face and mediated com
munication. In K. Finn, A. Sellen, & S. Wilbur (Eds.), Video-mediated communication (pp.
23–49). Mahwah, NJ: Erlbaum.
Williams, E. (1977). Experimental comparisons of face-to-face and mediated communica

tion: A review. Psychological Bulletin, 84, 963–976.
Yamashita, N., & Ishida, T. (2006). Effects of machine translation on collaborative work.
Proceedings of the Conference on Computer-Supported Cooperative Work, 11, 515–523.
Yamashita, N., Inaba, R., Kuzuoka, H., & Ishida, T. (2009). Difficulties in establishing com
mon ground in multiparty groups using machine translation. Proceedings of the SIGCHI
Conference on Human Factors in Computing Systems, 28 (1), 679–688.
Yamashita, N., Kaji, K., Kuzuoka, H., & Hirata, K. (2011). Improving visibility of remote
gestures in distributed tabletop collaboration. Proceedings of the Conference on Comput
er-Supported Cooperative Work, 14, 95–104.
Page 39 of 40

Zelenkauskaite, A., & Herring, S. C. (2008). Television-mediated conversation: Coherence

in Italian iTV SMS chat. Proceedings of the Forty-First Hawai’i International Conference
on System Sciences, 41, 145–156.
Zhang, Q., & Marksbury, N. (2011). Once broken, never fixed? The impact of culture and
medium on repairing trust in CMC. Lecture Notes in Computer Science: Internationaliza
tion, Design and Global Development, 6775, 341–350.
Zheng, J., Veinott, E., Bos, N., Olson, J. S., & Olson, G. M. (2002). Trust without touch:
Jumpstarting long-distance trust with initial social activities. Proceedings of the SIGCHI
Conference on Human Factors in Computing Systems, 21 (1), 141–146.
Susan R. Fussell
Susan R. Fussell, Department of Communication, Cornell University
Leslie D. Setlock
Leslie D. Setlock, Department of Communication, Cornell University
Page 40 of 40

The Role of Natural Language and Discourse Processing in Advanced Tutor
ing Systems
The Role of Natural Language and Discourse Process

ing in Advanced Tutoring Systems
Arthur C. Graesser , Fazel Keshtkar , and Haiying Li

Online Publication Date: Jan 2014 DOI: 10.1093/oxfordhb/9780199838639.013.003
Computer learning environments of the future need to be sensitive to the social, cogni
tive, emotional, and motivational (SCEM) states of students as they learn in their social
environment. Language and discourse plays a central role in tracking SCEM states and
influencing how the computer responds to promote learning. This essay describes a num
ber of computer tutors that are sensitive to these psychological factors and thereby help
students learn. Computer agents are central to the design of these systems. These sys
tems include one-on-one tutoring, conversational trialogs (a tutor agent and student
agent conversing with a human student), and a mentor agent interacting with students in
a multiparty serious game. All of these systems automatically analyze the language and
discourse of the students as they interact with the learning environments. The missions of
science, technology, engineering, and mathematics (STEM) learning are likely to fail with
out the integration of SCEM.
Keywords: tutoring, conversational agents, discourse processing, computational linguistics, intelligent tutoring
systems, emotions
The two questions that a social psychologist interested in language would want to ask
about this essay are “why tutoring?” and “why computers?” Some of the answers to these
two questions are obvious. Tutoring is an excellent way to help students learn various
skills and subject matters so the students function as good citizens in a society, are finan
cially secure, are healthy and happy, and teach their children well. Computers help peo
ple with either repetitive or exceedingly complex tasks that humans try to avoid. So, what
could be better? A computer tutor would result in less poverty and more satisfaction
among citizens, families, and the workforce.
However, it is the counterintuitive answers that are no doubt most illuminating to readers
of this Handbook. Tutoring intersects with social, cognitive, emotional, and motivational
(SCEM) mechanisms in unsuspected ways that make the phenomenon worthy of study.
The claim could be made that every speech act of a good tutor has tentacles to SCEM
processes. Human tutors frequently do not generate intelligent speech acts, and neither
Page 1 of 32

ing Systems
do computers, so we can explore the various ways that conversations succeed or fail dur
ing tutoring. The practical goal is to build SCEM-sensitive tutors, whereas the theoretical
goal is to understand tutoring as a social communication process that helps students
learn. Some illustrations of this are foreshadowed here:
1. The common ground (shared knowledge) between tutor and student is negligible
to modest, so there is a struggle to understand each other. Human tutors often do
not realize this, so they yammer on and on, believing mistakenly that they are under
stood by the student, much like a citizen giving directions to a lost visitor in (p. 492) a
city. The student and tutor need to struggle to achieve common ground because
there is a large gulf between what the student knows and the tutor knows (Chi, Siler,
& Jeong, 2004; Graesser, D’Mello, & Person, 2009; Graesser, Person, & Magliano,
1995). This gap in common ground between computer and student makes it more
feasible to build a reasonably smooth conversational computer tutor from the per
spective of the student because the student does not readily identify discontinuities
in coherence.
2. The roles of the tutor and student are frequently misunderstood, particularly when
the tutor is accomplished. The default assumption in our culture is that the student is
inadequate in knowledge and skills, so the tutor teaches these skills by supplying in
formation. But, in truth, a good tutor gets the student to do the talking, rather than
acting as an information delivery system (Chi, Siler, Jeong, & Hausmann, 2001; Millis
et al., 2011; Schwartz et al., 2009). Bad tutors give lectures whereas good teachers
get the student to talk. Bad tutors are know-it-all’s whereas good tutors ask ques
tions to encourage the co-construction of an answer (e.g., “Great question! How can
we figure out how to answer it?”). Some parents grumble when schools ask their
children to tutor other students, but they fail to realize that teaching is an excellent
way to learn. The good news for the computer scientist is that a computer can get a
student to do the talking by asking questions, changing topics, and throwing the con
versational spotlight on the student. Whenever the computer is uncertain about what
the student knows, it can finesse the conversation to get the student to talk and do
things.
3. A good tutor strategically violates pragmatic rules of everyday conversation. Par
ticipants in a normal conversation assume that whatever is said is part of the com
mon ground in the discourse space (Clark, 1996), that people ask questions when
they do not understand (Van der Meij, 1994), and that people should be polite and
avoid face-threatening acts (Brown & Levinson, 1987). However, there are a number
of conditions in which a good tutor should cast aside these assumptions and violate
these norms (D’Mello & Graesser, 2012a; Person, Kreuz, Zwaan, & Graesser, 1995;
Ogan, Finkelstein, Walker, Carlson, & Cassell, 2012). A good tutor is skeptical of the
student’s understanding of anything expressed, accepts the fact that the student is
reluctant to ask questions, and confronts the student to uncover his or her concep
tions of the material. There is a tradeoff between a tutor being confrontational and
being supportive. This makes it easier on the computer system: when the computer
Page 2 of 32

ing Systems
is in doubt, it can safely assume that the learner understands and knows very little
and that an impolite, probing question has added value.
4. A good tutor puts the student in cognitive disequilibrium. Students are in cogni
tive disequilibrium when there are obstacles to goals, contradictions, unexpected
events, unusual contrasts, and uncertainty in decisions. These events inspire thought
and are opportunities for learning even though cognitive disequilibrium triggers con
fusion, surprise, and sometimes frustration (Baker, D’Mello, Rodrigo, & Graesser,
2010; D’Mello, Lehman, Pekrun, & Graesser, in press; Graesser & D’Mello, 2012). It
is not always good for the tutor to be polite and supportive because problem solving
and the acquisition of deep knowledge tend to be accompanied by the negative affec
tive states of confusion, frustration, anger, and even rage (Barth & Funke, 2010;
D’Mello & Graesser, 2012b; Graesser & D’Mello, 2012). Politeness norms and deep
learning may be contradictory. Computer scientists may be reassured by this be
cause the best computer systems are good at presenting difficult challenges, where
as an empathetic computer tutor might not be convincing.
5. A tutoring interaction has a rich space of options rather than following a tightly
scripted delivery. There is a structure to tutoring, as will be described later in this
essay. But there is also a wide degree of flexibility and “optionhood” in what the tu
tor can say and do. The tutor can abruptly change its mind, introduce a new ques
tion, and end the immediate exchange without seriously disrupting the flow of con
versation. This is easier for computers to execute: when in doubt, the system can
start a new conversational thread and change its “mind”—just as human tutors do.
The point of this prelude is that tutoring is an interesting social phenomenon, that the
constraints of tutoring have complex interactions with SCEM mechanisms, and that many
folklore intuitions about tutoring are off the mark. This essay conveys how important it is
for social psychologists to play a central role in guiding research on human and computer
tutoring in the future.
Social and Pragmatic Characteristics of

(p. 493)
Human Tutoring and Ideal Computer Tutoring

This section reviews the social and pragmatic processes of human and computer tutoring.
Researchers in education and discourse processes have conducted detailed analyses of
human tutoring on a variety of subject matters, with tutors who vary in expertise and stu
dents who vary in age and abilities (Chi et al., 2001; Chi, Roy, & Hausmann, 2008; Evens
& Michael, 2005; Graesser & Person, 1994; Graesser et al., 1995; Lehman, D’Mello, Cade,
& Person, 2012; Lepper, Drake, & O’Donnell-Johnson, 1997; McArthur, Stasz, & Zmuidzi
nas, 1990; Person et al., 1995; Shah, Evens, Michael, & Rovick, 2002). These studies track
the speech acts, actions, and emotions of the student and tutor as they communicate turn
by conversational turn during the course of tutoring. Their detailed discourse and peda
gogical analyses have unveiled insights into SCEM mechanisms.
Page 3 of 32

ing Systems
A deep analysis of human tutoring was motivated by some meta-analyses that showed
learning gains of approximately 0.4 sigma when comparing one-on-one human tutoring
with classroom controls or other suitable controls over study time and content (Cohen,
Kulik, & Kulik, 1982). A sigma is a measure of the impact of an intervention that com
pares the difference between two means (experimental condition and control condition)
in standard deviation units. The expertise of the tutor does matter, but more evidence is
needed to support such a claim according to available research. Collaborative peer tutor
ing shows an effect size advantage of 0.2–0.9 sigma (Mathes & Fuchs, 1994; Topping,
1996), which is slightly lower than older unskilled human tutors. At the other end of the
spectrum, accomplished human tutors have an impact of 0.6–2.0 sigma (Bloom, 1984;
VanLehn, 2011; VanLehn et al., 2007), but the sample of accomplished tutors was small
and the effect sizes depended very much on the subject matter and targeted skills. These
studies showing the effectiveness of human tutoring prompted researchers to identify
patterns of language, discourse, pedagogy, and social interaction that might explain why
tutoring is so valuable in helping students learn.
The research on human tutoring convinced researchers in the computational sciences to

build intelligent tutoring systems to simulate tutoring. These systems help students learn
by holding conversations with them in natural language and by implementing tutoring
strategies that simulate either humans or ideal pedagogical strategies. Prominent exam
ples of these systems are AutoTutor (Graesser, D’Mello et al., 2012; Graesser et al., 2004;
VanLehn et al., 2007), CIRCSIM Tutor (Evens & Michael, 2005; Shah et al., 2002), Coach
Mike (Lane, Noren, Auerbach, Birch, & Swartout, 2011), Guru (Olney et al., 2012), iS
TART (McNamara, Boonthum, Levinstein, & Millis, 2007), ITSpoke (Litman & Forbes-Ri
ley, 2006), MetaTutor (Azevedo, Moos, Johnson, & Chauncey, 2010), Operation ARIES!
(Millis et al., 2011), PACO (Rickel, Lesh, Rich, Sidner, & Gertner, 2002), Tactical Lan
guage and Culture System (Johnson & Valente, 2008), and Why-Atlas (VanLehn et al.,
2007). These systems were made possible by landmark advances in the fields of computa
tional linguistics (Jurafsky & Martin, 2008), statistical representations of world knowl
edge (Landauer, McNamara, Dennis, & Kintsch, 2007), corpus linguistics (Biber, Conrad,
& Reppen, 1998), educational data mining (Baker & Yacef, 2009), and interdisciplinary
connections between psychology and computer science (McCarthy & Boonthum-Denecke,
2012; McNamara, Graesser, McCarthy, & Cai, in press; Pennebaker, Booth, & Francis,
2007).
This section identifies the tutoring strategies that prevail in normal tutoring and that
have been implemented in AutoTutor and a number of other computer tutors that help
students learn by holding a conversation in natural language. Whereas some human tu
toring strategies have been implemented in computer tutors, computer tutors also some
times implement ideal strategies not exhibited by human tutors. Whatever strategies are
implemented in a computer tutor, it is important to identify their foundations in SCEM
mechanisms. Two fundamental questions are asked in this program of research. First,
what discourse patterns occur in normal tutoring and ideal tutoring? Second, do these
Page 4 of 32

ing Systems
conversational patterns help students learn? The primary focus of this essay is on the first
question.
The Poverty of Common Ground

Successful communication requires the establishment of a common ground (or shared
knowledge) between speaker and listener or writer and reader (Clark, 1996; Holtgraves,
2002; Schober & Brennan, 2003). Common ground is minimal when there is a large gap
between the expert tutor’s knowledge and the student’s knowledge. Tutoring conversa
tions are very different when common ground exists or is achieved versus when it deteri
orates. This predicament requires a good tutor to (p. 494) detect and repair misunder
standings, but the vast majority of tutors are unable to do so (Graesser et al., 2009).
Graesser, D’Mello, and Person (2009) identified five “illusions” of tutoring that reflect the
likely deterioration of common ground and how the tutor might repair such deficits. The
DeepTutor computer tutor (Rus, 2010) has been designed to overcome these illusions and
thereby improve tutoring of physics.
1. Illusion of grounding. The tutor and student hold the mistaken belief that they
have shared knowledge about a word, referent, or idea being discussed in the tutor
ing session. Such grounding is often illusory. A good tutor tries to gauge the
student’s level of understanding and to troubleshoot potential breakdowns in com
munication.
2. Illusion of feedback accuracy. The tutor and student assume that their feedback to
each other is accurate during their communication. This often is a mistaken belief.
When tutors ask comprehension-gauging questions (e.g., “Do you understand?,” “Are
you following?”), many students mistakenly say “yes,” whereas those students with
better comprehension tend to say “no.” When the student gives error-ridden or
vague answers to a tutor’s question, the tutor’s short feedback (“good,” “maybe,”
“no”) is more likely to be positive than negative. A good tutor should not trust the
student’s answer to a comprehension-gauging question. Similarly, a student should
not trust the short feedback of a polite tutor.
3. Illusion of discourse alignment. The tutor and student often mistakenly assume
that they are connected on the discourse function of speech acts. For example, tutors
often give indirect hints to nudge the student to think in a particular direction
(“What about X?”), but students don’t interpret these speech acts as hints. Some
times students ask a question with a presupposed assertion (“Isn’t 17 a prime num
ber?”), but tutors answer the questions (“Yes. 17 is a prime number”) without giving
students credit for their contribution. A good tutor clarifies the epistemological sta
tus of the speech acts expressed in a tutoring session.
4. Illusion of student mastery. Both tutors and students assume that the student has
mastered much more knowledge than the student has really mastered. For example,
it is assumed that the student has mastered a complex conceptualization if the stu
dent can express a couple of words or ideas about it. However, this skimpy student
articulation hardly goes the distance in understanding a complex conceptualization.
Page 5 of 32

ing Systems
A good tutor asks follow-up questions to verify student understanding. A good stu
dent does not overestimate what he or she knows.
5. Illusion of knowledge transfer. The tutor believes that the student understands
whatever the tutor expresses; the student assumes that whatever he or she express
es is understood by the tutor. In fact, the student absorbs surprisingly little of what
the tutor says and vice versa. A good tutor assumes that nothing he says is under
stood unless the student expresses or does something that reflects understanding.
The following example illustrates a large gap in common ground between the tutor and
student. This came from a tutoring session on research methods.
TUTOR: Do you know how to get the main effect in a factorial design?
STUDENT: Yeah, and I understand that, like, its what the independent variable it
self…like a measurement of it, itself. And, um, I know, like, that looking at these, I
know how to get it.
TUTOR: Okay, right. So if you have a 3 by 2 factorial design, you look at the vari
ability among marginal means on one independent variable and divide that by the
error term.
STUDENT: I see.
Examples such as this illustrate the complexity of tutoring with respect to developing and
monitoring common ground between tutor and student. The student’s language is un
grammatical, vague, semantically ill-formed, and incoherent—in sharp contrast with the
tutor’s tight analytical construction. The student politely expresses “I see” after the
tutor’s description of a main effect, but the likelihood that the student truly understands
is near zero. This student’s contribution is representative of students’ language that a
computer tutor needs to handle if it converses with the student in natural language.
Expectation and Misconception-Tailored Dia

logue
The tutoring strategies of human tutors follow a systematic conversational structure that
is called expectation and misconception-tailored dialogue (p. 495) (EMT dialog; Graesser,
D’Mello et al., 2012; Graesser, Lu, Olde, Cooper-Pye, & Whitten, 2005; VanLehn et al.,
2007). Human tutors anticipate particular correct answers (called expectations) and par
ticular misunderstandings (misconceptions) when they ask their students challenging
questions (or problems) and trace their reasoning. As the students express their answers,
which are distributed over multiple conversational turns, their contributions are com
pared with expectations and misconceptions. The tutors give feedback to the students’
answers with respect to matching the expectations or misconceptions. The short feed
back consists of positive, neutral, or negative expressions either in words, intonation, or
facial expressions. Interestingly, sometimes the words are politely positive (good) or neu
Page 6 of 32

ing Systems
tral (okay), but the intonation or facial expressions are pedagogically accurate; for exam
ple, negative after an erroneous or vague student response (Graesser et al., 1995).
After the short feedback, the tutor tries to lead the student to express expectations (good
answers) through multiple dialogue moves, such as pumps (“What else?”) hints, and
prompts to get the student to express specific words (Graesser, Conley, & Olney, 2012).
When the student fails to answer the question correctly, the tutor contributes information
as assertions. These pump-hint-prompt-assertion cycles are frequently used in tutoring to
extract or cover particular sentence-like expectations. Eventually, all of the expectations
are covered, and the exchange is finished for the main question or problem. During this
process, the student occasionally asks questions, which are immediately answered by the
tutor, and the student expresses misconceptions, which are immediately corrected by the
tutor. A more accomplished tutor might try to get the student to answer her own ques
tions or correct her own misconceptions, but such self-regulated activities are extremely
rare in actual tutoring.
Expectation and misconception-tailored tutoring can be simulated on a computer, as in

the case of AutoTutor and Why/Atlas for introductory computer literacy and Newtonian
physics (Graesser et al., 2004; Graesser, D’Mello et al., 2012; VanLehn et al., 2007). The
success of the computer’s implementing the EMT dialogue mechanism depends on how
well semantic matches can be made between the verbal contributions of the student and
the list of expectations and misconceptions. Fortunately, advances in natural language
processing research have made major progress in the accuracy of these semantic match
es in computers. These semantic match algorithms have included keyword overlap scores,
word overlap scores that place higher weight on lower frequency words in the English
language, scores that consider the order of words, latent semantic analysis cosine values,
regular expressions, and procedures that compute logical entailment (Cai et al., 2011;
Graesser & McNamara, 2012; Graesser, Penumatsa, Ventura, Cai, & Hu, 2007; Rus &
Graesser, 2006). It is beyond the scope of this essay to describe these semantic match al
gorithms, but, briefly, in these assessments, a semantic match score (between 0 and 1) is
made between the verbal contributions of students and the sentence-like expressions of
expectations and misconceptions. The question is how well the computer match scores
correspond to the match scores of trained human judges. The performance data are im
pressive in most of these assessments. Cai et al. (2011) recently reported an algorithm
with computer semantic match scores correlated .67 with trained human judges, whereas
the trained human judges correlated .69 with each other.
In contrast to EMT dialogue, it is very difficult to implement a host of sophisticated peda

gogical strategies that humans also rarely exhibit, such as bona fide Socratic tutoring
strategies, modeling-scaffolding-fading, reciprocal teaching, frontier learning, building on
prerequisites, or diagnosis/remediation of deep misconceptions (Collins, Brown, & New
man, 1989; Palincsar & Brown, 1984; Rogoff & Gardner, 1984; Sleeman & Brown, 1982).
In Socratic tutoring, the tutor asks learners illuminating questions that lead them to dis
cover and correct their own misconceptions in an active, self-regulated fashion. In model
ing-scaffolding-fading, the tutor first models a desired skill, then gets the learners to per
Page 7 of 32

ing Systems
form the skill while the tutor provides feedback and explanation.The tutor eventually
fades from the process until the learners perform the skill unaided. In reciprocal teaching,
the tutor and learner take turns working on problems or performing a skill, as well as giv
ing feedback to each other along the way. Tutors who use frontier learning select prob
lems and give guidance in a fashion that slightly extends the boundaries of what the
learner already knows or has mastered. Tutors who build on prerequisites cover the pre
requisite concepts or skills in a session before moving to more complex problems and
tasks that require mastery of the prerequisites.
There are important implications to the discovery that EMT dialogue is both common in
human tutoring and can be implemented in computer tutors. One practical implication is
that these computer tutors can (p. 496) fill a critical need in providing human tutoring
when the human tutors are unavailable. Available empirical evidence indeed supports the
claim that computer tutors with natural language dialogue yield learning gains compara
ble to trained human tutors, with effect sizes of 0.6–1.0 (VanLehn, 2011; VanLehn et al.,
2007; Olney et al., 2012). A second implication is that these computer tutors may do even
better than human tutors if they weave in the more sophisticated ideal tutoring mecha
nisms just described. That is, the best computer tutor may be a hybrid of human strate
gies and ideal pedagogical tutoring strategies.
Speech Acts and Other Dialogue Moves

A speech act is a category of utterance, clause, or sentence that serves a discourse func
tion. Examples of speech act categories are questions, answers, statements (assertions),
requests, commands, short reactions, expressive evaluations, greetings, and meta-com
ments. The speech act categories adopted in the tutoring research are based on theoreti
cal schemes, with eight to ten categories that also can be reliably coded by trained judges
(D’Andrade & Wish, 1985; Graesser, Swamer, & Hu, 1997; Lehman et al., 2012; Olney et
al., 2003). These categories can be broken down into several dozen subcategories. Com
puter tutors need to segment the stream of words from the students’ contributions into
speech acts and then assign them to one of these categories.
The speech acts produced by the human or computer tutor have a somewhat different
classification scheme that is motivated by pedagogical considerations. These speech act
categories, or what are often called dialogue moves, include the following:
• Short feedback. The feedback to the student is either positive (“yes,” “very good,”
head nod, smile), negative (“no,” “not quite,” head shake, pregnant pause, frown), or
neutral (“uh huh,” “I see.”)
• Pumps. The tutor gives nondirective pumps (“What else?” “Tell me more”) to get the
student to do the talking.
• Hints. The tutor gives hints to get the student to do the talking or doing, but directs
the students along some conceptual path. The hints vary from being generic state
ments or questions (“What about X?,” “Why not?”) to speech acts that more directly
Page 8 of 32

ing Systems
lead the student to a particular answer. Hints are the ideal scaffolding move to pro
mote active student learning while directing the student to focus on important relevant
material.
• Prompts. The tutor asks a very leading question to get the student to articulate a par
ticular word or phrase. Sometimes students say very little, so these prompts are need
ed to get the student to say something.
• Assertions. The tutor expresses a fact or state of affairs.
Other categories of tutor dialogue moves are self-explanatory, such as answers to student
questions, corrections of student misconceptions, summaries, mini-lectures, or off-topic
comments.
One frequent discourse pattern occurs when the tutor tries to get the student to articu
late a sentence, such as the Newtonian law “Force equals mass times acceleration.” The
tutor starts with a pump, with the hope that the student articulates the law. If not, the tu
tor gives a hint, such as “What about force?” or “What does force equal?” If the student
doesn’t articulate “mass times acceleration,” then the tutor gives a prompt, such as
“Force equals mass times what?” If the student still comes up empty, the tutor simply as
serts the law by saying “Force equals mass times acceleration.” This pump → hint →
prompt → assertion cycle is common in human tutoring and is implemented in AutoTutor.
The distribution of tutor dialogue moves reflects the extent to which the tutor versus the
student is contributing to answering the question (or solving the problem). This is mea
sured in the correlations between students’ knowledge about the topic and the likelihood
that tutors express these categories. For example, there is a significant positive correla
tion between student knowledge and the tutor’s positive feedback, pumps, and hints, but
a negative correlation with the tutor’s negative feedback, prompts, assertions, and cor
rections (Jackson & Graesser, 2006). This finding suggests that a researcher can infer
what the student knows from the speech acts of tutors who intelligently interact with the
student. The tutors’ speech act distributions reveal the extent to which the information is
contributed by the student versus the tutor in the collaborative construction of an an
swer/solution.
Sometimes students misunderstand the discourse function of particular speech act cate
gories, such as hints. When hints are assertions, the student may view them as mere
statements of fact and not realize that the tutor is trying to steer the student to a better
answer. When hints are questions (such as “What about X?”) and the student does not see
what the tutor is driving at, then the student becomes confused and retorts (“Well what
about X?”). A good tutor minimizes confusion by preceding the hint with a discourse
marker that signals the (p. 497) discourse function, such as “Here’s a hint” or “Let me get
you back on track.”
Page 9 of 32

ing Systems
Five-Step Tutoring Frame

The tutor frequently implements a Five-Step Tutoring Frame to solve the problem or an
swer a difficult question (Graesser, Conley et al., 2012; Graesser & Person, 1994; Graess
er et al., 1995). The five steps are identified here:
1. Tutor asks a difficult question (or presents the problem).

2. Student gives an initial answer.
3. Tutor gives immediate, short feedback on the quality of the answer.
4. Tutor and student interact through a multiturn dialogue to improve the answer.
5. Tutor assesses whether the student understands the answer.
Step 4 is particularly important because this is when the student and tutor interact dy
namically in the collaborative construction of an answer. Step 5 normally involves a com
prehension-gauging question by the tutor, such as “Do you understand?” For a variety or
reasons, most students incorrectly say “Yes,” even when they do not understand. For ex
ample, they may have poor metacognitive assessments of their own knowledge, or they
may be reluctant to disappoint the tutor. The more knowledgeable student tends to say
“No, I don’t understand” (Chi, DeLeeuw, Chiu, & LaVancher, 1994; Graesser et al., 1995).
It may take knowledge to know what you don’t know (Miyake & Norman, 1979). For these
reasons, a good tutor should not rely on comprehension-gauging questions to assess the
students’ knowledge. Instead, he should ask follow-up question or give follow-up tasks to
further assess the students’ understanding.
The first three of these steps often occur in a classroom context, using easier short-an
swer questions. The teacher asks a question, a student is called on to give a quick an
swer, and the teacher gives positive or negative feedback (Sinclair & Coulthard, 1975).
Perhaps tutoring is better than classroom teaching because more difficult questions are
asked or because of the collaborative interaction in step 4. Research has not yet deter
mined which of these explanations is most plausible.
The Structure of a Tutor Conversational Turn

Human tutors and AutoTutor structure most of their conversational turns in three con
stituents or slots. The first slot is the positive, neutral, negative feedback on the quality of
the student’s last turn. The second slot advances the interaction with either prompts for
specific information, hints, assertions of expected answers, corrections of misconcep
tions, or answers to the student’s question. The third slot is a cue to shift the conversa
tional floor from the tutor to the student. For example, the human tutor cues the student
with a question, an intonation signal, or a gesture to encourage the students to talk. With
out this cue, there is a standoff of silence, as each waits for the other to speak. In summa
ry, the structure of many tutor turns follows this computational rule:
Tutor Turn → Short Feedback + Dialogue Advancer + Floor Shift
Page 10 of 32

ing Systems
Two examples of a tutor’s conversational turn are below:
• “That’s right. 17 is a prime number. Why is that important?”

• “Not quite. The difference between acceleration and velocity is [hand ges
ture].”
The short feedback that begins most tutor conversational turns has been investigated in
considerable detail (Evens & Michael 2005; Fox, 1993; Graesser et al., 1995; Person et al.,
1995). Humans are prone to give positive feedback after low-quality or error-ridden stu
dent turns, perhaps to be polite or not discourage the student from contributing. Experi
enced human tutors are allegedly more discriminating and less concerned with politeness
(Lehman et al., 2012), but systematic comparisons of tutors with varying expertise have
not been conducted on short feedback. Graesser, D’Mello, and Person (2009) discussed
the plausible hypothesis that students want their tutor feedback to be accurate, with po
liteness taking a back seat. AutoTutor sometimes does not accurately interpret student
contributions, so it occasionally gives incorrect short feedback. Students sometimes be
come annoyed when this occurs, occasionally dismissing the utility of AutoTutor altogeth
er (D’Mello, Craig, Witherspoon, McDaniel, & Graesser, 2008). Another observation from
these sessions with AutoTutor is that students want decisive feedback (yes vs. no) rather
than evasive or indecisive feedback (possibly, uh huh, okay). A polite or wishy-washy com
puter tutor does not seem to be as desirable as a decisive one. We speculate that the stu
dents’ assumptions about pragmatic ground rules and communication may be very differ
ent for human as opposed to computer tutors.
(p. 498) Role of Motivation and Emotions

In addition to cognition, emotions and motivation play a critical role in the learning
process. There is a rich literature on the role of motivational and attitudinal traits on
learning. Enduring traits that are relevant to learning have tapped constructs of motiva
tion, self-concept, and goal orientation (Daniels et al., 2009; Deci & Ryan, 2002; Dweck,
2002; Frenzel, Pekrun, & Goetz, 2007; Linnenbrink, 2007; Pekrun, Elliot, & Maier, 2006;
Schutz & Pekrun, 2007). For example, motivational theorists often contrast mastery-ori
ented learners who want to understand the material versus performance-oriented stu
dents who want good grades. Mastery-oriented learners are more persistent, even in the
case of failure, whereas performance-oriented learners have more negative emotions af
ter failures and tend to avoid such challenges. Self-concept theorists contrast students
who view themselves as having a trait of being good or bad at learning a subject matter
versus those who believe that effort will pay off at learning virtually any subject matter;
the former tend to get bored with learning activities much sooner than the latter. Another
contrast that has been made is between adventuresome learners, who want to be chal
lenged with difficult tasks, take the risks of failure, and manage negative emotions, and
cautious learners, who tackle easier tasks, take fewer risks, and minimize failure and the
resulting negative emotions (Clifford, 1991; Meyer & Turner, 2006). An ideal tutor should
Page 11 of 32

ing Systems
be sensitive to these psychological learner traits but, interestingly, this is an understudied
research area.
Lepper was among the first to directly investigate the role of motivation and emotions in
tutoring (Lepper et al., 1997; Lepper & Woolverton, 2002). An INSPIRE model was pro
posed to promote this integration. This model encouraged the tutor to nurture students
by being empathetic and attentive to students’ needs, to assign tasks that are neither too
easy nor too difficult, to give indirect feedback on erroneous student contributions rather
than harsh feedback, to encourage the students to work hard and face challenges, to em
power the students with useful skills, to give the students choices, and to pursue topics
that the students are curious about. One interesting tutor strategy is to assign an easy
problem to the student, but claim that the problem is difficult and encourage the student
to give it a try anyway. When the student readily solves the problem, she builds self-confi
dence and self-efficacy in conquering difficult material.
Research on intelligent tutoring has recently concentrated on the moment-to-moment

emotions that students experience during learning (Baker, D’Mello, Rodrigo, & Graesser,
2010; Calvo & D’Mello, 2010; Conati & Macclaren, 2010; Du Boulay et al., 2011; D’Mello
& Graesser, 2010; D’Mello, Craig, & Graesser, 2009; Graesser & D’Mello, 2012; McQuig
gan, Robison, & Lester, 2010; Picard, 2010; Woolf et al., 2009). The emotional states are
polled every few seconds by different methods of measurement, such as behavioral obser
vations, learner self-reports, expert judges, physiological recordings, facial expressions,
body movements, and language/discourse of the tutor–student interaction. This research
has revealed that a small set of learning-centered emotions dominate emotional experi
ences during learning: confusion, flow (engagement), boredom, and frustration, with de
light and surprise occurring less frequently and anxiety occurring under testing contexts.
These learner-centered emotions are very different from the six basic emotions investigat
ed by Ekman (1992), which are readily manifested in facial expressions: sadness, happi
ness, anger, fear, disgust, and surprise.
Interesting, investigations with AutoTutor have revealed that the language and discourse
of student and tutor are very diagnostic of the emotional states of the student (D’Mello et
al., 2008; D’Mello & Graesser, 2010, 2012c; Graesser & D’Mello, 2012). In fact, when all
of the learning-centered emotions are considered, language/discourse features predict
learner emotional states as well as facial expressions and better than body posture. The
positive emotion of flow tends to occur when the learner is quickly producing information
during his or her conversational turns and receives positive feedback from the tutor. The
negative emotion of boredom tends to occur later in the tutoring session, usually when
AutoTutor is presenting information (asserting, summarizing, lecturing); presumably, by
then, the student is getting bored from information overload and lack of activity. Frustra
tion tends to occur when students are producing information that they believe is on the
mark, but for which they receive negative feedback because AutoTutor does not give the
student due credit. Confusion tends to occur when discourse cohesion is low (e.g., Auto
Tutor uses a term or pronoun the student does not understand), the learner does not pro
duce much information, the student is slow to respond, the feedback is negative or con
Page 12 of 32

ing Systems
tradictory, and the student does not understand AutoTutor’s hints. At these moments,
(p. 499) the student is deep in thought and experiencing cognitive disequilibrium. As men
tioned earlier, confusion has been the highest predictor of learning gains from AutoTutor
compared with the other learning-centered emotions (D’Mello et al., in press; Graesser &
D’Mello, 2012). Confusion occurs when the learner experiences cognitive disequilibrium,
a state that occurs when learners face obstacles to goals, contradictions, incongruities,
anomalies, and uncertainty. Of course, persistent, hopeless confusion presumably has lit
tle pedagogical value.
The detection of student emotions is an important advance, but it is not particularly use
ful unless the tutor can intelligently respond to student emotions. Therefore, an emotion-
sensitive version of AutoTutor, called Affective AutoTutor, was developed to automatically
detect student emotions based on multiple communication channels and to respond to
students’ emotions by selecting appropriate discourse moves and displaying emotions
through facial expressions and speech (D’Mello & Graesser, 2012a, 2012b; D’Mello,
Craig, Fike, & Graesser, 2009). Student emotions are automatically tracked by facial ex
pressions, body posture, language, and discourse interaction (D’Mello & Graesser, 2010).
The primary student emotions that Affective AutoTutor tries to handle strategically are
confusion, frustration, and boredom because these are the emotions that run the risk of
leading to disengagement from the task. The tutor continues business as usual when the
student is emotionally neutral, in the state of flow/engagement, or is experiencing the
fleeting emotions of delight and surprise; there is no need to respond to these affective
states in any special way.
Affective AutoTutor uses a complex set of algorithms that determine how it responds to
student emotions. It is beyond the scope of this essay to cover these mechanisms, but a
few examples illustrate how this can be accomplished. When frustration is detected, the
tutor expresses supportive empathetic comments (“The material is difficult but I believe
you can get it”) to enhance motivation, in addition to supplying hints or prompts to help
the student contribute to answering the question. Frustration needs to be handled to pre
vent the student from transitioning to boredom and disengagement in a downward spiral.
When the student is bored, the tutor responds by giving more engaging material (some
razzle dazzle) or challenging the student with more difficult material (for high-knowledge
students) or easier material (for low-knowledge students). Regarding confusion, the cog
nitive disequilibrium framework predicts that confusion is a critical juncture in the learn
ing process, one that is sensitive to individual differences. Some students may give up
when experiencing confusion because they perceive themselves as not being good with
the subject matter or to avoid negative feedback. For these kinds of students, encourage
ment, hints, and prompts are allegedly the best strategy for helping them get over the
hurdle. Other students treat confusion as a challenge to conquer and expend cognitive ef
fort to restore equilibrium; these students need no special treatment. Affective AutoTutor
discriminates between these two types of students by the automated detection of confu
sion together with the quantity and quality of student contributions in the tutorial interac
tion.
Page 13 of 32

ing Systems
Affective AutoTutor has an emotion generator that displays a limited set of emotions
when the system issues a launch command. With the emotion generator activated, the
agent speaks with an intonation that is properly integrated with facial expressions that
display particular emotions. The agent nods enthusiastically with approval and expresses
positive feedback language after the student makes a correct contribution. The agent
shakes its head or has a skeptical facial expression when student contribution is of low
quality. There is an empathetic verbal message, kind facial expressions, and an encourag
ing demeanor when the student needs support. A small set of emotion displays like these
go a long way toward conveying the tutor’s emotions.
A study was conducted to test the impact of different versions of Affective AutoTutor on
learning gains (D’Mello & Graesser, 2012a). The study compared the original AutoTutor
without emotion tracking and emotional displays to the emotionally supportive Affective
AutoTutor version. The supportive Affective AutoTutor used polite and encouraging posi
tive feedback (“You’re doing extremely well”) or negative feedback after a low-quality stu
dent contribution (“Not quite, but this is difficult for most students”). When a student of
fered low-quality contributions, the tutor attributed the problem to the difficulty of the
material for most students rather than blaming the student. There was also a shake-up
version of Affective AutoTutor. This version tried to shake up student emotions by being
playfully cheeky and telling the student what emotion he or she was experiencing (“I see
that you are frustrated”). The simple substitution of this feedback dramatically changed
AutoTutor’s personality.
The different AutoTutor versions systematically influenced the learning of com

(p. 500)
puter literacy, but in a way that depended on the phase of tutoring and the student’s level
of mastery. During early phases of the tutoring session, the supportive Affective AutoTu
tor had either no impact (for low-knowledge students) or a negative impact (for high-
knowledge students) on learning. During a later phase of tutoring, the supportive AutoTu
tor improved learning, but only for low-knowledge students. These results suggest that
supportive emotional displays by AutoTutor may not be beneficial during the early phases
of an interaction when the student and agent are “bonding” but that a supportive tutor is
appropriate at later phases for students who have low knowledge and encounter difficul
ties. In essence, there may be an optimal time for emoting. The shake-up AutoTutor was
never fully tested because an early study indicated that learning gains were the same as
for the original AutoTutor. However, most adults have a positive initial impression of the
shake-up AutoTutor because the playful shake-up tutor is engaging, at least initially. Per
haps the shake-up tutor can be motivating when boredom starts to emerge for more con
fident, high-knowledge learners, but this needs to be tested in future research.
Some recent computer tutoring systems have given their agents a variety of personalities
that hold some promise in influencing motivation, emotions, and learning. Researchers
have explored the advantages and liabilities of a variety of personalities: polite, empathet
ic, assertive, cool, playful, and even rude (Baylor & Kim, 2005; Ogan et al., 2012; Person,
Burke, & Graesser, 2003; Wang et al., 2008). There presumably are times when a polite
and supportive agent is best; namely, when students lack confidence and need a boost.
Page 14 of 32

ing Systems
There are other times when a playfully rude tutor will keep the learner entertained while
serious content is smuggled into the material. The delicate balance between play and se
rious content is a frontier for future research.
Tutoring in Multiparty Conversations

Tutoring often occurs in a group context, so the logical next step is to have computer
agents participate in these groups. A human can learn vicariously by observing other
agents interacting in ways that model ideal learning processes and collaborative reason
ing. A human can interact with a peer during learning or even teach the peer agent while
the tutor agent intervenes periodically. A human can be tutored by an AutoTutor agent,
with a student agent periodically entering the conversation and illustrating ways that stu
dents can interact with the AutoTutor. A human can coach a student agent to help him or
her take a test administered by a teacher agent. All of these possibilities have been stud
ied in computer learning environments such as Operation ARIES! (Halpern et al., 2012;
Millis et al., 2011), iSTART (McNamara et al., 2007), and Betty’s Brain (Biswas, Jeong,
Kinnebruw, Sulcer, & Roscoe, 2010; Schwartz et al., 2009). Larger group interactions
with agents have developed in Tactical Language and Culture Training System (Johnson
& Valente, 2008), Crystal Island (Rowe, Shores, Mott, & Lester, 2010), River City (Ketel
hut, Dede, Clarke, Nelson, & Bowman, 2007), EcoMUVE (Metcalf, Kamarainen, Tutwiler,
Grozner, & Dede 2011), and other systems (Kim et al., 2009; Lane et al., 2011). The group
helps the student learn how to learn in a social world.
The following section describes two learning environments with multiple agents interact
ing with a human. Operation ARIES! incorporates trialogs that have a human agent hold
ing a conversation with a student peer agent and a tutor agent (Figure 1). AutoMentor is
a new system under development that has a computer mentor interacting with a group of
three to five students playing a simulation game on urban sciences. The use of agents in
group conversations can add considerable sophistication to the learning environment
and, in some ways, be easier to implement than one-on-one tutoring.
Conversational Trialogs with Operation ARIES!: Two Agents Are Bet

ter than One
Operation ARIES! (Halpern et al., 2012; Millis et al., 2011) teaches scientific critical
thinking through a series of game modules with two or more animated pedagogical
agents. In each of the modules, three-way conversations occur between the human, a stu
dent agent (named “Glass”), and a tutor agent (named “Dr. Quinn”). ARIES is an acronym
for Acquiring Research Investigative and Evaluative Skills; it was subsequently renamed
Operation ARA when it became commercialized by Pearson Education as a serious game.
It takes approximately 20 hours to complete ARIES and 7 hours to complete ARA.
Page 15 of 32

ing Systems
Fig. 30.1 Case study module in operation ARIES!

with three-way conversations among the human and
a tutor agent (named “Dr. Quinn,”) as well as two
new agents joining Dr. Quinn: a fellow student agent
(named “Tracy”) an alien defector (named “Broth”).
The human player and Tracy compete against each
other for the honor of going forward to the next mod
ule, where they will interrogate aliens. Broth wants
peace with Human “Bean” and is observing the ses
sions.
ARIES has three major phases: training, case study, and interrogation. The training phase
consists of an e-book, multiple-choice questions, and tutorial trialogs that teach scientific
concepts involved in psychology, biology, and chemistry. Students receive training on
twenty-one core scientific concepts (p. 501) while completing this phase and subsequent
phases in the serious game. Example core concepts are hypotheses, operational defini
tions, independent variables, dependent variables, random assignment, subject bias, and
correlation versus causation. In the case studies phase, players apply what they learned
in the training phase to realistic examples of flawed research. Specifically, they critique
research cases in the media (news reports, blogs, television) that exhibit bad science by
identifying which of the twenty-one core concepts are violated. For example, one case
study described an experiment that tested a new pill that purportedly helps people lose
weight, but with no control group (Figure 30.1). In the interrogation phase, players learn
how to ask scientists pointed questions about their research to inquire whether they are
violating the core concepts. The storyline is advanced by videos, news flashes, and e-
mails that are interspersed among the learning activities to build suspense, surprise, and
curiosity in an emerging narrative with a plot. It begins with the player joining the Feder
al Bureau of Science (FBS) as an agent-in-training and concludes with the player helping
save the world from aliens who are trying to take over the world by spreading bad sci
ence (Millis et al., 2011). Empirical studies have shown that ARIES helps students learn
research methods (Forsyth et al., 2012; Halpern et al., 2012; Millis et al., 2011).
Three types of trialogs are implemented in ARIES and launched under specific conditions.
They are (1) vicarious learning (learning by observing agents), (2) tutorial learning (learn
ing by being tutored by a tutor agent), and (3) learning through teaching (learning by
teaching a fellow student agent. The type of trialog that occurs for a particular core con
Page 16 of 32

ing Systems
cept is based on the level of knowledge exhibited by the player earlier in the game. Low
knowledge triggers vicarious learning trialogs, intermediate knowledge triggers tutorial
learning, and high knowledge triggers learning through teaching.
There is a research precedent for assigning these trialog categories to students on the ba
sis of their knowledge. Observational learning of tutorial sessions enhances learning, par
ticularly for low prior knowledge students when they watch a tutorial conversation and
comment on what is being learned rather than participating directly (Chi et al., 2008;
Craig, Chi, & VanLehn, 2009; Craig, Sullins, Witherspoon, & Gholson, 2006). In the
ARIES trialogs with vicarious learning, human students do participate but only minimally:
at one or (p. 502) two times during the interaction between the student agent and tutor
agent, the tutor turns to the human student and asks what he or she thinks by asking a
simple question (e.g., yes–no answer or a choice among three options). Drawing the hu
man in through the use of a question ensures that the human student stays connected
and engaged. Nevertheless, high-knowledge students do not benefit as much from watch
ing a tutorial dialog with minimal involvement. They need to be more active by generat
ing rather than observing information. Indeed, “playing teacher” is particularly effective
for high-knowledge students, which justifies their being assigned the learning-through-
teaching trialogs. Some learning environments contain “teachable agents” that require
the human student to teach a computerized agent. For example, Betty’s Brain (Biswas,
Leelawong, Schwartz, & Vye, 2005; Biswas et al., 2010; Schwartz et al., 2009) requires
the human student to teach the Betty agent about causal relationships in a biological sys
tem that are depicted in a conceptual graph. Betty has to get the conceptual graph cor
rect in order to pass a multiple-choice test. When she fails, the human interacts with Bet
ty to improve her conceptual graph and thereby improve her scores. Another mentor
agent guides this interaction through hints and suggestions. The learning-through-teach
trialogs are similar to these teachable agents in many ways. The human student teaches
the student agent, with the tutor agent periodically entering the conversation to guide a
productive interaction. It takes a sufficient amount of knowledge to produce the content
to teach the student agent, so only high-knowledge students receive such trialogs.
ARIES incorporates the EMT dialog mechanism just as the AutoTutor does. That is, the
goal of the trialog is to help the player articulate a specific expectation (sentence) in the
exchange, such as “A scientific hypothesis must have a prediction that can be tested.” The
accuracy of matching the students’ contributions in natural language with the expecta
tion is quite high (Cai et al., 2011), as discussed earlier. The trialog begins with a ques
tion (e.g., “What is a hypothesis?”). If the human student answers it correctly, then the
trialog quickly finishes and the student gets full credit (100 percent), indicating that this
core concept is mastered for the unit. However, if the student’s answer is off, then the tu
tor agent gives a hint, such as “What about testing a hypothesis?” If the student answers
correctly (“A prediction is tested when there is a hypothesis”) then the student gets par
tial credit (67 percent). Otherwise, the tutor agent gives a leading prompt question, such
as “What is tested when there is a scientific hypothesis?,” with the hope that the student
fills in the word “prediction” and thereby gets some credit (33 percent). If the human stu
dent is incorrect (receiving 0 percent credit), then the student agent can barge in and
Page 17 of 32

ing Systems
give the correct answer, after which the tutor agent gives positive feedback to the stu
dent agent. Instead of giving negative feedback to the incorrect human student, the tutor
gives positive feedback to the correct student agent. This promotes politeness and avoids
face-threatening negative feedback to the human. Thus, two agents are better than one.
The trialogs can be arranged to press the envelope of learning and social interaction even
further, as shown here.
• Requests for a summary. After the trialog covers the expectation, the tutor or stu
dent agent can request a summary from the human student, such as “Could you sum
marize what a hypothesis is?” This encourages the student to stay engaged and gener
ate information, in addition to providing a repetition of the learning experience. Atten
tion, generation, and repetition are strongly linked to learning.
• Requests for verification. The tutor or student agent can request the human student
to verify whether a statement is true or false, such as “Do hypotheses require a predic
tion?” This is another way to assess the student’s knowledge, keep the student en
gaged, and provide a repetition. These questions may also be sincere, in the sense that
the questioner does not know the answer. Whereas the tutor agent may understand
what a hypothesis is and believe the human agent does also, that does not mean that
the student agent understands what a hypothesis is. Again, two agents are better than
one.
• Student agent barging. At various points during the trialog, the student agent can
barge in and interrupt the thread of exchange between the human and tutor agent
with a question or other speech act. Similarly, the tutor agent can barge into the ex
change between the human and student agent. This is convenient when the main ex
change between two parties is deteriorating. The other agent can interrupt and say,
“Wait a minute, what is being tested?”
• Student agent echoing. There may be some uncertainty to whether the human
student’s contribution has a sufficiently high match to a particular expectation (or mis
conception). For example, the human student might say “A hypothesis predicts a test
outcome,” which (p. 503) does not perfectly match the expectation “A hypothesis has a
prediction that can be tested.” The computer is uncertain about what the human
means, so the student agent can echo what the computer thinks the human meant by
asking “Did you mean that a hypothesis has a prediction that can be tested?” and wait
ing for the human to answer. The tutor agent then gives feedback (positive or nega
tive) to the students if the human answers “yes” and neutral feedback (“okay”) if the
human answers “no.”
• Requests for clarification. Another approach to handling uncertainty in what the hu
man student is saying is for one of the agents to request clarification, such as “I don’t
understand,” “Could you rephrase that?,” “Could you be more precise?” Indeed, agents
are allowed to persistently say they don’t understand, just as people do: “I still don’t
understand but let’s move on.”
Page 18 of 32

ing Systems
• Assigning credit and blame. One possible scheme praises the human and blames the
student agent. The tutor can give positive feedback whenever human contributions
match expectations. When the human expresses something that matches a misconcep
tion, the student agent can partially echo that and the tutor then gives negative feed
back to the student agent rather than to the human. Consider the exchange below:
TUTOR: What is a hypothesis?

HUMAN: A claim that can’t be tested but predicted.
STUDENT AGENT: A hypothesis cannot be tested.
TUTOR: That is incorrect. A hypothesis can be tested.
HUMAN: But a hypothesis has a prediction.
TUTOR: Right. A hypothesis has a prediction.
In this fashion, the student agent gets the brunt of negative attributions, whereas the
human is blessed with positive feedback. This is also a way of decomposing complex
utterances that humans express. The student agent can echo pieces of the complex hu
man statement, and the plausibility of each piece can be evaluated by the tutor.
• Staging cognitive disequilibrium. Earlier, we discussed the impact of cognitive dise
quilibrium on confusion, thought, and learning gains. In ARIES, cognitive disequilibri
um is planted by manipulating whether or not the tutor agent and the student agent
contradict each other during the trialog or express claims that are incorrect (D’Mello
et al., in press; Lehman et al., 2011). In the case studies, the tutor agent and student
agent deliberate with the student on (a) whether there was a flaw in the study and (b)
if there is a flaw, which aspect of the study was flawed. The tutor agent expresses a
correct assertion, and the student agent agrees in the True-True control condition,
whereas the two agents agree on an incorrect assertion in the False-False condition.
The tutor expresses a correct assertion, and the student agent disagrees with an incor
rect assertion in the True-False condition, whereas the opposite is the case in the
False-True condition. As predicted, the contradictory conditions create cognitive dise
quilibrium, confusion, and often better learning than the control True-True condition.
These studies of ARIES illustrate ways in which trialogs can enhance the learning experi
ence and richness of the social interaction. They show how two agents can be better than
one. Other learning technologies have also shown some powerful advantages of multiple
agents. As discussed earlier, Betty’s Brain has a teachable agent in a learning-by-teaching
environment where (a) students teach a computer agent (Betty) to construct a causal map
through a visual representation of knowledge, and (b) a teacher agent (Mr. Davis) evalu
ates Betty’s learning performance by asking the agent to answer questions and take
quizzes. Betty’s Brain has prompts in the conversations between Betty, the teachable
agent, and Mr. Davis. The goal is to acquire a deeper mental model, better question-ask
ing skills, and enhanced self-regulated learning (Biswas et al., 2010; Leelawong & Biswas,
2008; Schwartz et al., 2009). Another system called iSTART (Interactive Strategy Training
for Active Reading and Thinking) is designed to help students learn metacomprehension
strategies that facilitate deeper comprehension in reading (McNamara et al., 2007). A
teacher agent and student peer agent interact with the human using natural language to
Page 19 of 32

ing Systems
model, provide feedback, and discuss reading strategies that improve comprehension of
difficult science texts. As in the case of ARIES, both Betty’s Brain and iSTART have shown
learning gains in a series of studies tested on thousands of students. However, this essay
focuses on language and dialogue rather than learning gains.
AutoMentor: Planting a Computer Agent in a

Multiparty Serious Game
Multiparty games and simulations have become more popular in recent years, so some re
searchers have attempted to develop serious games that build (p. 504) on these social me
dia. In some systems, a single student experiences a virtual world with several computer
agents that change over the course of several hours. Examples of serious games with a
single human interacting with multiple agents are Tactical Language and Culture Train
ing System (Johnson & Valente, 2008), Crystal Island (Rowe et al., 2010), and Coach Mike
(Lane et al., 2011). In other cases, students communicate on the web in a chat room or
some other form of computer-mediated communication as they interact in virtual worlds
with complex simulations. Notable examples of these are River City (Ketelhut et al.,
2007), EcoMUVE (Metcalf et al., 2011), and Urban Science (Shaffer, 2006).
Shaffer (2006) has argued that student learning is severely limited in simulation game en
vironments if there is no mentorship and expertise provided from professional stakehold
ers that guide students as they decide what to do and justify their decisions (Hatfield &
Schaffer, 2010; Shaffer, 2006). Shaffer’s epistemic games group has therefore collected
data with human mentors on some of his multiparty games (Urban Science, Land
Science). The game helps students understand the kinds of problems and problem solving
that socially valued professions routinely engage in. This includes how the development
of cities and suburbs are influenced by zoning, roads, parks, housing, and economic in
vestment. It includes findings in science that can be communicated in justifications of de
cisions. Land Science was designed to simulate a regional planning practicum experience
for students. During the 10-hour game, students play the role of interns at a fictitious re
gional planning firm (called Regional Design), where they make land use decisions to
meet the desires of virtual stakeholders who are represented by nonplayer characters
(NPCs). Students are split into groups and progress through a total of fifteen stages of
the game in which they complete a variety of activities including a virtual site visit of the
community of interest, where they explore the history of the site, the ecology of the area,
and the desires of different stakeholder groups. The students get feedback from the
stakeholders and use a custom designed geographic information system (iPlan) to create
a regional design plan. Throughout the game, players communicate with other members
on their planning team, as well as with a mentor (i.e., an adult who represents a profes
sional planner with the fictitious planning firm) through the use of a chat feature embed
ded in the game.
Page 20 of 32

ing Systems
We have recently collaborated with Shaffer’s group to build a computer mentor called Au
toMentor (Shaffer & Graesser, 2010). AutoMentor performs automated language and dis
course analyses on groups of 3–5 students who interact with each other through comput
er-mediated communication (chat) as they work on the simulation. AutoMentor periodi
cally makes suggestions on what the human mentor can say next to facilitate student col
laboration; the human decides whether to accept these suggestions. The eventual goal,
after multiple rounds of development and testing, is to have the AutoMentor replace hu
man mentors. It should be noted that AutoMentor is still in progress and has not been
tested empirically. Our aim here is to discuss some of the significant challenges that need
to be addressed in developing AutoMentor.
The first challenge was the handling of chat language (called chatese), which is replete
with acronyms, emoticons, telegraphic expressions, nicknames, figurative language,
spelling errors, ungrammatical expressions, and other deviations from standard English.
We were fortunate to learn that the rate of chatese is lower in group discussions than in
one-on-one chats, that conversational turns are short (about eight words), and that there
was no need to use our processing-intensive syntactic parsers on the language analysis.
Word spotting, word sequences, and statistical representations of semantics (such as la
tent semantic analysis) were adequate for performing the analyses of individual student
sentences and conversational turns. Conventional data mining procedures (Baker &
Yacef, 2009; Witten, Frank, & Hall, 2011) were adequate to classify student sentences in
to speech act categories (Moldova, Rus, & Graesser, 2011), personality of participants
(Keshtkar, Burkett, Li, & Graesser, 2012), and the epistemic categories of epistemic
games (Hatfield & Shaffer, 2010). This is an important engineering feat, but the task re
mains to align these results with theoretical claims.
The second challenge was detecting conversational patterns among students as they in
teract. We needed to have a good discovery methodology to do this. We turned to state
transition networks (STNs) to specify the conversational transitions both within and be
tween speech participants with respect to who speaks and the associated speech act cate
gories. Adjacent speech acts between speakers are known to follow conversational con
straints documented decades ago (Sacks, Schegloff, & Jefferson, 1974). Discourse acts in
educational contexts have been documented in great detail in the context of classroom
discourse (Gee, 1999; Nystrand, 2006; Sinclair & Coulthard, 1975) and human tutoring
(Cade, Copeland, Person, & D’Mello, 2008; (p. 505) Graesser, D’Mello, & Cade, 2011;
Graesser & Person, 1994; Graesser et al., 1997). For example, examples of adjacency
pairs in general conversation are (a) Question → Answer, (b) Promise → Acknowledgment,
and (c) Expressive Evaluation → Short response. In classrooms, a common sequence is
Teacher Question → Student Answer → Teacher Short Feedback, as discussed earlier. (See
also the discussion of the five-step tutoring frame.)
The sequences of speech act categories by different speakers were explored to identify
frequent conversation patterns in the multiparty interaction. More specifically, the com
puter classified sentences in the chat interactions into speech act categories and ana
lyzed sequences of the speech acts by various speakers. We identified those sequences
Page 21 of 32

ing Systems
that frequently occur over and above the base rate likelihood of speech acts and se
quences. One pattern that frequently occurred was within a mentor turn and was very
similar to human tutoring:
Basically, the mentor gives feedback, makes some justifications through statements/asser
tions, and then asks questions or makes requests to get the students to do something. An
other pattern that frequently occurred was the Mentor macromanaging the students’ re
sponses. Students ask “What do we do next?” and the tutor gives directives. We had
hoped to see other patterns reflective of students having an intellectual discussion, with
student sequences of statements, expressive evaluations, and questions, but that was not
statistically detectable. We had hoped that the mentor had an “uptake” of student activi
ties by responding to speech acts and actions initiated by students (Nystrand, 2006) and
to continue on their line of thinking, but that was also not apparent from our analyses. It
appears that considerable scaffolding is needed before intelligent mentors interact with
students in multiparty conversations. It does not come naturally.
The third challenge was to figure out when and how AutoMentor would respond when the
ideal moment arose. The “when” is nontrivial. When does the mentor intervene with a
turn in the sea of student turns in chat? Should it be during a pause of nonactivity, when
the students get into an argument, when the students are floundering, and so on? There
is no systematic research on these conversational dynamics. The “what” is also crucial
with respect to what the mentor expresses. Should students be encouraged with positive
feedback, challenged with adversarial remarks (devil’s advocate), or given mini-lectures?
Again, there is no systematic research base and theory on learning environments with re
spect to how the mentor should respond. These are questions for future research.
Conclusion
It should be apparent from this essay that SCEM mechanisms drive deep learning and
that language and discourse have a central role in these activities. The hard sciences
have declared an urgent need for citizens to pursue science, technology, engineering, and
mathematics (STEM) educational pathways. We would like to make an equal plea for
SCEM in the development of our learning environments. Without SCEM, the STEM mis
sion will die. It is time for a SCEM-STEM alliance.
Acknowledgments
The research was supported by the National Science Foundation (0325428, 633918,
0834847, 0918409, 1108845) and the Institute of Education Sciences (R305A080594,
R305G020018, R305C120001, R305A130030). Any opinions, findings, and conclusions or
Page 22 of 32

ing Systems
recommendations expressed in this material are those of the authors and do not neces
sarily reflect the views of these funding sources.
References
Azevedo, R., Moos, D., Johnson, A., & Chauncey, A. (2010). Measuring cognitive and
metacognitive regulatory processes used during hypermedia learning: Issues and chal
lenges. Educational Psychologist, 45, 210–223.
Baker, R. S. J. D., D’Mello, S. K., Rodrigo, M. T., & Graesser, A. C. (2010). Better to be
frustrated than bored: The incidence, persistence, and impact of learners’ cognitive-affec
tive states during interactions with three different computer-based learning environ
ments. International Journal of Human-Computer Studies, 68, 223–241.
Baker, R. S., & Yacef, K. (2009). The state of educational data mining in 2009: A review
and future visions. Journal of Educational Data Mining, 1, 3–17.
Barth, C. M., & Funke, J. (2010). Negative affective environments improve complex solv
ing performance. Cognition and Emotion, 24(7), 1259–1268.
Baylor, A. L., & Kim, Y. (2005). Simulating instructional roles through pedagogical agents.
International Journal of Artificial Intelligence in Education, 15, 5–115.
Biber, D., Conrad, S., & Reppen, R. (1998). Corpus linguistics: Investigating language
structure and use. Cambridge: Cambridge University Press.
Biswas, G., Jeong, H., Kinnebrew, J., Sulcer, B., & Roscoe, R. D. (2010). Measuring self-
regulated learning skills through social interactions in a teachable agent environment.
Research and Practice in Technology-Enhanced Learning, 5(2), 123–152.
Biswas, G., Leelawong, K., Schwartz, D. L., & Vye, N. (2005). Learning by teach
(p. 506)
ing: A new agent paradigm for educational software. Applied Artificial Intelligence, 19,
363–392.
Bloom, B. S. (1984). The 2 sigma problem: The search for methods of group instruction as
effective as one-to-one tutoring. Educational Researcher, 13, 4–16.
Cade, W., Copeland, J. Person, N. K., & D’Mello, S. K. (2008). Dialogue modes in expert tu
toring. In B. Woolf, E. Aimeur, R. Nkambou, & S. Lajoie (Eds.), Proceedings of the Ninth
International Conference on Intelligent Tutoring Systems (pp. 470–479). Berlin, Heidel
berg: Springer-Verlag.
Cai, Z., Graesser, A. C., Forsyth, C., Burkett, C., Millis, K., Wallace, P., et al. (2011). Tria
log in ARIES: User input assessment in an intelligent tutoring system. In W. Chen & S. Li
Page 23 of 32

ing Systems
(Eds.), Proceedings of the 3rd IEEE International Conference on Intelligent Computing
and Intelligent Systems (pp. 429–433). Guangzhou: IEEE Press.
Calvo, R. A., & D’Mello, S. K. (2010). Affect detection: An interdisciplinary review of mod
els, methods, and their applications. IEEE Transactions on Affective Computing, 1, 18–37.
Chi, M. T. H., DeLeeuw, N., Chiu, M. H., & LaVancher, C. (1994). Eliciting self-explana
tions improves understanding. Cognitive Science, 18, 439–477.
Chi, M. T. H., Roy, M., & Hausmann, R. G. M. (2008). Observing tutorial dialogues collabo
ratively: Insights about human tutoring effectiveness from vicarious learning. Cognitive
Science, 32(2), 301–341.
Chi, M. T. H., Siler, S. A., & Jeong, H. (2004). Can tutors monitor students’ understanding
accurately? Cognition and Instruction, 22, 363–387.
Chi, M. T. H., Siler, S. A., Jeong, H., & Hausmann, R. G. M. (2001). Learning from human
tutoring. Cognitive Science, 25, 471–534.
Clifford, M. M. (1991). Risk taking: Theoretical, empirical, and educational considera

tions. Educational Psychologist, 26(3/4), 263–297.
Cohen, P. A., Kulik, J. A., & Kulik, C. C. (1982). Educational outcomes of tutoring: A meta-
analysis of findings. American Educational Research Journal, 19, 237–248.
Collins, A., Brown, J. S., & Newman, S. E. (1989). Cognitive apprenticeship: Teaching the
crafts of reading, writing, and mathematics. In L. B. Resnick (Ed.) Knowing, learning, and
instruction: Essays in honor of Robert Glaser (pp. 453–494). Hillsdale, NJ: Lawrence Erl
baum.
Conati, C., & Maclaren, H. (2010). Empirically building and evaluating a probabilistic
model of user affect. User Modeling and User-Adapted Interaction, 19, 267–303.
Craig, S. D., Chi, M. T. H., & VanLehn, K. (2009). Improving classroom learning by collab
oratively observing human tutoring videos while problem solving. Journal of Educational
Psychology, 101(4), 779–789.
Craig, S. D., Sullins, J., Witherspoon, A., & Gholson, B. (2006). Deep-level reasoning ques
tions effect: The role of dialog and deep-level reasoning questions during vicarious learn
ing. Cognition and Instruction, 24(4), 563–589.
D’Andrade, R. G., & Wish, M. (1985). Speech act theory in quantitative research on inter
personal behavior. Discourse Processes, 8(2), 229–258.
D’Mello, S. K., Craig, S. D., Fike, K., & Graesser, A. C. (2009). Responding to learners’
cognitive-affective states with supportive and shakeup dialogues. In J. Jacko (Ed.), Hu
Page 24 of 32

ing Systems
man-computer interaction: Ambient, ubiquitous and intelligent interaction (pp. 595–604).
Berlin/Heidelberg: Springer.
D’Mello, S. K., Craig, S. D., & Graesser, A. C. (2009). Multi-method assessment of affec
tive experience and expression during deep learning. International Journal of Learning
Technology, 4(3/4), 165–187.
D’Mello, S. K., Craig, S. D., Witherspoon, A., McDaniel, B. T., & Graesser, A. C. (2008). Au
tomatic detection of learner’s affect from conversational cues. User Modeling and User-
Adapted Interaction, 18(1–2), 45–80.
D’Mello, S. K., & Graesser, A. C. (2010). Multimodal semi-automated affect detection from
conversational cues, gross body language, and facial features. User Modeling and User-
adapted Interaction, 20, 147–187.
D’Mello, S. K., & Graesser, A. C. (2012-a). AutoTutor and affective AutoTutor: Learning by
talking with cognitively and emotionally intelligent computers that talk back. ACM Trans
actions on Interactive Intelligent Systems, 2, 23:2–23:39.
D’Mello, S. K., & Graesser, A. C. (2012-b). Dynamics of affective states during complex
learning, Learning and Instruction, 22, 145–157.
D’Mello, S. K., & Graesser, A. C. (2012-c). Language and discourse are powerful signals of
student emotions during tutoring. IEEE Transactions on Learning Technologies, 5,
304-317.
D’Mello, S. K., Lehman, S., Pekrun, R., & Graesser, A. C. (in press). Confusion can be ben
eficial for learning. Learning and Instruction.
Daniels, L. M., Pekrun, R., Stupnisky, R. H., Haynes, T. L., Perry, R. P., & Newall, N. E.
(2009). A longitudinal analysis of achievement goals: From affective antecedents to emo
tional effects and achievement outcomes. Journal of Educational Psychology, 101, 948–
963.
Deci, E. L., & Ryan, R. (2002). The paradox of achievement: The harder you push, the
worse it gets. In J. Aronson (Ed.), Improving academic achievement: Impact of psychologi
cal factors on education (pp. 61–87). Orlando, FL: Academic Press.
Du Boulay, B., Avramides, K., Luckin, R., Martinez-Miron, E., Mendez, G. R., & Carr, A.
(2011). Towards systems that care: A conceptual framework based on motivation,
metacognition and affect. International Journal of Artificial Intelligence in Education,
20(3), 197–229.
Dweck, C. S. (2002). The development of ability conceptions. In A. Wigfield & J. S. Eccles

(Eds.), Development of achievement motivation: A volume in the educational psychology
series (pp. 57–88). San Diego, CA: Academic Press.
Ekman, P. (1992). An argument for basic emotions. Cognition & Emotion, 6, 169–200.
Page 25 of 32

ing Systems
Evens, M. W., & Michael, J. A. (2005). One-on-one tutoring by humans and computers.
New York: Rutledge/Taylor & Francis.
Forsyth, C., Pavlik, P., Graesser, A. C., Cai, Z., Germany, M., Millis, K., et al. (2012). Learn
ing gains for core concepts in a serious game on scientific reasoning. In K. Yacef, O. Za
ïane, H. Hershkovitz, M. Yudelson, & J. Stamper (Eds.), Proceedings of the 5th Interna
tional Conference on Educational Data Mining (pp. 172–175). Chania, Greece: Interna
tional Educational Data Mining Society.
Fox, B. A. (1993). Human tutorial dialogue. New Jersey: Lawrence Erlbaum.
Frenzel, A. C., Pekrun, R., & Goetz, T. (2007). Perceived learning environment and
(p. 507)
students’ emotional experiences: A multilevel analysis of mathematics classrooms. Learn

ing and Instruction, 17, 478–493.
Gee, J. P. (1999). An introduction to discourse analysis theory and method (2nd ed.). New
York: Routledge.
Graesser, A. C., Conley, M. W., & Olney, A. M. (2012). Intelligent tutoring systems. In S.
Graham & K. Harris (Eds.), APA educational psychology handbook: Vol. 3. Applications to
Learning and Teaching (pp. 451–473). Washington, DC: American Psychological Associa
tion.
Graesser, A. C., & D’Mello, S. K. (2012). Emotions during the learning of difficult materi
al. In. B. Ross (Ed.), The psychology of learning and motivation, vol. 57 (pp. 183–225).
New York: Elsevier.
Graesser, A. C., D’Mello, S. K., & Cade, W. (2011). Instruction based on tutoring. In R. E.
Mayer & P. A. Alexander (Eds.), Handbook of research on learning and instruction (pp.
408–426). New York: Routledge.
Graesser, A. C., D’Mello, S. K., Hu, X., Cai, Z., Olney, A. M., & Morgan, B. (2012). AutoTu
tor. In P. M. McCarthy & C. Boonthum (Eds.), Applied natural language processing and
content analysis: Identification, investigation and resolution (pp. 169–187). Hershey, PA:
IGI Global.
Graesser, A. C., D’Mello, S. K., & Person, N. K. (2009). Metaknowledge in tutoring. In D. J.

Hacker, J. Donlosky, & A. C. Graesser (Eds.), Handbook of metacognition in education (pp.
361–382). New York: Taylor & Francis.
Graesser, A. C., Lu, S., Jackson, G. T., Mitchell, H., Ventura, M., Olney, A. M., & Louwerse,
M. M. (2004). AutoTutor: A tutor with dialogue in natural language. Behavior Research
Methods, Instruments, and Computers, 36, 180–193.
Graesser, A. C., Lu, S., Olde, B. A., Cooper-Pye, E., & Whitten, S. (2005). Question asking
and eye tracking during cognitive disequilibrium: Comprehending illustrated texts on de
vices when the devices break down. Memory and Cognition, 33, 1235–1247.
Page 26 of 32

ing Systems
Graesser, A. C., & McNamara, D. S. (2012). Automated analysis of essays and open-ended
verbal responses. In H. Cooper, P. M. Camic, D. L. Long, A. T. Panter, D. Rindskopf, & K. J.
Sher (Eds.), APA handbook of research methods in psychology, vol. 1: Foundations, plan
ning, measures, and psychometrics (pp. 307–325). Washington, DC: American Psychologi
cal Association.
Graesser, A. C., Penumatsa, P., Ventura, M., Cai, Z., & Hu, X. (2007). Using LSA in AutoTu
tor: Learning through mixed-initiative dialogue in natural language. In T. K. Landauer, D.
S. McNamara, S. Dennis, & W. Kintsch (Eds.), Handbook of latent semantic analysis (pp.
243–262). Mahwah, NJ: Lawrence Erlbaum.
Graesser, A. C., & Person, N. K. (1994). Question asking during tutoring. American Educa
tional Research Journal, 31(1), 104–137.
Graesser, A. C., Person, N. K., & Magliano, J. P. (1995). Collaborative dialogue patterns in
naturalistic one-to-one tutoring. Applied Cognitive Psychology, 9(6), 495–522.
Graesser, A. C., Swamer, S. S., & Hu, X. (1997). Quantitative discourse psychology. Dis
course Processes, 23(3), 229–263.
Halpern, D. F., Millis, K., Graesser, A. C., Butler, H., Forsyth, C., & Cai, Z. (2012). Opera
tion ARA: A computerized learning game that teaches critical thinking and scientific rea
soning. Thinking Skills and Creativity, 7, 93–100.
Hatfield, D., & Shaffer, D. W. (2010). The epistemography of journalism 335: Complexity
in developing journalistic expertise. Paper presented at the International Conference of
the Learning Sciences (ICLS), Chicago, Illinois.
Holtgraves, T. (2002). Language as social action: Social psychology and language use.
Mahwah, NJ: Erlbaum.
Jackson, G. T., & Graesser, A. C. (2006). Applications of human tutorial dialog in AutoTu
tor: An intelligent tutoring system. Revista Signos, 39, 31–48.
Johnson, W. L., & Valente, A. (2008). Tactical language and culture training systems: us
ing artificial intelligence to teach foreign languages and cultures. In M. H. Goker (Ed.),
Proceedings of the 20th national conference on Innovative applications of artificial intelli
gence: Vol. 3 (pp. 1632–1639). Menlo Park, CA: AAAI Press.
Jurafsky, D., & Martin, J. H. (2008). Speech and language processing: An introduction to
natural language processing, computational linguistics, and speech recognition. Upper
Saddle River, NJ: Prentice Hall.
Keshtkar, F., Burkett, C., Li, H., & Graesser, A. C. (2012). Detecting players personality
behavior with any effort of concealment. In A. Gelbukh (Ed.), Proceedings of 13th Interna
tional Conference on Intelligent Text Processing and Computational Linguistics (pp. 503–
515). Heidelberg, Germany: Springer.
Page 27 of 32

ing Systems
Ketelhut, D., Dede, C. J., Clarke, J., Nelson, B., & Bowman, C. (2007). Studying situated
learning in a multi-user virtual environment. In E. Baker, J. Dickieson, W. Wulfeck, & H.
O’Neil. (Eds.), Assessment of problem solving using simulations (pp. 37–58). Mahwah: Erl
baum.
Kim, J., Hill, R. W., Durlach, P. J., Lane, H. C., Forbell, E., Core, M., et al. (2009). BiLAT: A
game-based environment for practicing negotiation in a cultural context. International
Journal of Artificial Intelligence in Education, 19, 289–308.
Landauer, T. K., McNamara, D. S., Dennis, S., & Kintsch, W. (Eds.). (2007). Handbook of
latent semantic analysis. Mahwah, NJ: Lawrence Erlbaum
Lane, H. C., Noren, D., Auerbach, D., Birch, M., & Swartout, W. (2011). Intelligent tutor
ing goes to the museum in the big city: A pedagogical agent for informal science educa
tion. In G. Biswas, S. Bull, J. Kay & A. Mitrovic (Eds.), International journal of artificial in
telligence in education (pp. 155–162). Heidelberg: Springer.
Leelawong, K., & Biswas, G. (2008). Designing learning by teaching systems: The Betty’s
Brain system. International Journal of Artificial Intelligence in Education, 18(3), 181–208.
Lehman, B., D’Mello, S. K., Cade, W., & Person, N. K. (2012). How do they do it? Investi
gating dialogue moves within dialogue modes in expert human tutoring. In S. Cerri, W.
Clancey, G. Papadourakis, & K. Panourgia (Eds.), Proceedings of the 11th International
Conference on Intelligent Tutoring Systems (pp. 557–562). Berlin Heidelberg: Springer-
Verlag.
Lehman, B., D’Mello, S. K., Chauncey, A., Gross, M., Dobbins, A., Wallace, P., et al. (2011).
Inducing and tracking confusion with contradictions during critical thinking and scientific
reasoning. In S. Bull, G. Biswas, J. Kay, & T. Mitrovic (Eds.), Proceedings of the 15th Inter
national Conference on Artificial Intelligence in Education (pp. 171–178). Berlin, Heidel
berg: Springer.
Lepper, M. R., Drake, M., & O’Donnell-Johnson, T. M. (1997). Scaffolding techniques of

expert human tutors. In K. Hogan & M. Pressley (Eds.), Scaffolding student learning: In
structional approaches and issues (pp. 108–144). New York: Brookline Books.
Lepper, M. R., & Woolverton, M. (2002). The wisdom of practice: Lessons learned
(p. 508)
from the study of highly effective tutors. In J. Aronson (Ed.), Improving academic achieve
ment: Impact of psychological factors on education (pp. 135–158). Orlando, FL: Academic
Press.
Linnenbrink, E. (2007). The role of affect in student learning: A multi-dimensional ap

proach to considering the interaction of affect, motivation and engagement. In P. Schutz
& R. Pekrun (Eds.), Emotions in education (pp. 107–124). San Diego, CA: Academic Press.
Page 28 of 32

ing Systems
Litman, D. J., & Forbes-Riley, K. (2006). Recognizing student emotions and attitudes on
the basis of utterances in spoken tutoring dialogues with both human and computer tu
tors. Speech Communication, 48, 559–590.
Mathes, P. G., & Fuchs, L. S. (1994). Peer tutoring in reading for students with mild dis
abilities: A best evidence synthesis. School Psychology Review, 23, 59–80.
McArthur, D., Stasz, C., & Zmuidzinas, M. (1990). Tutoring techniques in algebra. Cogni
tion and Instruction, 7, 197–244.
McCarthy, P. M., & Boonthum-Denecke, C. (Eds.). (2012). Applied natural language pro
cessing: identification, investigation, and resolution. Hershey, PA: IGI Global.
McNamara, D. S., Boonthum, C., Levinstein, I. B., & Millis, K. (2007). Evaluating self-ex
planations in iSTART: comparing word-based and LSA algorithms. In T. Landauer, D. S.
McNamara, S. Dennis, & W. Kintsch (Eds.), Handbook of latent semantic analysis (pp.
227–241). Mahwah, NJ: Erlbaum.
McNamara, D. S., Graesser, A. C., McCarthy, P. M., Cai, Z. (in press). Automated evalua
tion of text and discourse with Coh-Metrix. Cambridge, MA: Cambridge University Press.
McQuiggan, S. W., Robison, J. L., & Lester, J. C. (2010). Affective transitions in narrative-
centered learning environments. Educational Technology & Society 13, 40–53.
Metcalf, S. J., Kamarainen, A., Tutwiler M. S., Grotzer, T. A., & Dede, C. J. (2011). Ecosys
tem science learning via multi-user virtual environments. International Journal of Gaming
and Computer-Mediated Simulations, 3, 86–90.
Meyer, D. K., & Turner, J. C. (2006). Re-conceptualizing emotion and motivation to learn
in classroom contexts. Educational Psychology Review, 18(4), 377–390.
Millis, K., Forsyth, C., Butler, H., Wallace, P., Graesser, A. C., & Halpern, D. F. (2011). Op
eration ARIES! A serious game for teaching scientific inquiry. In M. Ma, A. Oikonomou, &
J. Lakhmi (Eds.), Serious games and edutainment applications (pp. 169–196). London, UK:
Springer-Verlag.
Miyake, N., & Norman, D. A. (1979). To ask a question one must know enough to know
what is not known. Journal of Verbal Learning and Verbal Behavior, 18, 357–364.
Moldovan, C., Rus, V., & Graesser, A. C. (2011). Automated speech act classification for
online chat. MAICS, 23–29.
Nystrand, M. (2006). Research on the role of classroom discourse as it affects reading

comprehension. Research in the Teaching of English, 40, 392–412.
Ogan, A., Finkelstein, S., Walker, E., Carlson, R., & Cassell, J. (2012). Rudeness and rap
port: Insults and learning gains in peer tutoring. In. S. A. Cerri & B. Clancey (Eds.), Pro
Page 29 of 32

ing Systems
ceedings of the 11th International Conference on Intelligent Tutoring Systems. Berlin:
Springer-Verlag.
Olney, A. M., D’Mello, S. K., Person, N. K., Cade, W., Hayes, P., Williams, C., et al. (2012).
Guru: A computer tutor that models expert human tutors. In S. A. Cerri & B. Clancey
(Eds.), Proceedings of the 11th International Conference on Intelligent Tutoring Systems
(pp. 256–261). Berlin: Springer-Verlag.
Olney, A. M., Louwerse, M. M., Mathews, E., Marineau, J., Hite-Mitchell, H., & Graesser,
A. C. (2003). Utterance classification in AutoTutor. In J. Burstein & C. Leacock (Eds.),
Building educational applications using natural language processing: Proceedings of the
Human Language Technology—North American Chapter of the Association for Computa
tional Linguistics Conference 2003 Workshop (pp. 1–8). Philadelphia: Association for
Computational Linguistics.
Palincsar, A., & Brown, A. (1984). Reciprocal teaching of comprehension-fostering and

comprehension-monitoring activities. Cognition and Instruction, 1, 117–175.
Pekrun, R., Elliot, A., & Maier, M. (2006). Achievement goals and discrete achievement
emotions: A theoretical model and prospective test. Journal of Educational Psychology,
98, 583–597.
Pennebaker, J. W., Booth, R. J., & Francis, M. E. (2007). Linguistic inquiry and word count.
Austin, TX: LIWC.net (www.liwc.net).
Person, N. K., Burke, D. R., & Graesser, A. C. (2003, June). RudeTutor: A face threatening
agent. Presented at the 13th Annual Meeting of the Society for Text and Discourse.
Madrid, Spain.
Person, N. K., Kreuz, R. J., Zwaan, R. A., & Graesser, A. C. (1995). Pragmatics and peda
gogy: Conversational rules and politeness strategies may inhibit effective tutoring. Cogni
tion and Instruction, 13(2), 161–188.
Picard, R. (2010). Affective computing: From laughter to IEEE. IEEE Transactions on Af
fective Computing 1, 11–17.
Rickel, J., Lesh, N., Rich, C., Sidner, C. L., & Gertner, A. S. (2002). Collaborative discourse
theory as a foundation for tutorial dialogue. In S. A. Cerri, G. Gouarderes, & F. Paraguacu
(Eds.), Proceedings of the Sixth International Conference on Intelligent Tutoring Systems
(pp. 542–551). Berlin: Springer-Verlag.
Rogoff, B., & Gardner, W. P. (1984). Adult guidance of cognitive development. In B. Rogoff
and J. Lave (Eds.), Everyday cognition: Its development in social contexts. (pp. 95–116).
Cambridge: Harvard University Press.
Rowe, J., Shores, L. R., Mott, B., & Lester, J. C. (2010). Integrating learning, problem solv
ing, and engagement in narrative-centered learning environments. International Journal
of Artificial Intelligence in Education, 20, 166–177.
Page 30 of 32

ing Systems
Rus, V. (2010). DeepTutor: An intelligent tutoring system that promotes deep learning of
science topics. Retrieved from University of Memphis, DeepTutor website: http://
deeptutor.memphis.edu/DeepTutor.Overview.V-RusEtAl.May-2012.pdf.
Rus, V., & Graesser, A. C. (2006). Deeper natural language processing for evaluating stu
dent answers in intelligent tutoring systems. In Y. Gil & R. Mooney (eds), Proceedings of
the American Association of Artificial Intelligence (pp.1495–1500). Menlo Park, CA: AAAI.
Sacks, H., Schegloff, E. A., & Jefferson, G. (1974). A simplest systematics for the organiza
tion of turn-taking for conversation. Language, 50, 696–735.
Schober, M. F., & Brennan, S. E. (2003). Processes of interactive spoken discourse. In A.

C. Graesser, M. A. Gernsbacher, & S. Goldman (Eds.), Handbook of discourse processes
(pp 123–164). Mahwah, NJ: Erlbaum.
Schutz, P., & Pekrun, R. (Eds.). (2007). Emotion in education. San Diego, CA: Academic
Press.
Schwartz, D. L., Chase, C., Chin, D., Oppezzo, M., Kwong, H., Okita, S., Biswas, G.,
Roscoe, R. D., Jeong, H., & Wagster, J. D. (2009). Interactive metacognition: Monitoring
and regulating a teachable agent. In D. J. Hacker, J. Dunlosky, & A. C. Graesser (Eds.),
Handbook of metacognition in education (pp. 340–358). London: Routledge Press.
Shaffer, D. W. (2006). How computer games help children learn. New York: Pal
(p. 509)
grave Macmillan.
Shaffer, D. W., & Graesser, A. C. (2010). Using a quantitative model of participation in a

community of practice to direct automated mentoring in an ill-defined domain. Workshop
proceedings at the Intelligent Tutoring Systems Conference on Intelligent Tutoring Tech
nologies for Ill-defined Problems and Ill-defined Domains. Pittsburgh, PA: ITS.
Shah, F., Evens, M. W., Michael, J. A., & Rovick, A. A. (2002). Classifying student initia
tives and tutor responses in human keyboard to keyboard tutoring sessions. Discourse
Processes, 33(1), 23–52.
Sinclair, J., & Coulthard, M. (1975). Towards an analysis of discourse. Oxford: Oxford Uni
versity Press.
Sleeman D., & Brown, J. S. (Eds.). (1982). Intelligent tutoring systems. Orlando, Florida:
Academic Press.
Topping, K. (1996). The effectiveness of peer tutoring in further and higher education: A
typology and review of the literature. Higher Education, 32, 321–345.
Van der Meij, H. (1994). Student questioning: A componential analysis. Learning and Indi
vidual Differences, 6, 137–161.
Page 31 of 32

ing Systems
VanLehn, K. (2011). The relative effectiveness of human tutoring, intelligent tutoring sys
tems and other tutoring systems. Educational Psychologist, 46(4), 197–221.
VanLehn, K., Graesser, A. C., Jackson, G. T., Jordan, P., Olney, A. M., & Rose, C. P. (2007).
When are tutorial dialogues more effective than reading? Cognitive Science, 31, 3–62.
Wang, N., Johnson, W. L., Mayer, R. E., Rizzo, P., Shaw, E., & Collins, H. (2008). The polite
ness effect: Pedagogical agents and learning outcomes. International Journal of Human-
Computer Studies, 66(2), 98–112.
Witten, I. H., Frank, E., & Hall, M. A. (2011). Data mining: Practical machine learning
tools and techniques (3rd ed.). Burlington, MA: Morgan Kaufmann.
Woolf, B., Burleson, W., Arroyo, I., Dragon, T., Cooper, D., & Picard, R. (2009). Affect-
aware tutors: Recognizing and responding to student affect. International Journal of
Learning Technology, 4, 129–164. (p. 510)
Arthur C. Graesser
Arthur C. Graesser, Department of Psychology, Psychology Department, University of

Memphis, Memphis, TN
Fazel Keshtkar
Fazel Keshtkar is an Assistant Professor in the Department of Computer Science at

the Missouri State University, Southeast.
Haiying Li
Haiying Li, University of Memphis
Page 32 of 32

Thomas Holtgraves - The Oxford Handbook of Language and Social Psychology-Oxford University Press (2014)

Uploaded by

Copyright:

Available Formats

Thomas Holtgraves - The Oxford Handbook of Language and Social Psychology-Oxford University Press (2014)

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Thomas Holtgraves - The Oxford Handbook of Language and Social Psychology-Oxford University Press (2014)

Uploaded by

Copyright:

Available Formats

Language and Social Psychology: Introduction and Overview

Language and Social Psychology: Introduction and

Print Publication Date: Sep 2014 Subject: Psychology, Social Psychology

Abstract and Keywords

Language pervades everything we do as social beings. When we interact with others, it is

Subscriber: UC - Berkeley Library; date: 31 December 2021

Subscriber: UC - Berkeley Library; date: 31 December 2021

Social Dimensions of Language Variation: So­

Subscriber: UC - Berkeley Library; date: 31 December 2021

Social Processes Involved in Language-Based

One important aspect of coordination is perspective taking, an attempt to consider

Much conversational coordination is accomplished nonverbally, by pointing, nodding, and

Subscriber: UC - Berkeley Library; date: 31 December 2021

Linguistic Underpinnings of Interpersonal

Subscriber: UC - Berkeley Library; date: 31 December 2021

Linguistic Underpinnings of Intrapersonal

Subscriber: UC - Berkeley Library; date: 31 December 2021

Language and Meaning

Subscriber: UC - Berkeley Library; date: 31 December 2021

Linguistic meaning is particularly challenging, of course, because of the possibility that

Although electrophysiological techniques have been used in psycholinguistic research for

Jean-François Bonnefon explores the meaning of reasoning terms, specifically quantifiers

Subscriber: UC - Berkeley Library; date: 31 December 2021

Subscriber: UC - Berkeley Library; date: 31 December 2021

The development of new electronic communications technologies such as texting, Twitter,

Clark, H. H. (1996). Using language. Cambridge, UK: Cambridge University Press.

Clark, H. H., & Brennan, S. E. (1991). Grounding in communication. In L. B. Resnick, J. M.

Grice, H. P. (1957). Meaning. Philosophical Review, 67, 377–388.

Subscriber: UC - Berkeley Library; date: 31 December 2021

Schegloff, E. A. (2007). Sequence organization in interaction: A primer in conversation

Searle, J. R. (1969). Speech acts. Cambridge, UK: Cambridge University Press.

Semin, G. R. (2000). Agenda 2000—Communication: language as an implementational de­

Thomas M. Holtgraves is a professor of Psychological Science at Ball State Universi­

Subscriber: UC - Berkeley Library; date: 31 December 2021

Language Attitudes: Social Determinants and Conse­

Print Publication Date: Sep 2014 Subject: Psychology, Social Psychology

Abstract and Keywords

Subscriber: UC - Berkeley Library; date: 31 December 2021

Socially Diagnostic Language Cues for Social

Subscriber: UC - Berkeley Library; date: 31 December 2021

The Power of Accents

Currently, we are witnessing a renewed interest in the methodology for assessing

Subscriber: UC - Berkeley Library; date: 31 December 2021

Subscriber: UC - Berkeley Library; date: 31 December 2021

Issues of Standardization and Nonstandardiza­

Subscriber: UC - Berkeley Library; date: 31 December 2021

Generally, it can be concluded that speakers of nonstandard language varieties—and es­

Subscriber: UC - Berkeley Library; date: 31 December 2021

Language Attitudes and Comprehensibility

Subscriber: UC - Berkeley Library; date: 31 December 2021

The Value of Nonstandard Accents

Subscriber: UC - Berkeley Library; date: 31 December 2021

Language Attitudes: Self-Categorization and

Subscriber: UC - Berkeley Library; date: 31 December 2021

Subscriber: UC - Berkeley Library; date: 31 December 2021

Language Attitudes: Beyond the Accent

Subscriber: UC - Berkeley Library; date: 31 December 2021

Subscriber: UC - Berkeley Library; date: 31 December 2021

Relatedly, using a second language can be affected by negative stereotypes (Paladino et

Subscriber: UC - Berkeley Library; date: 31 December 2021

Social Dimensions of Language Variation: So

Semin, G. R. (2000). Agenda 2000—Communication: language as an implementational de

Thomas M. Holtgraves is a professor of Psychological Science at Ball State Universi

Language Attitudes: Social Determinants and Conse

Issues of Standardization and Nonstandardiza

Generally, it can be concluded that speakers of nonstandard language varieties—and es

Toward an Interaction-Based Model of Lan

much more than [about] anonymous reports on questionnaires, produced on de

Ball, P. (1983). Stereotypes of Anglo-Saxon and Non-Anglo-Saxon accents: Some explana

Lindemann, S. (2002). Listening with an attitude: A model of native-speaker comprehen

Rubin, D. L. (1992). Nonlanguage factors affecting undergraduates’’ judgments of nonna

Howard Giles is Professor at the Department of Communication, University of Cali

Keywords: Communication Accommodation, Face Negotiation, Identity Negotiation, Intercultural-Intergroup Com

cation effectiveness. Likewise, when one promotes an effective mutual-interest goal-di

“Constructive communication skills” refer to our operational abilities to manage a com