The Language of Stories

How do we read stories? How do they engage our minds and create meaning?
Are they a mental construct, a linguistic one, or a cultural one? What is the
difference between real stories and fictional ones? This book addresses such
questions by describing the conceptual and linguistic underpinnings of narra-
tive interpretation. Barbara Dancygier discusses literary texts as linguistic
artifacts, describing the processes which drive the emergence of literary mean-
ing. If a text means something to someone, she argues, there have to be
linguistic phenomena that make it possible. Drawing on blending theory and
construction grammar, the book focuses its linguistic lens on the concepts of
the narrator and the story, and defines narrative viewpoint in a new way. Textual
use of pronouns and other referential expressions also comes under investiga-
tion, including their occurrence in constructions of speech and thought repre-
sentation. The examples come from a wide spectrum of texts, primarily novels
and drama, by authors such as William Shakespeare, Margaret Atwood, Philip
Roth, Dave Eggers, Jan Potocki, and Mikhail Bulgakov.

Barbara Dancygier is Professor in the Department of English at the

University of British Columbia, Canada. Her publications include
Conditionals and Prediction (Cambridge, 1998) and Mental Spaces in
Grammar (with Eve Sweetser, Cambridge, 2005).
The Language of Stories
A Cognitive Approach

Barbara Dancygier
cambridge university press
Cambridge, New York, Melbourne, Madrid, Cape Town,
Singapore, São Paulo, Delhi, Tokyo, Mexico City
Cambridge University Press
The Edinburgh Building, Cambridge CB2 8RU, UK

Published in the United States of America by Cambridge University Press, New York

www.cambridge.org
Information on this title: www.cambridge.org/9781107005822

© Barbara Dancygier 2012

This publication is in copyright. Subject to statutory exception

and to the provisions of relevant collective licensing agreements,
no reproduction of any part may take place without the written
permission of Cambridge University Press.

First published 2012

Printed in the United Kingdom at the University Press, Cambridge

A catalog record for this publication is available from the British Library

Library of Congress Cataloging in Publication data

Dancygier, Barbara.
The language of stories : a cognitive approach / Barbara Dancygier.
p. cm.
Includes bibliographical references and index.
ISBN 978-1-107-00582-2
1. Language and languages – Style – Psychological aspects. 2. Cognitive
grammar. 3. Discourse analysis, Literary. I. Title.
P301.5.P75D36 2011
8080 .036–dc22
2011015550

ISBN 978-1-107-00582-2 Hardback

Cambridge University Press has no responsibility for the persistence or

accuracy of URLs for external or third-party internet websites referred to
in this publication, and does not guarantee that any content on such
websites is, or will remain, accurate or appropriate.
for Jacek and Szymek
 Contents

List of figures page ix

Acknowledgements x

Introduction 1
1 Language and literary narratives 4
1.1 Where does narrative meaning come from? 6
1.2 Literary analysis and linguistic analysis 8
1.3 Literature, language, and human nature 11
1.4 Literary texts and communication 16
1.5 Why is fiction special? 21
1.6 Narrative and grounding 23
1.7 Approaching narratives 29

2 Blending, narrative spaces, and the emergent story 31

2.1 Applying blending to fictional narratives 32
2.2 Narrative spaces as mental spaces 35
2.3 Narrative spaces – an example 40
2.4 Emergent story 53

3 Stories and their tellers 58

3.1 Narrators, narrative spaces, and viewpoint 60
3.2 Types of teller and epistemic viewpoint 64
3.3 Second-person narratives 76
3.4 The teller, the author, and the character 79
3.5 Multiple tellers 80
3.6 Narrative space embedding 84
3.7 Narrative viewpoint and narrative spaces 86

4 Viewpoint: representation and compression 87

4.1 Viewpoint and representation 91
4.2 Viewpoint compression 96
4.3 Decompression for viewpoint 100
4.4 Fictive vision, causation, and change 102
4.5 The micro level, the macro level, and viewpoint compression 106
4.6 Speech, thought, and multiple levels of representation 108
4.7 Narrative thought and intersubjectivity 112
viii Contents

5 Referential expressions and narrative spaces 117

5.1 Compression, decompression, and cross-space mappings 118
5.2 Proper names, frame metonymy, and the status of a character 119
5.3 Role-value mappings as cross-space connectors 121
5.4 Common nouns 128
5.5 Personal pronouns, viewpoint, and the narrator 129
5.6 Deictic I and the construal of subjectivity 136

6 Fictional minds and embodiment in drama and fiction 139

6.1 Deictic ground in literary discourse 139
6.2 Mental spaces, physical spaces, and dramatic narratives 141
6.3 Materiality of the stage and fictional minds 146
6.4 From dramatic narratives to novelistic narratives 164
6.5 Fictional minds, bodies, and brains 168

7 Speech and thought in the narrative 171

7.1 Types of discourse spaces in the narrative 172
7.2 Speaking for thinking 174
7.3 Levels of embedding in thought representation 178
7.4 Viewpoint compression and constructional compositionality 183

8 Stories in the mind 195

8.1 The linguistics of literature 195
8.2 The storyworld reality 200
8.3 Blending and narrative analysis 201
8.4 A bridge to the truth 203

Notes to the text 205

References 210
Literary works cited 223
Index 225
 Figures

2.1 Narrative spaces (example 2) page 39

2.2 Narrative spaces in The Blind Assassin 41
2.3 Referential connections across spaces in The Blind Assassin 52
3.1 Intrusive narrator (example 2) 65
3.2 Ego-viewpoints 68
3.3 Present time and present tense (example 7) 71
3.4 Ego-viewpoint in the present tense (example 8) 72
3.5 Second-person narrative (examples 11 and 12) 77
3.6 Second-person narrative (examples 13–15) 78
4.1 Reported voices and viewpoint (example 1) 91
4.2 Representation and compression 1 (example 11) 98
4.3 Representation and compression 2 (example 11) 99
5.1 Narrative spaces and roles in American Pastoral 127
6.1 The mental space schema of a theatrical performance 145
6.2 Acting/Speaking blend 156
6.3 Feeling/Speaking blend 157
7.1 Discourse blend (examples 4 and 5) 176
7.2 The ‘crow’ blend (examples 9, 10, and 11) 180
7.3 Third-person report (example 15) 184
7.4 First-person report (example 18) 188

ix
 Acknowledgements

A project like this one requires that the author finds friends in unexpected
places, and I was lucky to meet many, and usually when I needed them most.
I was writing this book feeling cautious about what various groups of people
might say about it, and they invariably gave me their support and encourage-
ment. I thus want to thank all of those who could have dampened my enthusi-
asm and didn’t, and who gave me good advice instead. The list is too long to
quote here.
I owe special thanks to colleagues who read major fragments of earlier drafts
and helped with very specific responses: Uri Margolin, Eve Sweetser, Sean
McAlister, Patsy Badir, Lieven Vandelanotte, Margery Fee, Suzanne Fagel,
Maarten van Leeuwen, and my Cambridge University Press reviewers. Their
careful consideration of my ideas and structural choices has led me through
many hours of doubt.
I owe much to the scholarship and inspiration of Mark Turner, Gilles
Fauconnier, and Eve Sweetser. I also want to thank David Herman, who
encouraged me and trusted in the final outcome before I had the courage to
call this project a book, and Alan Palmer, who listened and shared thoughts
when I needed it badly. I am grateful to Elena Semino and Paul Simpson for
their sustained support of my various projects and their interest in my work. One
of the chapters of this book could never have come into existence if not for the
inspiration and friendship of Michael Booth and Amy Cook. I have also been
lucky to know Lieven Vandelanotte, whose work, guidance, and support were
with me all along.
I have been given much support by the literature and linguistics colleagues
in my own department – the Department of English at the University of British
Columbia. This book was conceived against the background of their vast
knowledge and wonderful open-mindedness, and I am very grateful to them
all for their encouragement, friendship, and sense of humor.
My graduate students have been my day-to-day audience, and have sat through
many hours of workshops in which we discussed their ideas and mine. They have
asked excellent questions, and have helped me focus when I was slipping. I can
only mention some of them here: Jennifer Schnepf, Sean McAlister, Mike
x
 Acknowledgements xi

Borkent, Slade Stolar, Tiffany Johnstone, Janet Lermitte, Robin Steen, Eve
Preus, and Mark Deggan.
Last, but not least, my family has helped me in more ways than I can count.
My husband, Jacek, supported me with his impeccable literary taste and his
boundless computer expertise – both were indispensable in the completion of
this book. I also want to thank him for the motto to Chapter 8. My son, Szymek,
infected me with his admiration for Melville and Poe, and also his passion for
traditional Japanese theatre. Many ideas in this book owe much to his careful
reading and willingness to share thoughts.
This project has been supported financially with a grant from the Social
Sciences and Humanities Research Council of Canada, number 410-2005-
1644. I am grateful to SSHRC officers for making the funding available and
providing prompt answers to all my questions.
Last but not least, I want to thank the editorial staff at Cambridge University
Press, and especially my editor, Andrew Winnard. His professional insight
helped me see the big picture and guided me through various rapids.
Finally, I owe much to the writers of all the brilliant stories I got a chance
to read again and again. Thanks to them, writing this book was an exciting
experience. It would not be fair to say that all these great novels were just
excellent materials to work on, as on occasion I had to leave my analyst persona
aside, and let the text speak to me. The stories just wouldn’t let me do anything
but read. I am truly grateful for these moments.
I have no doubt that the book has many faults, which my critics will be quick
to point out. But I am sure I would have made many more mistakes if not for all
the people who were willing to listen, read, and talk. I am deeply grateful to all
of them.

Excerpts from American Pastoral by Philip Roth. Copyright © 1997 by Philip

Roth. Used by permission of Houghton Mifflin Harcourt Publishing Company.
All rights reserved.
Excerpts from The Human Stain by Philip Roth. Copyright © 2000 by Philip
Roth. Used by permission of Houghton Mifflin Harcourt Publishing Company.
All rights reserved.
Excerpts from The Blind Assassin by Margaret Atwood are reproduced with
permission of Curtis Brown Group Ltd., London on behalf of Margaret Atwood.
Copyright © Margaret Atwood 2000
Details of all the literary texts referred to are given in the ‘Literary works
cited’ section of this book.
 Introduction

The relationship between language and literature is a contentious issue. On the

one hand, it may simply be described as a relationship between raw material
and a finished product – language provides the basis on which creative and
unique works of literature emerge. On the other hand, once we look at meaning,
the dividing lines begin to fade – it is difficult to define a sharp boundary
separating the meaning of literary works and the meaning of other texts. One
way of downplaying the obvious links is to claim that fiction engages knowl-
edge in a much broader and much more culturally specific way than everyday
use of language does. But that would be an exaggeration. One could not follow
an ordinary discussion of, say, climate change if one did not have any prior
knowledge of the issue. Terms such as global warming or carbon neutral
mean something only in the context of the debate and the science underlying
it. It is also not true that literature is different simply because it is not intended
to describe facts which have actually occurred. Linguistic constructions such
as conditionals do that too – they set up imaginary scenarios and explore
their consequences. The dividing line is either too fine to see, or shifts too
much, depending on the assumptions.
In this book I attempt not to debate those differences, but instead to talk about
literary texts as linguistic artifacts, focusing on the description of language
processes which participate in the emergence of literary meaning. My point of
departure is the belief that if a text means something to someone, there have
to be linguistic phenomena that make that possible. I found much inspiration
and help in two cognitive theories of language: construction grammar, with
its interest in the interaction between lexis and grammar and its trust in the
meaningfulness of syntactic choices, and conceptual integration, with its focus
on the mechanism of the emergence of meaning, regardless of the means that
lead to the initiation of the process.
Both theories have been applied to various works of fiction. However, the
goal of this book is not to discuss specific interpretations of specific texts in
terms of blending and construction grammar, but rather to explain how such
interpretations might be arrived at. In other words, this book explores the
processes of meaning construction which yield (all the varied) interpretations

1
2 Introduction

of literary texts. Crucially, I attempt to start with sentence-level phenomena

and study the ways in which they contribute to text-wide meanings, relying on
blending through all the levels. This project is thus an attempt to propose a
blending interpretation of fictional storytelling, along with the linguistic strat-
egies it involves.
In order to talk about linguistics and literature, I had to accept the necessity
of zooming in and out all the time. Some of the fragments in this book are closer
to linguistics, especially when the analysis focuses on specific sentence-level
choices, while other parts discuss the structure of entire novels, with limited
attention to the sentence-level language they rely on. Such shifts of focus
may seem somewhat distracting, but I also hope they will keep various kinds
of readers engaged. This book is addressed to cognitive linguists, stylisticians,
narratologists and literary critics alike, though these groups will surely find
different parts of it relevant to a different degree. It also attempts to facilitate a
conversation across these divisions.
The structure of the book also reflects its multiple foci. Chapter 1 discusses
recent work which sheds some light on the role of narratives in our under-
standing of language, literature, and culture; it contextualizes the ensuing
discussion against a broader range of research. Chapter 2 briefly introduces
blending theory and constructs some of the main tools of my methodology.
Among others, the chapter discusses primary mechanisms of story construction
and explains my understanding of the story as a conceptual construct. In
Chapter 3, I propose the analysis of the concept of a ‘narrator’ in terms of
epistemic viewpoint. Going against the tradition which assumes that every story
needs a teller, I discuss ‘narratorship’ as a cluster of epistemic and linguistic
choices which organize the text-wide viewpoint structure of the story. Chapter 4
investigates viewpoint configurations at lower levels of narrative structure,
and argues for an understanding of narrative discourse as directed primarily at
prompting conceptualizations of various kinds. In this chapter I also introduce
the primary mechanism of story emergence – viewpoint compression. Then,
Chapter 5 discusses the story-constructing role of referential expressions, espe-
cially pronouns and role descriptors.
Chapter 6 briefly introduces narrative underpinnings of the discourse of
theatre, arguing for tight conceptual connections between the text of the play,
the space of the theatre, and the bodies of actors. I also show how the cognitive
specificity of the theatre set-up affects linguistic choices, especially the emer-
gence of specialized constructions. In the remainder of the chapter I attempt
to trace the route which leads from on-stage narratives, where all the modalities
contribute to the emergence of the story, to novelistic, or on-page, narratives,
relying exclusively on linguistic forms. In this context, I also discuss the
primary importance of the forms through which fictional minds are represented
in narratives. In Chapter 7, I continue the discussion of the ways in which
 Introduction 3

representing characters’ thoughts and emotions (not just words) lies at the core
of narrative discourse. I give a brief overview of the constructions used to
represent conceptualizations held by various story participants, to conclude
that such textual choices are subordinate to the overall mechanisms of view-
point compression. Finally, Chapter 8 provides some closure to the main themes
of the book.
While there may be only limited explicit attention given to the issue, the
ensuing text also argues for a mode of analysis which treats the goals of literary,
linguistic, and cognitive studies as possibly overlapping or even shared. There
was much uncharted territory and a vast volume of research to contend with, but
some of the central questions have emerged clearly. How does the mind
negotiate the divide between everyday discourse and fiction? How does our
creativity work in these two contexts? What assumptions about the nature of
mind and language yield results potentially relevant to the humanities? What
exactly can literary work contribute to the discussion? There are clearly more
such questions than any single book can raise. This book attempts to identify
and discuss the few that seemed important.
1 Language and literary narratives

There is nothing so strange it cannot be true, and no story so unlikely it cannot

be told. No story is a lie, for a tale is a bridge that leads to the truth.
(The Arabian Nights, retold by Neil Philip)

This book is about how the centuries-old thoughts quoted above are indeed
true. Why are stories not lies, even though they don’t tell the truth? How do
they help us to learn from our experience and the experience of others? And how
does language support the meaning of stories? The structure of that “bridge …
to the truth” is what I will try to understand.
The human ability, or even desire, to tell, understand, watch, and create
stories has engaged a number of disciplines, each of which poses a different
set of questions about the core of the phenomenon. Why do we enjoy stories?
What’s in it for us as a species? Could our culture exist without stories? Are they
a mental construct, a linguistic construct, or a cultural construct? Is there a
difference between real stories and fictional stories? These are just some of the
questions of interest to anthropologists, psychologists, narratologists, philoso-
phers, linguists, and literary scholars. The answers have been, by necessity,
partial, and directed at the interests of the disciplines they emerge from, but it
is becoming increasingly clear to all concerned that some cross-disciplinary
dialogue is necessary. This book is an attempt to bring together at least some of
the questions out there while focusing on one central aspect of storytelling: how
do stories construct meaning?
The answer to a broad question like this requires making some assumptions
about the general mechanisms of meaning construction and about the corre-
lations between formal choices and the meanings prompted. This approach
would naturally define the project as a basically linguistic or stylistic one,
if not for the fact that while meaning construction is a process prompted
by forms (linguistic, visual, artistic, and otherwise), it is in fact a cognitive
process, involving various faculties of the mind. Elucidating the emergence of
meaning in various storytelling contexts thus requires adopting a method-
ology which is sensitive enough to fine-grained linguistic details, while open-
ing itself up to questions about cultural context, our cognitive capacities, and

4
 Language and literary narratives 5

the simple fact of there being various ways to interpret any storytelling
artifact.
The very term ‘meaning construction’ requires a brief explanation, so as to be
clearly distinguished from literary and anthropological work reliant on ‘social
constructivism.’ Throughout this book I will be using ‘meaning construction’ to
refer to the processes which yield meanings of language expressions based on
the frames evoked, constructional meaning patterns, and, among others, blend-
ing. Meaning construction in this sense relies on the specific word and grammar
choices. Even in the simplest case such as a choice of I will do it, which commits
the speaker, versus I would do it, which does not commit her at all, the construal
of the former expression as a promise and of the latter as advice relies on
meaning-construction processes crucially dependent on the local language
choices. The construction depends on all the linguistic forms used, from the
choice of the first person I (rather than third person he, which would not commit
the speaker in any way), and of the modal will (rather than might, which would
not involve a commitment), to the choice of tense. The present tense form will
prompts a more committed meaning than the past form would. Also, while a
social constructivist reading of a text is a matter of a critic uncovering frames
possibly (but not necessarily) existing in the writer’s mind, the framing of past
tense is a different phenomenon. Past tense forms are commonly used to mark
epistemic or emotional distance, and thus are often used for politeness (so that
Could you call me tomorrow? is more polite than Can you call me tomorrow?).
The framing of the past tense such that it marks the speaker as less committed is
a matter of standard language use, independent of any particular language user.1
The methodological framework to be used in this book relies primarily
on two theories – mental spaces theory and its extension known as blending,
or conceptual integration.2 Both theories assume that meaning is not a set of
discrete conceptual packets neatly correlated with linguistic forms, especially
lexical forms, but that in each case it emerges as a result of one’s use of formal
signals (verbal or not). It also relies on processes which economically use the
available complexity of possible aspects of meaning to yield an interpretation
which optimally fits the needs of ongoing discourse. In actual discourse,
though, interpretations are often negotiated rather than simply communicated,
and thus any account of meaning needs also to consider signals built into the
form being communicated which address the argumentative and inferential
aspects of meaning construction. The fact that such mechanisms take the form
of specialized grammatical constructions (cf. Verhagen 2005, 2008; Dancygier
2008c) should further convince us of the centrality of meaning negotiation to
any meaning-emergence processes. As this book argues, understanding narra-
tives relies on very similar processes: emergence, construction, and negotiation
of meaning through specific language choices. The choices may be found at the
lowest level of linguistic structure (such as using the first or the third person in
6 Language and literary narratives

the representation of narrative voice) or they can pertain to text-wide decisions

on narrative structure (whether the story is presented chronologically or in a
fragmented manner), but they are language choices readers rely on in the
construction of narrative meaning.

1.1 Where does narrative meaning come from?

Meaning is often talked about in terms which seek to divorce it from the
situation in which it emerges, and to propose formal constraints on what counts
as a meaningful utterance. In the context of stories, this approach would not
yield any understanding of their meaning, simply because stories do not open
themselves to this kind of analysis. The same can be said about attempts,
especially in structuralist narratology, to arrive at clear taxonomies and termi-
nological distinctions. These taxonomies range over a very broad spectrum
of phenomena, starting from representations of time flow to a typology of
‘narrators.’ While this kind of work has alerted linguists and stylisticians to
phenomena which can now conveniently be talked about as ‘flashbacks’ or
‘intrusive narrators,’ it has not explained adequately why these forms would be
used and how they are understood. Trivializing things a little, one could ask why
flashbacks are so common in novels or movies (and, yes, also in natural
conversations and oral stories), if most narratologists claim at the same time
that the story is (roughly) a temporal sequence of linked events? And is a
narratorial voice which (who?) addresses the reader directly just a frill, a
naïve stage in the development of narrative form, or a charming attempt to
make written fiction feel like a conversation? Admittedly, recent work in
cognitive narratology (cf. Herman 2002, 2003a), Fludernik’s natural narratol-
ogy (1997, 2003) and stylistics3 marks a move away from taxonomies and in the
direction of cognitive explanations, but there is still much to say about the
connection between the lowest-level language choices and the narrative, and,
primarily, about how this connection is mediated by cognitive mechanisms
involved in the emergence of meaning.
It seems to be an accepted view in many disciplines (including some linguis-
tic and cognitive theories) that language is a separate module of the mind and
thus cannot be considered in the context of art, literature, or culture. The fallacy
is compounded by a common belief that linguistic analyses are too formal and
too focused on representing structural patterns to be useful in an enquiry into
meaning in any context broader than a sentence. While this is true in many
cases, it is not always true. The linguistic models to be relied on in this book,
within the broadly understood discipline of cognitive linguistics, assume that
meaning construction is a process which affects numerous levels, and that the
overall meaning does not arise as a result of simple addition or compounding of
the lower-level constructs. On the contrary, meaning-construction processes
 1.1 Where does narrative meaning come from? 7

invariably involve selection and narrowing on the one hand, and emergence of
new configurations on the other. There is a significant range of linguistic work
using mental spaces and blending to account for meanings of nominal com-
pounds and sentential or phrasal-level constructions,4 and such analyses typi-
cally involve uncovering the composite constructs at the lower levels. Research
has also been done on the emergence of meaningful blends in discourse and
conversation (cf. papers in Oakley and Hougaard 2008; Pascual 2002, 2006a,
2006b, 2008). What the analysis proposed in this book aims at is describing
the consequences of lower-level linguistic choices at the higher levels of
narrative discourse. While there are some limits to treating the whole text as a
linguistic artifact, the focus on processes rather than forms opens longer and
more complex stretches of discourse to linguistic analysis. Thus the intended
contribution of the analyses proposed in this book is to offer a view of meaning-
construction processes in multidimensional artifacts such as novels or plays.
The processes of selection, narrowing, and emergence of new meanings are
complex and depend on more than the understanding of specific language items.
As Deacon argues (2006), blending is an interpretive process which relies on
our generic symbol-processing and referential capacities. That is, the ability to
use linguistic symbols in new combinations, to juxtapose them or compose
them into new configurations opens our capacity for coining new meanings to
nearly limitless possibilities. One might read this to mean that the potential for
our linguistic abilities to be used in the construction of new meanings is not
inherently limited and relies on the fact that manipulation of symbolic concepts
is the daily bread of our cognition. Indeed, even a very simple expression can
illustrate that potential. For example, the expression emotional roller-coaster
cannot be explained without an analysis of the meaning of the component
concepts, especially roller-coaster. It is a fully conventionalized expression
which refers to a machine found in amusement parks which takes people on a
seemingly dangerous ride involving abrupt ups and downs, lack of control of
one’s movements, as well as high speed. While any such experience in real life
would be unwelcome, it is treated as ‘fun’ in the context of an amusement park.
The addition of the adjective emotional changes the meaning significantly.
There is no longer any physical object involved, there is nothing suggesting
entertainment, or even an experiencing person’s conscious decision to be
engaged in the activity. An emotional roller-coaster evokes negative (rather
than amused) emotional response parallel to what roller-coasters do (jerking
people up and down at a breakneck speed); it applies to a person’s psychological
state, unpleasantly and abruptly altering between elation and depression. The
composition process of this blend creates a concept not available in the compo-
nent expressions, and, even more important, evokes an emotional response
opposite to the original. This is also one of the points Deacon makes when he
stresses the fact that emergent reference may involve conflicting emotions.
8 Language and literary narratives

The brief discussion above does not do justice to the complexity of the blend
described, even though it remains at a relatively low level of complexity, which
is the lexicon. And yet, the analyses throughout this book will attempt to look at
narrative structure in its entirety through the lens of such low-level choices.
While longer texts will be the target, sentence-level phenomena such as pro-
nominal choice, the use of tense, the impact of referential expressions, or
representation of discourse will be considered as the lowest level of meaning
construction. In general, the assumption of this book (and other cognitive
linguistic texts) is that grammar is as meaningful as lexis, and thus grammatical
choices impact the meaning of texts. That is, a text’s reliance on interior
monologue or on direct discourse is a stylistic choice which, via the interpre-
tation of the specific constructions, impacts the way the narrative is understood.
This analysis will also go a step further in arguing that a use of a specific form,
associated with a meaningful construction, may ‘infect’ other aspects of the
expression and carry constructional meaning while only one bit of the con-
struction is present. The linguistic process whereby chunks of constructions
spread constructional meaning has been described as the mechanism of con-
structional compositionality (Dancygier and Sweetser 2005; Dancygier 2005b).
Examples are common. The simplest instance I can recall is an advertisement
for Schweppes tonic in the London underground, many years ago. It consisted
primarily of the letters Sch . . ., running on the wall along the tracks. The
combination of letters, unusual in English, brought up the brand name, and its
flagship drink. It seems a long way from this to a novel, but I will argue that
strategic (rather than random) linguistic choices, such as the choice of first-
person or third-person narration, do propagate through narrative structure to
orchestrate narrative viewpoint throughout the levels. In other words, anaphoric
and cataphoric use of pronouns is a concept which explains a host of referential
meanings affecting sentences, but in a longer text these concepts do not suffice
(cf. Emmott 1997). Similarly, a consistent choice of past tense as the narrative
form affects not only the meaning of every sentence used as ‘narration,’ but also
the viewpoint construction in any sentence elsewhere in the narrative which
could be described as representing speech or thought. For these aspects of
meaning to be accounted for, we need to assume that grammatical choices can
single-handedly construct (not just participate in) meaning configurations that
they are ordinarily a part of.

1.2 Literary analysis and linguistic analysis

To be able to make its point, this book will engage with a number of storytelling
artifacts. I will be relying on material mainly from novels, but also from travel
narratives and plays; there will also be some short examples from poetry. Texts
will be chosen to represent particular phenomena, though some questions
 1.2 Literary analysis and linguistic analysis 9

regarding the historical emergence of various narrative strategies will also

be addressed. Crucially, none of the literary periods will be specifically
foregrounded.
This evokes the question often raised in the discussions of cognitively
motivated attempts to talk about literary texts. As Weber 2004 (among others)
points out, cognitive methodology often downplays or disregards the historical
and cultural embedding of the texts, or differences among readers. If it is indeed
the case, then does it mean that such analyses are lacking something essential?
Certainly not, and there are many reasons for this. First, it does not seem to be
the case that everything that can be said about a historically situated text is
period specific. If anything at all is indeed wrong, it is the suggestion that only
period studies can reveal frames and cognitive patterns. What cognitive meth-
odology intends to capture is patterns of thought, and not exclusively patterns
of Victorian thought or modernist thought. Still, this approach does not preclude
cognitively informative work looking at a specific period. In fact, quite a few
Renaissance scholars (Hart and McConachie 2006; Cook 2007, 2010;
Moschovakis 2006) are doing cognitively informed work themselves, and the
conversation between the generic and the specific is most fruitful for both sides.
Second, it is not true that ‘cognitive’ equals ‘culture neutral.’ Some of the
most thorough cognitive analyses of literary texts (cf. Sweetser 2004; Turner
2004; Canning 2008) are also deeply immersed in the beliefs of their respective
periods. Still, there are probably many cognitive analyses where writers ignore
literary/cultural/historical expertise, either because it is not relevant to their
topic or because they have different goals, and it would seem unreasonable to
assume that all cognitive work on literary texts has to be supported with in-
depth considerations of their historical context.
Also, there is the question of bridging the gap across different ways in
which language is talked about in linguistics and in literary studies. Current
approaches to language in the work of philosophers such as Derrida or Lacan
are not easy to reconcile with routine linguistic analysis. However, cognitive
approaches to meaning are less formal than most and also more open to
questions of cultural or emotional framing, so it seems possible that the linguis-
tic and literary scholars might find shared interests here.
Linguistic work on literature is often seen as restricting the freedom to offer a
new interpretation (cf. Jackson 2005). Yet, while meaning is not entirely
indeterminate, it is also not determinate. It is a perfectly natural reaction on
the part of those engaged with various sources of interpretation to feel that
someone arguing for a single interpretation is missing the point. And it is true
that many stylistic and cognitive analyses can be easily read to claim that what
they discuss is the ‘prescribed’ interpretation (even though I am not aware of
research where such claims are made explicitly). The assumption of this book,
for comparison, is that while it is impossible to predict all reader responses and
10 Language and literary narratives

possible interpretations of a text, the focus on cognitively based processes of

interpretation should in principle open the floor to this discussion. Indeed,
what is it that counts as an interpretation? Are there perhaps interpretations
out there which depart excessively from optimal avenues of text processing?
Certainly. And, also, are there perhaps language-based interpretations out there
which are too restrictive? Of course, there are many. The point in these cases
seems to be that the interpretive tools employed are either not specific enough or
too rigid, and, I argue, these limitations are the result of broader assumptions
about . . . yes, about language. Whether one believes that linguistic conventions
are irrelevant to meaning or that they determine meaning entirely via grammar
and discrete lexical packets, the distortion is almost equally dangerous.
The model of language which at least attempts to avoid these kinds of traps
is a model which assumes that all meaning is constructed, regardless of the level
we are looking at. The semantics which would be satisfied with the character-
ization of the word such as a rock as [−animate] [+concrete] [+count] will not be
able to open itself to the contexts where a rock is viewed as a geological
specimen, material suitable for sculpture, material suitable for construction,
a weapon, a tool, or a symbol of stability and certainty, or to the possible
contextual ambiguities. By the same token, if a linguistic model assumes that
the present tense is simply an aspect of the morphology of English and that it is
different from the imperative form, it will not even address the possible inter-
pretation of a common usage where these forms are difficult to distinguish, as in
an airport BMW ad featuring an attractive picture of a car and the line Miss your
flight; still make the meeting. The slogan is obviously not issuing an imperative,
and it is not describing a situation which can be construed as the use of
‘present tense.’ What is required is an understanding that a similar use of a
present form appears in non-predictive clauses of predictive constructions
(If you miss the flight, you can still make the meeting), and that the verb form
is to be understood as neither future nor present nor imperative in meaning,
because the specific constructional meaning evoked is ‘non-predicted assump-
tion to be used in further reasoning.’ Dancygier and Sweetser (2005) propose a
broader discussion of how a verb form may take its meaning from a construction
it participates in and how the use of the form may prompt the meaning of a
whole construction. The point I am trying to reiterate here is that grammatical
and lexical choices may carry more meaning than some grammars allow.
On the other hand, from the language perspective, an interpretation may
appear unnecessarily open-ended. Results vary, but there are clear cases where
interpreting passes over central aspects of narrative meaning because of over-
reliance on cultural concepts. In Margaret Atwood’s novel The Blind Assassin
the actual plot, or ‘what really happened,’ is not narrated directly, but requires
construction by the reader. The crucial question of the plot is whether a man
accused of arson actually did commit it. He is wanted for the crime, and the
 1.3 Literature, language, and human nature 11

consequences of the fire ripple through the lives of all the characters, but no
answer is provided. It is in fact available in the text, but only if the reader
assumes that all its narrative strands, even if apparently incongruous, are present
for a reason and that the form of the novel can be interpreted in a way which
reveals the whole plot. However, the indirectness of the narrative actually
distracted quite a few critics (Wilson 2006; Perrakis 2006; Roy 2006) into
thinking that the surface incongruence is in fact yet another instance of the
‘instability’ of meaning, and into focusing on one of the themes common in
contemporary criticism – violated women’s bodies. The unfortunate conse-
quence is that they failed to consider crucial aspects of the plot and essentially
misrepresented the plight of the main character. In cases like these, the problem
is not that the ‘interpretation’ simply lacks depth and convincing power, but that
the interpretive avenues chosen impoverished the result, while an analysis using
the complexity of the structure to reveal the complexity of meaning might have
worked better.
Some critical work might be read to suggest that writing challenging liter-
ature intends to background the very idea of a coherent interpretation. From the
linguistic perspective, such a suggestion would be difficult to accept. Within the
framework I’m assuming, there is no reason to believe even the most challeng-
ing literary texts to be beyond our interpretive strategies by their very nature,
although in fact readers do give up on texts which do not engage them enough.
As long as a reader is prepared to do what she does in colloquial contexts, that is,
read out of a text but also read into it, the most difficult narratives can be read
and coherently interpreted. With this assumption, fragmented and ostensibly
incoherent stories are indeed ultimate exercises in language use, because they
stretch the cognitive abilities of ‘making sense’ to their limits. It has been my
experience, as well as that of many of my students, that some practical expertise
in cognitive analysis opened higher levels of our abilities as readers, by making
explicit the strategies which remain implicit in less challenging instances of
reading.

1.3 Literature, language, and human nature

While the nature of literary and linguistic meaning is still a hotly disputed
subject, more recent arguments revolve around a much broader set of concepts
involving human nature, evolution, the origin of culture, et cetera. There is a
growing interest in various areas of the humanities in the study of the individual
mind, independent thought, genetic endowment, cognitive processes, and the
bodily basis of cognition, but we are far away from an agreement on how these
questions are to be incorporated into the existing body of scholarship. One
proposed approach relies heavily on the findings in evolutionary psychology
and related disciplines. For example, Steven Pinker’s book The Blank Slate
12 Language and literary narratives

(2002) is a passionate defense of what we used to call ‘human nature,’ our

shared as well as individual cognitive and evolutionary heritage which is not
subject to cultural or social molding. While the evidence he presents to talk
about, for example, education is clear and convincing, the argument becomes
less clear when it comes to the final discussion of the arts. Pinker criticizes the
current state of the arts rather harshly, and further points out that esoteric
concepts such as ‘artistic value’ have direct correlates in ‘human nature.’ To
conclude his argument, he discusses some literary examples.
Unfortunately, it is not likely that the way Pinker chose will convince anyone
seriously interested in literature or language. Several anecdotal examples not
supported with any specific argument cannot count as evidence for the above
rather strong claim which the reader is expected to accept. Intuitively, his claim
that literature reflects human nature is equally unsubstantiated as insisting that it
only reflects human culture, and the earlier extensive discussion of the tensions
between nature and nurture do not suffice in supporting this very specific claim.
An off-hand statement that reading initiates a cascade of neuronal and cognitive
events does not explain how this is different from the avalanche of such
events prompted by an appetizing smell or a beautiful view, or how humans
have come to enjoy interacting with fictional artifacts of various kinds and keep
on perfecting the means of creating them. While I agree with the general gist of
Pinker’s story, it is somewhat too cursory to be satisfactory. I appreciate, though,
that the arts were not the main focus of his project.
However, another claim implied in much of Pinker’s work is convincing – the
humanities should return to an attempt to understand humans as they are. In
view of serious advances in the sciences the goal seems to be somewhat more
realistic, and there is more and more work pointing in that direction. For one,
Slingerland’s argument (2006) for what he calls ‘vertical integration’ (of arts
and sciences) presents much of the already available evidence for the viability
of such endeavors. Also, the growing interest in cognition within the study of
music, visual art, performance, film, et cetera suggests that there is an awareness
of the need for the humanities to at least interact with the explanations science
offers and to prompt new questions in science.5
There is, however, the tricky issue of how exactly science fits in the human-
ities and what humanities need to do to expand their current understanding of
what culture contributes to literary reading. In an issue of Criticism, for exam-
ple, Evans (2007) makes specific suggestions for a more refined view of culture,
such that would reflect the complexity of our times and fit better into the
interests of criticism (and not cultural anthropology). In the same issue, van
Oort (2007) discusses the potential for interaction between humanities and hard
sciences and argues that a close methodological fit, such as the one proposed by
Slingerland (2006), is not desirable, because humanities need to focus on the
specifically human ability to use symbols in communication and culture, even
 1.3 Literature, language, and human nature 13

at the cost of downplaying more traditional critical interests. While this is an

interesting argument, I am not yet sure that the nature of symbolic representa-
tion is a rich enough base to enable the ‘after-theory’ humanities to flourish. It is
also interesting that van Oort expects the humanities to include the discussion of
language (I couldn’t agree more), but, surprisingly, puts cognitive linguistics
among hard sciences, and thus outside the humanities. Further discussion of
how cognitive linguistics might be useful to the humanities will be taken up
throughout this book, but van Oort’s suggested dividing lines will probably
surprise researchers on both sides.
The point where I agree with van Oort is when he suggests that the humanities
need their own methods, whether science-based or not. However, the call for
the integration of disciplines is sometimes read as an encouragement to make
humanities look like sciences. One such effort is represented by a group of
literary critics now known as ‘Literary Darwinists.’6 Their claim is that liter-
ature is an evolutionary phenomenon and should be studied as such. Their main
argument is that literature is a direct reflection of the human evolutionary set-up,
so that literary texts reflect concepts such as mate selection, dominance, and
other concerns of species survival and reproductive psychology. I am reading
their work as a response to Pinker’s call for putting human nature first, but the
specific choice made by this group is tantamount to substituting ‘mother nature’
with ‘mother evolution.’ As such, it is a misrepresentation of literary tradition
and its cultural and humanistic role, and another bold attempt to find one key
concept to explain a complex issue. The swing of the pendulum from culture
to evolution does not even ask how culture is linked to evolution and glides
gracefully over everything in between. Also, while the literary Darwinists focus
on the few relevant aspects of human nature, and do not include historically
determined cultural concepts in their discussion, they are also avoiding any
recognition of what schools of psychology other than evolutionary psychology
have to say. This also includes theories already used by other literary scholars,
such as the theory of mind, cognitive psychology, or the psychology of emo-
tions (Gerrig 1993; Zunshine 2006; Hogan 2003).
The conclusion is that this specific attempt at marrying science and the
humanities is not a particularly promising avenue. My own understanding of
the problem is that building such bridges requires more than one step. One
cannot go from evolution into literature directly, because too much is missing:
human cognition, including the cognitive roots of culture; the cultural and
cognitive role of art; and the role of language in representing both. The gap is
not easily filled, but some attempts have been made. Throughout this book, I
will rely on some of the concepts formulated by Donald (1991, 2001, 2006) and
Deacon (2006). They both refer primarily to art, but many of their claims are
naturally applicable to literature (some are not, but they do not undermine the
use of those which do apply).
14 Language and literary narratives

Donald’s approach assumes that there is a dual explanation to the emergence

of art. One is mimesis, as the primary and evolutionarily oldest mode of
representation, while the other is self-reflection. He argues that art, which relies
on both, is an activity intended to influence other minds, and it functions in the
context of distributed cognition. It thus has a dimension that is both individual
and social, as mental representations are also necessarily shared and requiring a
mode of transmission (such as the early mimetic forms involving the body or
purposefully created representation, and, in the later stages, a specialized form
of communication, that is, language). The self-reflective nature of art applies not
only to the individual, but also to the society as a whole, and, in later stages, to
human nature. Donald also stresses that art does not emerge solely from an
individual mind, but that the source of creativity is in fact in the distributed
network of cultural memory and its technological development. Thus a cave
drawing and a cathedral are both rooted in the cultural system the artists
participate in, and even though they rely on currently available technology,
they serve the reflective needs of the society as well as creating specific
cognitive outcomes – monitoring collective memory, giving salience to cultural
values, reinforcing the social network. Narratives can naturally be viewed
similarly (cf. Herman 2006).
The self-reflective nature of art is further discussed by Deacon (2006), who
sees the role of art in terms of the way it affects individuals emotionally.
Interestingly, Deacon proposes a category of ‘emergent emotional states’
which rely specifically on the availability of more than one cognitive agent or
representation. Candidates for such emotions are irony, nostalgia, righteous
indignation, or humor, because they imply an interaction of various cognitive
constructs. The source of nostalgia, for example, is being able to access two
mental representations at the same time – the one reflecting the current state, and
the other rooted in the emotionally loaded memory. To put it in cognitive
linguistic terms, the tension between different construals or mental spaces
creates an emotional/cognitive reaction which is different from basic emotions
(see Damasio 1999, 2003) such as fear. Deacon’s emergent emotions can only
appear in the network of distributed cognition, so that both the foregrounded
construal and its competitor are available (as when I feel indignant but the
cultural norm approves). The concept also seems to tap into Damasio’s theory
(1999) postulating the emergence of human consciousness out of just such
complex emotions.
While Donald’s and Deacon’s work gives us a top-down view of how the
cognitive architecture of our species affects human production and reception of
art (including literature), there are more issues to be considered in the bottom-up
view. Donald makes it clear that art works in a two-way interaction, where the
artistic object offers its structure and framing and the viewer/reader adds her
or his own, and Deacon also stresses the symbolic roots of art, where the
 1.3 Literature, language, and human nature 15

representation emerges in the recipient’s mind. There is, however, a whole set of
questions pertaining to the cognitive work done in the close interaction with
narratives. For example, the work in cognitive psychology reported in Gerrig
and Egidi (2003) focuses on the postulated automatic and necessary types of
strategies of narrative comprehension, emerging out of experimental work.
There seem to be two major possible strategies for reading narratives: either
text processing is primarily memory-based, or it relies on a search for coherence
which prompts the construction of causative chains. While testing memory
mechanisms may seem a safer and more minimalist route, researchers working
closer to actual literary narratives (Emmott 1997) are inclined to propose read-
ing strategies which are constructivist and coherence-based. Evidence from the
studies of the therapeutic role of narratives (Klein 2003) also supports this
approach, as subjects find that coherent accounts of traumatic events prevent
recurring uncontrollable emotional responses to the memories of trauma.
Causal coherence is one of several strategies proposed in Herman’s (2003a)
discussion of the role of narratives in the organization of experience. Herman
follows Bruner (1991) in viewing narratives as tools for thinking. The specific
strategies Herman discusses are: chunking experience, imputing causal rela-
tions among events, framing events as instances of a certain type of occurrence,
and construction of event sequences. Jointly, the strategies allow one to con-
struct events as bounded concepts appearing in a sequence and linked causally
(‘he proposed, she accepted, they got married’). One might add that these
strategies are supported by a number of linguistically available frames (such
as, what counts as a marriage proposal), scenarios (marriage requires prior
consent of both parties), and attribution of intentionality (his proposal counts
as the expression of his wishes), but the fact that such conceptualizations in fact
rely on each other makes it difficult to determine which of the prompts is the
crucial one. It seems that such instances of ‘organization of experience’ are in
fact a kind of feedback loop: the emergence of narrative structure may crucially
rely on such lower-level cognitive operations, while at the same time these
operations support categorizations of behavior which make narrative structure
possible. Furthermore, events can be categorized as participating in a narrative
chain only when they can naturally be construed as having causal force –
otherwise, they do not participate in narrative structure. It is thus clear that
causality plays a crucial role in the concept of narrativity, but the way in which it
participates in narrative structure is more difficult to determine. In any case,
these low-level cognitive processes are an important aspect of the construal of
narratives.
To sum up, the resulting picture of art-based cognition requires that the
properties of the brain and our core skills of symbolic representation be put in
the context both of a social network which gives validity and purpose to the
creative acts, and of the lower-level cognitive operations such as chunking and
16 Language and literary narratives

adding causal connections. Moreover, while art relies on our mimetic imagi-
nation, it becomes more complex as humans’ technological, cultural and com-
municative abilities are honed in the context of a human culture understood as a
cognitive distributed network. It is perhaps possible now to see how literature
owes its existence to our general cognitive abilities, to our ‘artful’ mind, and to
the communicative needs of a group, all of which gave rise to more and more
sophisticated use of language (see Dancygier [in press a] for a broader dis-
cussion). The combination of human cognitive set-up, the cognitive roots of
culture, and the cognitive roots of language is what might be an account of
literature which uses science without simply repeating its claims in a context
which is not reducible to the concerns of scientific research. As more and more
researchers notice, being able to see the human brain at work through fMRIs and
scans will not tell a researcher in the humanities much beyond the obvious
observation that something is happening in the brain as we listen to Mozart or
read a novel. The humanities need their own tools to represent the ultimate
forms of human language and cognition, though these tools cannot ignore what
science has to offer. In this book, I will attempt to clarify one end of the
spectrum – the contribution of language as a cognitive tool to the formal and
interpretive complexities of literary texts.

1.4 Literary texts and communication

If questions of ‘evolution’ are relevant to this enquiry, it is to the degree to which
we can attempt to answer the question of how literature has emerged in its
present form. I am not at all interested in speculating whether literature is an
adaptation or an evolutionary by-product,7 because the way such questions are
posed assumes that once it has emerged it continues playing its evolutionary
role and whatever changes in literary form is no longer relevant to the question.
I believe that the questions which can be considered profitably are different
ones: what form of communication has fiction evolved out of? what can the
emergence of different genres and poetic forms teach us about both literature
and language? did language have to change to accommodate literature, and if
so, how? how do literary conventions depend on our folk understanding of the
human mind and human emotions? These questions cannot be answered
through evolutionary theory, but they are, in my view, directly pertinent to the
question which seems to be the evolutionary Holy Grail: why do humans create
and read fiction?
In what follows I will assume that in order to understand the emergence of
literature we have to place it in the context of communication. Broadly speak-
ing, I find the picture emerging out of Tomasello’s (1999, 2008) argument about
the origins of human communication to be a good starting point, especially in its
interesting elaboration of the concepts proposed by Donald and its reliance on
 1.4 Literary texts and communication 17

the concept of joint attention, recently broadly discussed by Oakley (2009).

Tomasello sees the emergence of narrative as relying on the human basic need
to share information and to arrive at social, and also linguistic, norms in the
process. Further, Tomasello claims, telling stories gives rise to specialized
linguistic constructions which allow tellers and listeners to connect events
across different temporal planes, to keep track of individuals, and to take
another person’s perspective. Indeed, the emergence of elaborate pronominal
systems and of constructions reporting thoughts and feelings while clearly
keeping track of whose thoughts and feelings they are suggests some form of
feedback loop between narrative and linguistic forms.
Tomasello’s model can also be useful as a starting point in an attempt to
track the emergence of fiction as a step in the emergence of forms of sharing
information. An attempt to establish social norms through storytelling requires
at some point either that the narratives used to establish those norms become
more salient than what is available in the immediate environment, or that the
actual true stories are told in an exaggerated form. Perfect examples of this can be
found in folk myths and stories. For example, Scalise Sugiyama’s (2005; Scalise
Sugiyama and Sugiyama in press) work, including her account of stories of
cannibalism told in various oral traditions of foraging peoples, poses some crucial
questions regarding the history of storytelling as an information-sharing device.
Do the vivid and horrifying descriptions of cannibalism she discusses actually
represent the past of all of those peoples? In other words, did it actually happen
that children could be cooked and eaten by their parents in time of famine? Or are
these stories the products of generations of inventive storytellers who made the
stories more and more vivid to make their messages more memorable? Have
children actually been abducted and devoured when they left the safety of the
settlement, or did they need to be discouraged in exaggerated ways from leaving
the safety of the group? From the point of view of Tomasello’s account the question
is as interesting as it is unanswerable, since it is impossible in these cases to
separate the account of events which occurred in the past from the narrative
selection and presentation of events which play the normative cultural role better.
Since the question about the fictionality or truthfulness of early narratives
cannot be resolved, it is probably more useful to consider the ways in which
their messages may have been structured. One such way relies on linguistic
markers of evidentiality and hearsay (similar to English adverbs like reportedly,
but often taking inflectional form, as in Central Pomo, Makah, or Pawnee).
While evidential markers appear in some languages to distinguish first-hand
from second-hand accounts of events, other languages develop special story-
telling forms such as the ‘mythic’ or ‘ancient’ past (in Upper Chinook, or
Kiksht’). While evidential markers are primarily available in spoken commu-
nication, the ‘mythic past’ has developed specifically for narrative purposes.
In other words, the more salient the cultural role of the narrative, the less salient
18 Language and literary narratives

the question of its truthfulness or the specific source of information. Once the
text becomes a shared cultural artifact, it participates in the system of distributed
cognition and is no longer treated as one speaker’s communicative contribution.
Presumably, culturally salient stories are tools of distributed cognition and as
such may develop a linguistic status of their own.
It is also important to note that such stories are subject to the process which
blending theory refers to as ‘compression’ – the primary process leading to new
blends. An example of compression could be a very simple expression, such as
the generations of tomorrow, where the deictic form referring to the day following
the current one (from the point of view of the speaker) is used to refer to all such
days in our future. The compression of all specific ‘tomorrows’ into one down-
plays their temporal or deictic location, and highlights the speaker’s generic
intent. In the stories under discussion here, various potentially factual events
are compressed into the one the story actually tells, moving again from the
specific to the generic. In the process of compression, the story may undergo
various embellishments, so that the generic aspects (the ones prompting ‘narrative
transport,’ as Gerrig 1993 defines it) are more salient. Thus the gruesome image
quoted by Scalise Sugiyama and Sugiyama, with the mother stirring the pot in
which the children’s little feet and hands are still visible, is too strikingly detailed
to be an actual description of an actual event from distant past. Its salience and
descriptive detail seem to serve the purposes of the generic message and effective
compression, not the accuracy of the storyteller’s account. One can thus conclude
that such narratives are true and not quite true at the same time, because the
singularity of the event recounted is not important from the point of view of the
cognitively distributed goals of the group.
Looking at oral storytelling from this perspective may help us see the
connection between early narrative forms and highly elaborate contemporary
ones. First, oral stories are limited by the capacity of human memory, but, at the
same time, they expand that capacity by relying on the storyline as the glue
which holds various events, times, people, and locations together. Written
narratives capitalize on these narrative constructions, but they are gradually
freed from the worry about anyone’s capacity to retain the story in its entirety,
and narrative form becomes more and more complex. Second, the stories such
as the ones discussed by Scalise Sugiyama (also in her earlier work) are clearly
intended to achieve a certain effect in the listeners – compassion with victims,
fear of predators, increased sense of dependence on social and family structures.
But these effects are achieved as a kind of ‘narrative uptake,’ not because they
are directly communicated. The mechanism whereby an elaborate linguistic
construct can prompt ‘narrative transport’ (cf. Gerrig 1993) must have been
present in the minds of the foraging peoples and it is present, in an enhanced
form, in our minds. Still, regardless of the form and context of the story, the
‘point’ of the story is only available through constructing its meaning – not
 1.4 Literary texts and communication 19

only by tracking the sequence of events, but, more important, by exploring all
the ‘what-if’ scenarios which could have prevented overly confident children
from being abducted or might help us understand ourselves and our world
better. There is no limit to what a narrative can do, but it requires a mind
which uses the story as it is told or written to construct its meaning.
Apart from this broad context in which stories achieve their status of mean-
ingful artifacts there is also the question of how a literary text is similar to or
different from any other text, spoken or written. I will assume that language and
its basic communicative context are the basis on which literariness is built. I
treat the basic communicative context similarly to what cognitive linguists call
the Ground (Langacker 1991, 2001; Coulson and Oakley 2005; Verhagen
2005) – the immediate surroundings, the presence of the speaker and the hearer,
shared knowledge, shared visual and aural field, shared understanding of
bodily experience, and, last but not least, the surrounding discourse context.
The Ground makes various aspects of communication possible, since meaning
can be derived from ostensive aspects of communication, from objects manifest
in the environment, and from the interlocutors’ body language, gesture, facial
expression, tone of voice, et cetera. In my understanding, the Ground conceived
in this way also provides access to shared cultural and conceptual framing, so
that the interlocutors can assume access to culturally salient concepts. The
difference between the Ground of any specific exchange and the default com-
municative context is that the latter does not have to profile any individual
interlocutors, but rather a generic concept of what engaging in communication
is about. Similarly, what is required is a concept of a shared visual field, and not
a concrete visual field accessible in a specific exchange.
Another aspect of what is required for literature to build on is the assumption
that linguistic expressions prompt conceptualizations. As Verhagen (2005)
argues, any exchange may involve two aspects of conceptualization: the ‘sub-
ject’ of conceptualization (the speaker, the hearer, other participants, and salient
aspects of the Ground) and the ‘object’ of conceptualization (the actual subject
matter of the exchange). While the potential disagreement between the speaker
and the hearer as to the actual conceptualization of some subject matter is
always a possibility, what Verhagen also shows is how much of the actual
linguistic structure is addressed at negotiating those conceptualizations. If a
child visiting a zoo calls to the parent “Look, Dad! A dragon!” this counts as an
attempt to conceptualize a phenomenon manifest in the Ground and in the
shared visual field, but when the father answers “No, Jimmy, it’s a lizard/a
crocodile/a snake . . .” this counts as negotiating a different conceptualization,
with its attendant cultural consequences (‘we cannot start pretend play whereby
you are a knight and I’m a wizard and we slay the dragon, but we can learn more
about reptiles’). Naturally, there is no such potential exchange between the
participants in the act of reading (reader versus author/narrator/text), but the
20 Language and literary narratives

negotiation of conceptualizations is built into the narrative structure. This is,

I think, what Tomasello talks about when treating narrative as an engine to
produce more complex linguistic constructions. These structures not only
maintain temporal, referential or viewpoint links, but allow the construction
of narrative meaning to happen. The greater part of this book is devoted to
revealing the various ways in which narratives perform this task.
It is important to be specific about how narrative context deviates from the
default communicative context. First, ordinary communication features two
interlocutors, the speaker (I) and the hearer (you), who, in the most standard
set-up, have visual and aural access to each other and exchange their deictic
roles in every turn of the conversation. In this default set-up what counts as
communicated has to be said or done (or, from the point of view of the receiver,
heard or seen). Naturally, this would never suffice, and that is how languages
develop various forms of establishing categorizations, encouraging inferences,
reaching beyond the available Ground, reporting past situations, et cetera. But
the default, as I will try to show in chapters 6 and 7, remains surprisingly
persistent in structuring different literary forms.
Deviations from the default are common, and in most cases discourse is
maintained via appropriate constructions, such as back-channeling in phone
conversations (all the yeah and hmm expressions signaling continued presence),
emoticons in e-mail, page layout and paragraphing in journalistic discourse,
intonation patterns, rhetorical devices and repetitions in public speaking,
et cetera. More interesting, however, there are instances of discourse which
seem to assume the default in obvious absence of its crucial ingredients. People
routinely speak to animals and plants, and also to cars, computers, and other
gadgets (mostly phrases like Don’t do this to me!), while animals and objects
commonly speak in stories, comics, and poems. Not only does this deviation
not cause incomprehension, but most of the time it also goes unnoticed. It
seems that the very fact of something being said evokes the default set-up,
and thus ostensible communication is treated (linguistically) as bona fide
communication.
Such a ‘constructed’ communicative set-up is just an example of the fact that
something being communicated suggests intentional communicative agents.
Dennett (1997) talks about our ‘intentional stance,’ whereby we attribute
agency to objects and animals, but it seems to be only a part of what Deacon
calls a ‘representational stance,’ whereby we tend to treat things occurring
around us as meaningful. In the context of a written narrative, both stances
are involved. The presence of a text is treated by default as interpretable via our
symbol-reading abilities, and as presented to us in its form by an intentional
agent. Reading is thus a communicative act which has an addresser and an
addressee, even though the communicative intent is available only through the
form of the text in front of us. It is interesting to note how natural it is for readers
 1.5 Why is fiction special? 21

to talk about ‘books’ or other texts as ‘saying’ something, and attributing

intentions to authors (‘the author is saying X,’ ‘the author meant X,’ et cetera).
These ‘fallacies’ seem to rely on the default understanding of communication.
This seems also to be the case with extended narratives. There used to be a
teller of an oral story, and the listeners may have assumed that the com-
munication mostly goes one way, but these assumptions did not affect the
communicative impact. Contemporary fiction is mostly written, and so the
‘communicator’ is hidden behind the text, but the concept of a narrator has
substituted that of a teller. As I will argue throughout this book, the narrator
(but not the author) may be no more than a linguistic construction (or con-
structions, since there are many kinds of narrators), and yet a written work of
fiction is naturally seen as conforming to the communicative default. Also,
earlier narrative forms are more preoccupied with mimicking the ‘somebody
told me’ or ‘I saw it’ pretense, while less narrator-mediated forms emerge
much later. Furthermore, visual prompts participate in storytelling alongside
words, and often in ways which not only support words, but sometimes give
them different meanings.

1.5 Why is fiction special?

Or is it? There are several recent attempts to address the question. Zunshine
(2006) sees reading fiction as practice for our theory of mind, assuming, with
Palmer (2004), that the access to characters’ minds is more important than the
story itself. Oatley and his co-workers,8 for comparison, argue for some kind of
cognitive gain, especially if the material read is sophisticated and has some
psychological depth. Similar claims, though based in an evolutionary view,
have been made in Boyd (2009), who argues for the development of fiction as
driven by cognitive play. Researchers in the cognition of art look at fiction as a
form of art, and try to view it alongside the path leading from cave drawings to
Van Gogh. Others seek the value of fiction in entertainment, in crossing the
boundaries of time and space, in exercising compassion, et cetera. However,
there are many questions which are not addressed through these enquiries.
First, why do people write fiction? The view from neuroscience, lucidly
presented by Flaherty (2004), suggests that creative urges to write are to a degree
a part of our biology. From the cultural perspective, there has never been a
problem with low supply of aspiring writers, though we do not seem to
question the fact that only some writers make a difference to our view of the
world and have pushed the boundaries of narrative and poetic skills to new
heights. There is plenty of fiction which seems no more than chewing gum for
the eyes, and yet finds plenty of committed readers, and definitely counts as
literature. There is also the new phenomenon of writing as a job. When Haruki
Murakami, a very prolific and popular author, talks about his career, he simply
22 Language and literary narratives

says that he thought at some point in his life that he could make a living by
being a writer (Murakami and Gabriel 2008). My point is that while we tend to
focus on the writers who make a difference and build on that, a lot of story-
telling seems to have much less sophisticated forms or less high-flown goals.
It is a form of linguistic expression which builds on culturally supported
cognitive needs.
Furthermore, we ‘talk fiction’ much more than we are willing to admit, and
our languages have elaborate structures allowing speakers to explore different
scenarios. Conditionals and other hypothetical constructions in languages are
varied and complex, no less complex than a fictional story, if not more, simply
because of the compactness of actual constructions. There are many ways for
speakers to mark evidentiality, epistemic stance, or emotional stance, and most
of these linguistic forms are either lexicalized, or grammaticalized, or construc-
tionally determined. Linguistic forms are typically understood to mean more
than the facts described (consider implicature, linguistic politeness, indirect
speech acts, and the like), and open themselves to reasoning more often than just
stating the facts (hence so many attempts to treat language as based on some
kind of logic, however misguided some of these attempts may be). My point is
that the idea that colloquial language use is about facts while fiction is not is
simply wrong. Language is very well equipped to deal with ‘what ifs’ and
hidden meanings and is usually very clear in representing the speaker’s attitude,
knowledge and the source of information. Contrary to common belief, state-
ments that logicians love to work with, such as Grass is green or Mary is in
London, are much less common than ones starting with if, I think, he said, or
I wish.
The question is, then, why treat fiction as . . ., well, pure fiction? In his
delightful personal account of narrative techniques, Vargas Llosa (1997/2003)
talks about the author as resembling “the Catoblepas” – a mythical creature
which consumes itself, starting from the feet. As he argues, the themes an author
will explore have to emerge from his own experience, though he can then give
them any narrative form he chooses.
There is thus a grain of truth in any fictional narrative, however weird it may
be. Setting the case of realism in fiction aside, genres departing from ‘realism’
also are real to a degree. Even a cursory look at the ‘ultimately’ fictional genre of
science fiction shows that while trying to construct storyworlds as remote from
planet Earth as they can, authors cannot but rely on our reality in many of its
aspects. At the other end of the spectrum, texts which are treated as novels but
have been described (by their authors as well) as primarily autobiographical
are versions of actual happenings. A perfect example of this phenomenon is
J. G. Ballard’s acclaimed novel Empire of the Sun, which is a blend of fact and
fiction. As a child, Ballard did spend three years in an internment camp in
China, and the novel is based on his story. In his autobiography, The Miracles of
 1.6 Narrative and grounding 23

Life, Ballard claims that most of the events described in the book happened in
reality, though not to the people actually portrayed in the novel. He also explains
the reasons for his choices, for example, for placing his character, Jim, in the
camp without his parents, while in fact they were there with him. Ballard wanted
to focus on a child’s experience without the protective layer which could only
blur the message – a choice as much related to compression and narrative
salience as the one in the cannibalism story. Fiction thus cannot be divorced
from facts, some as basic as the existence of oceans, dogs, and bacteria, and
some as specific as the conditions in internment camps in China during World
War II.
Furthermore, the need to categorize any given story as either true or not true
seems to be important for cultural rather than linguistic reasons. It is important
to know whether, let’s say, a politician accepted a bribe or not, because it is a
legal issue. But whether a man named Darcy actually married a woman named
Elizabeth Bennett, or whether they never met, or never even existed, is not
relevant to the story. Pride and Prejudice is not a true account of what happened
(as far as we know), but it can be read as such and may not ‘feel’ different in
terms of factual plausibility from a story about distant cousins one can hear from
one’s grandmother. Crucially, the latter story would have a different form: it
would be shorter and use colloquial linguistic constructions. It thus matters little
whether the story presents facts, but it is crucially important how it is told, and
here is where the issue of literary discourse becomes interesting.
To conclude, fiction has to be based on reality to have its effect. Reading
biographies is often as moving and interesting as reading fiction, because the
textual matter and the attendant constructed meaning or emotional response is
triggered through the same medium and the same conceptual mechanisms.
Uncovering the linguistic specificity of the medium and the conceptual mech-
anisms leading to narrative uptake is the goal of this book.

1.6 Narrative and grounding

One of the themes in the discussion of narratives which this book engages with
is the cognitive status of narrativity. While definitions of narratives vary, they
consistently stress temporality, causality, representation of minds, relationship
to spatially and temporally defined storyworlds, problem solving, and the
construction/resolution of some conflict (see Ryan 2007 for an interesting
discussion). From the cognitive perspective, Herman (2003a) argues further
that narratives are ‘tools for thinking,’ which help in the organization of
experience, through the establishment of chunks of experience and causal
links among them, recognition of types of events, et cetera. These lists include
some more general concepts such as causality, and some which are narrative-
specific, such as the ‘storyworld’ or ‘conflict.’ And yet there are also attempts
24 Language and literary narratives

to reach to the very conceptual core of the narrative and establish the role of
narrative structures in thought processes in general. One such attempt is the
book by Turner (1996), which sees basic narrative organization as a primitive
concept and the source of the emergence of both linguistic and literary forms of
expression. As Turner argues, perception of basic ‘spatial stories’ such as
opening the door or throwing a rock is the starting point of higher levels of
conceptualization. In this section, my goal is not to arrive at some definition of
narrative which would reconcile all these descriptions, but to consider some
problems in relating narrativity to conceptualization.
Herman’s approach (2002, 2003a) identifies aspects of narrative which
participate in conceptualization of events as linked into a story. Generally,
Herman sees the story as a composite concept which provides a link between
raw experience and its ‘storylike’ conceptualization. Turner, for comparison,
views ‘storiness’ as a basic perceptual tool, leading to conceptualization of all
kinds of experience as well as to projections of one ‘story’ onto another, leading
to metaphor, parable, and more complex narrative forms. Essentially, Turner’s
concept of a primitive narrative is a configuration of image-schematic concepts
rooted in sensorimotor experience in which (causal) force or sequence plays a
central role. Turner’s approach to narrative, and, by the same token, the entire
theory of conceptual metaphor and blending, was extensively criticized in van
Oort (2003), primarily on the grounds of what he perceives as the assumed
direct link between perception and conceptualization. Since the issue, which I
will refer to as the ‘grounding problem’ (after Harnad 1990), is clearly at the
core of much of the current discussion of the narrative, I will attempt to review
the questions briefly.
In much of the work in cognitive linguistics it is assumed that meaning is
grounded in perception and sensorimotor experience.9 Basic spatial configu-
rations called image schemas (such as containment, force, balance, bounded-
ness, et cetera) lead to the emergence of linguistic expressions describing
concrete aspects of experience, but can then be projected into abstract domains
of human activity. Thus most concrete categories as well as many abstract
meanings are grounded in human perception and experience, via levels of
projections which also distinguish multiple meanings of lexical items. For
example, the preposition in can be used to describe concrete cases of contain-
ment, but also abstract states such as being in despair. Van Oort criticizes this
explanation by arguing that it assumes a direct connection between percep-
tion and linguistic categorization, while what is needed is an explicit inclusion
of a separate intervening level of symbolic representation. Van Oort refers
extensively to Deacon (1997), who sees symbolic representation as the core of
human cognition. He also evokes the claim that words, as symbols, are
primarily in relationship to other words, and not to the speaker’s perceptions
or actions. Finally, van Oort reads Turner’s discussion of the basic role of
 1.6 Narrative and grounding 25

narrative as a hidden claim of the ‘story’ being an innate concept, very much
like Chomsky’s universal grammar.
It might be helpful to put this discussion in the context of recent experimental
work in cognitive psychology. Surely, the idea of representation and catego-
rization is at the centre of the discipline’s interest, and some of the research done
in the last few years explicitly addresses the grounding problem – what does it
mean for perceptions to drive conceptualization? how are categories formed
in the brain? how do abstract concepts emerge? First of all, research starting
as early as the 1990s (Barsalou 1999; Harnad 1990) suggests that treating
meaning in terms of links within a network of abstract symbols in the mind
runs into problems. One, the approach does not explain the observed emergence
of arbitrary symbols from experience and perception, and, two, it offers no
solution to the commonly asked question about the ways in which symbols
connect to reality, since without some connection to the real world, symbols
cannot perform any cognitive function. Experiments which ensue suggest that
perception and action are central to higher cognition, without postulating
intervening mental representations.10 Moreover, one of the claims is that expe-
rience can be used in the conceptualization of situations in other times and
places, and the idea led to further investigation of how basic perceptions and
actions can participate in conceptualizations without necessarily going through
the configurations of arbitrary symbols.
There seem to be many reasons not to treat discrete symbolic representation
of categories as the basis on which meaning emerges. Recent work by Prinz
(2002, 2004, 2005) makes a very strong case for the experientially grounded
emergence of abstract concepts, including emotions. First, Prinz shows how
even complex concepts such as ownership can be naturally explained through
configurations of basic perceptual and sensorimotor concepts. Second, he sees
concepts as necessary for action, and not for static categorizing, so that con-
ceptual representations are variable and context-sensitive rather than participat-
ing in a stable system of symbols. Furthermore, experiments reported in Spivey,
Richardson, and Gonzalez-Marquez (2005) show the smooth and pervasive
interface between on-line language use and sensorimotor processes. Broadly
speaking, research reported, among others, in Spivey (2007) and Pecher and
Zwaan (2005a, 2005b) supports some early suggestions made in Gibbs and
Matlock (1999) that tracing complete conceptualizations in the brain may not be
possible because of the dynamic and continuous nature of such processes.
At the same time, the research gives experimental support to two major claims
of cognitive linguistic theories – one, that meanings emerge out of the basic
levels of embodiment, and, two, that meaning is constructed on-line, for the
needs of the specific act in which the conceptualizer is immersed, which does
not preclude conceptualizations of situations not directly available in embodied
experience. In other words, meaning is embodied and dynamic, even when what
26 Language and literary narratives

is being conceptualized is abstract, decoupled from the circumstances of the act

of conceptualization, or leading primarily to an emotional response.
Crucially, theories of embodied meaning do not require that arguing for a
connection between experience and conceptualization puts one automatically
on the side of extreme empiricism. While the authors who argue for embodied
roots of abstract thought or emotions from a philosophical (rather than psycho-
logical) perspective (Johnson 1987, 2007; Lakoff and Johnson 1980, 1999;
Damasio 1999, 2003; and Prinz 2002, 2004) agree that mind/body dualism is a
convenient fiction, they are centrally interested in the emergence of conceptu-
alizations, rather than in defending some form of empiricism. Even when
experimental work confirms the role of image schemas in higher cognition,
what is properly being described is the role of basic spatial configurations
emerging from experience in concept formation. In other words, the upshot of
the research is explaining the reason why an abstract word like status evokes the
upper end of a vertical scale and a spatial expression like up top does too. Here is
where van Oort appears to misinterpret the cognitive paradigm, by assuming
that what is being said is that a use of a spatial preposition to describe a spatial
configuration relies directly on perception without symbolic conceptualization,
while its abstract meaning does not, whereas what is argued is that all kinds of
meanings emerge on the basis of perception concepts.
It is thus indeed possible to argue that basic spatially and temporally situated
actions can serve as sources in conceptualization of abstract or displaced
concepts, such as the ones underlying basic narratives – which has been
Turner’s claim, formulated at a level closer to language than to cognition. The
claim does not need to involve any hidden or explicit assumptions of innateness,
because conceptualizations which emerge are not static or determinate enough
to be candidates for such status. In general, claims of an innate mechanism
related to language are not corroborated by experimental research. Second,
while the symbolic nature of language is beyond dispute, the participation of
symbols in various relationships with other symbols does not have to assume
either discreteness or systematic correlations (such as binary oppositions,
semantic features, et cetera). Also, our model of meaning has to capture the
processes whereby new meanings emerge out of contextually prompted combi-
nations of old meanings, and frameworks such as construction grammar and
blending have made significant progress in that direction. Perhaps most impor-
tant, though, treating symbols only in opposition to indexes and icons misses
the point, for it creates an impression that only indexes and icons are linked
directly to experience. While they have their own roles to play, symbolic
concepts, as has been amply demonstrated, rely even more specifically on
embodied experience. In other words, symbols are not detached from the real
world; on the contrary, they are immersed in it via their reliance on image
schemas and action. Categories and symbolic concepts are thus related to other
 1.6 Narrative and grounding 27

categories and concepts not on the basis of the system of related meanings,
but on the basis of underlying shared experiential concepts. To give just one
example, if Spivey’s eye-tracking experiments reveal minute eye-movement
upwards as a shared and consistent response to words such as up, tall, and
respect (in clear contrast to down or submit), the connection thus revealed is
among various applications of the basic image schema of verticality, from
movement or location, through scalar measurement specific to cultural norms
(a tall pygmy is a short Masai), to a social relationship. As an additional
bonus, the proposed experiential treatment of meaning does not exclude
cultural differences from the actual conceptualizations.
Returning now to the concept of ‘narrative,’ we might say that in the context
of the discussion above, it is not substantially different from any other form of
conceptualization in that it relies on the dynamic emergence of representations
prompted by the text (in terms of both the events portrayed and the emotional
response to those events), and that it relies on embodied cognition. The speci-
ficity of how narrative chains of events are held together has much to do with the
processes identified by Herman (2003a), but these processes are also essentially
supported by experiential concepts. If I think of a portion of a story as ‘giving
offence,’ I have to rely on the force-dynamic interaction between the partici-
pants, the emotional impact of their words and actions, the need to restore
balance, et cetera. The fact that the events to be conceptualized are fictitious is
not an obstacle to conceptualization. Naturally, narratives make heavy demands
on cultural knowledge, memory, empathy, and other aspects of cognition, but
they are not encumbered by the need to refer to discrete symbolic representa-
tions as independent steps in the process.
Turner’s claim about the centrality of narrative thought and projection to
conceptualization of any kind should not be read to suggest some evolutionary
process which led to the emergence of the narrative module in the mind. In fact,
claims about the evolutionary history of specific human abilities are becoming
common. Since Fodor’s early work (1983), followed by the Chomskyan para-
digm in linguistics and Pinker’s (1994) ‘language instinct,’ the search for
modules in the brain to explain essentially human abilities seems to continue,
even if jocularly. For example, a recent book by Dutton (2009), The Art Instinct,
adds to the ‘instinct’ metaphor by considering art as a specifically human trait,
doubtless originating in some evolutionary event or other. While any discussion
of whether such innate modules in general are indeed a defensible explanation
of the central aspects of human cognition is beyond the scope of this book, there
is clearly no need to postulate a narrative module.
Is there, then, any connection between basic narratives and evolution? While
some recent work (Carroll 1995, 2005; Gottschall and Wilson 2005) connects
evolution and literature, as well as evolution and art (Boyd 2005), these
researchers make no claims about narrativity as such, and yet ‘narrative
28 Language and literary narratives

thought’ seems to pervade conceptualization. And as van Oort suggests,

Turner’s idea of the ‘literary mind’ might be read as saying something similar.
But while the pervasiveness of narrative is a fact, I hope to have shown that there
is no need to postulate a narrative module or ‘instinct.’ Narrative-based con-
ceptualizations may be indispensable (so I tentatively agree with Turner here),
but they are readily explicable in terms of lower levels of embodied cognition.
Also, while language is naturally treated as the basic environment in which
stories exist, it is not indispensable to narrativity, since a story can be ‘told’
through visual means. The concepts required for a narrative to emerge (sequen-
tiality, causation, chunks of experience, cultural framing of such chunks, image-
schematic force-gestalts of conflict and restored balance, et cetera) are the same
concepts which are required for other language conceptualizations to emerge,
so we should perhaps advocate a stronger claim, such that narrative form relies
on a specific cluster of such concepts.
Moreover, recent work in both cognitive psychology and evolutionary theory
allows one to talk in a sufficiently rigorous manner about human abilities which
are not only cognitively central, but also culture-based. The work by Donald
and Deacon mentioned above speaks of ‘coevolution,’ and the idea is given a
more specific meaning in the theory of gene-culture co-evolution.11 The focus
on cognitive sources of culture can at least suggest ways in which a centrally
cultural non-primitive concept such as the ‘narrative’ may have become crucial
to higher cognition, while remaining reliant on more primitive image schemas
of force, sequence, or balance, on specialized language forms (extended struc-
tures which connect sentences in a variety of complex ways), and on the
culture’s need to represent events and cultural norms in forms which guarantee
their temporally extended availability.
This discussion has not yet addressed a related but largely independent
question concerning the cultural and linguistic emergence of narrative fiction.
Zunshine’s work (2006) locates the explanation in evolutionary psychology and
more specifically in the so-called ‘theory of mind’.12 The term is best known in
the context of Baron-Cohen’s (1995) generally accepted explanation of com-
municative deficiencies summarized in the term ‘Autism Spectrum Disorder.’
At least some of the difficulties of autistic speakers are attributed to their lack of
an understanding of how language and behavior can be read in order to grasp
people’s mental and emotional states and adjust one’s language and behavior to
meet their expectations. Zunshine builds on this ability to ‘read’ other minds to
argue that fiction offers a form of ‘exercise’ in using it.
There are some problems in such a use of the concept. First of all, ‘mind-
reading’ includes (and is tested with the use of) understanding of communica-
tive components such as eye contact, facial expression, or body posture, and
not exclusively language. It is also being used in explaining a variety of
behaviors, including appreciation of humor, irony, and figurative language.
 1.7 Approaching narratives 29

Reading cannot test all these skills, or serve as practice, because it relies on a
narrative rendering, not lived experience, and is mediated entirely through
language. If there is any disposition commonly associated with good mind-
reading which can also be applied to fiction, it is the ability to grasp the
emotional impact of people’s words and actions. Interestingly, appreciation of
mental and emotional acts has already been claimed to be the single most salient
feature distinguishing narrative fiction from other narratives. Palmer (2003)
argues that ‘fictional minds’ are the core of narrative discourse, even if narration
presents mostly actions – this is what fiction is about, and the plot is in fact
secondary to this primary focus; a similar point has also been made in the
philosophical context by Currie (2007). I agree with Palmer because his obser-
vations elucidate the role of much of what needs to be narrated in a novel, but
also because they explain how fiction stimulates an emotional response in the
reader without relying on the whole constellation of abilities roughly summar-
ized under the rubric of ‘theory of mind.’ Palmer’s claims can also be supported
with recent research on the role of embodiment in representing mental states
(cf. Zlatev et al. 2008), and I will return to the question in Chapter 4.
However, accepting Palmer’s argument does not require a sharp distinction
between, let’s say, narratives in oral traditions and narrative fiction as we
know it. Any narrative form assumes the readers’ or listeners’ ability to feel
for the characters and interpret the events from the characters’ point of view – to
feel sympathy, envy, disgust, anger, et cetera. While some of these emotional
responses may be complex, or even conflicted, they are better explained through
our ability to feel empathy, and also to attribute emotional meaning to linguistic
expressions – even simple experiential concepts such as up (alert, conscious,
confident, in good mood) or down (tired, inactive, sick, unconscious, dejected).
To conclude, I will treat extended fictional narratives – the primary focus of
the project – as relying on our abilities to construct meaning and respond to it
emotionally. But I will also assume that these abilities emerge, more or less in
the ways described by Tomasello (2008), from the simplest linguistic forms
based in embodied cognition, through more and more complex narrative forms,
all the way to the complexities of contemporary stories. I will also argue, mainly
in Chapter 6, that narrative ways of representing mental and emotional states of
characters have evolved through literary periods and genres towards a more and
more direct, and thus narratively more complex, access to fictional minds.

1.7 Approaching narratives

As the discussion above has indicated, narratives and narrative fiction have been
subject to enquiry in a number of different disciplines, from psychology and
cognitive science, through philosophy, literary criticism, stylistics and discourse
analysis, to linguistics and narratology. Each of the disciplines contributes
30 Language and literary narratives

much to the discussion, while also having its own goals and methods. Where
does this project locate itself then?
The simplest answer would be – towards the linguistic and cognitive end of
the spectrum. I want to propose an analysis which is informed by current
research in cognitive science and which clearly addresses the linguistic phe-
nomena specific to extended narrative texts. I will refer repeatedly to the
concepts proposed by narratologists, but I also want to put this discussion in a
context which is more open and more restricted at the same time, as it seeks to
explain processes of conceptualization by looking at the fine-grained linguistic
prompts for these conceptualizations. At the same time, the tools proposed
should be useful in a discussion within stylistics, where language is discussed in
terms of specific textual choices and their impact on the special quality of the
individual text or genre. Like some previous work in cognitive discourse
analysis (Emmott 1997), this project will also attempt to look at features of
textual matter which contribute to text-wide aspects of interpretation.
The hardest relationship to define is between this work and literary criticism.
While I hope to open the linguistic tools to the role of cultural values or
historical facts, I cannot make them central to the discussion, and thus much
of the interests of literary critics will be downplayed. However, my interest in
paths of interpretation is what critics should recognize as relevant to their field.
There is also the question of pedagogy. In my work with students of literature,
I found that introducing them to the concept of a frame or an image schema
often gave rise to new and innovative interpretations, not because the concept
could be found in one text or another (it would probably be found in any), but
because of the special salience of some basic cognitive configuration in a text. In
other words, they noticed things that required interpreting because they were
highlighted once a cognitive concept became available. So cognitive analysis
did not seem to restrict the playing field, it in fact opened it up.
This book will thus focus on the link between language and cognition and the
ways in which such correlations affect the construction of narrative meaning.
First, I will outline the ways in which tools developed in the theory of con-
ceptual integration can be applied to longer texts. I will start the discussion of
narratives with situating the text-wide construal of narratorship in the concep-
tual integration model. The central part of the book will focus on discussion of
viewpoint, looking at some ways in which sentence-level lexical choices and
referential expressions affect narrative viewpoint. I will then move on to the role
of representation of fictional minds in various storytelling genres, including the
discussion of various types of discourse in the narrative, with special focus on
the role of direct speech and thought. Finally, I will resume more general
discussion in the concluding chapter.
2 Blending, narrative spaces, and the
emergent story

When he woke up, the dinosaur was still there.

(The Dinosaur, Augusto Monterroso)

Stories can, naturally, be treated as linguistic constructions. They are attested

examples of natural written discourse, and, while it is possible to treat linguistic
choices in the narrative as primarily rhetorical, there is also crucial reliance on
very basic facts of language structure. Stories should also be talked about in
terms of meaning. They are particularly challenging in this respect, if only
because of the sheer length of the text and the variety of narrative options
available. This project proposes a framework which makes the discussion of the
meaning of stories possible. It relies heavily on analytic tools established within
the discipline of cognitive linguistics (mental spaces theory, construction gram-
mar, and frame theory), but it is primarily focused on the theory of blending, or
conceptual integration.
Fictional narratives rely on a variety of linguistic phenomena, but adapt
them to their needs. To give just one example, they use personal pronouns
and deixis in ways which exploit the default spoken setting, but set it against
a different kind of deictic ground. Thus a pronoun like I is often used in novels
to represent the narrator – a narrative construct, rather than an actual com-
municator. The discourse setting is that of a communicative act occurring
across a spatial and temporal distance, with no clear addressee available
(other than some generic reader). The consequences of such a set-up are far-
reaching, as they create the need for a different treatment of referential
expressions designating characters and put the concepts such as the ‘present’
or the ‘past’ in a different context. These peculiarities will be discussed
throughout this book.
With respect to meaning complexities of fictional narratives, blending
(cf. Fauconnier and Turner 1998a, 1998b, 2002) seems to be particularly
well-suited to proposing an account of narrative structure. The concepts and
mechanisms proposed in blending work translate naturally into the needs of an
account of narratives, but also open the door to the discussion of blending
mechanisms and properties which seem specific to the emergence of meaning

31
32 Blending, narrative spaces, and the emergent story

of longer texts. In what follows, I will elaborate the blending framework in ways
which best respond to the needs of accounting for narrative discourse.

2.1 Applying blending to fictional narratives

The essential claim of blending is that conceptual packets (mental spaces or
frames, here called inputs) can be activated in a person’s mind by verbal as well
as nonverbal prompts and integrated into a new conceptual configuration (called
a blend) to construct new meanings as discourse develops. The conceptual
structure of the blend, while relying on projection from the inputs, achieves
its own coherence by selecting only the relevant parts of the inputs and by
compressing vital relations, such as time, causation, or identity. For example,
while the inputs are two mental spaces one of which is present and the other
past, the temporal distance between them may be compressed to yield a blend
which is either past or present. The inputs remain activated and retain their
separate conceptual status, but a specific blend may also project structure back
to the inputs (backward projection), and thus reconstrue them for the purposes
of the ongoing exchange. The blend is characterized by its own structure
(emergent structure), but can then become an input to another blend, or series
of blends. While blending can account for ongoing processes of meaning
construction (as in following the course of a conversation or reading a comic
strip and accumulating content until the joke gels in the final frame), they can
also become entrenched as new expressions, such as compounds or idioms.
Blending also accounts naturally for multimodal contexts, where visual, lin-
guistic, and aural prompts work in combination, as in film.
The way the blend emerges is largely dependent on the topology (internal
structure) of the inputs. While in a number of cases inputs are mental spaces,
they may also be structured by frames (as introduced in Fillmore 1985, 2006;
Lehrer and Kittay 1992). For example, a blended construction such as When
I learn to surf, I will move to California assumes the availability of two mental
spaces: the future space in which I am a surfer and I move to California, and the
present reality space where it is not the case. Each of these constructs has the
structure of a mental space, with participants, temporal and spatial features, and
internal structure (such as a connection between California coast and surfing
opportunities), and they yield a blended space wherein being a surfer in
California also retains other aspects of my identity: family ties, life story, et
cetera. Crucially, the imagined future situation remains a mental space precisely
because it has been constructed on-line, that is, within ongoing discourse, to
indulge in an ‘alternative life’ for the duration of the conversation about surfing,
and it will be deactivated when the discussion of perfecting surfing skills stops
or when the topic is changed.
 2.1 Applying blending to fictional narratives 33

Frames are significantly more stable, in that they structure our understanding
of reality beyond the conversational context and are often associated with
specific lexical items. While we all have the general concept of dying, we
apply very rich cultural framing to concepts such as killing, murder, infant
mortality, accidental death, or genocide. Each of these includes the fact of a
person or persons dying, but they also include complex scenarios specifying the
circumstances, causes, the degree of agency, cultural and social realities, et
cetera. We can thus say that they participate in different frames, and that these
frames are associated with assignment of guilt, legal consequences of guilt,
moral values, medical problems, et cetera. Using any of the expressions listed
above evokes all the associated framing, and, by the same token, appropriate
social actions.
The most important aspect of framing is the possibility of accessing the
entire frame when only one aspect of it is mentioned. In the most obvious
case, the discussion of buying a new car automatically implies the aspects of the
frame which are not explicitly mentioned: selling, transfer, object transferred,
price of transfer, et cetera. In the less obvious cases, when an aspect of the frame
is chosen in a less predictable manner, we can also talk about frame metonymy –
linguistic usage whereby an expression is used to stand for an associated aspect
of the frame, while also evoking the entire frame. Cases of such pragmatically
established metonymies were discussed, among others, in Fauconnier (1994
[1985], 1997). Fauconnier analyzes examples such as Room 4 has a visitor,
which can be said by a member of hospital staff to refer to the patient in room 4
and announce that a visitor has arrived. In Dancygier and Sweetser (2005), the
concept of frame metonymy has been expanded to include constructional
frames, so that, for example, a specific use of the present tense to mark an
unpredicted premise in a predictive reasoning signals predictive meaning even
if no other constructional features are marked. Thus, a sentence such as You miss
one more meeting, you’ll be fired counts as a prediction of dire consequences if
an as-yet-unpredicted event of being late occurs, without using a conditional
predictive construction in its entirety – the verb forms are enough for the frame
to be evoked.
In recent work (among others, Dancygier 2009), I argued that frame meton-
ymy at the lexical level, when paired with frame metonymy at the constructional
level (termed constructional compositionality), creates blended constructional
patterns in which meaning emerges based on the specific blend between the
evoked lexical frame and the constructional frame. For example, a sentence
such as Iraq is a new Vietnam seems to follow the format of a copular
construction in which the subject is a specified entity and the predicate is a
non-specific name of a category (as in John is a teacher), but both proper names
are used to evoke metonymically frames of US interventions and extended
military conflicts (rather than ‘countries’), and thus the adjective new signals the
34 Blending, narrative spaces, and the emergent story

transfer of some of the framing of Vietnam (high cost, lack of success) into the
framing of Iraq. As a result, Vietnam does not become a common noun (even
though the frame is detached from its standard referential meaning and an
indefinite article is used), while Iraq counts as metonymic for the Iraq war,
rather than the country. The interaction among the frames evoked by both
proper names, the adjective new, and the construction yields a blended meaning
not available outside of the construction. I will discuss framing and reference in
more detail in Chapter 5, but the connection between frames and blends has thus
been briefly illustrated, and will be relied on in further discussion.
A simple example of a blend is a compound expression such as computer
literacy. The two inputs are frames of being (one) a computer user and (two) a
person able to read and write. The frames are rich culturally. While literacy used
to be considered the cornerstone of participation in contemporary society,
computer use until recently was not considered an indispensable skill. The
compound expression blends the two concepts to refer to a set of computer
skills which are now considered essential. They are not restricted to reading and
writing, but involve also an understanding of basic computer functions. While
the ‘computer’ input contributes the understanding of which skills are basic, the
‘literacy’ input further qualifies those skills as required of anyone who wants to
be considered minimally educated. The blend also informs both inputs, by
defining computer skills as basic, and expanding the scope of literacy to
match contemporary needs. Interestingly, the expression is then cleverly
reframed in the name of a well-known Silicon Valley institution – Computer
Literacy Bookshop, where the expression computer literacy refers to interaction
with written language again, even though only computer professionals are the
intended readers. It is thus a good example of how subsequent blends manip-
ulate both the inputs and the earlier emergent meanings.
A more complex blend is prompted by a clausal construction such as If I were
your mother, I would teach you how to cook. It sets up an imagined situation
which is a blend of the if-space (the speaker is the hearer’s mother) and the
reality in which the speaker and the hearer participate (the hearer cannot cook,
her mother never taught her, et cetera.). The emergent structure also creates a
new imagined identity blend, wherein (under one of the possible interpretations)
the hearer’s mother shares the commitment of the speaker to teach the hearer to
cook. The blend is complex, but the mechanism that allows it to emerge is the
same simple mechanism, and, crucially to the subject of this book, operates on
inputs which are rich in contextual information, feature various participants, and
present events and human dispositions as causing other events and dispositions.
We can assume that to the speaker and hearer of the conditional all the
contextual implications are known – the desires and personalities of the partici-
pants, the reasons why knowing how to cook is useful, and the reasons why the
speaker cares about the hearer’s skills.
 2.2 Narrative spaces as mental spaces 35

Similar meaning construction processes occur in longer narratives. Frames

and mental spaces structure inputs, which then become integrated, possibly in
ways specific to a reader, into the emergent blend. The process continues
throughout reading, until the complete blend of the story emerges. In what
follows, I will consider some details of the process.

2.2 Narrative spaces as mental spaces

Mental spaces were defined (Fauconnier 1994[1985], 1997) as conceptual
packets set up primarily by linguistic expressions (though visual or gestural
prompts are also common), and being used in the on-line mental processes of
meaning construction. Thus a mental space may be a recall of a past event in the
course of a conversation about the current situation. For example, if the speaker
says I was really tired last night, but I’m better now so do tell me what you think
about my book she calls up a past space in the first clause, describes her present
state in the next, and evokes the spaces of the content of the book, the content of
the hearer’s thoughts regarding the book, and the hearer’s expected conversa-
tional turn. All these are mental spaces, activated for the duration of this part of
the conversation, and then becoming latent until evoked again, when, let’s say,
the speaker later complains about the hearer’s bitter criticism to someone else. It
is clear even from this example that mental spaces are constructs of varying
complexity (from tiredness to the content of an entire book), attributable to
various conceptualizers (here, the speaker or the hearer), and including mental
representations of concrete objects, conversational moves, opinions, et cetera. It
is often asked whether it is possible to confirm the presence of mental spaces in
the brain, but the question seems to misconstrue the usefulness of the concept.
Even if we could monitor someone’s spontaneous engagement in a conversation
and discover which areas of the brain are involved (some have to be!), we could
at best capture some of the brain structure relevant to the content of the
conversation, but apparently not the thought processes prompted by the dis-
course. Also, as Spivey argues (2007), the flow of thought, when traceable in the
brain, creates a trajectory which maintains a number of open options through-
out – very much in agreement with what mental space theorists claim.
Mental space analyses have shown how mental space networks can represent
the on-line correlation between choice of expression and the resulting mean-
ing.1 But mental spaces are not a strictly linguistic phenomenon, as recent work
on gesture, art, comics, theatre, or film shows;2 rather, the theory is a useful tool
in teasing out the details of various forms of meaningful interaction and
discovering the correlations between such prompts and meaning.
The above discussion of mental spaces and blending makes it clear that
the framework has been mostly used to elucidate the emergence of meaning
of expressions, discourse fragments, or pictorial/verbal artifacts. Fictional
36 Blending, narrative spaces, and the emergent story

narratives, for comparison, are much more complex and also simply longer. It
does not seem possible to provide a clear explanation of how individual
sentences may be subjected to the overall storytelling goals. To address this
concern, I will argue that narrative construction proceeds through several
levels of narrative organization, involves an emergence of partial narrative
structures called narrative spaces, and is primarily organized by various forms
of narrative viewpoint.
A narrative space has most of the features of a mental space.3 It is set up
through language expressions, often ones not unlike ordinary space builders –
such as Once upon a time . . . or In a hole in the ground there lived a hobbit . . .
and further structured by grammatical choices such as person or tense. It has its
topology, such as space, time, cultural norms, or language spoken. It also has
participants, some directly involved in the telling, called narrators, some being
directly affected by the events, called characters, though there are also partic-
ipants (intradiegetic narrators) who play both roles at the same time. Like
mental spaces, narrative spaces can be elaborated in much detail, while others
are barely mentioned. Both types of spaces are structured by viewpoint phe-
nomena, though narrative spaces seem to feature more elaborate constructions
of viewpoint, to be discussed throughout this book. To sum up, a narrative space
is a mental construct participating in the emergence of the story, having
distinctive topology and narrative status, and linked to other narrative spaces
in ways which prompt story construction.
Narrative spaces are different from other ‘worldlike’ concepts such as possi-
ble worlds (as applied to narrative) which are discussed in relation to the ‘real’
world as its alternatives.4 Narrative spaces are defined in parallel to and in
correlation with lower-level linguistic phenomena participating in meaning
construction, and, while they may apply to ‘worldlike’ constructs, they do not
apply to texts in their entirety. An emergent story results from the blending of all
of the text’s narrative spaces. The structure of narrative spaces is thus deter-
mined by the specific organization of the text, and no claims are made about
their necessary completeness or coherence.
The concept of a mental or narrative space also shares some features with the
construct which has been termed a ‘text world’ (introduced by Werth 1999, and
developed by Gavins 2003, 2007). Text world theory shares some of its
assumptions with the mental spaces framework – for example, it also talks
about linguistic forms being used to prompt conceptual constructs and accu-
mulating as discourse progresses. It is primarily a theory of the interpretation of
discourse, and has been applied to a number of discourse contexts – verbal and
nonverbal – including the context of the narrative. The primary difference is
possibly that the text world theory is much more text-driven than the mental-
spaces-and-blending framework, while also being more focused on the analysis
of discourse, rather than the elucidation of linguistic concepts. Mental spaces
 2.2 Narrative spaces as mental spaces 37

theory has originally been conceived as a theory of meaning, and, while it is

developing its textual applications, it still relies more heavily on semantics than
on discourse. Also, mental spaces analyses are naturally extended into work on
blending. I will be relying on mental spaces specifically, but I assume most
claims can be ‘translated’ into the text world approach.
A narrative space is thus a construct which is set up through linguistic means
and continues being elaborated through some parts of the text (possibly all). It is
also subsequently enriched through blending and gradually starts functioning in
the network leading to an emergent story. In these respects it is thus similar to
a mental space, which participates in extended discourse. What constitutes a
crucial difference, however, is the nature of the discourse, since a narrative
requires that its primary spaces be maintained and elaborated until the comple-
tion of the reading process, until their role in the text is fulfilled. To give a simple
example, if a flashback space is set up, it can then play different narrative roles,
not necessarily dependent on time, though in each case it would constitute an
independent, temporally distinguished space. Setting up such a space can
constitute background to the main story, or the beginning of the story proper,
or a digression which explains some form of behavior. In any such instance the
space will be incorporated into the story, though not only temporally.

2.2.1 Sentential level versus textual level

It may seem counterintuitive to claim that a narrative space may in some cases
take the form of a sentence, but there is no reason to draw a specific line at any
level of linguistic structure, because what matters is the narrative structure. For
example, the text which won the contest for the shortest story is also a sentence.
It was composed by Augusto Monterroso, and it is entitled The Dinosaur:
(1) When he woke up, the dinosaur was still there.
The story has all the ingredients necessary, but asks the reader to do much of
the story-construction work. There is a narrator (probably omniscient), and a
character, plus a dinosaur. The when clause profiles a temporal location wherein
the character awakens, which suggests that he was asleep before (in Spanish it
might have also been a ‘she’). The word still in the main clause further suggests
that before the character fell asleep, the dinosaur was also present. There is thus
a sequence of events, there are characters in some relationship, there is a
narrator, and there is suspense – how come there is a dinosaur in the presence
of a human being? What has often been called ‘suspense,’ in my view, is the
need to engage with the text and have the intention to complete the gaps in the
story. In other words, the reader is required to use extra-textual and linguistic
knowledge to complete the emergent structure of the blend being prompted by
the narrative inputs.
38 Blending, narrative spaces, and the emergent story

In the case of Monterroso’s story, there are in fact several options. The
dinosaur could be a real T. rex or triceratops, and then the story qualifies as
science fiction, and the character is possibly in grave danger (hence the fact that
he could sleep in the presence of a monster is intriguing indeed). Or it is a
fantasy about a friendly dinosaur, who guards the character when he sleeps. If
the spaces set up are not intended to represent the story’s reality but the
character’s delusions (think ‘crazy scientist’), then the story might be a variety
of ‘stream-of-consciousness,’ following the inner visions of the character. If the
dinosaur appeared in a dream, when the character slept, then the reader is faced
with a mystery – did the dinosaur cross over from dream to reality? or is the
character still asleep, but dreams that he is waking? In fact, there is an impres-
sive number of possible readings of the story, all dependent on how the reader
(not the text) construes the spaces set up, and how the spaces are then linked by,
for example, the presence of the same characters (‘he’ and dinosaur in the same
‘reality’ or not, dinosaur projected from dream to ‘reality’, et cetera). All the
interpretations above are blends – a reality/fantasy blend, a reality/hallucination
blend, or a reality/dream blend. What is more, as the ‘dream’ version suggests, it
may not be possible or even interesting to resolve what kind of blend it is, so
long as the story construction process engages the reader.
Even in a longer narrative, the sentential level of the text may crucially
contribute to the structure of narrative spaces. Consider example (2), where
the main character in Eggers’s novel A Heartbreaking Work of Staggering
Genius, Dave, is in his car, struggling with guilty thoughts about having left
his small brother with a babysitter:

(2) I will come home and the door will be open, wide. The baby-sitter will be gone and
there will be silence . . . Blood on the walls, handprints soaked in blood. A note to
me, from Stephen, taunting; . . . There will be a hearing, a trial, a show trial –
How did you come to meet this man, this baby-sitter?
We found a posting, on a bulletin board.
And how long did your interview of him take?
Ten, twenty minutes. (AHWOSG, p. 126)

The narrator is assigned to an independent space (story-viewpoint space; SV-

space), which has the entire story in its scope. The story itself is contained in the
main narrative space (MN-space), which consists of a number of narrative
spaces (NS). The details of this set-up will be explained in Chapter 3, but
example (2) illustrates what it means for lower-level expressions to participate
in narrative space construction.
The main narrative space in (2) is that of Dave’s life with his brother, Toph.
As Dave is driving across Bay Bridge, his mind starts conjuring up an elaborate
scenario wherein the babysitter kills Toph, and Dave is then put on trial for
negligence. The sentence I will come home . . . thus sets up an imagined future
 2.2 Narrative spaces as mental spaces 39

SV-space Narrator: Dave

present

MN-space NS1
Narrator: Dave / ‘I’
Narrator: Dave / ‘I’ driving to San Francisco
present tense
Dave’s life NS2 – crime
Narrator: Dave / ‘I’
imagined future
babysitter (killer) NS3 – trial
Toph (victim) Narrator: Dave / ‘I’

NS3 – Discourse space

S1: Dave
S2: prosecutor

Figure 2.1 Narrative spaces (example 2)

space, then developed with the gory details of the crime. Then, from the point of
view of the imagined ‘crime’-space, another space is built, that of an imagined
trial. Each one of these spaces is set up at the sentence level, and even though
their activation does not extend beyond several sentences, they do play a role in
the narrative construction (see Figure 2.1).
First, the scenario developing in Dave’s mind is one of numerous instances in
the novel where Dave’s unrealistic fears and his lack of confidence lead to
erratic and exaggerated responses. The ‘crime-and-trial’ sequence of spaces is
thus a part of a broader narrative construct representing Dave’s life. While
seemingly local, it participates in the higher-level story construction process,
which does not have to proceed linearly, in the way the text does, but elaborates
spaces incrementally, throughout the text.
Second, the example represents one of the key issues of story construction –
the management of viewpoint. In terms of mental spaces theory, the spaces just
described represent familiar patterns of viewpoint and focus. The space of
Dave’s life is the viewpoint space from which the ‘crime’ space is being set
up and elaborated, as Dave’s fantasy – it is thus the focus space. As discourse
progresses, the ‘crime’ space takes the role of viewpoint space, and the ‘trial’
space is now in focus, and being elaborated. This viewpoint/focus structure
explains why the trial is construed as the result of the imaginary crime, and not,
say, of Dave’s driving habits.
40 Blending, narrative spaces, and the emergent story

The viewpoint/focus configuration explains how spaces form chains, such

that a higher-level space is the viewpoint, and a lower-level space is the focus.
However, in order to explain the processing of narratives as complex as novels,
we also need an explanation of the mechanism which links various lower-level
spaces into clusters contributing to the construal of higher-level spaces. I will
discuss this mechanism as cross-input projection. At the same time, the content
of the lower-level spaces contributes to the construal of the higher-level space,
so that Dave’s fantasy is not only projected downwards from the highest space
of his narrative, but also used in the construal of the narrative itself – for
example, it gives the reader access to an understanding of what kind of narrator
Dave is, and how his narrative as a whole relies on fact and fantasy alike. These
higher-level construals result from what I will refer to as viewpoint compres-
sion, an integration mechanism which allows lower-level viewpoint to contrib-
ute to the higher level. The phenomenon also has its linguistic correlates, such
that lower-level spaces are presented without linguistic indications of their
being embedded in higher-level spaces. The entire fragment partially quoted
in (2) is narrated alongside the description of Dave’s drive, and thus, stylistically
at least, it is presented as belonging to the core of the story being narrated. The
effect of engaging multiple levels of spaces and presenting them as one is the
primary sign of viewpoint compression. I will discuss the phenomenon in more
detail throughout the book.
Beyond the sentential level, fictional narratives set up various configurations
of spaces, distinguished by basic space topology (such as time and space, or
participants). Some contemporary narratives rely on an even more fragmented
structure, wherein major narrative spaces are presented as independent texts –
there can be a novel within a novel, as in Atwood’s The Blind Assassin, different
narrators, or stories ostensibly consisting of separated narratives, to be linked at
the end, as in Alex Garland’s The Tesseract or in the movie Babel. Accounting
for the astounding variety of such narrative choices would not be possible in any
analysis, but I will discuss one text, The Blind Assassin, in more detail, to
illustrate the processes involved.

2.3 Narrative spaces – an example

Margaret Atwood’s novel The Blind Assassin is a good example of how the
multiplicity of narrative space operates in a narrative. The novel’s framing
narrative is a memoir written by a now aged and lonely woman, Iris Griffen
(née Chase). She lives in the small town where she grew up and writes the story
of her life, to be read after her death by her estranged granddaughter. The main
narrative space thus has a first-person narrator, Iris, remembering the past.
Consequently, large parts of what we are reading are descriptions of Iris’s
daily struggle to go on writing, but the main story of the novel is reconstructing
 2.3 Narrative spaces – an example 41

Iris’s entire life. These two narrative spaces of ‘Iris’s present’ and ‘Iris’s past’ are
the core of the narrative structure.
Also embedded in the text is a novel, titled “The Blind Assassin,” ostensibly
written by Iris’s sister, Laura. It is written in the present tense, with just two
characters never identified by name – just a ‘he’ and a ‘she.’ These grammatical
choices (no I narrator, present tense) help the reader maintain the embedded
novel’s relative independence, even as it is gradually incorporated into the story.
In the embedded novel, the two anonymous characters compose a pulp science-
fiction story, taking place on the planet Zycron, and featuring a blind assassin
as one of the (again nameless) characters. This story is never completed, and
thus its role in the entire text is difficult to see. Finally, the text is sometimes
interrupted with short newspaper clippings, giving an official version of some of
the events. The clippings often contradict the information provided by Iris’s
memoir. They jointly represent a narrative space telling the official version of
some of the events, reappearing throughout the text. The space is anchored by
the format of the clippings and is reactivated as the need arises. Taken together,
the clippings construct a narrative space which tells a story different from the
one narrated in the rest of the text.
All these major spaces are presented as independent texts, with different
authors and characters, occurring in different locations and times. As Figure 2.2
shows, they are initially set up as independent spaces – and the set-up changes

SV-space Narrator: Iris

present

MN-space

Narrator: Iris
Author: Iris Clippings Sci-fi story
past tense Author: journalist Author: he
Mrs. Richard E. Narrator: 0
Memoir Griffen present tense
Narrator: Iris mute girl (she)
past tense blind assassin (he)
Embedded novel
Laura Lord of the Underworld
Author: Laura
Alex King
Narrator: 0
Richard present tense
Father
he
she

Figure 2.2 Narrative spaces in The Blind Assassin

42 Blending, narrative spaces, and the emergent story

as the text progresses. Crucially, they all have to be treated as subordinate to the
space of Iris narrating (they are within the scope of the viewpoint of the SV-
space), and are gradually identified as complementary to the memoir she writes.
Coherence links are suggested throughout and the fabric of the story emerges
through various links across the narrative spaces. Even more important, all the
viewpoints of all the spaces are finally compressed with the viewpoint of the
initial set-up with Iris as the narrator.
The case of The Blind Assassin illustrates all the central processes of story
construction. I will attempt to elaborate them below, starting with the concept of
narrative anchors.

2.3.1 Narrative anchors

I have introduced the concept of a narrative anchor in earlier work, focusing
mainly on The Blind Assassin and F. Scott Fitzgerald’s The Great Gatsby
(Dancygier 2007, 2008a). In both of these cases I argued that the fragmented
nature of most narratives requires specific narrative mechanisms which provide
coherence links across different narrative spaces. I defined narrative anchors as
expressions which set up or suggest the availability of narrative spaces, but do
not elaborate them right away. Such ‘place-holders’ may activate new narrative
spaces and allow them to remain active, but the spaces are elaborated gradually
as the text unfolds, and often contribute to the topology of other spaces
constituting the story.
The second function of anchors is ‘link-building’. The links they set up may
prompt what I have called cross-input projection – spreading of topology from
one input to another and building the coherence and completeness of the
emergent story. To refer to the ‘clippings’ again, they are constantly cross-
linked with other spaces, adding topology, foreshadowing events to be narrated
later, but, first of all, presenting a viewpoint not available in other narrative
spaces. The usefulness of anchors becomes clear when we realize that it is
difficult to imagine a fictional narrative which would explicitly provide the
entire content of its narrative spaces. Narrative coherence crucially depends on
selection of the content needed at a given point in the story, and on suppression
of all information which would reveal too much or drown the important
narrative elements in a chaos of unbounded verbosity. Narrative anchors are
difficult to define formally because they are not restricted to any specific aspect
of narrativity which is typically discussed (plot, characterization, et cetera), but
they may participate in all. They are best defined in terms of the processes they
trigger which lead to the construction of the story.
Various expressions can play the role of anchors, and the choice depends on
the type of cross-space connection to be established. One can thus expect that
well-selected referring expressions would lead to connections in terms of
 2.3 Narrative spaces – an example 43

finding counterparts in different spaces and establishing identity (see 2.3.2 for
more discussion), but descriptive expressions may play the same role. For
example, the man named Alex in the main narrative space of The Blind
Assassin has to be recognized as being the same as the one referred to
only as ‘he’ in the embedded novel space, even though the two counter-
parts are characters in different stories, taking place in different space and
time:
(3) He stubs out his cigarette, reaches for another, thinks better of it. He’s still smoking
ready-mades, a luxury for him. (The Blind Assassin, p. 133, embedded novel)
(4) He produced a packet of cigarettes from his shirt pocket – Craven A’s, as I recall. He
tapped one out for himself. I was a little surprised that he smoked ready-mades – it
didn’t go with his shirt. (The Blind Assassin, p. 222, Iris’s memoir)

Making the connection is prompted by careful deployment of the same bits of

information in both narratives, and by repetition of some of that information.
Bits of characterization, such as (3) and (4), are spread throughout the narrative
spaces via cross-input projections and jointly create the character’s identity
which fits coherently in all the narrative spaces. Crucially for this example, as
the page numbers show, the reader first sees the man in the embedded novel
smoking ready-mades, but is also alerted to this trait by the use of still – which
indicates a time prior to the scene when the same trait was observed. So when
Alex Thomas is first introduced in Iris’s memoir (in a scene temporally prior to
that of [3]) and the same information is presented as surprising, the reader can
make the connection between the two characters. In fact, guessing Alex’s
identity is made easy, but the identity of the woman remains elusive till almost
the end of the novel.
Anchors can also rely heavily on frames. As shown by Lermitte (2010),
anchors may also be images which form an entire network of concepts and
jointly give meaning to an abstract and difficult text. In Annie Dillard’s non-
fiction text Holy the Firm the meaning is not so much a story as a reflection on
religious and artistic passion and a life centered around such deeply felt
emotions. The crucial image is that of a moth being attracted to the flame of a
candle, dying, and acting as the candle’s wick as a result, allowing it to burn
longer: She burned for two hours without changing, without bending or lean-
ing – only glowing within, like a building fire glimpsed through silhouetted
walls, like a hollow saint, like a flame-faced virgin gone to God. Passages such
as this recur through the text, highlighting various dimensions of the emerging
blends.
The recurring images of burning as a sacrifice which in the end prolongs life
allow Dillard to tell a poignant story of creative endeavors and religious
devotions as acts in which an individual needs to ‘die’ in order to ‘keep the
flame burning’ and go on living through what has been created. The anchors do
44 Blending, narrative spaces, and the emergent story

not simply ‘construct the story’ here, they construct its meaning through a
network of blends and frame metonymies. Without them, the text achieves no
narrative coherence.
These various examples of anchors pose questions as to the actual cognitive
nature of anchoring. It seems that it relies on a variety of possible construals and
mappings. In the work done so far, I have distinguished several types. First,
there are anchors which are quite naturally space builders. When the text of The
Blind Assassin repeatedly refers to ‘notebooks’ a now dead character left in
hiding, the content of the notebooks constitutes a narrative space which prom-
ises to fill huge gaps in the story. Naturally, when the notebooks are found, they
do reveal crucial facts.
Another type of anchor consists in evocation and repeated re-activation. A
mention of Alex as wearing a blue, worker’s shirt but smoking ready-made
cigarettes is a salient descriptive detail, and its reactivation in a different
narrative space prompts the cross-input projection linking Alex to ‘him’ and
making this cross-space identity available in both spaces. This type of anchoring
relies as much on the reader’s attentiveness and memory as on the salience of the
frames thus constructed and on the sheer number of anchors establishing and re-
establishing the same cross-input links. Crucially, such anchors not only link the
spaces, but also prompt projections of topology from input to input (narrative
space to narrative space). The identity of the character named Alex in one input
is cross-linked to the identity of an anonymous man in another, but the infor-
mation from one input is thus also available in the other.
Below I will present an extended analysis of a pair of anchors in The Blind
Assassin which are central to the narrative meaning of the entire novel.
Throughout the novel, these anchors play all of the roles described above –
narrative space building, prompting cross-input mappings, establishing identity,
metonymy, and metaphor. While this may be a special case, it does show the
potential of anchoring in a complex story.

2.3.2 Anchoring: representation blends and frames

The main narrative space of the novel is the life of two sisters, Iris and Laura,
and their involvement with a man, Alex Thomas. Iris is the narrator, while
Laura, who has committed suicide, is not given a narrative voice. Even though
one of the narrative spaces is a novel she presumably wrote, it turns out in the
end that the novel was also written by Iris and published under Laura’s name.
Throughout the reading, the text playfully alternates between suggesting Laura
or Iris, leaving it to the reader to construct the true identity of the narrator of
‘Laura’s novel’ (I will refer to it as ‘embedded’). The authorship, in turn, also
identifies the writer as Alex’s secret lover. While there are many salient anchors
 2.3 Narrative spaces – an example 45

leading the reader to a conclusion, the main anchors are a photograph and a
hand. They are initially mentioned in the first lines of the embedded novel:
(5) She has a single photograph of him . . . the photo is of the two of them together, her
and this man, on a picnic . . . he’s holding up his hand, as if to fend her off in play, or
else to protect himself from the camera . . . Over to one side – you wouldn’t see it at
first – there’s a hand, cut by the margin. (The Blind Assassin, p. 7)

The fragment is full of unanswered questions – who are the people? whose hand
is visible at the edge? what happened to the man? and many more. They all
emerge because of the fact that the concept of a photograph (similarly to
picture, painting, drawing, et cetera) prompts a blend connecting two mental
spaces: the representation space (distinguished by its form, colour, materiality,
boundaries, et cetera) and the represented space (the actual reality it captures,
with its temporal and spatial features, identity of participants, as well as,
crucially, topology available beyond what the representation shows). In the
fragment, only the representation space is provided, while none of the features
of the represented space are clarified – hence all the unanswered questions
which frame the entire ensuing narrative. It is thus a perfect example of a
narrative anchor. It opens up a narrative (represented) space which is necessary
to complete the representation blend prompted by the description of the repre-
sentation space, and the anchor remains active until the blend is complete – until
the end of the novel.
The narrative of The Blind Assassin uses this anchor throughout to keep the
reader guessing. It is reactivated regularly, and each time a new piece of the
represented narrative space becomes available. The next mention, in Iris’s
memoir, is as follows:
(6) Then he took a picture for the paper with his camera . . . Alex Thomas raised his hand
as if to fend him off. (The Blind Assassin, p. 223)

This appears in a fragment describing a factory picnic when the sisters first met
Alex. The identity of the man in the photograph is suggested by the second
mention of his gesture, but since both women were present at the picnic, the
woman telling the story in the embedded novel is still mysterious. Finally, the
newspaper report of the factory picnic is published:
(7) One of the pictures was of Alex Thomas, with the two of us – me to the left of him,
Laura to the right, like bookends. Both of us were looking at him and smiling too, but
he’d thrust his hand up in front of him . . . The caption was, “Miss Chase and Miss
Laura Chase Entertain an Out-of-Town Visitor.” (The Blind Assassin, p. 241)

The anchor of Alex’s raised hand is repeated again, but the identity of the
woman in the photograph in (5) becomes even harder to discover. The fragment
makes it clear that while the represented space is now known (Laura, Iris, and
46 Blending, narrative spaces, and the emergent story

Alex at the picnic) the representation in the photograph mentioned in (5) is

incomplete. One of the women – Laura or Iris – is missing from the photograph,
and the hand at the edge is thus a token of the presence of ‘the other woman.’ At
this point, the identity of the narrator of the embedded novel has to match the
woman in the photograph, and would also point to the one whose hand is
visible. It is clear that the writer is playing this ‘now you see it, now you
don’t’ game on purpose. In fact, the remaining (very rich) narrative is playfully
providing ‘helpful’ information which alternatively points to Laura or Iris, and
thus the text offers an abundance of anchors which do not materialize into
complete spaces or unambiguous cross-input links until the end of the novel.
In this context the ‘hand’ becomes another important anchor, and carries with
it a rich array of meanings. At the most basic level, the hand in the photograph is
simply an indication that there was a person sitting next to Alex – it is thus an
anchor to the undisclosed identity of the other woman at the picnic. At the same
time, it suggests that someone (the narrator of the embedded novel?) did not
want the other sister in the picture – in the literal and figurative sense. The
‘hand’ is not only the hand of someone who was present in the represented
space and time, but the hand that was left after someone tried to erase that person
from the memory of the events. There are several levels of frame evocation here:
‘hand’ evokes the entire body and an entire person, the deliberate distortion of
the original scene evokes someone’s intent, and the incomplete photograph is
also an anchor to the missing part of the narrative, where the character forcefully
removed from the story tells what had really happened. The double anchor
of the photograph and the hand is a token of what the narrative needs in order
to achieve coherence – the missing part of the story connected to the other
woman.
Later in the text we learn that Laura had access to the negative and produced
two prints from it, each with one sister absent (save the ‘hand’); what is more,
following a childhood habit of colouring black and white drawings, she stained
each of the ‘hands’ a different color: Laura’s hand is yellow, Iris’s hand is blue.
(8) But she’d cut herself out of it – only her hand remained. [yellow-Laura]
...
“I have another one [photograph], for me.”
“And I’m not in yours?”
“No,” she said. “You’re not. None of you but your hand.” [blue-Iris]
This was the closest she ever came, in my hearing, to a confession of love for Alex
Thomas. (The Blind Assassin, p. 277)

This explains how there could be two photographs in the possession of two
sisters – evoking two stories of love and loss related to Alex Thomas. However,
both stories are in a sense written into both pictures, for even though the
narrative might focus on one of the women, it will not be complete without
the other.
 2.3 Narrative spaces – an example 47

The hand/photograph anchor begins to interact with other anchors later in the
text. In the embedded novel the photograph is kept in a gardening book, which
also appears in the main narrative space (Iris’s autobiographical story), where
she describes ways of occupying herself in the difficult period of her married
life. “Perennials” is thus another anchor, identifying the narrator of the embed-
ded novel as Iris:
(9) She has a single photograph of him. She tucked it into a brown envelope . . .
between the pages of Perennials for the Rock Garden, where no one else would
ever look. (Prologue: Perennials for the Rock Garden) (The Blind Assassin, p. 7)

In the final chapters, Iris finally acquires her sister’s notebooks, mentioned in
the very first paragraphs of the novel as a source of crucial clues to Laura’s
suicide.
(10) I riffed through the other notebooks. History was blank, except for the photograph
Laura had glued into it – herself and Alex Thomas at the button factory picnic, both
of them now colored light yellow, with my detached blue hand crawling towards
them across the lawn. (The Blind Assassin, p. 627)

The hand’s suggested motion (crawl) evokes intended interference in the

lovers’ encounter, as the body image of someone crawling evokes bent posture,
such that the couple might not even notice anyone approaching; this in turn
could be read as signaling evil intentions. The ‘crawling hand’ is narratively
equal to a crawling person, someone who might then hurt the couple on the
lawn. There are several important points to note here. First, this very low-level
narrative choice has text-wide consequences, precisely because it appears in
connection with the novel’s primary anchor. One of the points this analysis
makes is that a narrative achieves its meaning through the compounded effect of
meanings emerging at the lowest lexical and grammatical level.
Second, it is important to note that the view of Iris as a destructive intruder
which (10) suggests comes at the end of the novel, when Iris realizes that she
unknowingly caused her sister to commit suicide. For Iris, the narrative voice of
the story, to describe herself as my detached blue hand crawling towards them
across the lawn, is to complete the untold story of the photograph. The
representation blend has not only been fully integrated, through a match of
the representation space and the represented space, but it achieves its narratively
satisfying status through the interpretive viewpoint of one of the participants.
The facts of the picnic as such would not tell the story that led to Laura’s death,
but Iris’s realization of her own role in her sister’s life is the story being told
throughout the book. In the end, ironically, the true story of the novel is told
through the truncated photograph, not the photograph published in the paper,
with all three of the main characters sitting together. In a sense, the way this
anchor shapes the story in The Blind Assassin is by adding viewpoint and
48 Blending, narrative spaces, and the emergent story

causation to the events which otherwise would not have been framed to achieve
any psychological effect.
The role of the ‘hand’ anchor does not end there. While it structures much of
what happens in the novel, it also participates in the construal of narratorship and
narrative viewpoint. Narratorship here is not only the question of an epistemic
viewpoint (who knew enough to tell the story in this way, a crucial consideration
with respect to the embedded novel). Iris is ostensibly telling her own story, but it
becomes clear as the text progresses that she is, perhaps primarily, telling the story
of her sister’s life, as she finally understands her own role in it. In the last sections
of the novel she talks about herself as the ‘hand’: Laura’s odd, extra hand,
attached to no body – the hand that passed her on. The frame of the ‘hand’
now refers to a common understanding of the ‘hand’ as the body part capable of
manipulating and doing things. The detached hand in the picture now stands for
Iris’s decision to attribute the novel to Laura. The embedded novel tells the story
of Iris’s secret relationship with Alex, but granting Laura the authorship has ironic
consequences in view of the fact that the characters in the embedded novel are
never referred to by proper names, only as ‘he’ and ‘she.’ By giving Laura the
authorship, Iris’s ‘hand’ in a way also gives the participation in the story to her
sister – something that could have been the case and which was clearly Laura’s
deepest desire. In fact, people in Iris’s family understood the novel as Laura’s
confession of a secret affair, so from their point of view, Laura was Alex’s lover.
By referring to herself as Laura’s ‘extra hand,’ Iris removes herself from the story
as its full-bodied participant – again, something the anchoring photograph sup-
ports. The anchoring image of the severed hand is reframed as the narrative voice.
Iris explains this further:

(11) I can’t say Laura didn’t write a word. Technically that’s accurate, but in another
sense – what Laura would have called the spiritual sense – you could say she was
my collaborator. The real author was neither one of us: a fist is more than the sum of
its fingers. . . . Laura was my left hand, and I was hers. We wrote the book together.
It’s a left-handed book. That’s why one of us is always out of sight, whichever way
you look at it. (The Blind Assassin, p. 644)

These fragments provide crucial metanarrative comments. The tricky structure

whereby a myriad of anchors keep on pointing to first one sister and then the
other is now described in its full meaningfulness. Throughout the reading, one is
tempted to keep on guessing which one of the sisters is the ‘she’ in the
embedded novel, and the reader’s detective skills are rewarded when Iris admits
that she wrote it and she was Alex’s lover. But the final comments quoted in
(11) suggest that the shimmering narrative has a different purpose. The story is
now seen as having two emergent structures, not just one. There is the narrative
Iris constructs to set the record straight and explain her sister’s suicide, but there
is also the narrative where she and her sister become one narrative agency
 2.3 Narrative spaces – an example 49

and thus share in the experience of the story as well as the telling of it. The
narrative thus reconstrues Laura’s life, by giving her a chance to experience
what could not happen in the novel’s reality and to become an agent in the way
the story is told.
In the final sections the double photograph/hand anchor plays its last narra-
tive trick. The embedded novel is spread chapter by chapter throughout the text,
with other parts of Iris’s narrative taking much space in between. The Epilogue
starts very similarly to how the embedded novel begins – see the fragments
quoted in (5) and (9) for comparison:

(12) The photo has been cut; a third of it has been cut off. In the lower left corner there’s
a hand, scissored off at the wrist, resting on the grass. It’s the hand of the other one,
the one who is always in the picture, whether seen or not. The hand that will set
things down. (Epilogue: The other hand) (The Blind Assassin, pp. 649–650)

There is one striking difference, though. In the Prologue ([9]), the woman is
clearly identified as Iris through the Perennials anchor (available later in the
text), and the description of the photograph does not specify whether it is the left
or the right hand that is seen next to the man. In the Epilogue, the hand is
identified as the left hand, which clearly suggests that it belongs to Iris (when we
imagine looking at the photo as described in [7], with Iris seen to the left of
Alex). So the woman looking at the photo in the Prologue is Iris, but the woman
looking at it in the Epilogue is Laura. This also explains why the ‘hand’ in (12)
is the one that will ‘set things down,’ that is, Iris. What is more, the Epilogue
also reconstructs the embedded novel to match the entire story of the photo-
graph. The Prologue introduces Iris’s viewpoint, the text in the middle focuses
on Alex and his lover, whoever she is, and the Epilogue adds Laura’s viewpoint.
To make the analogy clearer, the women in the Prologue and the Epilogue are
construed as ‘viewing’ their respective photographs, so that the concept of
narrative viewpoint is clearly evoked. The anchor has now achieved its full
potential and it explains what Iris means by saying that both sisters had their
‘hand’ in the embedded novel.
The anchor described here requires a careful reader. Interestingly, critics (Roy
2006; Perrakis 2006; Wilson 2006) treat The Blind Assassin as just a post-
modern exercise in instability of meaning. But the acknowledgement of the
anchors, as I hope to have shown, does not trivialize the novel – on the contrary,
it helps reveal its double bottom and its poignant view of what narrative fiction
does. Just telling the story of what happened is not good enough, it seems to be
saying, because the narrative may be the story that should have happened, but
never did.
The crucial role of the anchors discussed above depends largely on the nature
of the concepts they metonymically evoke. The idea of a photograph is in itself a
rich source of mental space topology. The other anchor, the hand, is also craftily
50 Blending, narrative spaces, and the emergent story

chosen. Being associated with the body but also with the human ability to
manipulate, make new things, destroy things, et cetera, the ‘hand’ is easily seen
as metonymic for a number of actions the person thus represented can undertake
and perform. Moreover, all the actions associated with the hand also require a
will and a subjectivity, thus in spite of not bearing any agency in itself, this body
part can naturally stand for the subjectivity of a person. Atwood’s use of these
anchors capitalizes on all of these aspects of the framing of photos and hands.
The naturalness of the story construction processes triggered by the anchors
supports Deacon’s (2006) discussion of how art works. He argues that in a
broader context a single word such as hawk is much more than a symbol for an
actual species of a bird, but rather a way to activate a network of concepts in the
mind of the hearer/reader (bird of prey, powerful beak and claws, ruthless
hunter, linguistic form, emotional response, et cetera). Furthermore, activating
the image of the claws and predation may simultaneously evoke the concept of a
victim, fear, and pain, or, for comparison, power, courage, attack, and pride. I’m
reading Deacon’s claim as an acknowledgement of the aspects of the process of
reading which rely not only on the constructing power of the words alone, but
also on the richness of experiential frames already present as the background to
reading. The anchors in Atwood’s novel seem to support this. While some are
relatively simple (for example, Alex’s ready-mades), others interact with seem-
ingly unrelated frames present in the reader’s mind. Also, all such frames evoke
emotional responses needed in processing texts.
Relying further on Deacon’s work, one can claim that anchors are an ultimate
way of building on what he refers to as our ‘representational stance’ – the
natural assumption of human cognition that something stands for something
else. It is a more straightforward (though not simple) process in our use of
language as a system of symbolic representations, but it is more complex in the
narrative. Naturally, the narrative expressions are symbolic in the ordinary sense
of the word, but they may be deployed in the text in a way which poses a
challenge to our representational stance. We do know what a photograph is and
we expect it to represent a recognizable reality. If that reality is not provided, it
can still be constructed. We can thus say that narrative anchors are narratively
salient expressions which rely metonymically on frames and exercise our
representational abilities, but create suspense by providing necessary informa-
tion only bit by bit.
To sum up, the cognitive impact of anchors on the reading of narratives is
multidimensional. At the simplest levels, anchors are frame metonymies (pho-
tograph for the representation blend it constructs, fire for pain, destruction,
purification, light, et cetera). Moreover, the textual situation of anchors guaran-
tees the emotional response Deacon sees as indispensable in the appreciation of
literature. Perhaps more important, the patterns according to which anchors are
reactivated and gradually enriched with new framing are also the patterns of
 2.3 Narrative spaces – an example 51

narrative construction at every level. They participate in establishing temporal

sequences of events, in revealing identities of characters, and in supplying
missing elements of the plot or prompting their reconstruction. Overall, they
provide the skeletal system of interacting frames which, as they are enhanced
with new topology, point to crucial narrative spaces and connections between
them, and thus lead to the emergent story.
Narrative anchors capture more than the progressive accumulation of content
and context, as described, for example, in Werth’s study of text worlds (1999) or
Emmott’s discussion of ‘contextual frames’ (1997). They are also a more
specific mechanism than what literary critics refer to as ‘motifs’, especially in
the mental-space setting function. They exploit the mechanisms of frame
metonymy, frame evocation, mental space set-up or evocation, and cross-
space projection. At the same time, they capture the complex interaction
between specific expressions used, emergent connections across various parts
of the text, and the reader’s processing of it.

2.3.3 Reference and story construction

The discussion above shows how narrative anchors of various types play a role
in prompting cross-input projections and the emerging cohesion of the story.
Here, I will consider the ways in which referential expressions contribute to the
emergence of the story.
The central piece in the narrative-space structure of The Blind Assassin is the
embedded novel, featuring only two characters, referred to through pronouns
only. The pronouns he and she let the reader assume that these forms refer to
subjectivities identified elsewhere in the novel, and thus a large part of the
reading process consists in attempts to establish cross-space identity links
which would explain who the characters are. He is soon identified as Alex
Thomas, while the decision about the identity of the she oscillates between two
sisters, Iris and Laura Chase.
However, another narrative space of the novel, a science fiction pulp story
Alex is composing, also features no names, using only descriptions of roles
instead. The roles are specific to the text and not very inventive (‘blind assassin,’
‘mute girl to be sacrificed,’ ‘king of the underworld’), but their presence in the
overall novel and the metaphorical understanding of their characteristics (blind,
mute, hidden) provide important clues to the understanding of the story
(cf. Schnepf 2006 for more discussion). The story these characters participate
in explains the facts from their lives which are not represented anywhere in the
text, but the role of these facts is appreciated only through the matching of the
roles with characters in other narrative spaces. Crucially, the entire construal of
the sci-fi story represents the viewpoint of the in-novel writer, Alex, who is thus
revealing the nature of his participation in the story. He was not given his own
52 Blending, narrative spaces, and the emergent story

voice anywhere else in the narrative, and thus the pulp sci-fi space is where his
viewpoint is construed.
Without the reliance on the role of referring expressions The Blind Assassin
could not be imagined in its actual form. It forces the reader to construct
dynamic frames to represent the identities signaled by proper names – other-
wise, the entire reasoning of the choice between Iris and Laura as the ‘she’ of the
embedded novel could not take place. It builds on the assumption that personal
pronouns do have fully framed counterparts within the spaces of the text. It
constructs roles which allow the reader to place characters in the appropriate
frames. And it also relies heavily on the construal of the SV-space as profiling
the narrator matched with a character (Iris), so that the spaces construed from
another viewpoint are constantly matched with the SV-narrative.
Crucially to the point of this project, the establishment of cross-input projec-
tions based on the links between names, role-descriptors, and pronouns is an
exploitation of low-level linguistic mechanisms on the level of the extended
discourse, such as a narrative. Knowing that a pronoun requires an antecedent in
some mental space or other is what prompts readers into the attempts to seek
referents of the expressions in the embedded novel in other parts of the
narrative. Similarly, the frames set up in the sci-fi text, with their attendant
roles, create the need for further referential links. All these referential connec-
tions are not simply available in the text of the novel: they need to be con-
structed (Figure 2.3 shows the final resulting set of referential links). In a sense,

SV-space Narrator: Iris

present

MN-space

Clippings
Narrator: Iris
Author: journalist
Author: Iris
Mrs. Richard E.
past tense
Griffen
Sci-fi story
Memoir
Author: Alex
Narrator: Iris Embedded novel Narrator: 0
past tense
Author: Iris present tense
Laura
Narrator: 0 Lord of the Underworld
Richard
present tense King
Father
Alex he blind assassin (he)
she mute girl (she)

Figure 2.3 Referential connections across spaces in The Blind Assassin

 2.4 Emergent story 53

referential frames (profiling slots for a name, a pronoun, and a role) serve as
prompts for the filling in of all the slots, even though there are no clear anaphoric
or cataphoric patterns to rely on. Such a text-wide deployment of reference
seems to be a characteristic of narrative discourse.
This form of reliance on referential expressions is used in cross-input con-
nections and the resulting enrichment of input narrative spaces with content.
But we should also note that in the process of reading, the distance between the
referential slots in individual spaces needs to be compressed, so that the final
complete identity of characters emerges as a result. Blending and compression
are thus participating in story construction, even at the level as basic as identity
of characters and reference.
Referential expressions in the narrative pose a complex and multidimensional
problem. Starting with the distinction between first- and third-person narration,
continuing to the less usual forms of second-person narratives and an occasional
use of we, the range of usage of pronouns in narratives poses complex questions
about the internal coherence within a text. In Chapter 3, I will consider pronouns
used as text-wide narrative choices; Chapter 5 will look further at the question
of proper names and roles, while Chapter 7 will consider the use of pronouns in
Speech and Thought Representation (STR).

2.4 Emergent story

In much of the narratological literature the concept of the story is defined in
relation to text.5 The text is the form (some also mention discourse as an
additional element, cf. Toolan 2001) while the story is what the text represents.
While the text may be fragmented, incoherent, temporally disorganized, the
story is a temporal sequence of causally linked events leading to a resolution of
some conflict or problem. The contrasting concepts seem very much like an
elaboration of the form/meaning pair, but there is an additional claim, made with
special clarity by Chatman (1978, 1990), arguing that the story as the abstracted
chain of events remains the same when it is told through a different text or a
different medium, such as a film. This may be related to the linguistic tradition
of maintaining that propositional meaning remains constant under various
formal changes, so that it is natural to see the ‘story’ as the meaning of the
‘text.’ Regardless of the source of the distinction, the ‘story’ is one of the most
salient concepts in narratology, but it is typically not discussed in terms of its
cognitive status emerging in reading. While Herman (2003a) talks about cog-
nitive aspects of story construction such as chunking of experience or attribut-
ing causation, the concepts he evokes refer to the general concept of a story as a
way to structure experience. The ‘story’ under discussion here, on the other
hand, is the semantic correlate of a specific text.
54 Blending, narrative spaces, and the emergent story

The text, however, should not be seen as simply the words and syntactic
structures, as it also relies on cultural availability of frames structuring the
events described via the choice of words. The text is not just a collection of
linguistic forms, it is also a rich resource of framing. To explain the emergence
of the construct called a ‘story,’ I will treat textual choices as prompts guiding
layers of blends leading to the sequential, causally structured configuration.
Finally, the story is the final result of several modes of interaction with the
text – reading the words, activating the frames, searching for correlates in one’s
experience, making cross-space connections, blending narrative spaces, estab-
lishing identities, constructing tentative scenarios, storing them in memory,
revising them as new events are narrated, responding emotionally, et cetera,
et cetera, et cetera. The richness of the story as a construct can best be explained
if we attempt to discover the processes that lead to the final result. Much of this
work is already being done in various contexts – consider Gerrig’s work in
psychology (1993), Emmott’s studies combining cognitive psychology and
discourse analysis (1997, 2002, 2003; Emmott et al. 2007), or Herman’s
renewed vision of narratology (2002, 2007). The goal of this study is to give
the view afforded by cognitive tools used in the analyses of linguistic meaning.

2.4.1 Sequence of events and the story

One of the most contentious aspects of ‘storiness’ is its idealized status which,
as narratologists argue, remains the same in different tellings. The claim calls for
an explanation of where such a construct resides. It cannot be an overall
interpretation arising in a single reader’s mind, since such a construct would
be too specific. If, on the other hand, the concept is expected to remain constant
across choices of a medium, across readers, or even across readings (as readers
may change their perception in time), then it has to be ‘boiled down’ to the
sequence of events and some conflict to be resolved. It would then be a very
minimalist view of the story.
The sequentiality of the story raises questions about the temporal disconti-
nuities in texts, which are a norm rather than an exception. In fact, the need to
disrupt temporal sequence appeared important in the most celebrated of epics –
the Odyssey. It starts in the middle (with the episodes when Telemachus
attempts to take control of his father’s household, while Odysseus is held
captive on Calypso’s island), changes narrators, and does all the ‘modern’
narrating besides. But knowing why Odysseus is missed on Ithaca puts his
entire journey in a different light and leads into the conclusion where the point is
not that the journey ends, but that balance returns at last.
Also, temporal infractions serve specific narrative purposes. Even in classical
texts, not as experimental as contemporary ones, flashbacks are used for specific
reasons. The story in Pride and Prejudice starts just before the ‘eligible
 2.4 Emergent story 55

bachelors’ appear on the scene and proceeds, through various complications,

until the two couples are happily married (it is unfair to Austen’s talent to
summarize her masterpiece in this way, but time is all that is being considered
here). In fact, there are important flashbacks in the novel: one when Wickham
tells a false story of his youth which puts Darcy in a bad light, and the other
when Darcy tells the true story of what happened between them. So it is by far
not enough to describe these stories as flashbacks, since they present two
opposite images of the main character, and entirely change the course of the
story – the first makes marriage with Darcy impossible for a person with
Elizabeth’s sensibilities, the other makes it very desirable. They do so not
because of what ‘really’ happened in the story’s past, but because the versions
of the past are crucial to characterization. The little deviation from sequentiality
is thus not just acceptable, but central to the construction of characters’
dispositions.
Broadly speaking, the list of temporal infractions in narratives is so long and
varied that insisting on the centrality of this dimension of cognitive organization
seems questionable. What is more, it is also questionable on experiential
grounds. First of all, sequential interpretations of events profiled in linguistic
constructions often emerge from construal, rather than from straightforward
meaning of linguistic forms. They had a baby after they got married or They first
got married and then had a baby clearly indicate the sequence of events with
after, first, and then, but They got married and had a baby or They had a baby
and got married only allow one to construe the sequence one way or the other
based on assumptions of the clause order reflecting the order of events. But in
real situations we only follow the sequence to the degree to which we experi-
ence the events ourselves or are exposed to their results. Temporal sequence is
rarely relevant to our understanding of events, but knowing their consequences
is crucial. The sequence is a questionable criterion even in the most sequential
of stories, but epistemic stance and understanding of causation seem to matter
much more (see Dancygier in press b for more discussion).
The popularity of the ‘sequential’ core of narratives can be explained if we
consider the cognitive construction of time. Stories develop through time and
our conceptualization of time assumes something like a linear path progressing
from prior events to the events that follow. The discussion of metaphors of time
and their instantiations (Lakoff and Johnson 1999, Evans 2004, Nuñez and
Sweetser 2006) postulates the emergence of our temporal concepts from our
experience of motion through space (which not only changes our position but
also takes time). Motion through time also assumes a specific orientation, such
that the past is ‘behind’ us and the future ‘ahead’ of us. Stories also assume a
direction, following a path from the story’s past to its future and its conclusion,
among other reasons because the characters’ experience of the story’s events is
modeled after our experience of time, but primarily because a sequential
56 Blending, narrative spaces, and the emergent story

organization is best suited to the representation of the causal organization,

which seems more central to narrativity. In other words, even if the events do
not appear in a sequence, the causal chains suggest such a sequence, and, as a
result, the entirety of the story is best viewed as sequential, even though its
telling may not follow the temporal chain. Also, even the spontaneous oral
stories, such as the ones analyzed by Labov (1972) and Labov and Waletzky
(1967), include components such as evaluation, where events from the story’s
past or even future can be evoked to justify the telling of the main line of
‘complicating action.’ To conclude, sequentiality is useful as the final construal,
and as such it is also the most salient one.

2.4.2 Story construction, vital relations, and optimality constraints

A ‘story’ can thus be discussed as a cognitive construct, a blend, emerging
through the process of meaning construction triggered by reading. The role of
the text is crucial in providing such prompts, but the emergence of the story
relies to a comparable degree on the frames evoked in the reader’s mind, and on
the construction of double-scope blends, integrated into the mega-blend. The
story is the mega-blend arrived at in the interaction with the text.
The processes of integration, as described above, rely on all the primary
aspects of blending. Narrative spaces, such as the ones in The Blind Assassin,
are set up as independent inputs, but then selective projection utilizes the
elements needed in the blend. For example, the story of The Blind Assassin
does retain all the hesitation about the authorship of the embedded novel, and
retains Iris as the actual author (Figure 2.3). Also, subsequent blends throughout
the process affect the understanding of the inputs, via backward projection.
Finally, the story-blend itself undergoes the required integration and elaboration
(so that, for example, Alex and he are fused into one participant, but the under-
standing of his identity relies on setting up a coherent view of his persona).
Vital relations play a role in the construction of the story as well. Because of
the two inputs, Alex and he are not really connected at all at first, then become
connected by the relation of Analogy, until they fuse into the Identity relation in
the emergent story, and in the final stage he is also linked to the ‘blind assassin’
via a Role–Value link. Similarly, the relation of the role of the author of
the embedded novel is not clearly attributed to any of the Chase sisters, but in
the final blend Iris is linked to that role, as well as the role of the lover in the
embedded space and the role of the mute girl in the sci-fi story (she cannot tell
anyone either about her relationship with Alex, or about her sister). The vital
relation of Time also affects the blend, as temporally unconnected spaces (the
memoir and the embedded novel) are compressed into the sequential organiza-
tion of the story; the relation of Change becomes the relation of Causation as we
 2.4 Emergent story 57

discover why Laura committed suicide, et cetera. Enumerating all the vital
relation changes would take far too much space, but these examples support
the claim that story construction progresses along the lines proposed in the
blending framework.
Construction processes described above also satisfy all the optimality con-
straints, as described by Fauconnnier and Turner (2002). The integration prin-
ciple allows representations in the blend to be manipulated as a single unit, and
that is what appears necessary in the story of The Blind Assassin, after all the
identity relations are established. The blend maintains the inner structure of
the inputs (topology and web principles), while tightening metonymies (as in
the case of the hand and the photograph discussed above). Finally, the under-
standing of the entire story can be unpacked into an explanation of how the
blend integrates the inputs, and the blend is necessary (good reason principle) to
explain the events and the role of characters.
Although The Blind Assassin seems to be an unusual example of the diffi-
culties of story construction, it is not more than a useful one. One of the reasons
to argue that the story is an emergent structure is to consider differences among
readers. The emergent construct is not the same for all readers as different
aspects of integration can come into focus on different occasions – even the
critics’ responses to novels show that foregrounding certain aspects of the story
yields a different blend in the reader’s mind. Thus the stability of the story in
various media (such as film versus text) is an idea requiring further investiga-
tion. While it may not matter much if some bits of plot are omitted from the
movie or if secondary characters are fused, the process of arriving at the story
through visual means may highlight or downplay frames which play different
roles in the reading. There is enough similarity in most cases, but because the
process of arriving at the final blend is different, the results may also not be fully
compatible.
3 Stories and their tellers

I am the man who comes and goes between the bar and the telephone booth.
Or, rather: that man is called “I” and you know nothing else about him . . .
(If on a Winter’s Night a Traveller, Italo Calvino)

The preceding chapter has outlined the concept of a narrative space as a textually
prompted construct used in story construction. Narrative spaces can thus be
distinguished through their temporal dimension, a consistent subplot, or the con-
struction of a specific epistemic point of view. I have also explained how narratives
achieve coherence through levels of blending leading to an emergent story.
However, reading a work of fiction invariably assumes a text-mediated
contact with the constructed fictional subjectivity often referred to as ‘the teller’
or ‘the narrator.’1 The concept is textually constructed, which has been stressed
repeatedly, primarily in order to avoid simplifications whereby the narrator is
identified with the author. At the same time, the illusion is compelling in many
narratives, and the reason why the author is so naturally understood to be
responsible for the way the story is told has not been answered in sufficient
depth. While the ‘death of the author’ (as a viable narrative concept) has been
widely publicized, the news may be exaggerated, as some flesh-and-blood
authors begin to claim their right to be considered legitimate participants in
the narrative exchange.2
Moreover, the ‘demotion’ or backgrounding of the author also avoids any
interest in the question that is just as important as the question of the reader
response and narrative transport, namely, the ability of a writer to construct a
textually available subjectivity. What one might call ‘authorial transport’ is a
cognitively unique ability whereby a writer can construct a mind other than his
or her own and use it as a medium in making a story accessible to the reader.
While the interest in the theory of mind has highlighted the psychological
aspects of reader response (Iser 1974; Zunshine 2006), these approaches do
not cover the mechanisms of story creation. My tentative suggestion is that
creative storytelling requires the construction of very elaborate blends in which
actual experience is embodied in a different context, after a process of selection
and compression. Selection is necessary in order for the most salient aspects of

58
 Stories and their tellers 59

experience (whether physical, historical, or psychological, first-hand or

observed) to be used as the conceptual skeleton of the emerging blend. The
cognitive ‘padding’ which adds specific character traits, circumstances, events,
et cetera is the matter of the emerging narrative spaces. However, the crucial
mechanism is that of compression (Fauconnier and Turner 2002; Turner 2006a;
Dancygier 2004b).
I argued in Chapter 1 that compression is the mechanism whereby specific
stories of the oral tradition acquire their generic power to shape behaviors and
maintain distributed cognitive networks. The crucial effect emerging out of the
compression is also an enhancement of the central themes and increased
salience of the intended ‘narrative uptake.’ In the written fictional narratives
these tasks are equally important, but are also subject to the author’s choices.
Naturally, any narrative selects the facts to be narrated and leaves much of the
assumed story untold, but the kinds of selection would depend on the authorial
choice of themes. The central compression mechanism, however, is the one
whereby the entire story is packaged into a text where only some of the story’s
events are told, and the specific compression choices lead to narratives of
different kinds. At the very basic level, a fictional story affords a compressed
view of the events, of the kind which is typically not available in actual
experience. Only a novel gives one access to a picture as complete as is required
in a form which is manageable, because it is compressed, and providing the
required multitude of viewpoints. The kinds of compressions will affect all
aspects of the narrative – time, viewpoint, characterization – but without the
compressions the text could not achieve its salience.
The processes of selection and compression play an important role in narra-
tive form. In the actual narrative organization they are often subordinated to the
choice of some subjectivity as the source of the narration, whichever term is
chosen to describe it. While it is clear that such a subjectivity is a narrative
construct, the reader has the impression that there is an intentionality and
epistemic status behind the discourse, which is typically talked about as a
narrator (Gibbs 1994, 1999). The intentionality is crucial in that the very act
of storytelling assumes the intention to use the story in its proposed form to
communicate some content, even if various forms of narrative experimentation
disrupt the impression of consistency and purpose. The epistemic status is
equally important as a constitutive feature of ‘narratorship,’ since, regardless
of the scope of knowledge the narrator displays (ranging from omniscient,
constituting a narrator who knows everything including characters’ thoughts,
to unreliable, with a narrator who has a very limited view of the facts), a narrator
gives the reader access to crucial narrative facts.
In this chapter, I will focus on the aspects of ‘narratorship’ that involve
specific configurations of narrative spaces, and also consider the macro-level
narrative viewpoint in its relation to the textual profiling of storytellers. As
60 Stories and their tellers

Chapter 2 made clear, narrative spaces as cognitive constructs depend on

various kinds of choices, and may not always take the form of narratives, not
even skeletal ones. Here, however, I will focus on how various conceptualiza-
tions of the story’s teller(s) affect the narrative as a whole.

3.1 Narrators, narrative spaces, and viewpoint

While the focus of this book is not primarily aligned with the interests in
narratology, nevertheless the distinctions of types of narration introduced by
Genette (1980) and later developed in numerous studies (Rimmon-Kenan 1983;
Bal 1985; Simpson 1993; Herman 2002) should briefly be mentioned here. The
two questions which seem to be most pertinent to both the original distinctions
proposed by Genette and their later expansions seem to seek a positioning of the
narrating fictional subjectivity within the story or outside, and aligning it with
that of a character or not. While these are indeed crucial distinctions, they fit the
framework proposed here only to some degree. First of all, the dual contrast
between ‘inside’ or ‘outside’ of the story cannot naturally be maintained in a
number of narratives where there are in fact different storyworlds set up,
possibly narrated through different types of narrator. Furthermore, as in the
case of the ‘Clippings’ space in The Blind Assassin discussed in Chapter 2, there
may be no single ‘narrator’ present in any form, while the coherence of the
narrative space is achieved through an inexplicit but shared epistemic and moral
viewpoint. While it is true that even the most fragmented narrative builds a
conceptual structure which I have termed ‘emergent story,’ the boundedness of
this construct is difficult to establish in a way which would clearly determine the
position of the narrating voice. It seems intuitively convincing to claim that if a
narrator is not at the same time a character, then he/she/it must have some
knowledgeable position in a different domain, but it is interesting to ask also
what that domain is and how the narrator gets access to the ‘story’ if he/she/it is
positioned outside of it. I will thus attempt to talk about the ‘type of narration’ or
‘narrator’ as a feature of the narrative space configuration in any given case, and
show how the spaces are linked via blending and viewpoint compression.
As regards the cases where a narrating voice is aligned with that of a character
within the bounds of the storyworld, the status of that voice also depends on the
role of the specific character in the text. If the character/narrator tells his or her
own story, in the first person, it is still typically the case that he/she relies on
facts reported by others as well, and the balance between ‘narrating-the-self’ or
‘narrating-the-story’ is different in different texts. For example, the narrator/
character in Eggers’s autobiographical novel is almost solipsistic in his focus on
his own consciousness (though even he relies on others in some fragments of the
story), while the narrator/character of Chatwin’s Utz is in fact telling the story
of another character, very much as an omniscient narrator would. While
 3.1 Narrators, narrative spaces, and viewpoint 61

narratologists have made all the required distinctions, I will assume here that
they can be elaborated and clarified in a broader context.
The categories to be discussed below emerge out of the understanding of the
concept of narrative space, as outlined above. Primarily, I will rely on the
concept of narrative space as a source of narrative structure (among others,
the features of the so-called storyworld), but, more important, as a source of
viewpoint configurations. I will understand narrative space viewpoint as a
particular use of the structure of the space. While the narrative may profile a
number of characters, and temporal and spatial structures, and involve knowl-
edge of certain facts, et cetera, the specific viewpoint, whether local or global, is
a matter of selecting an aspect of narrative structure within the space as fore-
grounded or specifically profiled. In other words, viewpoint does not neces-
sarily rely on a viewing subject, so that it is not necessarily ‘someone’s’ point of
view, but it relies crucially on a selection of an aspect of space topology which
provides a filter through which the events are narrated. Viewpoint may thus be
spatial or temporal in the basic sense of aligning the narration with a given
spatial or temporal dimension of the narrative space. But it can also be epis-
temic, when access to certain facts, but not others, is assumed (which includes
any range of narrating voices, from omniscience to unreliable narrator), or
ideological, when the very selection of facts presents the story in a specific
light.3 In this approach, trying to come up with a classification of viewpoint
types will not lead to a better understanding of the concept, though it may add to
an appreciation of its complexity. In fact, narrative spaces, even more than
mental spaces, can, and indeed have to, profile a number of possible viewpoint
options in order to construct a topology suitable for narrative processing.
While the ‘narrating fictional subjectivity’ may represent aspects of view-
point which remain constant during most of the narrative, there are also local
shifts of viewpoint which may affect expressions as small as individual senten-
ces or phrases. As I will try to show in this chapter and the next one, viewpoint is
best talked about as a configuration of viewpoint aspects, ranging across
narrative spaces involved, and not as a uniform feature of a narrative fragment
or the whole text. Even in a basic act of spoken communication one constantly
operates in a network of viewpoint dimensions – the speaker’s and the hearer’s
(with their epistemic and emotional stances), the time of conversation, the time
of events talked about, the spatial locations experienced and mentioned, epis-
temic or other viewpoints available in the conversational record, in the broad
context, or in the artifacts available, et cetera. Choosing to foreground one of
these aspects in a construction does not make other viewpoint angles disappear,
and shifts are common. If a speaker says in the course of a gossiplike exchange
I’m not saying I expect he will propose, but he might, she is manipulating a
number of viewpoints in just one sentence, even if we disregard the time and
place of the conversation for the sake of clarity. There is her epistemic viewpoint
62 Stories and their tellers

(I expect), the hearer’s viewpoint (you are saying he will propose), the speaker’s
rejection of attributed viewpoint (I’m not saying it), the man’s emotional view-
point (he might be in love enough to propose), the distanced viewpoint marked
by the past tense might (instead of may), etc. There is no reason to assume that
narratives are any less complex in this respect.
However, narratives can be talked about in terms of levels of viewpoint
structure, which I will refer to as ‘macro-level’ and ‘micro-level.’ First, there
seems to be a crucial difference between the consistent, narrative-wide choice of
the way in which the narrating subjectivity is profiled, and the sentence-level
shifts and manipulations of viewpoint. While the former emerges out of specific
kinds of narrative space configurations, the latter elaborate the topology of
individual spaces to give depth to the story. Second, macro-level phenomena
reflect the text’s assumed epistemic viewpoint and contribute directly to the
construction of the emergent story, while micro-level forms have more to do
with the construction of specific events and conceptualizations. The two levels,
however, are not independent, as the aggregate effects of micro-level shifts
affect the macro level all the time, especially with respect to the epistemic
viewpoint of a narrative space. Let me illustrate the interaction between levels
with an example, before moving on to the general discussion of macro options
(in this chapter) and micro phenomena (in Chapter 4).
Example (1) comes from the autobiographical novel by Dave Eggers A
Heartbreaking Work of Staggering Genius, where he tells the story of the
years when he was bringing up his young brother, after their parents had died.
The text follows the inner thoughts of the narrator/main character throughout. In
the text preceding (1) he talks about a night spent with a woman, while Toph, the
brother, slept at a friend’s house.

(1) I wanted to be home in case he came back early . . . made it in time . . . The house was
empty, and I dove into bed, fell back asleep, and when he came back home his
brother was there, of course had been there the whole time, of course had never left.
(AHWOSG, p. 112)

The first sentences, as the rest of the novel, refer to Dave (the sole narrator of the
entire text) as I, while he refers to Toph. But when reference to Dave shifts from I
to his brother, both levels of viewpoint are affected. At the micro level, the
viewpoint no longer focuses on Dave’s intentions (make Toph believe that he
spent the night at home), and shifts to the perception attributed to Toph (that
indeed Dave never left the house). But it clearly involves the macro level as well,
since the expression his brother assumes Toph’s view of Dave’s family role
(available throughout the text), and Dave’s view of himself as ‘Toph’s/his’ brother
(if the viewpoint totally shifted to Toph, my, not his, would be the required
pronoun). The fragment thus manipulates viewpoint to let the reader appreciate
both Dave’s intention to mislead, and the fact that Toph was convinced by the
 3.1 Narrators, narrative spaces, and viewpoint 63

situation Dave had staged, both of which are relevant to the macro-level narrative,
even though the effect is achieved through a minute shift at the micro level.
Discussing viewpoint as the specific presentation of the content of a given
narrative space makes it possible to acknowledge the reasons why a text may
profile several spaces. In the case of The Blind Assassin, macro-level viewpoint
varies significantly from space to space, while in other narratives the choice of
macro viewpoint may be more consistent. Even though in all cases the narrative
space is structured in terms of time, space, and the epistemic or emotional
disposition of the teller, et cetera, macro viewpoint foregrounds some of the
aspects of space topology and backgrounds others. For example, the consistent
first-person narration of Eggers’s novel allows him to focus on the main charac-
ter’s emotional states and obsessions, rather than on the consistent presentation of
events; thus the actual story, even though basically sequential, has to be pieced
together in the reader’s mind. In other texts the narration of events may take
priority, but then a specific temporal focus will be chosen while other pieces of the
timeline will be rendered through flashbacks or omitted altogether.
In what follows, I will argue that the macro-level viewpoint is determined by
various configurations of three types of spaces. First, there is story-viewpoint
(SV) space, the space where the ‘narrator,’ or some features of narratorship, are
located. The overarching narrative structure (the primary plot) can be referred to
as main narrative (MN) space, which may profile its own narrator(s), but may
rely entirely on the SV-space structure. Finally, a character in the MN-space can
be selected as the narrator, and we can then talk about Ego-viewpoint, such that
the character’s knowledge or intentionality provides the primary focus to the
narrative mode. I will consider various examples of these categories below.
I will also rely on two additional concepts introduced in cognitive grammar
(Langacker 1990a, 1991) to clarify the kinds of possible correlations between
construals prompted and the linguistic forms prompting them. While the verb may
require a construal including an agent, a selected grammatical form may not profile
the agent specifically. In an active voice construction, such as Tom broke the
window, both the agent (Tom) and the patient (window) are on-stage participants,
explicitly profiled by the form of the sentence; in the passive form The window was
broken, however, the agent is off-stage, but nevertheless remains a part of the
construal, since the act of breaking requires someone doing it. Various construc-
tions can manipulate the ways in which construals are prompted by expressions,
and narrative constructions can be claimed to do the same, with different elements
of the space topology on-stage or off-stage for different lengths of time.
I will describe some examples of various narrative constructions of tellers in
terms of their relationship to the MN-space of the story being constructed. My
goal will not be to cover exhaustively all available types, but rather to outline
ways in which such constructions can be talked about. I will also attempt to rely
on the linguistic correlates of these narratorial constructions, primarily tense and
64 Stories and their tellers

person, and refer to certain general ways in which these categories influence the
narrative. For example, the use of past tense as the primary form of narration
creates some distance between the narrator and the story, thus often opening
more room for various degrees of omniscience. The use of present tense, for
comparison, suggests a narration following the events as they develop and more
immediate access to the minutiae of the story, but leaves less room for reflection
and commentary. Furthermore, the choice of first person reinforces the effect of
immediacy and direct access, while the choice of third person is often correlated
with some distance and various degrees of ‘omniscience.’ These general obser-
vations will be further substantiated through the examples discussed below. In
what follows, I will look at some of the common narratological constructions
rendered by these choices and discuss examples which relate these concepts to
configurations of narrative spaces.

3.2 Types of teller and epistemic viewpoint

The so-called omniscient narrator is a common choice of narrative form – from
Henry Fielding’s Tom Jones to most Victorian novels, and all the way to
contemporary literature. In many earlier texts, as in Tom Jones, the narrator is
also ‘intrusive’ (addresses the reader directly), but as the novel form develops
this form is used less and less. ‘Omniscience,’ however, remains a common
feature of narrators. Considering that the concept in fact applies to both Henry
Fielding’s and Virginia Woolf’s work, it does not seem to be a category salient
enough to distinguish narrative techniques with any reliability. What it does
refer to, though, seems to be the narrative space set-up whereby the content of
all the narrative spaces of the text is accessible from an independent space in
which all other spaces are embedded. This ‘outer’ narrative space, which I
introduced above as ‘story-viewpoint space’ (SV-space), usually has little or no
topology of its own in terms of time, setting, events, or subjectivities, but it
houses a vantage point with the rest of the narrative in its scope. Because of the
SV-space, the story is presented as if there were a deictic ground for the overall
structure. I argue that the deictic elements such as time, speaker/hearer, or
location are profiled even if the narrator is not. The vantage point is cognitively
available in all cases, but the degree to which such a vantage point is associated
with a narrating subject or not accounts for the various choices made.4

3.2.1 On-stage narrators

For example, an intrusive narrator, as in Tom Jones, speaks as follows:
(2) Reader, I think proper, before we proceed any farther together, to acquaint thee that I
intend to digress, through this whole history, as often as I see occasion, of which I am
myself a better judge than any pitiful critic whatever. (Tom Jones, p. 17)
 3.2 Types of teller and epistemic viewpoint 65

Narrator: ‘I’
SV-space
Reader: ‘you’
Time: present

MN-space Character: he/she

Tense: past
Character: he/she
Character: he/she

NS1 NS2 NS3 NS4

Metanarrative space

Figure 3.1 Intrusive narrator (example 2)

Here is an on-stage (fully profiled) narrator who addresses the reader directly
and comments on expected reactions to the narrative (see Figure 3.1). The SV-
space thus has a certain topology – the narrator is on-stage, the time is later than
story-time. All characters are referred to in the third person. One of the results of
this set-up is that there is clear deictic distance between the SV-space and the
MN-space, including each of the narrative spaces (NS) profiled. While perhaps
artificial, this narrative set-up constructs its SV-space as close as possible to the
default discourse space, with the subjectivities of the speaker and the hearer, and
as such is also the closest to what might be considered the most salient narrative
set-up. In fact, the narrator’s mention of critics also sets up a metanarrative
space, where the entire novel and its discourse are available for independent
evaluation.
This kind of set-up is also interesting with respect to its epistemic viewpoint.
The source of all the information the story provides is localized in one space.
There is no need for the story’s narrative spaces to be evaluated from the
epistemic perspective or matched with appropriate subjectivities – all is in the
hands and mind of the narrator. This narrative form is not often used now,
although it is worth noting that it was also used by Bulgakov (to be discussed
below), with a highly humorous effect.
66 Stories and their tellers

The narrator may further attune his/her presence to the needs of the narrative
construction. In the first pages of Henry James’s novel The Portrait of a Lady
the narrative starts with an “I” narrator, commenting on what he intends to
render to the readers. The “I” voice introduces the scene through a fine and
extended description of the sunlit lawn where tea is served, but then the
narrator’s presence gradually blends away as the characters begin to speak
and take their roles in the story. The narrator’s voice in the first paragraphs is
only used to give the setting and the atmosphere of the first scene, but the
narrator’s presence is thus defined for the duration of the novel. Though this
narrative choice has little influence over the text as a whole, it is an interesting
confirmation of the need for the SV-space as a separate cognitive construct
available.

3.2.2 Off-stage narrators

It is interesting to compare these examples with the more contemporary form of
omniscience, as in Anne Tyler’s novel Ladder of Years. The main character,
Delia, comes to a job interview, and is waiting in the office:
(3) She crossed to the desk, which was bare except for a telephone and a typewriter. She
lifted a corner of the typewriter’s grey rubber hood. Manual; not even electric. (She
had worried she would find a computer.) She gave a small, testing spin to the swivel
chair behind it.
Good afternoon, she would say. I’m here to ask if . . .
No, not ask. Ask was too tentative. (Ladder of Years, p. 94)

This is a more common type of omniscient narrator, who knows Delia’s spatial
and temporal location and follows her actions (she crossed to the desk, lifted a
corner of the hood, et cetera). More important, the narrator knows what Delia
registers consciously in the environment (bare desk, manual typewriter), and
what her thoughts are. The narrator is thus off-stage (does not speak as an
independent subjectivity), but all that is said about Delia is relayed through the
SV-space, which is located temporally later with respect to the main narrative
space. The only glimpse of a narrator’s subjectivity is the comment in paren-
theses: She had worried she would find a computer. It is less important that the
sentence flashes back to Delia’s thoughts before her arrival, but it is crucial that
it is an explanation of the reasons why Delia checked the kind of typing
equipment she would be using if she were to get the job, and why the phrase
not even electric is to be interpreted as an expression of relief rather than dismay.
The almost invisible narrator is thus present here and guides the reader in
interpreting Delia’s words and actions. In other words, while off-stage, the
narrator’s subjectivity is still present as a source of coherence and organization
of the story.
 3.2 Types of teller and epistemic viewpoint 67

The next example comes from Virginia Woolf’s modernist novel Mrs.
Dalloway:
(4) But – but – why did she suddenly feel, for no reason that she could discover,
desperately unhappy? . . . no, it was not Sally Seton saying that Richard would
never be in the Cabinet because he had a second-class brain (it came back to her); no,
she did not mind that; nor was it to do with Elizabeth either and Doris Kilman. (Mrs.
Dalloway, pp. 120–121)
Here is an ostensibly similar third-person, past-tense narration (which
Humphrey [1954] describes as ‘indirect internal monologue’), but the narrator
does not say much about the character’s actions and focuses on the flow of
thought in the character’s mind, with just a minimal level of commentary – it
came back to her. Throughout the novel, narration of events is minimal, and
transitions between episodes are primarily transitions from one consciousness
to another:
(5) “The time, Septimus,” Rezia repeated. “What is the time?”
He was talking, he was starting, this man must notice him. He was looking at them.
“I will tell you the time,” said Septimus, very slowly, very drowsily, smiling
mysteriously at the dead man in the grey suit. As he sat smiling, the quarter
struck – the quarter to twelve.
And that is being young, Peter Walsh thought as he passed them. (Mrs. Dalloway, p. 70)
At first, the reader is given access to Septimus’s visions, occasionally interrupted
by scraps of conversation between Septimus and his wife. But in the last sentence,
the viewpoint abruptly shifts to Peter Walsh, and it remains in his thoughts for
some paragraphs. These are but a few examples of the way in which a narrative of
this type includes much more than the specific viewpoint of a specific character at
any given point in the story. There are many Ego-viewpoints represented
(Clarissa, Rezia, Septimus, Lady Bruton, Miss Kilman, et cetera) through reports
of the characters’ acts and thoughts, but the narrating voice is minimal and the
distance between the SV-space and the MN-space is also much reduced.
The set-up with Ego-viewpoints is represented in Figure 3.2. While the case
of example (3) is different in that one Ego has been selected (Delia), and (4) and
(5) represent a multiplicity of parallel Ego-viewpoints, the general narrative
space structure is similar.
Also, the dotted lines which link characters to the zero narrator represent the
fact that the viewpoint of a character becomes the narrative viewpoint for so
long as that character’s viewpoint structures the narrative – the entire novel, as
in (3), or short fragments, as in (4) and (5). This ‘uploading’ of viewpoint to a
higher level constitutes the narrative effect which will be discussed throughout
the text – that of viewpoint compression. It is this compression to SV-space
which yields the effects wherein the narrator is off-stage and yields the story-
viewpoint to a character.
68 Stories and their tellers

SV-space Narrator: 0

Reader: 0

Time: present

MN-space
Character2: he/she

Character1: he/she
Tense:
Character3: he/she past

NS1 NS2 NS3 NS4

Figure 3.2 Ego-viewpoints

3.2.3 Omniscience and narratorship

There are thus many ways in which the omniscient narrator participates in the
story construction. The narrator may be on-stage, and thus be given an inde-
pendent voice, or off-stage, orchestrating the narrative. But in both cases the
subjectivity talked about as a narrator is part of the topology of the SV-space.
The third person in all of the instances discussed above is a device setting up a
deictic centre of the narrating subjectivity outside of the main narrative space.
Also the choice of the past tense serves similar purposes – it distances the main
narrative space from the SV-space. The concept of ‘omniscience’ relies on the
setting up of a deictic center independent of the story’s main space, which then
provides a bird’s-eye view of the structure of the story. The choice of the
subjectivity being on- or off-stage has stylistic consequences, but exerts much
less influence over the way the story is read.
At the same time, the narrating subjectivity, whether intrusive or completely
backgrounded, is a source of a variety of blends, involving both the characters
and the reader. The intrusive narrator requires that the reader or ‘narratee’ blend
him/herself with the position of the addressee in the fictional deictic set-up of
the SV-space. The interaction still proceeds only in one direction (from the
 3.2 Types of teller and epistemic viewpoint 69

narrator to the reader, but not the other way round), but the reader is given the
illusion of a personalized narrative account which addresses the reader’s story-
construction needs directly and guides the reading process. In a sense, the
intrusive narrator style is an interesting (even if now rare) narrative form
which openly acknowledges that the story told is not necessarily sufficient for
the reader to grasp all that is required.
It is also interesting to point out that even the most intrusive of narrators is not
restricted to the SV-space as if it were an ivory tower, and may actually signal
some degree of involvement in the main narrative space.
(6) He now lived, for the most part, retired in the country, with one sister, for whom he
had a very tender affection . . . This lady was now somewhat past the age of thirty.
(Tom Jones, p. 17)

In (6), the narrator tells a story which is presumably past with respect to the SV-
space, and yet occasionally uses the adverb now in the way which is mostly
found in free indirect discourse (see Nikiforidou 2010; also Chapter 7). In the
case of (6) such constructions are found in narration, and not in represented
speech and thought fragments. This is a result of the narrative set-up in which
the narrator is explicitly communicating things to the reader, so that narration in
this case turns into a form of represented speech, but of the narrator, not of a
character.5 The narrator of Tom Jones does not act as a character as well, and yet
needs to mark a presence both in the MN-space and in the SV-space – as a
witness-like participant in the former and as a narrator in the other. Crucially, the
use of now here is referring to the time of storytelling, not the time of the story’s
events, so it helps establish the state of affairs in the MN-space with respect to
the moment when the narration begins. It negotiates the relationship between
the SV-space and the MN-space.6
The omniscient narrator in Tyler’s Ladder of Years is different. The reader is
not an overt addressee, the narrator is off-stage, and the narration does not
explicitly signal the SV-space. The narrator can naturally blend with the char-
acter and report the thoughts (Manual; not even electric.), but also comments on
these same thoughts by evoking another moment in the story, whether previ-
ously narrated or not (She had worried she would find a computer.). Such
narrative moments suggest that the narrator can access any point in the MN-
space at any point in the narration. In a way, the narrator’s viewpoint alternately
blends with the character-viewpoint in MN-space, or returns to the SV-space
viewpoint, and this alternation between spaces constructs a specific form of
omniscience. But the blend favors one character, so that the viewpoint alter-
nation affects two stable viewpoints in two spaces.
The difference between Tyler’s narrator and that of Mrs Dalloway is mainly in
the proportion between the SV-space viewpoint and Ego-viewpoint. While Tyler’s
narrator reports various actions and movements, Woolf’s alternates from character
70 Stories and their tellers

to character and intervenes only occasionally. There are no transitions from one
Ego-viewpoint to another, or between the inner and outer states of characters,
so that narrative spaces are naturally compressed into one flow of the narrative.
The primary blend of this kind of narration is compressing the SV-space viewpoint
with subsequent Ego-viewpoints in MN-space, and backgrounding the story-
viewpoint.
To sum up, these ‘omniscient’ blends are varieties of a similar initial set-up,
and there are probably other varieties possible. What remains constant is some
degree of profiling of the SV-space, while differences result from various
degrees of compression of the distance between the narrator viewpoint in SV-
space and any specific character viewpoint in the MN-space. The higher the
degree of compression, the less narrating the narrator does and the closer the
narrative gets to some variety of the ‘stream-of-consciousness’ technique.

3.2.4 Tense, person, and distancing

It seems that the distance between the SV-space and any of the narrative spaces
is most saliently marked by the choice of the third person, not of past tense.
Some texts, such as Neal Stephenson’s Snow Crash, maintain the third person
and shift to the present tense (see Figure 3.3), and the effect may be similar to
some of the narration modes described above:
(7) The Oldsmobile keeps popping. Jason can’t think of what it is until he realizes that
people are shooting at him. Good thing he let his uncle talk him into springing for full
armor! When he figures that one out, he actually gets psyched. This is the real thing,
man! He’s driving around in his Olds and the bastards are shooting at him, and it just
don’t matter. (Snow Crash, p. 147)

The Ego-viewpoint here is maintained for some length of the narrative, and the
characters whose viewpoints are thus represented change very often. The use of
the present tense signals temporal viewpoint compression up to the SV-space,
and the effect is that the narrative now favors keeping record of the on-line
character’s thoughts and perceptions (The Oldsmobile keeps popping; This is the
real thing, man!). In spite of the present tense, on-line form, the narrator still uses
varieties of STR (he realizes that people are shooting at him), and narration, but
the consistency with Ego-perspective is achieved differently. Even in the nar-
rated parts (When he figures that one out, he actually gets psyched.) the narrator
chooses expressions which would seem more appropriate to the character whose
Ego-viewpoint is used here. It is thus an interesting blend of the narrator’s
epistemic viewpoint and the character’s discourse style, which makes the sub-
sequent Ego-narrator choices less conspicuous and gives coherence to longer
narrative fragments. At the same time, viewpoint compression here affects not
only the character viewpoint (as in Figure 3.2), but also temporal viewpoint.
 3.2 Types of teller and epistemic viewpoint 71

SV-space Narrator: 0

Reader: 0
Time: present

MN-space
Character2: he/she

Character1: he/she
Tense: present
Character3: he/she

NS1 NS2 NS3 NS4

Figure 3.3 Present time and present tense (example 7)

For comparison, first-person/present-tense narratives (typically not consid-

ered ‘omniscient’) go still further in compressing the SV-space and the MN-
space, so that in effect the SV-space viewpoint and the MN viewpoint become
the same (see Figure 3.4):

(8) I open the window, turn up the volume. I pass two cars at once and get on the
highway . . . Through the toll, the light, onto a ramp, onto the bridge. Now I can’t turn
back. The Oakland shipyards to the left, a billboard encouraging the saving of water.
(AHWOSG, Eggers, p. 126)

In (8), the narrator and the character subjectivities and viewpoints are fully
blended, but the SV-space does not disappear. The record of events such as
opening the window or passing other cars narrates the driver’s actions and
measures the passage of time since Dave (the Ego-narrator) left home, inde-
pendently of his on-line record of perceptions (shipyards, billboard), and
thoughts (I can’t turn back). What seems to be the most salient characteristic
of this narrative mode is the completeness of the narrator/character blend, so
that all parts of the story, including the events which took place prior to the
storyline in the MN-space, are told with the same voice.
72 Stories and their tellers

SV-space Narrator: ‘I’

Reader: 0
Time: present

MN-space

Main Character: ‘I’

Character2: he/she
Tense: present
Character3: he/she

NS1 NS2 NS3 NS4

Figure 3.4 Ego-viewpoint in the present tense (example 8)

There are small parts of the story told in the past, but the difference is that the
‘past’ parts are not intended to be an on-line account of one person’s partici-
pation in the events, but are a more ordinary distanced narrative. An example
from the same novel is given in (9):
(9) When my father was in intensive care, about a day and a half from throwing in the
towel, a priest was sent, presumably to administer last rites. After meeting him and
ascertaining the purpose of the visit, my father quickly dismissed him, sent him out.
(AHWOSG, p. 40)

Characteristically, the style of passage (9) is markedly different from (8), as the
events are rendered in distanced form, including past tense, markers of epis-
temic distance (presumably), passive voice (was sent) and formal vocabulary
(ascertaining the purpose of the visit). These brief moments when Dave leaves
his frantic on-line persona aside and assumes a narrating role with all serious-
ness of purpose mark viewpoint shifts, but also reconfigure the type of narrator
Dave is and can be. While most of the text is in the frantic present-tense, those
composed past-tense parts may be read not just as shifts of temporal perspective,
but also as glimpses of a different ‘narratorship.’ While both (8) and (9) belong
to the book’s MN-space, their style is adjusted to what is being narrated – past
 3.2 Types of teller and epistemic viewpoint 73

family events can be presented in a reflective and serious way, but the ongoing
roller-coaster of bringing up a young brother calls for a different narrative style.
The multiplicity of smaller spaces set-up is also represented in Figure 3.4 by the
smaller circles around more major narrative spaces.
The types of narration sketched above are distinguished, via constructional
means, as reflecting a specific relationship between the SV-space and the MN-
space. The compressions along the lines of temporal distance (past or present)
or the distance between a deictically determined communicator in the SV-
space and a character in the MN-space (first or third person) yield several
types, even though they are not always clearly distinguishable. The important
point is that in most cases the specific viewpoint configuration remains as the
default set-up throughout a novel, so that various specific parts of the narrative
structure are subjected to the specific SV/MN relationship. For example, all
the numerous Ego-viewpoint blends in Mrs. Dalloway are naturally assumed
to be subjected to the MN-space of the day of Clarissa’s party. Similarly, the
narrative of Snow Crash blurs the Ego/narrator boundaries through consistent
stylistic form. Also Eggers’s narrative, which is less stable in its viewpoint set-
up, remains within the bounds of an Ego-narration. Features of style, however,
contribute significantly to how the narrative space set-up will be framed.
Eggers’s narration is scattered and shifts narrative-space-viewpoint all the
time, and thus, by comparison, the characters in Mrs. Dalloway seem to
‘think’ in voices too similar and stylistically colorless to represent abrupt
viewpoint shifts, even though the Ego-viewpoint changes all the time. The
stylistic dimension (choice of register, degree of distance) thus seems to create
the quality of narration which obliterates or highlights the viewpoints profiled.
The smoother the transition from narrator to Ego, the more cohesive the
narrative seems.
The scattered quality of Eggers’s story persists even when the Ego-
viewpoint is maintained. Example (10),7 which follows (8) immediately,
without even starting a new paragraph, has to be read from the point of view
established in (8) (Dave, driving to San Francisco, after he left his brother with
a babysitter):

(10) I will come home and the door will be open, wide. The baby-sitter will be gone and
there will be silence . . . Blood on the walls, handprints soaked in blood . . . There
will be a hearing, a trial, a show trial –
How did you come to meet this man, this baby-sitter?
We found a posting, on a bulletin board.
And how long did your interview of him take?
Ten, twenty minutes. (AHWOSG, p. 126)

Most readers would agree that Dave’s worries about his brother are exaggerated,
so that he could be viewed as an ‘unreliable’ narrator. The concept was
74 Stories and their tellers

introduced by Booth (1961), and was recently discussed by Semino (2007) as

related to what stylisticians refer to as ‘mind style.’ Whichever term is used, the
point seems to be that some narratives are relayed through an Ego who is either
unaware of or unable to grasp the nature of the events (for example, Benjy in
Faulkner’s The Sound and the Fury). As Semino rightly points out, there are a
variety of ways in which the specificity of the constructed mind influences the
narrated content. The so-called unreliable narrator is reliable enough for the
reader to construct the story, but the narrative viewpoint is indeed expressed
somewhat differently. Primarily, unreliable narrators are presented as only
capable of a very specific epistemic viewpoint, so that the narrowness of the
understanding of events strikes the reader as inaccurate. Whether such a narrator
is accessible through the first-person or third-person style, other viewpoints are
not directly available. It seems that such cases constitute another blend, such
that the SV-space is entirely blended with the MN-space relayed through one
specific Ego. Such a narrative space naturally calls for the reader to construct
alternative viewpoints, prompted only by the bizarre consequences of following
the viewpoint available. In the case of (10), the readers understand that a parent
would most likely not be put on trial for leaving a child with a babysitter, but
does it make Dave ‘unreliable’? He is, after all, narrating his obsessions and
worries, which are a crucial part of the story told. It seems that the viewpoint is
entirely subjected to the goals of the narrative, and such goals may not include
giving a reliable account of events. As the examples above suggest, the goals are
more often focused on representing the character’s epistemic and emotional
perspective.
The examples above all point to the specificity of distance in narrative
discourse. Dancygier and Vandelanotte (2009) and Vandelanotte (2010) dis-
cuss the varieties of distancing devices in literary texts – specific forms of
represented speech and thought, temporal distance, metalinguistic distance,
epistemic and emotional distance – mostly available through textual access to
more than one viewpoint-marked narrative space. In the fragments considered
above, the concept of narratorship itself seems to emerge from the varieties of
temporal and epistemic distance between the major spaces involved (SV-
space, MN-space, Ego-space). While the temporal distance (e.g., present
SV/past MN) and epistemic distance (SV access to story/Ego access to
story) seem to be the most common devices, the stylistic features marking
emotional, evidential or social distance contribute substantially to the impact
of the text, including perception of viewpoint. Ironically, Mrs. Dalloway may
seem more coherent, in spite of its galaxy of narrating Egos, while Eggers’s
text feels scattered, even though there is only one Ego narrator, because its
emotional temperature keeps jumping up and down and his mind wanders in
most unlikely directions. Overall, the viewpoint texture of a narrative as a
whole depends as much on the general space set-up as on the stylistic profiling
 3.2 Types of teller and epistemic viewpoint 75

of various aspects of viewpoint in those spaces and the foregrounding or

backgrounding of various types of distance.

3.2.5 Constructional compositionality

Viewing the narrative options described above as constructions highlights
important facts about interpretive consequences. There seem to be four con-
structional parameters involved: narrator on- or off-stage, person, tense, and the
profiling of Ego-viewpoint (or several). All these parameters represent inter-
actions between the SV-space and the topology of the MN-space, suggesting
various degrees of distance and compression.
The degree of profiling of the SV-space correlates with the degree of
distance between the SV-space and the MN-space. If the narrator is an on-
stage part of the SV-space, the degree of compression with the MN-space is
rather low, as the storyworld and the character are not seen as participating in
the narration. For comparison, an off-stage narrator invites less distance and
more compression, especially with many Ego-viewpoints profiled. The yield-
ing of narrative voice to Egos (as in the fragment from Woolf) downplays the
role of the SV-space, reducing the distance to the temporal dimension, and
yielding a narrative where the character perspective dominates and the organ-
izational role of the narrator is not foregrounded. Intrusive narration and
multi-Ego narration are thus two extremes along the spectrum of narratorial
involvement. In this context, the choice of first-person or third-person narra-
tion is also dictated by the degree of distance involved – with first-person
narration being fully compressed into the SV-level (so that the independent
narrating voice is not construed by the set-up), and the third person giving the
narrator opportunities to intervene and comment.
The choice of tense is less consequential. I have assumed that the SV-space is to
be treated as ‘present,’ since the narrating is always done in the narration’s present.
This yields a rather straightforward set-up in the case of narration in past tense, as it
only underscores the distance between the SV-space and the MN-space. Narration
in the present tense, however, does not suggest that the time of the story’s events is
the same as the time of the story-telling. Rather, it diminishes the distance between
the spaces, and allows the story’s time to be compressed into the SV-space – thus,
even if the SV-time is present time, it does not mean it is the same as the story’s
present tense, which will still be past time, wih respect to SV. There is still
conceptual distance between the SV- and MN-spaces.
This brief overview suggests that the actual form of any given narrative
emerges out of a combination of these constructional parameters, and is pri-
marily defined by the degree of remaining distance between SV- and MN-
spaces. Crucially, not many narratives are entirely consistent in the choice of
the narrative space configuration, and such shifts can be explained only against
76 Stories and their tellers

the template of the SV-space. The SV-space thus needs to remain separate from
other narrative spaces, so that the overall coherence of the story can be estab-
lished at some stable narrative level.

3.3 Second-person narratives

Some of the issues raised above can further be clarified with a look at the less
common form of narratives – with second-person pronoun as the primary SV-
space-related choice. These instances are related to the SV set-up described
above, but exploit it in different ways. In fact, there are more significant
differences across examples than, let’s say, among third-person narratives.
One of the texts where the second person dominates is Oriana Fallaci’s novel
The Man, describing the life of Alekos Panagoulis, her close friend. She
dedicates the book to him (For you), then, in the introduction, promises to tell
your tale and describes Panagoulis as my only possible interlocutor. The story
follows, with Panagoulis consistently referred to as you.
(11) You thought about the amnesty all night, sometimes believing it, sometimes not,
and when you didn’t believe it, you felt serene, when you did believe it, your
conscience was split in two. (The Man, p. 132)

Interestingly, the second-person form is used here to represent the inner thought
of the main character, so that the text contains all the narrative elements the
reader expects. There is also an I in the novel, referring to Fallaci herself, also
framed as the narrator:
(12) As if stung by a wasp I stood up. I said I had to leave you and go find a hotel. You
answered categorically: “You’re not going anywhere. You stay here.” (The Man, p. 150)

The novel thus explicitly sets up a deictic frame, with the writer as an
equivalent of the speaker, and the main character as the addressee (see
Figure 3.5). The addressee position here is not that of the reader or narratee,
who remains a standard recipient of what the text communicates. To accom-
modate the you, we need an extra space of the reality, in which Fallaci and
Panagoulis were friends, and we need to establish a cross-mapping between
the two of them as people and as narrator/character and character/addressee
respectively. The novel’s set-up thus explicitly comments on reality and
fiction, while also blending the two grounds – the imagined deictic ground
wherein Pangoulis is an addressee, and the novel’s narrative ground, where the
addressee is the generic reader.
The set-up is different in another novel which relies heavily on you –
Calvino’s If on a Winter’s Night a Traveller. The structure of the text is very
interesting, with the novel itself partially embedded in it, and with various
excursions into other (presumably nonexistent) novels. You in the text refers
 3.3 Second-person narratives 77

SV-space
Discourse space
Speaker: Oriana Narrator: ‘I’
Hearer: Alekos Reader: 0
Time: present
Reality-space
Oriana
Alekos

MN-space Tense: past

Character: Oriana
Character: Alekos

NS1 NS2 NS3 NS4

Figure 3.5 Second-person narrative (examples 11 and 12)

to the reader, but the reader is present not only in the SV-space, as is usually the
case, but also as a character whose experience of reading and interacting with
the text is the primary theme of the novel:
(13) The novel you are reading wants to present to you a corporeal world, thick, detailed.
Immersed in your reading, you move the paper knife mechanically in the depth of
the volume. (If on a Winter’s Night a Traveller, p. 42)

The present tense and the progressive aspect in (13) also suggest that the
intended construal is that of the ongoing process of reading (with the exception
of fragments from embedded imaginary novels, which are differently framed).
The MN-space is initially structured as taking place in a railway station, with
the reader as one of the characters and the narrator as the other. The text also
implies that the scene is developing within the reader’s field of vision:
(14) I am the man who comes and goes between the bar and the telephone booth. Or,
rather: that man is called “I” and you know nothing else about him . . . (If on a
Winter’s Night a Traveller, p. 11)

However, the reader’s observations are relayed through the narrator’s voice, as
in (15).
(15) It is obvious that I am a subordinate, I do not seem the sort of man who is travelling
for personal reasons or who is in business for himself. (If on a Winter’s Night a
Traveller, p. 14)
78 Stories and their tellers

SV-space
Discourse space
Speaker: Narrator Narrator: ‘I’
Hearer: Reader Reader: 0
Time: present

MN-space Tense: past

Character-Reader: ‘you’ Character-Narrator: ‘I’

NS1 NS2 NS3 NS4

Figure 3.6 Second-person narrative (examples 13–15)

Similarly to The Man, the novel profiles the narrator and the addressee as
characters, but neither of them is correlated with reality, so that the cross-
mappings reach across MN-space and SV-space (see Figure 3.6). The two
characters are thus compressed up to the SV-space, but the narrator retains his
ability to reach into the mind of the character, even when the character’s
thoughts are about the narrator as character. The identity of the I character
is not developed to include an independent subjectivity, and he is reduced to his
narratorial function and the perception of the reader-character.
Second-person narratives (cf. Fludernik 1994) thus rely on mechanisms
described above for first- and third-person texts. The presence of the narrator
and reader slots in the SV-space is exploited for cross-mappings. However,
while the reader is only profiled in the text in intrusive narration, in the examples
above other cross-mappings are set up – with a character in the MN-space, or
with an addressee in the reality space. Furthermore, the viewpoint configuration
such as in (14) and (15) includes a blend wherein the I-narrator does not report
his or her own story, but another character’s, cross-mapped with the reader.
Finally, such a set-up does not invalidate the added layer of narrative space
structure wherein there is an actual reader reading Calvino’s novel.
 3.4 The teller, the author, and the character 79

This discussion has not mentioned the narrative use of you which is very
similar to the colloquial discourse usage – the generic you, as found in sugges-
tions, how-to descriptions and other contexts where a general pattern of behav-
ior is being described. Because this generic form is not specific to narratives,
I will not discuss it any further.
Thus far, I have tried to trace some of the constructional correlates of general
viewpoint mechanisms in a fictional story. The tense and pronoun choices
discussed here are specifically those which can be naturally explained in the
overall narrative space set-up. There are also more local, micro-level pronomi-
nal choices, sometimes affecting the concept of narratorship, but these will be
discussed in Chapter 5. At the same time, there are also text-wide consequences
of how the teller of the story is construed, especially with respect to its author, as
tellers can be discussed from the point of view of the overall coherence of the
text. The question of how the conceptualization of the teller affects the story will
be considered in the remainder of this chapter.

3.4 The teller, the author, and the character

Recent work by J. M. Coetzee explicitly addresses the complexities of the role
of the teller in the understanding of a text (see Dancygier [2010] for broader
discussion). Coetzee himself is an author of fiction as well as nonfiction, and his
recent novel, Elizabeth Costello, combines the two. In the text, he includes
various articles and lectures he wrote or delivered as J. M. Coetzee and attrib-
utes them to a fictional author, his alter ego Elizabeth Costello. The resulting
blend represents authorship from two perspectives: on the one hand there is the
subjectivity named on the cover, who only speaks through the authored text,
while on the other hand there is the actual person, having reasons to write what
she does. Costello is different from any other author we know only from a
vignette on the dustjacket in that she also represents the desire to be understood
and heard, strong personal beliefs, and tiredness with fame. It is natural to read
her as a blend of the fictional subjectivity and that of the actual author.
In the next novel, Slow Man, Costello appears again. She walks into the
apartment (and life) of a character, Paul, who loses his will to act after a
crippling accident. This narrative construction, whereby there is an author
inside the narrative space, gives a view of authorship which questions a number
of default assumptions. While other characters in the novel as well as its critics
suggest that Costello’s interventions are those of an author writing the novel
who wants to put Paul in her book, this seems to be an oversimplification. True,
Costello tries to prompt Paul to action, but not in order to use the events as
narrative material, but rather in order for the narrative to ‘take place.’ As she (or
Coetzee) sees it, it is the character’s responsibility to ‘create’ the story by living
it. The story of Slow Man does not exist in Costello’s (or anybody else’s) mind,
80 Stories and their tellers

and the reader thus reads a novel as existing in a narrative space independent of
the author’s or narrator’s or character’s subjectivities, but driven directly by
their agency as characters and writers. The presence of Costello as an embodied
narrative choice (to live a life worth writing about and let the novel emerge from
it) constantly reminds Paul that a different story could be lived and told, but if
we recall that she is in fact Coetzee’s alter ego, the choices she offers are choices
Coetzee might be associated with. In effect, the character is not ‘put in a book,’
but the author is.
In the next novel, A Diary of a Bad Year, Coetzee’s next alter ego, Señor C.,
speaks more openly of his role and claims his right not to be seen as ‘merely’ a
set of rhetorical tricks. Coetzee himself is a critic, but in his own case, he speaks
through his novels on his own authorial viewpoint and his understanding of
what fiction is. He seems to find the pronouncement of the author’s death (even
if it is meant just to be provocative) to be an insult to his personal and
intellectual involvement in the process of writing and, while he is willing to
yield most of the power to the character and the logic of the story, he does not
like the idea of giving it all to the reader.
To sum up, Coetzee’s recent work suggests a re-evaluation of the narrative
which goes in two directions. On the one hand, fiction needs to be freed from the
overpowering subjectivity of a teller, and given its own place as a representation
of a story. On the other hand, stories and characters need their authors to take
shape and be convincing. While the author cannot write a story that a character
refuses to live, life is full of stories which are lived in a way worthy of an
author’s attention. The viewpoints which thus emerge are all subjected to the
viewpoint of the story as such.

3.5 Multiple tellers

The question of narratorship and the cognitive status of a teller becomes
particularly interesting in the cases where the text as a whole profiles more
than one teller, while ostensibly telling the same story. The case of narratives
which rely on various focalizers is not the one under scrutiny here; instead, I will
focus on cases where the text designates different characters or different iden-
tifiable narrators to tell different partial stories.
I have discussed one such case in an earlier work (Dancygier 2008a), Jan
Potocki’s The Manuscript Found in Saragossa. The text has an interesting
history, and a still more unusual composition. It was written in French by Jan
Potocki, a Polish aristocrat living and writing in the late eighteenth and early
nineteenth century, whose life story could make a fascinating novel of its own.
The text was first published in two different parts, so that even reconstituting the
entire book in its intended form was a historical and critical challenge. The ‘novel’
consists of 165 stories (or so one story-counter claims), connected in different
 3.5 Multiple tellers 81

ways through embedding and digression, told by a number of different tellers, and
yet, in some way, constituting a coherent narrative of the story of the main
character, Alphonse van Worden, and of the Moorish family of the Gomelez, to
whom he is related through his mother. The intricacies of the narrative in this case
challenge the reader in many ways, as even keeping track of which level of
embedding a given story belongs to is not a straightforward task.
As I argued elsewhere (2008a), the book does form a coherent whole not
because the stories naturally form one overarching storyline (they don’t), but
because various stories contribute to the establishment and maintenance of
viewpoints. In other words, tellers are not necessarily responsible for a piece
of narrative structure, but their stories represent ways of interpreting the overall
story being told, and thus the whole is not only a fascinating narrative, but also a
reflection on honor, religion, tolerance, the value of wealth, as well as mathe-
matics, philosophy, and history. One of the most salient questions is the choice
between religion and faith on the one hand (this is the time of the Inquisition!)
and honor on the other hand, as principles guiding a moral man. Indeed, these
are the choices the main character, Alphonse, faces throughout the text, but the
options are presented through different stories, told by different people in
different places and times. Thus there are a number of stories whose conclusions
suggest that faith is the best foundation, while others could be read as mocking
religious devotion, and still others reinforce the idea that honor is the best guide.
Within his own story, Alphonse chooses honor, and gains respect as well as
riches. The power of the Manuscript, then, is in the setting up, elaboration, and
reinforcement of viewpoints, while their continued presence and salience in the
text are constructed through a variety of smaller narratives, none of which can
sustain the subtlety and multidimensionality of the moral and philosophical
debate. Thus the number of ‘tellers’ and narrative contexts is the source of
variety and interest, but the main viewpoints guiding the story and its main
character are established over and above the various narrators, through the
construction and selection of the stories themselves.
There are also cases where the story profiles more than one teller but in the
end remains unspecific as to the subjectivity which the story is to be attributed
to. There are numerous examples of this in contemporary literature, but I will
discuss one in detail here – Mikhail Bulgakov’s The Master and Margarita
(M&M) – to propose also an analysis of how the misleading structure fits the
viewpoints represented.
The novel has two major narrative spaces: the story taking place in Moscow,
in the late 1920s, told by an omniscient intrusive narrator (M&M1), and a novel
about Pontius Pilate and a wandering philosopher Yeshua who is crucified at the
end (PP&Y). The narrator of the PP&Y novel is also omniscient. M&M1 is a
story of frightening, but also hilarious events which occur in Moscow as a result
of the visit by Satan (traveling as Mr. Woland) and his team. Among other
82 Stories and their tellers

things, Woland organizes a ball where all the sinners and souls committed to
eternal damnation have a day’s break. He needs a lady to accompany him, and
an appropriate woman is soon found. Her name is Margarita.
The PP&Y novel was written by a man living in Moscow, who only appears
in the text as the Master (his name is unknown). In the novel, the story involving
Pontius Pilate, Jesus (here Yeshua), and Judas, and leading to the Crucifixion is
told in an entirely realist mode, with no apparent allusions to the religious
context of the story. The writing of the novel and the reasons why it was never
published (in fact, the manuscript was burned) form one of the narrative themes
in M&M1, where the Master is one of the characters.
From the narrative perspective, the most interesting aspect of the PP&Y
novel is how it is given to the reader. The first installment is told by Woland
(Satan) to two incredulous men – the critic Berlioz and the poet Ivan Bezdomny.
However, it is presented as an eyewitness account – Woland describes what he
saw and heard in Pilate’s palace. Ivan the poet is shocked by this encounter and
eventually ends up in a lunatic asylum, where the next fragment of the novel
appears to him in a dream. Finally, Margarita recovers the entire manuscript
(Satan gives it back to her, adding that ‘manuscripts don’t burn’) and reads it
(throughout the remainder of the M&M1 story), which is how we get to the end
of the text. The last words, as announced earlier by Margarita, are “procurator of
Judea, the equestrian Pontius Pilate.”
The text of the PP&Y novel is thus presented at once as a novel on a subject
entirely divorced from the reality of Moscow in the post-revolution years, and
as an account of Pilate’s encounter with Jesus and of the crucifixion, which is
fictional (it is a novel), but also factual (eyewitness report). On yet another level,
Ivan and the Master are both insane, and Margarita, by usual standards, loses her
sanity as well – so perhaps these are just ravings of lunatics – but then, how
come they give a very sane account of events which are typically seen in more
otherworldly terms? The choice of narrators is the only source of such
questions.
The search for authorship within the storyworld is equally elusive. PP&Y is
the novel written by the Master (even though it exists beyond the manuscript or
the author’s consciousness). But in the final scenes the Master is allowed to free
Pontius Pilate and is thus still responsible for his characters and has the power to
change their fate – as we expect a true author to do. But now he not only ‘writes
them,’ he can change their lives in the reality the novel represents, which is now
a part, though an invisible part, of the reality as we know it.
Also the authorship of M&M in its entirety becomes questionable. The final
words of the Master’s novel, “procurator of Judea, the equestrian Pontius
Pilate,” also end the story of Ivan the poet, and the entire book. The words
become a powerful narrative anchor which marks the limits of authorship
as based on magical insight into the truth, not on the power to create fiction.
 3.5 Multiple tellers 83

In this sense, the Master, the poet, and the actual writer Bulgakov are all
candidates for authorship-as-insight, while the Master and Bulgakov are
blended as possibly actual writers of the words. Indeed, there is much in
Bulgakov’s biography that matches the misfortunes of politicized criticism
which destroy the Master, but, first of all, the final words seem to be the ultimate
wink of this self-ironic and at times amusing, at times deeply moving novel. It
does not really matter who wrote what, just in case some eager critic will want to
use it against the writer.
The way in which Bulgakov plays with the concept of authorship and with
the boundary between fact and fiction is of course deliberate. The truth or
fictionality of the story of the Crucifixion is crucial to the worldviews which
are competing in the historical context throughout Bulgakov’s life. The political
impact of Bulgakov’s text is beyond the interest of this analysis, but there are
important narrative lessons to be learned. There are things authors can talk about
which are just for the enjoyment of their readers, but there are also things in so-
called-fiction which are deeply true, even though there is no proof of their truth
in any historical document. It is the truth not owned by the author, but perhaps
shown to him in a moment of reflection, or in a dream. The blurring of all crucial
narrative categories in this important novel is deliberate, and it is achieved
through manipulation of narrative structure worthy of a true Master.
From the point of view of the emergent story, the case of M&M is also
interesting as a result of its choice of tellers. The two primary narrative spaces
(M&M1 and PP&Y), which start out as separate (the ‘Woland-in-Moscow’
story and the Pilate/Yeshua story) go through interesting changes of status as the
text develops. Initially, they are linked through Woland, who is a character in
M&M1 and the first teller of the PP&Y part. Later, the link is further developed
through Ivan’s dream. Then Ivan meets the Master and the Master’s novel first
enters the narrative scene. The fact that the reader has thus far had access to
PP&Y through major characters in M&M1, and can only then reconstrue the
story as the Master’s novel has crucial importance for the narrative structure.
The progression is from eyewitness report to a dream, to a novel, and thus the
truth-or-fiction issue is set up from the start as impossible to resolve. Having
different narrative spaces as homes for construals of the Crucifixion story makes
it possible for these spaces to maintain three different viewpoints throughout the
text, even though the sequence of events told is not disrupted at all – Woland
starts the story, Ivan’s dream continues it, and then the Master’s novel takes care
of the rest. This example supports the suggestion made earlier that the sequen-
tiality of the emergent story is of secondary importance to the meaning of the
narrative, while available viewpoints shape the emergent structure in substantial
ways. To put it simply, if the Crucifixion story is true, then Satan exists and can
interfere in contemporary affairs, while Ivan and the Master are sane and have
been given access to the truth. If it is a delusion or fiction, then we have read a
84 Stories and their tellers

funny fantasy story. This choice is established textually at the beginning, since
the first scene of the novel has Ivan and the critic, Berlioz, discussing the
question of whether Jesus really existed or not. The whole conversation can
thus be viewed as a narrative way of establishing the two viewpoints which
drive the ensuing narrative – similarly to Potocki’s text. The viewpoints, not the
facts told, decide on how this book is read and understood, and both viewpoints
remain active till the very end.

3.6 Narrative space embedding

Thus far, I have looked at instances where the SV-space either has little structure
or is only a source of temporal and epistemic viewpoint, but plays little role in
the story. However, there are also cases of straightforward space embedding,
where the SV-space contributes elements of the story and often changes the
emergent story significantly.8
The role of embedding in emergent narrative structure can be illustrated with
the narrative space configuration of Ian McEwan’s novel Atonement. The text, a
typical third-person/past-tense omniscient narrative, consists of three parts. It
starts when the main character, thirteen-year-old Briony Tallis, writes a play,
The Trials of Arabella, a rhymed story full of kitsch emotions and childish ideas
of love. Although the play does not participate in the story proper, it provides an
opportunity to characterize Briony as an aspiring writer, a person prone to using
big words, and treating fiction as seriously as she treats life. The play fore-
shadows the coming events, where Briony witnesses sexual encounters between
her sister, Cecilia, and a man named Robbie. She interprets these events as
Arabella probably would have, as signaling brutality rather than romance. After
another character helpfully provides the term ‘maniac,’ Briony has her vocabu-
lary to describe Robbie, and when another girl is raped, she blames it on him.
This part of the narrative clearly defines Briony’s over-excited confusion of the
adult world around her with the fiction she writes. In a sense, Briony tells two
fictional stories – of Arabella and of the rape, both structured with the same
authorial disposition.
The story develops dramatically. Robbie is sentenced, then freed to join the
army, while both sisters take part in the war effort as nurses. Later in the war,
Robbie returns to England, and Briony meets with him and Cecilia to acknowl-
edge that her accusations were wrong. This ends the main narrative space of the
novel, and the text is signed “BT, London 1999.” This brief addition, however
inconspicuous, changes the entire viewpoint structure of the novel. The narrator
is no longer a creation of the (implied?) author, McEwan, but the main character,
Briony. The novel we read is not the entire story constructed by McEwan, but
only its part. And we also learn why the text had to be written, as the last words
are: She knew what was required of her. Not simply a letter, but a new draft, an
 3.6 Narrative space embedding 85

atonement, and she was ready to begin. A “new draft” of what? This narrative
anchor, as well as that of the date, are crucial to the final shape of the emergent
story. The fourth part of the novel, titled London, 1999, mentions various drafts
of the novel (1940, 1947, 1999) and acknowledges that Briony has been
changing the facts from draft to draft until she finally decided to tell the
whole truth of her crime, but not say what really happened to Cecilia and
Robbie. She suggests a possibility that they died in the war and never saw
each other after the police took Robbie away, but we can only guess that that
was what ‘really’ happened.
Throughout the majority of the book the text constructs a viewpoint of a
narrator who knows what happened and does not judge anyone, but in its last
twenty pages it changes everything. An additional narrative space is set up
where the true author writes a story which will count as ‘the truth,’ including the
name of the real rapist, now married to the girl he raped. It is in fact an explicit
construction of an SV-space. The ‘embedded novel’ was developing throughout
the first 350 pages as an ordinary set-up with a narrator off-stage, and a temporal
and epistemic viewpoint yielding a third-person/past-tense narrative. But the
final revelations add a lot of topology to the SV-space, which now has a first-
person/present-tense narrator who also identifies herself as the author of the
story told in the MN-space. The MN-space thus becomes an embedded space, in
the sense typically used in mental spaces theory, which means, among other
things, that the stance and viewpoint of the higher SV-space are projected into
the lower MN-space. This means that the topology of the ‘embedded novel’
space now inherits the relevant bits of the topology of the SV-space, so that,
among others, there is a cross-space connection linking Briony-the-character to
Briony-the-author. But the epistemic viewpoint of the MN-space had to change,
in that we now know that the ‘true’ story resides in the SV-space, along with all
the previous drafts where different names were used and different circumstances
were described.
The new set-up also crucially changes the intended sense of atonement,
which now has three different meanings. In the embedded MN-space, atone-
ment for Briony’s guilt is achieved by clearing Robbie’s name and accepting the
blame. The novel itself, as a text to be read, counts as atoning for the crime by
changing its result and giving Robbie and Cecilia a life together. In effect, for a
reader in the space in which the embedded novel exists as an artifact, there is a
happy ending. For the author, however, atonement is only to be sought. The
various drafts she alludes to, with different versions of what happened, are
attempts, and she finally settles for the version which is the closest to the truth,
but changes the ending – and that is how atonement might be achieved.
To conclude, the establishment of the SV-space is consistently used to
structure the story’s general viewpoint and to put the MN-space story in
perspective, or even change its structure. Regardless of whether the set-up is
86 Stories and their tellers

textually obvious, or whether the embedding space is smoothly blended with

the embedded space, as in the Tyler or Woolf examples, the narratives discussed
here rely on the embedded set-up and its effects – primarily on the projection of
SV-space viewpoint into the MN-space.

3.7 Narrative viewpoint and narrative spaces

What the examples above suggest is that one might consider narrative viewpoint
in terms independent of the subjectivity of a specific teller, but instead as a
feature of the narrative space. Major narrative spaces, being mental constructs
and basic units of narrative structure, are equipped with rich topology, including
spatial and temporal grounding, the record of events and/or conversations, and,
first of all, the characters. In the various cases of profiled narrators, as described
in Section 3.2, the viewpoint is at least partially relayed through the fictional
subjectivity of the narrator. The need for the ‘narrator’ construct, whether
independent of the characters in the space or blended with one of them, emerges
out of the default communicative set-up, whereby content communicated
should be attributed to a communicator. But the presence of the construct
does not do much more than provide a focal point which gives the reader access
to the topology of the space.
What we have seen in the texts briefly described above is that narrative
viewpoint can be complicated by creative manipulation of spaces, rather than
reconstrual of narrators. There may be additional spaces set up, or viewpoint
may be spread over a number of different narrative spaces and tellers, as in
Potocki’s novel, or the space topology and epistemic viewpoint may alterna-
tively match one character or another or remain shared across two subjectivities.
Finally, as in Bulgakov’s text, distributing narratorship across different sub-
jectivities can contribute to the establishment of multidimensional or competing
viewpoints. In none of these cases is the viewpoint uniformly attached to a
subjectivity of a narrator; rather, it is a characteristic of the narrative space.
While it is perhaps more typical for narratives to profile tellers, they remain a
structural solution to a conceptual problem – how does the reader get access to
the facts s/he needs in order to comprehend the story and see its broader
implications. At the level of the reader’s construal of the emergent story, the
selection of an aspect of narrative topology to serve as the focal point is a way to
direct the story-construction processes.
4 Viewpoint: representation and compression

There were two men in my father’s chair.

(Coasting, Jonathan Raban)

The preceding chapters have outlined ways in which narrative viewpoint is

construed and linguistically expressed at the level of major narrative spaces of a
story. However, viewpoint phenomena of more partial nature pervade narrative
discourse at every level. There are several terminological issues which require
clarification at this point.
‘Viewpoint’ and ‘perspective’ are the broadest terms in use, with respect to both
macro-level phenomena (such as narratorship, temporal and spatial anchoring of
the narrative, et cetera) and micro-level phenomena, which are often talked about in
terms of Genette’s (1980) concept of ‘focalization,’ or ‘who sees.’1 Much of this
terminology evokes visual access to situations, though the nature of focalization
may include other ways in which the narrative is aligned with a fictional sub-
jectivity, whether of a narrator or of a character. To avoid the visual implications,
Toolan (2001) proposes to talk about ‘orientation’ instead, which in turn evokes
spatial configuration of some kind. All three terms, ‘focalization,’ ‘orientation,’ and
‘viewpoint,’ in fact cover a broad range of phenomena having to do with the
specific use of narrative space topology – temporal, spatial, perceptual, and the like.
Much of the relevant discussion is also complicated by the assumption that
there is a difference in narrative discourse between stylistic forms which ‘tell’
(such as narration) and those which ‘show’ (those which are assumed to represent,
rather than describe) (cf. Booth 1961; Rimmon-Kenan 1983). The distinction,
while useful in many ways, calls for a clarification of other concepts. Narration, in
this sense, is the domain of the narrator, who (which?) is a narrative construct
only – so if ‘telling’ the story is what is indeed talked about, this involves rather
strong assumptions at play about a narrating subjectivity or a ‘teller.’ In fact,
‘telling’ seems to be a more complex concept. In Chapter 3, I suggested that there
are three sources of the conceptual set-up underlying textually determined con-
struction of the narrator’s ability to tell the story: the viewpoint configuration
involving the SV-space and the MN-space, the reliance on the default communi-
cative set-up, and the choice of grammatical forms such as tense or person. In

87
88 Viewpoint: representation and compression

such a set-up, the ‘showing’ would also be the domain of the narrator, except that
the ‘voice’ is temporarily given over to a character. But how can a character have
a ‘voice’ independent of the narrative choice of the specific situation and word-
ing? The so-called direct speech in narrative fiction should not be taken as a token
of plausible conversational discourse, as any discourse analyst knows very well.
Natural discourse is full of interruptions, hesitations, overlaps between speakers,
unfinished sentences, discourse markers, et cetera, in sharp distinction to the
narrative representation of discourse. The idealization does not consist only in
smoothing out the performance, but also in very careful choice of which
exchanges between characters will be represented and how. Direct discourse in
fiction is as ‘narrative’ as narration or free indirect discourse, but it is designed to
‘tell’ (not really ‘show’) the story through a different conceptual medium (for
further discussion of direct discourse see Chapter 7).
‘Viewpoint’ as a feature of a narrative space, rather than of a fictional mind, is
also a generic concept which underlies much of the discussion of represented
speech and thought, especially forms such as ‘free indirect discourse,’ which
have been claimed to represent a ‘dual voice’ (Fludernik 1993). One of the
‘voices’ then would be that of a narrator, and the other of the character, but both
types of fictional minds are primarily a source of viewpoint, rather than dis-
course as such. The difference between a fictional mind which is ‘seeing’ and
one which is ‘speaking’ may be partly tenable with respect to represented
speech, which would then be an instance of ‘speaking,’ but does not seem
justified in application to thought, given that the unspoken content of charac-
ters’ minds definitely involves those characters’ viewpoints (or ‘seeing’) and is
to a much lesser degree aligned with specific forms of expression. The forms of
speech and thought representation (STR) are usually considered to be an
independent narrative phenomenon, and are definitely distinguishable as lin-
guistic constructions (see Vandelanotte 2005, 2009 for a very thorough analy-
sis), but they do assume a viewpoint of a speaker or thinker. What is more, the
indirect types of discourse described within STR are often (misleadingly) talked
about as representing actual words or thoughts, while they can at best be
claimed to be narrative-specific representations of their content.2
Yet another question is the distinction between ‘speech’ and ‘thought’ in STR,
although in most cases both are assumed to be represented by almost identical
constructions, whether described as direct or as indirect.3 In natural language use,
the distinction is crucial, in terms of addressivity, accessibility, and the specificity
of propositional content. In narrative fiction, for comparison, the distinction is
almost nonexistent, and I will discuss more specific exmples in Chapter 7.
Generally, I will assume that both have the role of representing conceptualiza-
tions, with little need to distinguish sharply between communicated and non-
communicated ones, since the overarching goal is to communicate them both to
the reader.
 Viewpoint: representation and compression 89

It seems that the treatment of ‘thought’ in narrative discourse emerged out of

the possibility to use similar ‘reporting’ constructions to handle both speech and
thought. In spoken discourse, it is equally possible to say Patsy said she was in
love or Patsy thought she was in love, and Patsy’s disposition thus described is
the same. What is different is the management of epistemic commitment of all
the subjectivities involved. In the ‘say’ example Patsy went on record about her
feelings, or the ‘narrator’ is going on record about correctly interpreting Patsy’s
words (which may have been different). In the ‘think’ example there are more
possibilities. Patsy may have said something suggesting she suspected her
‘symptoms’ indicated being in love (so the speaker is interpreting); or she may
have communicated what her suspicions were (such as I think I am in love,
though I’m not sure); or the speaker indicates that Patsy did think she was in love
but no longer does; or the speaker means to suggest that Patsy was wrong, even if
Patsy didn’t realize it at the time. All these interpretations arise out of various
epistemic stance configurations prompted by the constructions, but the possi-
bility which is in fact least likely is that the speaker is ‘reporting’ Patsy’s
thoughts as they were, simply because in natural discourse we do not assume
access to other minds, we can only construct our understanding of them and
communicate that. Why, then, would narrative discourse be so different? It is
only special in one respect – that it excludes the possibilities wherein the
‘narrator’ was in communication with Patsy, but otherwise it leaves the door
open to options favoring ‘construal’ rather than ‘reporting.’ Thus narrative
discourse inherits some of the constructional meanings available in spoken
discourse, but it also adds the possibility that the narrator had access to Patsy’s
mind without talking to her. This, in turn, is a viewpoint phenomenon, which can
only be defined in terms of viewpoint configurations discussed in Chapter 3.
Moreover, the idea of narrative access to minds calls for an explanation going
beyond the usual treatment of STR (see Chapter 7 for more discussion).
Before I move on to the sections detailing specific instances of micro-
viewpoint besides STR, I will look briefly at the viewpoint potential afforded
by seemingly straightforward representation of speech of characters. Example
(1) illustrates some of the issues raised above. In the fragment, the Ego-narrator
of Margaret Atwood’s The Blind Assassin, Iris, accounts for the information
she received about the death of her sister, whose car went off the bridge:
(1) I was informed of the accident by a policeman: the car was mine, and they’d traced
the license. His tone was respectful: no doubt he recognized Richard’s name. He said
the tires may have caught on streetcar track or the brakes may have failed, but he also
felt bound to inform me that two witnesses – a retired lawyer and a bank teller,
dependable people – had claimed to have seen the whole thing. They’d said Laura
had turned the car sharply and deliberately, and had plunged off the bridge with no
more fuss than stepping off a curb. They’d noticed her hands on the wheel because of
the white gloves she’d been wearing. (The Blind Assassin, p. 3)
90 Viewpoint: representation and compression

Ostensibly, the fragment is an elaborate report of various stretches of speech and

thought. There are thoughts of the narrator, as she guesses how the police got to
her and why the policeman sounded respectful, there are the policeman’s words
describing his interpretation of events and the information he had gathered from
the witnesses, and there are the eyewitness reports which include descriptions of
what happened, and in what manner. At this level, the fragment is saturated with
STR constructions, with words and thoughts attributed to four different
individuals.
The fragment appears on the first page of the novel, and introduces some of
the major narrative spaces and viewpoints to be elaborated. There are several
levels determining how the fragment shapes the reading of the remaining parts
of the text. First, there are several participants mentioned, but only three are
given salient narrative roles: Iris (the Ego-narrator), Laura, her sister, whose
death begins the story, and Richard (as we learn later, he was Iris’s husband).
The actual speakers – the policeman and two witnesses – are not referred to by
name and are never mentioned again. They are, in a sense, sources of informa-
tion about the circumstances of Laura’s death, rather than characters proper.
Primarily, their reports are used in establishing two possible interpretations of
what happened – as an accident or as suicide (see Figure 4.1). Jointly, they
construct the crucial point of the entire narrative, rather than having a specific
role as characters whose speech and thought need to be represented.
Second, the reported utterances, independently of their speakers, give differ-
ent accounts of what happened. The policeman, who knows that Iris’s husband
is a prominent businessman and politician, who wouldn’t appreciate an ugly
case of suicide in his family, helpfully provides a possible explanation of
Laura’s death as an accident (he reports it to Iris in this way). But he sounds
speculative (may have caught, may have failed), and admits that the witnesses
told a different story. The witnesses are very clear, report what they ‘saw’ and
‘noticed’, and use words like deliberately – clearly describing a suicide.
Between the policeman’s hesitant epistemic stance and the witnesses’ strongly
positive epistemic stance (supported by visually, and thus directly, acquired
evidence), it is clear that suicide is the more likely explanation, but there may be
reasons why it won’t be described that way. In effect, the reported words play no
role other than the construction of two takes on the facts – suicide versus
accident – and frame the rest of the novel as an attempt to explain both the
suicide and the efforts to hide it. As such examples suggest, reporting utterances
in narrative discourse may have goals pertaining to the overall structure of the
story, and are not necessarily local attempts to render the ‘scene’ (Rimmon-
Kenan 1983). Contrary to the narrative form chosen, the reported discourse
fragments here count as a kind of proleptic ‘summary,’ or, as I would prefer to
call it, as narrative space builders. In order to consider them in this way, though,
we have to assume that viewpoint spreads (in Fauconnier’s 1994 [1985], 1997
 4.1 Viewpoint and representation 91

SV-space Narrator: Iris vpoint: vpoint:

present accident suicide

Narrator: Iris vpoint: vpoint:

MN-space
past tense accident suicide

NS1 vpoint: vpoint:

Laura dies accident suicide

DS 1
DS 4 S: policeman
Clipping H: Iris
vpoint: accident
DS 2 DS 3
S: policeman S: witnesses
H: Iris H: policeman
vpoint: accident vpoint: suicide

Figure 4.1 Reported voices and viewpoint (example 1); S = Speaker,

H = Hearer, vpoint = viewpoint

sense) through compression across narrative levels, from the lower to higher
levels of the story, and thus participates in story construction. This understand-
ing of viewpoint will be explored throughout the chapter.
The issue of STR and the various constructions involved will be taken up in
Chapter 7. In what follows, I will attempt to clarify the conceptual under-
pinnings of micro-level viewpoint, starting with examples involving visual
perspective and modes of visual representation, and leading towards under-
standing micro-level viewpoint as a representation of experiential phenomena
and of conceptualization. I will then return briefly to the concept of focalization.

4.1 Viewpoint and representation

While the general tendency in narratology is to treat Genette’s ‘seeing’ meta-
phorically and move away from the visual understanding of viewpoint,4 I will
start out here by showing that even literally visual viewpoint has specific
discourse functions pertaining to narrative viewpoint in general, and helps in
explaining viewpoint phenomena on the micro level of the narrative. The
potential of visual viewpoint and visual representation is well illustrated by
the extended fragment, from the first pages of Jonathan Raban’s travel narrative
Old Glory, to be discussed here. The fragment, titled The River, shows a series
92 Viewpoint: representation and compression

of micro-level viewpoint shifts and compressions which use the narrative as a

way of constructing a broad representation of a concept. Thus, what seems to be
a non-narrative series of visual representations is a story of the emergence of a
concept which is crucial in framing the entire text.
The book is a travel story in which the narrator/traveler sails down the
Mississippi in a small boat. It starts in the present tense, as a description:
(2) It is as big and depthless as the sky itself. You can see the curve of the earth on its
surface as it stretches away for miles to the far shore. Sunset has turned the water to
the color of unripe peaches. There’s no wind. (Old Glory, p. 11)

The description suggests an observer taking in the view, even though it is

phrased with the viewer off-stage, only as a conceptualizer. In the next sen-
tences, the description becomes less literal:
(3) A fish jumps. The river shatters for a moment, then glazes over. The forest which
rims it is a long, loping smudge of charcoal. You could make it by running your
thumb along the top edge of the water. (Old Glory, p. 11)
In the text, the forest is not compared to a “smudge of charcoal,” it is one. And in
the very next sentence it is suggested that the description is not that of an actual
view, but of a pictorial representation being created by someone. In fact, it is
impossible to decide what the intended interpretation is, as the representation in
the observer’s mind and the representation in a picture are blended, and thus the
viewpoints are also blended. The compressed ‘representation/reality’ viewpoint
introduces the story as one of both real travel and of narrativized aesthetic
experience. However, the next paragraph changes the viewpoint (and the status
of the “river”) once again:
(4) It is called the Mississippi, but it is more an imaginary river than a real one. (Old
Glory, p. 11)
“It” refers to the blended construct introduced in the first two paragraphs, but it
also acknowledges its being mostly “imaginary” – so that the ‘objective’
existence of either the river or the painting is questioned. The next paragraphs
shift into the past tense and profile an on-stage ‘I’ narrator, but the story is still
not about events the narrator participated in, but about the gradual emergence of
the concept of ‘the Mississippi’ in the conceptualizer’s mind. In the following
paragraphs the reader is taken through various pictorial representations of the
river which gradually shaped the traveler’s conceptualization and his decision to
sail the length of the entire river and finally experience it beyond his imagi-
nation. First, there was the book he read as a child (I had first read Huckleberry
Finn when I was seven. The picture on its cover, crudely drawn and coloured,
supplied me with the raw material for an exquisite and recurrent daydream.),
then a map (I found the Mississippi in the family atlas . . . I looked at the
 4.1 Viewpoint and representation 93

Mississippi, wriggling down the middle of the page . . .), which both support
childhood imaginary games played on the shore of a local stream in the
traveler’s native Norfolk (. . . if I concentrated really hard, I could see the
Mississippi there. First I had to think it twice as wide, then multiply by two,
then again . . .).
Further details are added as years go by – there was a picture in the National
Geographic (very disappointing in its literalness), then a painting by George
Caleb Bingham (It showed the Missouri, not the Mississippi, but I recognized it
immediately as my river.), all of which add salience to the conceptualization
which has started in a child’s mind. With one paragraph of text devoted to each
of the subsequently available representations and its relevance to the childhood
dream which remains the point of reference, the text is effectively constructing
the representation in the writer’s mind, which is marked with a blended, and yet
very specific viewpoint. The initial image of the calm river filling the entire field
of vision is the image which has emerged out of years of transforming actually
encountered representations into the private one. The whole fragment is an
example of viewpoint construction through a series of specific representation
spaces. The cumulative result, though, is that the represented object, the actual
Mississippi river, never enters the stage, as the cross-space connections link
various representations in paintings, atlases, photographs, and childhood
dreams to the concept in the writer’s mind. The rest of the book is devoted to
the real river, but it is to be viewed through the lens of the concept narratively
constructed in the first paragraphs.
The first sentence of the ensuing travel story is Now, thirty years later, the river
was just a hundred miles ahead. Interestingly, the now marks a viewpoint shift
from the story of the imaginary river, to the story of the trip, which is just about to
start. The sentence, even though it is not part of the “imaginary” narrative, also
completes the compression of all the representational ‘encounters’ mentioned
earlier into the river, the concept originating in childhood, thirty years earlier.
Even though reading the National Geographic or viewing Bingham’s art had to
happen in adulthood, the time compression requires that its focal point remains in
the childhood years in the English countryside. “The river” is the one which he
had dreamed about, and the same one he will travel on.
Here is, then, a case where what could presumably be talked about as reported
perception, or aesthetic viewpoint, is an elaborate exercise in constructing
viewpoint through blending and compression of a number of representation
spaces into a narrative space which can naturally be viewed as the SV-space.
When the MN-space of the trip is then introduced, the narrator there (very much
on-stage and interestingly sarcastic and opinionated) is not only a traveler
reporting the events, he is also the ‘owner’ of the dream concept of the river
which propels him forward. The fact that the initial description is given in the
present tense (even though the childhood fantasies had to be past, and even
94 Viewpoint: representation and compression

though the travel narrative itself is written in the past) further supports the
suggestion that the initial image of the river is the SV-space of the entire
narrative – it is the generic state of mind guiding the narrator. The ensuing
paragraphs in the past tense construct the history of this narrative vantage point.
In the fragments discussed above the question of perception and representa-
tion is put in an interesting context. On the one hand, the images described are
presented from a visual perspective – looking at a picture on a book cover, a
painting, or a little stream. On the other hand, as a result of all the compressions,
the viewpoint is not strictly visual and would better be described in terms of a
mental representation. As a result, these narrative renderings of pictorial repre-
sentations rely on a triple, rather than double, configuration of mental spaces.
First, any pictorial representation such as a photograph or a painting evokes
two mental spaces – the represented space (the assumed reality portrayed) and
the representation space (the visual form it is given, which includes selection of
the field of vision, choice of focus, or color, the literalness level, etc.). This
distinction is well established in cognitive literature (Jackendoff 1975;
Fauconnier 1994 [1985], 1997), through the discussion of examples like In
the picture, Mary is a strawberry blonde, not a redhead, where the phrase in the
picture is a builder of a representation space, while the description of Mary as a
redhead comes from the represented reality space. The aspect of this set-up
which needs to be stressed here is the fact that the two spaces are two inputs to a
representation blend, which then allows one to talk about either input or the
whole configuration. Thus, as long as the picture is activated in the conversa-
tion, one can talk about Mary as a blonde or a redhead, and both mental spaces
will be used in arriving at the interpretation.
The ‘river’ example does much more than that. It adds a third space – the
viewer’s mental representation, which arises out of gradually blending the
representations described. Narrative viewpoint here is coaligned with this view-
point space, even in the initial paragraphs where there are no signals of a specific
narrator or conceptualizer. Here, then, is another instance of narrative viewpoint
first defined in terms of the narrative space being set up or elaborated, and only
then attributed to a narrator.
Focusing on a representation blend (with its two inputs, even if only one is
described) is different from narrating a representation of some narrative reality
in a character’s mind. In these cases it is possible for the narrative to render the
mental image only, even if it seems incongruent, and let the reader guess what
the reality is. In (5), from Paul Theroux’s novel The Picture Palace, the on-stage
Ego-narrator, a photographer, describes her experience with her first camera:
(5) It was a summer afternoon in 1917. My father hung upside down in the little lozenge
of glass; my mother’s chair was stuck in a canopy of flowers where my beautiful
brother Orlando’s toes were planted, . . . (The Picture Palace, p. 45)
 4.1 Viewpoint and representation 95

In early cameras the view available through the lens was always upside down,
which is what (5) describes. The frame of ‘picture-taking’ is evoked, and the
reader knows that when the shutter clicks the represented family scene will be
reproduced in a photograph, which will not be upside down. Thus what is
described in the fragment is the image recorded on the photographer’s retina,
but the image makes sense only in relation to the representation blend prompted.
The image is aligned with the narrative viewpoint recording the sense of dis-
connect between the narrator and her family, which is elaborated later in the story.
The effect can also be achieved when the narrative acknowledges the possi-
bility of a common optical illusion. When one is sitting on a moving train, one
can perceive the world outside the window as moving, and the train as sta-
tionary. In (6), the narrator is sailing along the shore, and represents the view
from the boat in a similar way:
(6) Yet I could see the land creeping past the wheelhouse window. I shut my left eye and
squinted, lining up a coppice of dead elm trees against the steel rigging of the mizzen
shrouds. The trees were making definite but slow progress while the boat stayed still.
(Coasting, p. 92)

As in the cases above, the writer ‘constructs’ his field of vision and allows the
illusion to take over, while knowing that the reality matches the undistorted
representation whereby the boat, not the land, is in motion.
There is also a category of narrative constructions building on representation
blends prompted by mirrors. Mirrors are a source of a variety of blends, where
the reflection takes on a life of its own and is in some way decoupled from the
object reflected. The examples below are just a tip of the ‘mirror’ iceberg in
literature and art, but what seems to be common across the examples is that the
reflection is treated as meaningful in its own right, and not merely in its relation
to the original. In (7), a woman is dressing up as a man to be able to participate in
a men-only event, while in (8) the narrator is trying on a new suit, to fit better
into the gentlemanly style of the town of Guntersville where he temporarily
resides:
(7) She gave me a man’s jacket, a pair of striped trousers, a derby hat. I put them on and
looked in the mirror. I was a man. (The Picture Palace, p. 97)
(8) “Hey, John, you look real good!” sang out a passing salesclerk as I stood in front of a
tarnished mirror, inspecting Rayburn in his new Guntersville uniform. (Hunting
Mister Heartbreak, p. 196)
When the woman in (7) says I was a man, she is not deluded into thinking she
has changed sex, but establishes an independent point of view whereby anyone
looking at her in disguise (not at her reflection) would have to believe her to be
a man. The power of the disguise allows her to decompress her sense of self
from her outer appearance and view the latter – as she perceives it in the
96 Viewpoint: representation and compression

reflection – from the perspective of someone who has no access to her selfhood
(any stranger at the event that night). This also means that the woman in her
disguise IS a man from the viewpoint of anyone but herself.
Example (8) is somewhat similar. The man in front of the mirror is the author,
the travel writer Jonathan Raban, but his name has been charmingly revised by
the inhabitants of Guntersville into “John Rayburn,” apparently fitting better
into their expectations. The reflection in the mirror does not match the writer’s
sense of self, but he can now see himself as Rayburn – someone meeting the
expectations of the people in the town.
The next two examples are again striking in how similar they are, despite
coming from different texts (Raban’s novel Foreign Land and Carol Shields’s
Larry’s Party):
(9) Putting the phone down, he noticed his face reflected in the dark uncurtained
window . . . What was upsetting was that, at first glance, it wasn’t his own face.
The hair and the beard were his, but not the plummeting cheekbones, the sunken
eye sockets . . . They were his father’s. (Foreign Land, p. 215)
(10) His father’s solid, ruddy presence . . . He met it each morning in the shaving mirror of
the various modest hotels where they stayed . . . He’d turn his eyes slowly toward the
mirror, creeping up on his face, and there the old man would be. (Larry’s Party, p. 22)

In both cases, the characters looking into the mirror are men whose relationships
with their fathers are strained. However, the crucial part is not that they see
themselves as so different as to be unable to understand their fathers, but that
they see themselves as worryingly similar. Their own reflections in the mirror
are no longer simply representations of their own faces, suggesting a unique
identity; rather, their facial features are twisted to resemble those of their fathers.
While inheriting facial features is considered normal, these examples propose
an uncanny loss of identity: a son is looking into a mirror, and sees someone
else – his father. The viewpoint thus represented is not visual at all, but relies on
a common assumption which sees physical appearance as a token of inner
character. Here, however, it is reversed – the inner self’s similarity to that of
another person (the father) is narrated through the perceived (though possibly
nonexistent) similarity in physical appearance. For this reversal to work, the
narrative has to establish a representation blend of the father and his reflection in
the mirror and cross-map it with the analogous blend of the son and his
reflection. With respect to viewpoint management, this narrative solution is
another case of relying on representation to manipulate viewpoint.

4.2 Viewpoint compression

Further discussion of the astounding variety and narrative power of representa-
tion blends is beyond the scope of this analysis, but the viewpoint effects of the
 4.2 Viewpoint compression 97

examples discussed above call for an explanation. In earlier work (Dancygier

2004a, 2004b, 2005a, 2008a, 2008b) I described the blending pattern involved
under the term viewpoint compression. As I argued, the multiplicity of view-
points in narrative discourse is conceptually manageable because of a series of
compressions bringing micro-level viewpoint up to the macro level of narrative
spaces. Thus various partial and very local viewpoints are interpreted as con-
tributing to or blending with the viewpoint of the narrative space currently being
elaborated.
Many of the examples above provide an illustration of how such compres-
sions work. For example, the fragment quoted in (5) presents the narrator’s
family members as hanging upside down. The description represents the image
in the narrator’s immediate field of vision, not the state of affairs in the narrative
space (MN-space), but it is narrated as if it were a part of the MN-space, rather
than the narrator’s raw perception, not mediated by conceptualization.
Typically, there would have also been a mention of the camera, and there is a
mention of it in the MN-space indeed, but not directly in connection with the
upside-down image (except the “lozenge of glass” – the lens). The lower-level
viewpoint of direct perception was thus compressed with the MN-space view-
point of the scene describing the narrator receiving her first camera as a present.
What is more, the compression gives a salient sense of the Ego-narrator’s first
glimpse of the intricacies of photography, which could not have been given in
such a poignant manner were the illusion to be dispelled. The mechanism of
viewpoint compression thus also explains how a strikingly ‘inaccurate,’ or even
false description achieves meaningfulness in the narrative.
Most of the examples throughout the chapter can naturally be explained
through compression, including the STR examples in (1), where highly local-
ized perceptions of witnesses and the policeman are compressed up to the MN-
space where the competing viewpoints of the ‘suicide’ and the ‘accident’ are
elaborated in the course of a narrative of over 600 pages. In what follows I will
illustrate the phenomenon through another visual example.
In (11), a fragment of Raban’s novel Foreign Land, the main character,
George, is on his boat, reflecting on his past life:
(11) He got the charcoal stove going, poured himself half a tumblerful of Chivas Regal
and laid out food in tins. Vera watched him from her photograph on the bulkhead.
She said: “Oh, George – you eating chickenshit again?” (Foreign Land, p. 298)
The novel is structured as the most typical past-tense/third-person omniscient
narrative. There is thus the stable SV-space, while the MN-space covers
George’s return to England after many years abroad and his subsequent decision
to sail away on his boat. The narrative space in which George prepares his meal
is thus within the MN-space. Vera is George’s former friend, who belongs to
the story’s past, described through numerous flashbacks, in which she often
98 Viewpoint: representation and compression

SV-space
Narrator: 0

MN-space

Photo-blend
NS1
George on the boat Representation (Vera)
present
Represented past (Vera)

DS1 NS2
Speaker: Vera George and Vera
Hearer: George past

Figure 4.2 Representation and compression 1 (example 11)

admonishes George for his unhealthy dietary habits. In the fragment quoted in
(11), Vera is talked about as if she were in the same space and time in which
George is – in the MN-space. But she is not, since he is on the boat alone, and
the only actual sign of her presence is a photograph. There are thus several
levels of mental-space structure here, all of which are compressed to ostensibly
participate in the general MN-space viewpoint.
First, there is the SV-space (past tense, third-person narrative) and narrative
space NS1 (George alone on the boat) and NS2 (George’s past relationship with
Vera). Vera is introduced via a representation blend, consisting of Vera’s photo-
graph on the bulkhead and the represented reality of the past relationship
between George and Vera, evoked from the NS2 space of the past relationship.
This initial configuration of spaces is shown in Figure 4.2. The representation
blend is a part of NS1, which is the space currently being elaborated, while NS2
also has a discourse space in it, where Vera comments on George’s diet. Two
connectors make the ensuing compression possible: the connector linking the
represented past in the representation blend to the past in NS2, and the cross-
space identity connectors profiling Vera’s presence in all the spaces.
Crucially to the emerging blend, Vera is presented as ‘watching’ George, so
that she has him in her field of vision, instead of him looking at her photograph,
which naturally falls within his field of vision. How is this possible? The reader
probably solves the puzzle by assuming that the words spoken by Vera are only
in George’s mind or are a memory of a past situation when indeed she said
something like this. But it is still puzzling to consider why the narrative does not
use explicit expressions directing the reader, something like Vera would say so if
she could see him or He remembered Vera saying X. The fact that there is a
 4.2 Viewpoint compression 99

SV-space
Narrator: 0

MN-space

NS1 NS3 Vera

George George’s
on the boat thought
present DS1
Speaker: Vera
Hearer: George
photo on
bulkhead
NS2
George and Vera

Figure 4.3 Representation and compression 2 (example 11)

solution available suggests that there is a standard interpretive procedure which

makes finding such solutions possible. I argue that the procedure relies on
viewpoint compression.
In the specific example, several things are happening. First, the representa-
tion blend of the photograph evokes the reality of George’s past relationship
with Vera. The evocation, though, emerges in George’s mind upon looking at
the photo, not in the narrator’s space. The entire past space (NS2) is now the
‘present’ in George’s thoughts. The part of the space structure where George is
looking at the photo is compressed to leave only the content of his perception
(Vera, as represented in the photo) in the wording. The perception further
prompts the decompression of the blend matching the photo and the reality it
represents, still in George’s mind. He mentally decouples the physicality of
the photo from the memories it stands for, and the representation blend is
compressed in the way which brings the space of the past relationship (and
not, say, the quality of the photo) in focus, along with its attendant discourse
space. The narrative space of the relationship (elaborated earlier in the novel)
is thus brought into the meal-on-the-boat scene within the scope of the
MN-space via George’s thought-space. Consequently, the photo (i.e. the
representation input) level is compressed out of the construal. Now that
the relationship is in this sense ‘brought back,’ Vera can play her part of
watching George and commenting on his habits. The entire temporal and
narrative distance between the boat scene and the past relationship has been
compressed. The effects of the compression, such as Vera’s vocal presence in
NS1, are represented in Figure 4.3.
100 Viewpoint: representation and compression

The series of compressions has thus established a number of interpretive

possibilities of a fragment like this one. At the MN-level, George is getting his
food ready, while glancing at Vera’s photo and remembering her criticism of his
dietary ineptitude. At the flashback-to-George’s-past level, evoked as the rep-
resented reality of the photo, there is some characterization of Vera, and perhaps
a memory of a specific occasion or occasions when she criticized George. At the
very basic level, there is the objective reality of the boat, with food and drink on
the table, and a photo on the bulkhead. All of these levels of the story structure
are available at the same time and open the various ways of interpreting
the fragment (as a flashback, as an expression of a sense of loneliness, as the
character’s awareness of his own faults, or as an ironic self-comment, et cetera).
These interpretations emerge out of the availability of all the narrative spaces
outlined here, but are present not as exclusive options, but rather as various
viewpoints available from different positions in the narrative space network.

4.3 Decompression for viewpoint

One more aspect of viewpoint construction has been illustrated by some of the
examples above, though it was not explicitly discussed – decompression for
viewpoint. For example, in examples (7) and (8) the identity of the person
looking into the mirror undergoes a crucial change. While any person normally
conceptualizes herself or himself as one entity, whose physique, mental ability,
style of clothing, et cetera are blended into one unique whole, there are situations
when we see various aspects of our identity as independent. The dividing lines
may run across different dimensions, as was shown in the discussion of the so-
called split-self metaphors, but the most common expressions of split self
assume a difference or a conflict between two major aspects of our personality
called the ‘subject’ and the ‘self.’ As Lakoff argues (1996), expressions such as
I’m not myself today, I have to reward myself, I’m at war with myself represent
various ways in which our identity may be compartmentalized.
In earlier work (Dancygier 2005a, 2005b), I discussed a number of such
examples as instances of decompressions of identity. One’s sense of uniqueness
is a result of a highly compressed blend, but it is natural to decompress that
whole when need arises, if only to be able to recognize the changes that
inevitably occur. Decompression is thus the flip side of compression in
that our need to achieve a holistic understanding of complex phenomena has
sometimes to give in to the need to appreciate their inner complexity.
Examples (7) and (8), however, do more than decompress the concept of
identity. While in both cases the person’s sense of identity remains unchanged
(the photographer is a woman, the travel writer does not become a Southern
gentleman), the reflection in the mirror highlights those aspects of identity
which have been changed to adjust to other people’s expectations. The
 4.3 Decompression for viewpoint 101

reflections in both cases are thus representative of how others in the narrative
view the character. In both cases, then, the decompression is double – on the one
hand, the reflection becomes a token of an identity different from the person
standing in front of the mirror (decompression of representation blend), and on
the other hand, the person’s identity is decompressed into two personas who will
play different roles in the narrative.
Example (12) is a more complex example of decompression for viewpoint. In
the scene, the narrator meets his father after years of feeling distant and resentful
towards him. His perception is based on childhood memories of a serious and
conservative man, and that is his concept of the father’s identity. Now they meet
again and the father has changed into someone much more likeable:
(12) There were two men in my father’s chair. One was my contemporary; a cheerful,
plainclothes, bearded, radical debunker. I could only see the other, a far older man,
if I squinted hard . . . While my father talked I tried and failed to get the two men to
coalesce into one person, but they wouldn’t go. (Coasting, p. 173)
The “two men in [his] father’s chair” are two decompressed parts of his
identity – the one from his younger years, when he actually seemed older
because of his manner and appearance, and the one at the time of the encounter,
markedly different and in some way younger. These two decompressed ‘iden-
tities’ are there as anchors to two different framings which the writer is trying
(unsuccessfully) to reconcile: the perception of his father as a rather unlikeable
person and the pleasant man in front of him. The decompression constructed in
the narrative serves the need of allocation of different viewpoints to these two
irreconcilable frames of the father’s identity. Furthermore, the story here is told
exclusively from the writer’s viewpoint, so that the two perceptions are
described as physically present in the scene. Here, as in (11), the distance
between the construal in the Ego-narrator’s mind and the events in the MN-
space has been compressed, so that two conceptualizations are talked about as
two story participants, even though there is just one.
Similar decompressions are possible in the case of places. In (13), the narrator
comes to Cairo from Yemen, and, by comparison, it seems to him to belong to the
Western rather than Arabian civilization, while the tourists arriving from England
see it as exotic. Cairo (as a concept constructed by a visitor) is thus decompressed
into two differently framed locales, while each frame participates in the con-
struction of a different viewpoint (Cairo feeling-like-home/Cairo as exotic).
(13) My fellow-diners and I had come at Cairo from different angles, and we’d arrived at
different places. They’d flown from Gatwick to the land of the Pharaohs, while I had
made a homecoming of sorts from Sana’a. (Arabia, Jonathan Raban, p. 267)

All the examples above decompress concepts into two independent entities
based on different frames applied. Furthermore, all these decompressions allow
102 Viewpoint: representation and compression

first-person narrators to represent events from more than one viewpoint.

Consequently, these single-viewpoint Ego-narratives are enriched with a multi-
plicity of perspectives on the events narrated. While these viewpoints could
have been represented through reporting and discussion, the lower-level decom-
pressions of identity allow other compressions to take place across narrative
space levels – so that, for example, the past and present images of a man can
both coexist in the MN-space, even if only in the eye of the narrator. The term
‘decompression for viewpoint’ can thus be further elaborated into ‘decompres-
sion for viewpoint compression,’ since the decompressions at the lower narra-
tive level are then subject to viewpoint compressions which allow the
Ego-narrator to naturally represent, rather than simply report, different perspec-
tives on the events narrated, at a higher level. Thus when Raban comments on
Cairo, he is also narrating from the Ego-perspective, as the comparison of
frames he makes could not have been made by the tourists whose viewpoint
he contrasts with his own. ‘Decompression for viewpoint compression’ is thus a
narrative technique allowing a single Ego-narrator to represent other view-
points, constructed via decompression, by compressing them with his/her own.

4.4 Fictive vision, causation, and change

It is worth noting that in most of the examples discussed above there is a distinct
‘visual’ flavor to how the viewpoint is constructed. While this may seem a
simple consequence of the fact that many of the examples are representation
blends, it is not a sufficient explanation. The initial paragraphs of the ‘River’
fragment are a description of a conceptualization primarily through visual
means. The image does not include the well-known non-visual framings – the
dangerous power of the river, its energy, or its flow. Even though constructed
out of other representation blends, this image is so intensely visual as possibly to
fool the reader into thinking that someone standing on the shore is describing
what he/she sees.
The examples with mirrors ([7], [8], [9], and [10]) behave similarly – in all
four cases the person looking into the mirror ‘sees’ someone else – not ‘imag-
ines,’ not ‘thinks about,’ but simply sees. Also, when the traveler in (6) writes
Yet I could see the land creeping past the wheelhouse window, he is simply
describing what he sees and not what seems to be the case. In all of these cases
what is reported as visual perception is in fact conceptualization, similarly to the
situation in (11), where what is presented as the words said by Vera belongs to
one of the many aspects of the narrated state of George’s mind. There seems to
be a consistent strategy at play here whereby mental representations and con-
ceptualizations are narrated as perceptions. It is probably most salient in (12),
where an actual act of looking at one’s father is narrated as seeing two of them,
each one of different age and appearance. Crucially, the ‘viewer’ tries to have
 4.4 Fictive vision, causation, and change 103

the two fathers merge into one, but cannot make it happen. In this case, the entire
complex conceptualization is narrated as literally seeing two men instead of
one. Conceptualizing here is represented as seeing. In fact, the persistence of
this narrative construction suggests a usage pattern which I will refer to as fictive
vision. The term intends to mimic Talmy’s (1996) idea of fictive motion,5
represented by examples such as The curtain runs all the way to the corner,
where mental scanning of the length of the curtain is talked about as if actual
motion was taking place (Matlock 2004 talks about it as an instance of cognitive
simulation). Fictive vision is similar, in that visual simulation or a visual
construal is used to support conceptualization. While further discussion of the
role of fictive vision in language conceptualization is beyond the scope of this
project, the examples discussed throughout this chapter suggest that the usage is
pervasive.
There seems to be much in linguistic usage to support this narrative strategy.
We often talk about metaphorically ‘seeing’ as ‘understanding’ or ‘knowing’ (as
in I see what you mean, The lecture was clear) or even as ‘conceptualizing’ (as
in Look at it from my perspective, I don’t see it that way). Given that visual
perception is our primary cognitive source of conceptualizing the world around
us, relying on expressions of visual perception in talking about mental processes
seems natural, beyond the primary metaphors linking seeing to knowing or
understanding. These correlations give at least a partial explanation of why
narrating thoughts can be done through narrating perception, even though the
perception is not literal. The other part of the explanation is probably the central
question of contemporary narratives – how are thoughts to be narrated?
However, I want to argue further for a distinction between ‘thought,’ as
discussed in the context of STR, and experiential conceptualization and/or
perception. In most discussion of STR, ‘thought’ appears to apply to various
inner processes, expressing anything from propositional thought (He thought
she was pretty) to narration of an emotional state (He liked her), and while some
classifications are very subtle and well supported with corpus evidence (Leech
and Short 1981, Semino and Short 2004, Short 2007, Vandelanotte 2005, 2009),
it is still not clear how they are to be treated as complete. At the other end of the
spectrum, where the entire varied collection of narrative choices is talked about
as falling under the rubric of the ‘stream-of-consciousness’ techniques
(Humphrey 1954), what seems to be assumed is that narratives such as the
monologue of Molly Bloom give the reader direct access to characters’ thoughts
or consciousness. But there is a difference between thoughts and experiential
conceptualizations which in the context of the narrative is not trivial at all, and
in the cases discussed here conceptualization is at stake, not thought. The point
of using mock-perception as a means of narrating conceptualization is not to
give the reader access to a character’s thought processes, or to verbalized
statements which count as thought-content, but to allow the reader to experience
104 Viewpoint: representation and compression

the narrated reality through the eyes of the narrating Ego. In other words, using
visual construals as a means of conceptualization may give the reader the kind
of insight which results not necessarily from access to thought processes, but
from immediate access to experience. A sentence such as There were two men in
my father’s chair cannot naturally be treated as representing thought, but it is an
experiential rendering of a sense of disconnect between one’s memory of a man
and the reality of the person sitting in a chair opposite. The fictive vision here is
a simulation of experiential on-line conceptualization (as opposed to stable
categorization).
Experiential aspects of textual choices, in combination with visual ones,
seem to play a central role in the narrative function of examples such as (14)
and (15):
(14) We were shrinking . . . With nothing to measure itself against now except the open
Atlantic, the ship, so enormous in Liverpool, so lordly on the Irish sea, was
dwindling into a dot. (Hunting Mister Heartbreak, p. 32)
(15) The cops are huge. They make the room tiny, filling it with black. (AHWOSG,
p. 266)
Both fragments are examples of fictive vision, or perception-as-
conceptualization. In (14), the perception of the ship getting smaller is the result
of the change of the background against which it is viewed, such that it seems
smaller than it was at the quayside. There are two additional aspects of the
construal here – the choice of the pronoun we, and the progressive form. The
pronoun typically aligns the speaker with a group which she is now perceived to
be a member of. In this case, it is not clear that the speaker includes other people
in the we choice, but it seems more likely that we refers to the ship and all her
passengers, with the ship being the core of the group. The choice of the
progressive also signals the change in progress – the change in perception, in
this case. Consequently, this example is not only about perception as conceptu-
alization, but also about the change of perception as change of
conceptualization.
Example (15) is similar. Whether the policemen are in fact more than average
height or weight is not the issue, but the change in the perception of the room
(which now seems tiny) is what is being described. Additionally, the change in
perception is also conceptualized causally (they make the room tiny), as a result
of the appearance of the cops.
The perfect combination of visual viewpoint with change and causality can
be found in (16). The narrator is driving down south, noticing the gradual
change of climate:
(16) The trees that had been skeletal and grey the day before were coming into leaf
this morning . . . The harder I stepped on the gas, the faster I could make things
grow. I made the first magnolia burst suddenly into flower, woke the first snake
 4.4 Fictive vision, causation, and change 105

from hibernation . . . At the rate I was going, it would be fall by Tuesday morning.
(Hunting Mister Heartbreak, p. 112)

Rather than saying something to the effect of ‘spring has already come to the
southern parts of the US,’ the Ego-narrator attempts to render the gradual
perceived change. As in (14), the choice of the progressive (were coming, I was
going) supports the conceptualization of gradual change. There is also the sense
of accelerated passage of time – if the time elapsed since the day before brought
trees into leaf, then spring will change into fall in a matter of days. In fact, change
is perceptually correlated with the passage of time (it normally takes weeks or
months for trees to go into leaf, and here the change occurred overnight, so it must
have been due to some fantastic compression of time – weeks into hours, months
into days). Finally, there are two constructions adding causality into the concep-
tualization of this trip. The X-er, the Y-er construction (as in The more, the better,
cf. Fillmore et al. 1988) sets up two causally linked scales, where the increase
along one scale (more people at the party) causes an increase of value along
the other scale (the degree of fun). In (16), making the car go faster also makes
nature go through the seasons faster. The resulting conceptualization is really
complex. The driver is going south, seeing more and more signals of spring as he
moves through space. The faster he goes, the more space he covers, and the more
spring-like nature becomes. But from the perspective of the car the movement
through space is less in focus, while the changing view along the road is high-
lighted. The perception of change is thus narrated as if the narrator’s actions
causing motion through space were the cause of the change outside. While visual
viewpoint is still at the core, the conceptualization adds causality to the picture
and re-construes the narrator from an observer to an agent – a more interesting
role from the perspective of the narrative. In any case, this mock-perception of
causation related to speed of motion does a good job of highlighting the intensity
of perceived changes and compressing the narrative time.
The combination of visual compression and causation seems to be more
likely than the originality of (16) suggests. The sentence which follows the
description in (5) is I had stood them on their heads, but nothing dropped out of
their pockets. The narrator here also compresses perception with causation, so
that the awkward position seen through the lens (a little lozenge of glass) is in
fact a consequence of what the narrator does.
To conclude, visual structuring of viewpoint may be what strikes us as the
most salient aspect of the examples discussed here, but it is more appropriate to
consider these cases as instances of experiential viewpoint – which uses visual
perception to render experience (of change, agency, lack of coherence between
the mental and the real, et cetera). These cases are thus instances of representing
conceptualization through the lens of its experiential aspects. As a result, the
MN-space events are rendered through compression of lower-level mental and
106 Viewpoint: representation and compression

experiential construals upwards to the MN level. In other words, the ‘event’ of

driving down south is narrated through the experience it offers, and not through
objectively measurable aspects (distance covered, cities or states passed, time
elapsed, et cetera). Viewpoint compression thus seems to be the mechanism
which makes such experiential narration possible.
Furthermore, such narrative choices seem natural in the context of the dis-
cussion of experiential grounding of conceptualization and emotion (see section
1.6). It is not at all surprising that narrative choices would capitalize on the links
between experience and higher cognition by appealing to the reader’s experi-
ential abilities, rather than rely on the ‘telling’ technique. It is thus possible to
redefine and expand the idea of ‘showing’ by applying it to narrative choices
which prompt experiential alignment.

4.5 The micro level, the macro level, and viewpoint compression
In most of the examples considered above decompression and compression
serve the needs of aspects of the narrative being elaborated in the MN-space
at the time (photographer’s preparation to appear as a man, George’s meal,
et cetera). Even though in (11) the narration of the relationship with Vera
(scattered throughout the text) is being evoked, it serves the needs of the
George-on-the-boat-space, rather than George’s-previous-life-space. There are,
however, interesting instances of viewpoint compression which also provide links
across different narrative spaces. I will discuss two such examples here.
Characteristically, these fragments appear at the beginning of their chapters.
The first, example (17), comes from a chapter of Foreign Land which focuses
on Diana, George’s new friend, who lives in Harmony Cottage. Earlier in the
story (though not in the immediately preceding chapter), George took Diana for
a ride on his boat:
(17) Harmony Cottage was out at sea. Diana felt the floor roll away under her feet and
steadied herself by leaning on a joist as her kitchen tilted and yawed. (Foreign
Land, p. 186)

The initial sentence, describing the cottage as being at sea, has to strike the
reader as strange and contrary to common experience. Only in the second
sentence the verb felt makes it clear that it is Diana’s experiential perception
that is being narrated, not the actual facts. What the fragment describes is a
common sensation, especially among inexperienced sailors, whereby being on
the boat feels stable, while the land feels like it is swaying after one gets off. The
fragment thus narrates what Diana feels as she gets off the boat and comes back
home, but does so through evoking the sequentially (not textually) preceding
narrative and using the space of Diana’s experience as the current narrative
viewpoint space. As a result of all the compressions, the continuity of the
 4.5 Levels and viewpoint compression 107

narrative is not disrupted, and Diana’s perspective is firmly established as the

one on which the chapter will focus.
Let me consider one more example to illustrate the point. Example (18),
coming from Alex Garland’s novel The Tesseract, opens a new chapter. It is
important to note that at this point in the narrative the fact that Doming had died,
in his house, is already known to readers, so that the first sentence of (18) has to
be understood with respect to a narrative space other than the MN-space.
(18) Doming died at the gates of the graveyard. Everywhere else – the house, the road to
the church, the church itself – he had been alive. In the same way that nobody is
about to leave until they reach the bus depot and see the bus: alive. And character-
istically quiet. But at the gates of the graveyard he died, suddenly, and Rosa was
overwhelmed by the understanding. (The Tesseract, p. 169)

Only in the last sentence of (18) is the reader told that the fragment talks about
Rosa’s (Doming’s daughter) mental state. She ‘knows’ that her father died, but
the realization only hits her when the funeral procession reaches the gates of the
graveyard – as if he were symbolically crossing from the world of the living into
the world of the dead. Consequently, one can read the first sentence as repre-
senting Rosa’s state of mind, and the rest of the fragment as the narrator’s
elaboration of it. But even the first sentence here is phrased as a narration of a
momentary event, a sudden and complete change of state. This is appropriate,
since both ‘death’ and ‘realization’ are events of that kind, though one describes
a change in bodily state, and the other a change in mental state. In fact, the
connection is even more interesting since the change in Rosa’s mental state in a
sense completes the change in Doming’s bodily state – not only is he now dead,
but he is also dead in the minds of the people who loved him.
The fragment poses a question similar to the one discussed in connection with
(17) – how is the access to Rosa’s mind rendered in the narrative? It may be a
rather unusual example of a free indirect report, though FID would typically
require enough distance to see the mental event in its entirety, and some
narrative structure wherein Rosa’s thoughts belong. It would not be appropriate
to rephrase the sentence as Rosa thought (or even realized) that Doming died at
the gates of the graveyard or some such. The point is that the sentence renders
the emotional event ‘as it happens,’ without any distance, in progress. The
narrative construction here gives the reader unmediated access to Rosa’s emo-
tional viewpoint, without referring to her specific thoughts or words.
The question these examples pose is parallel to the fictive vision examples
above: the micro-level viewpoint is rendered through immediate access to a
character’s bodily or emotional state. The compression goes all the way from
the top space to what could have been rendered as she felt as if the house were
swaying or she was shocked by the impact of her father’s death, that is, a
character’s bodily or emotional experience. The expressions actually used
108 Viewpoint: representation and compression

skip over all those levels of narrative structure, but the reader is expected to rely
on all the mediating narrative spaces even if some of them have not yet been set
up (such as Rosa’s feelings and Diana’s dizziness). These experiential compres-
sions seem to be in line with the examples above (especially [14] and [15]), but
they render third-person conceptualizations, not first-person Ego-narrator’s
conceptualizations. The compression is thus more complex, but direct experi-
ence remains in focus.
Throughout the chapter, I have been discussing instances of viewpoint which
rely on visual, experiential, emotional, and other types of conceptualization.
While most of these instances could probably be categorized as examples of,
broadly speaking, focalization, or, more specifically, represented speech and
thought, represented perception, or some variety of narration, any such descrip-
tion would miss important aspects of meaning construction. The examples seem
to be more coherently described as instances of narrative viewpoint leading to
representation of conceptualizations which, in turn, participate in the construc-
tion of the events in the narrative space being elaborated. While the concept of
viewpoint may often correlate with visual descriptions, that in itself does not
explain their role, since they may participate in representation blends, in the
construction of experiential viewpoint, and in the construal of narrative events.
Whether we see the source of a conceptualization in visual perception, or
attribution of causality, or framing, or representation of speech and thought,
the fact remains that the manipulation of viewpoint is primarily the manipu-
lation of narrative spaces, not of the filtering through one fictional conscious-
ness or another. Processes such as decompression for viewpoint or compression
to higher viewpoint lead primarily to the construal of the narrative space.
Furthermore, viewpoint construction may require that many narrative spaces
are at play – decompressed, blended, embedded, or evoked from other parts of
the narrative. As a result, the complexity of viewpoint configurations cannot be
consistently described either in terms of focalization with respect to a given
character or in terms of STR.
The review of examples quoted throughout this chapter suggests that rather
than attempt to refine classifications of instances in which fictional minds are
used as sources of communicated conceptualizations, we might consider mech-
anisms which allow various conceptualizations to emerge and contribute to the
overall richness and coherence of the narrative. Viewpoint decompression-plus-
compression seems to be one such mechanism.

4.6 Speech, thought, and multiple levels of representation

As the examples throughout the chapter suggest, viewpoint phenomena range
over much more than focalization, and are also more specific, in that they include
various types of conceptualizations. At the same time, they also involve
 4.6 Speech, thought, and levels of representation 109

interlocking levels of representation. In recent work by Zunshine (2006), meta-

representation was considered a crucial aspect of fictional prose, one which tests
our evolved ability to represent other minds to its limits. In fact, Zunshine also
argues that narrative discourse can sustain more complex configurations of meta-
representation than spontaneous discourse does, as it is natural in fiction for a
narrator to represent the story told by someone, who also represents the words and
thoughts of characters, who in turn represent the words and thoughts of other
characters. Zunshine’s use of the concept ‘meta-representation’ was criticized,
among others, by Boyd (2006). Boyd observes that Zunshine’s framework is a
much-impoverished version of either ‘theory of mind’ or evolutionary psychol-
ogy, and also that the use of the concept of meta-representation in the analysis of
texts is in some cases inaccurate and misleading. From the linguistic perspective,
Zunshine’s claims seem striking in that they postulate a qualitative difference
between the number of levels humans can process in speech and in writing.
While it is difficult to prove that there is indeed an imbalance between speech
and writing in this respect, the fact remains that the multilayered nature of
fiction calls for an explanation. In what follows, I will argue that narrative levels
of meta-representation do not cause much difficulty, as they do not need to be
maintained as separate layers of representation, but are compressed in a way
which makes the lowest space participate in the higher one(s). The example of
The Manuscript Found in Saragossa discussed in Chapter 3 is a perfect test
case – there are so many levels of meta-representation that the characters who
are listeners sometimes try to recapitulate where they are in the web of stories
and who is telling what to whom. And yet, maintaining the exact meta-
representation relations is not the primary task in narrative comprehension,
and establishing the current viewpoint trumps other processing tasks. When
several stories in the Manuscript, told by different people to different addres-
sees, suggest the idea that beautiful women could turn out to be dangerous
demons, the reader quickly forgets who told which story to whom, but empa-
thizes with the main character, Alphonse, who keeps on meeting the same
women (or devils, both options appear equally possible) and watches him
deal with the puzzle. As Herman argues (2006), having additional tellers
could also be explained in terms of the explicit evocation of distributed cogni-
tion, and the suggestion seems plausible in that compressed viewpoint is more
useful to a group than the details of who said what to whom. Oral tradition
seems to support this further, since in that context the person of the actual teller
is not directly relevant to how the story is received.
It thus seems plausible to claim that narrative discourse manipulates numer-
ous viewpoints by definition, but may bring one kind of viewpoint into focus in
any given fragment, whether one sentence long or ranging over longer stretches
of the text. At a minimum, such a fragment is embedded in the scope of the
SV-space, and within the viewpoint of the MN-space, while also reflecting the
110 Viewpoint: representation and compression

viewpoint of the specific narrative space being elaborated. In the course of

reading, the reader has to come up with construals of all these viewpoints, and
can then add the viewpoint of a specific character to it. Out of context, though, it
may not be clear which viewpoint spaces are activated at any given time. In
example (19), from Sara Paretsky’s Hard Time, the first sentence is naturally
construed as narration, the second one could be the Ego-narrator’s reflection on
the injuries sustained, until in the last sentence the comments are explicitly
attributed to her doctor, Lotty:
(19) The gun had dug a deep bruise into my side when I tumbled from the boxcar. I’d be
sore for four or five days, but if I was careful I’d be okay . . . Lotty dispensed that
verdict at her clinic Sunday afternoon. (Hard Time, p. 233)
With respect to STR, the passage distinguishes levels of meta-representation –
the narrator, V. I., represents the words uttered by Lotty. But under the inter-
pretation including viewpoint compression, the prediction about being sore for
about five days is presented from two viewpoints at the same time – what Lotty
said after examining V. I. on Sunday, and what V. I. is thinking/saying later,
when she is recapitulating the situation. The compression, as in all other cases
examined, brings the space which marks the lowest and most local level of
viewpoint (Lotty on Sunday) up to the narrative space where the relevant events
are being elaborated (V. I. planning her actions in the space where she is
conducting her investigation – the MN-space). But because V. I. is also the
narrator, the compression goes higher, into the SV-space, where V. I.’s health is
a factor in the entire story being told. As a result of all the compressions, the
diagnosis communicated at the lowest viewpoint level by one person partici-
pates in the thought processes of another person, who is also the narrator at the
higher level. This example shows clearly how the concept of viewpoint com-
pression possibly obliterates the need to distinguish clearly between V. I.’s
represented thought and Lotty’s represented speech.
A somewhat more elaborate example, from Anne Tyler’s The Accidental
Tourist, is presented here as (20). The main character, Macon, who writes
guidebooks for people who don’t enjoy travel, is preparing to leave home:
(20) For his trip to England, he dressed in his most comfortable suit. One suit is plenty,
he counseled in his guidebooks, if you take along some travel-size packets of spot
remover. (Macon knew every item that came in travel-size packets, from deodorant
to shoe polish.) The suit should be a medium gray.
When he’d finished packing, he sat on the couch to rest. Or, not to rest, exactly,
but to collect himself – like a man taking several deep breaths before diving into a
river.
The furniture was all straight lines and soothing curves. Dust motes hung in the
slant of sunshine. What a peaceful life he led here! If this were any other day he’d be
making some instant coffee. (The Accidental Tourist, p. 25)
 4.6 Speech, thought, and levels of representation 111

There are three primary spaces at play here: the MN-space of Macon’s life after
his son’s death, the narrative space of his difficult period after separation, and
the space of the guidebooks he writes (in italics). Interestingly, there is also the
characterization of Macon outside of the narrated events, in the SV-space where
the narrator-viewpoint resides (in parentheses). The last paragraph of the frag-
ment in (20) is naturally described as a ‘free indirect thought’ account of
Macon’s perception and reflection. Even though it is not mediated by any
expression suggesting access to Macon’s mind, it is naturally compressed
with the ‘leaving home’ part of the narrative as another aspect of it.
What is more important, though, is that all the spaces are not felt to be
separate and they smoothly blend into one. It is easy to see the first sentence
as pure narration, but in fact it may also be an account of Macon’s decision to
wear a specific suit. The quote from the guidebooks may be the narrator’s ironic
comment, or Macon’s self-reflection on how he follows his own advice, but it
has to be evoked as relevant to some aspect of the narrated events. There are thus
several lower-level spaces, but there is also one higher-level space, compressed
into one narrative viewpoint of Macon’s experience. As a result, there is no
natural interpretation here which would clarify the levels of meta-
representation, other than the narrator’s access to all the spaces at play.
It appears, then, that STR phenomena such as FID constitute but a part of the
viewpoint or meta-representation phenomena displayed in the fragment. The
questions do not seem to be restricted to ‘dual voice’ or even dual deixis
(proximal deictics here and this in the distal past context),6 because the fragment,
especially in its last paragraph, compresses the character’s experiential viewpoint
with the narrator’s focus on the impact of all the details of the day. In this view, the
sentences describing the furniture and the light represent Macon’s perceptions, as
relayed through the narrator’s privileged position, and not simply omniscient
narration. I hesitate to talk about such instances as representing a ‘dual voice’;
instead, I understand them as cases of viewpoint compression wherein the
narrator ‘speaks in lieu of’ the character. The character’s viewpoint is clearly
signaled, but compressed upwards, to let the narrator go on telling the story.
It should be clear from the discussion above that the concept of viewpoint
compression makes somewhat more realistic assumptions about the degree to
which a narrative can be viewed as transforming the characters’ words into
various forms of discourse (direct, indirect, or free indirect). It is not necessary
to assume that Lotty said You’ll be sore for four or five days, but if you’re careful
you’ll be okay (rendered as I’d be sore for four or five days, but if I was careful
I’d be okay), or, even more important, that Macon thought What a peaceful life I
lead here! The expressions in examples (19) and (20) are renderings of a
character’s possible thoughts or words, but narrated from the viewpoint of the
MN-space or the SV-space (depending on the type of narrator), to best represent
the information from the space’s point of view.
112 Viewpoint: representation and compression

The question of meta-representation also comes up in examples like (1)

above, where there are indeed different speakers. The witnesses told the police-
man what they saw and he is summarizing their testimonies to Iris, who is now
narrating them to the reader. But the question of a verbatim repetition or meta-
representation does not really come up, since all represented speakers have
different goals in saying what they do. What really matters is only the highest
level – Iris’s intended narrative account of what happened to Laura and the
highest-level viewpoint which calls for suggesting an accident and a suicide at
the same time. Thus the compression again goes all the way up, and the levels of
meta-representation do not need to be maintained.
To conclude, viewpoint compression is a blending mechanism which
attempts to account for the fact that zillions of low-level facts, observations,
or thoughts are compressed into more manageable viewpoint spaces and used in
the processing of the narrative as a whole. There is no need to mentally keep
track of each subtle change of mood or voice throughout the reading, because
these specific bits of narrative structure are quickly compressed upwards to
participate in the construction of major narrative spaces.

4.7 Narrative thought and intersubjectivity

The discussion of viewpoint proposed in chapters 3 and 4 raises a number of
questions with respect to the actual paths of conceptualization required or used
in such contexts. At this point in the discussion it might be useful, then, to go
back to some ideas formulated in section 1.6, and relocate the concept of
viewpoint developed above in the context of recent proposals in cognitive
science and philosophy.
Section 4.6 also evokes the approach suggested by Zunshine, whereby
enjoyment of fiction as well as our meta-representation abilities depend on the
module known as the ‘theory of mind.’ I suggested above that the need to
maintain specific meta-representation hierarchies seems to be an impediment
rather than an aid to narrative comprehension. Quite independently of meta-
representation, though, the application of theory of mind to narrative was
intended as an explanation of our ability to engage with narrative characters
in ways somewhat similar to our engagement with other people. However,
recent research into the abilities summarized under the ‘theory of mind’ caption
suggests a picture relying more on embodiment, activation of neuronal patterns,
and, first of all, intersubjective engagement.
In a recent edited volume on perspectives on intersubjectivity (Zlatev et al.
2008) a strong argument is made against ‘theory of mind’ in the form we know
it. The arguments raise a number of issues, ranging from evolutionary implau-
sibility to rejection of the model of the mind which makes such strong
 4.7 Narrative thought and intersubjectivity 113

assumptions about inferential processes leading to specific formulation of other

people’s beliefs and desires, or, simply speaking, their reasons for acting the
way they do. Jointly, the papers in the volume present a variety of perspectives
(from philosophy, psychology, cognitive science, and language) leading
towards the concept of a ‘Shared Mind’ – the mind relying on embodied and
experiential contact with others. Reviewing the entire argument is beyond the
scope of this book, but some of the suggestions made support the general
theoretical claims advocated here.
First, it is proposed (primarily in Zlatev’s contribution, 2008), that the ability
to adopt a view of another person emerged even before the modern linguistic
ability, so that at the evolutionary level which Zlatev calls ‘triadic mimesis’ a
person (a hominid?) could deal with viewpoint complexities represented by an
example such as “I expect that you know that I know that X.” This stage is
argued to be crucial to the development of intersubjectivity in hominid evolu-
tion. While initially such constructs must have depended on joint attention,
based primarily on seeing and shared context, they evolved into a ‘shared
mind.’ Interestingly for my purposes, the development had to start with what
is still the default natural understanding of viewpoint (seeing things from a
certain perspective), and then progress from there towards more abstract con-
cepts – just as the embodiment hypothesis predicts.
The idea of an intersubjective mind also highlights the embodied aspects of
cognition and the immediacy of our responses to others. The aspect of our
neural set-up which is now often discussed in the context of empathy or emo-
tional response to the situation of another person is the existence of mirror
neurons. While they have been claimed to have a crucial role in language (cf. the
Neural Theory of Language described in Gallese and Lakoff 2005), they are
clearly highly pertinent to the concept of intersubjectivity. Crucially, as
Gallagher and Hutto (2008) point out, mirror neuron activation does not involve
the distinction between first-person or third-person experience, does not require
a setting up of models or representations, and is entirely natural, so that it cannot
be fabricated to ‘pretend’ anything or gain access to another mind. In this light,
the model of ToM, as it has been used, seems significantly over-engineered. As
the intersubjective model claims further, contextualized action is more impor-
tant to the ‘shared mind’ concept than the inner mental processes. The theory
essentially sees ‘others’ as agents whose actions are framed pragmatically and,
consequently, sees us as sentient participants prepared to act in situations
including others, situations which necessarily rely on the recognition of (some-
times conflicting) viewpoints.
The ‘shared mind’ approach seems to explain the viewpoint phenomena
discussed above much better than other approaches available. Although the
narratives I have looked at represent viewpoint mediated through linguistic
expressions, the mechanism of immediately appreciating the consequences of
114 Viewpoint: representation and compression

what is being described seems essentially similar. If language evolved out of

embodied action, and if it still carries the embodied core in its abstract concepts,
there is no reason why language should not be a sufficient prompt to the type of
processing the ‘shared minds’ assume. Moreover, as Tomasello (2008) points
out, language develops along with the cultural artifacts and socially distributed
cognitive needs and it increases its formal complexity in response to the
increasing complexity of distributed cognitive networks. A more specific lin-
guistic discussion of the ‘constructions of intersubjectivity’ can be found in
recent work by Verhagen (2005, 2008). For example, Verhagen discusses forms
of represented speech and thought (He said . . ., He thought . . .) as examples of
such constructions, since they explicitly evoke conceptualizations by partici-
pants other than the speaker and the hearer. Certainly, Verhagen’s idea could
naturally be taken further to explain free indirect discourse, but there is perhaps
a different way to view the issue of intersubjectivity in narrative discourse.
It seems plausible to claim that fictional narratives develop our intersubjec-
tive skills to their limits, since they rely on the extended, though not unlikely,
construal of the very communicative context from which the narratives emerge.
Narrative can be treated as a communicative act of a specific kind (and it has
been, among others in Pratt 1977 and Lanser 1981), but it assumes a lot more
than the text as a communicative artifact. Its very basic forms (as described in
Chapter 3) assume a concept of ‘someone communicating something about
someone else from some point of view’ – which is the basic intersubjective set-
up, but its specific micro-level choices manage all the viewpoints potentially
available in the storyworld. The ‘shared mind’ concept assumes that actions,
their motivations, and the emotional response to those actions are parts of the
same cognitive package, situated in a pragmatically rich context. If that is true,
then narratives fit the description. They do provide the context in which actions
occur, they situate actions in ways which make reasons for those actions
possible to explain, and they do coordinate a multiplicity of viewpoints, of all
kinds. And they evoke emotional responses without necessarily attributing
beliefs or desires to the characters.
What the ‘shared mind’ approach postulates is that ‘understanding reasons’ is
the central ability afforded by the intersubjective set-up. As Gallagher and Hutto
argue (2008:27): “understanding reasons for action demands more than simply
knowing which beliefs and desires have moved a person to act. To understand
intentional actions requires contextualizing these, both in terms of cultural
norms and the peculiarities of a particular person’s history or values.” The
idea was first proposed in Hutto (2007b), as the ‘Narrative Practice
Hypothesis,’ which argues that the intersubjective (or ToM) skills develop in
the process of understanding actions in context, and not in attempts to get into
people’s heads. The hypothesis is further put in the context of research findings
(see Decety and Chaminade 2003) which show that understanding of oral
 4.7 Narrative thought and intersubjectivity 115

stories is impeded when facial expressions and body language are in emotional
discord with what is said. Although Hutto’s work is not intended to explain
fictional narratives, the textual equivalents of bodily behavior play an analogous
role. Expressions describing a character’s gaze, tone of voice, expressions of
represented thought, strategic use of reporting verbs like moan or shout are all
signals of reasons why characters act the way they do and why they respond the
way they do. Even though no such theory is evoked there, Palmer’s work (2004)
describes the many ways in which fictional minds are at the center of narrative
acts and in fact proposes an understanding of narrative discourse as constructing
access to other minds and viewpoints.
The role of imitative and emotional responses in narrative comprehension is
further discussed in Currie (2007), who proposes that reading requires adopt-
ing a ‘framework’ through which we can identify with the characters’ stance.
Such a framework or stance is independent of the comprehension of the plot,
and would seem to be a natural extension of the Narrative Practice Hypothesis.
Together, these approaches explain how sentences such as Doming died at the
gates of the graveyard or Free Indirect Discourse fragments can at all be
processed by a reader. They rely on everything else the text makes available –
different characters and their stance, the characters’ actions and reasons for
those actions, the entire context, with its conflicting viewpoints, but they also
rely on our ability to use all these prompts to construct an intersubjective
configuration wherein we can perceive characters’ acts and experiences as
indicative of the way in which they, in turn, perceive and understand others.
There are many textually implied minds at play, from the narrator to the most
lowly character, and together they create an intersubjective constellation we
can grasp without analyzing each one separately, based on their actions and
interactions.
One might claim that postulating a line of development from triadic mimesis
to narrative fiction is a bit of a stretch, but it is less far-fetched than it seems if
one considers the degree to which narratives rely on multiple viewpoints. There
have been numerous attempts in the literature to solve the question of viewpoint
by multiplying categories or expanding them to cover race, gender, honesty,
self-consciousness, et cetera (as in Lanser 1981). But these enhancements only
show that what may be properly understood as viewpoint is all that human
perception, interaction, culture, and social situatedness can suggest. I have here
proposed a different approach, whereby narrative viewpoint is a function of
textually and culturally available narrative space topology, including different
perceptions of that topology.
I will also claim further in the chapters to come that narrative form has
continued to develop in the way which makes human minds more and more
central to it and has also kept on revising its ways of making such fictional
minds accessible. Oral tradition narratives mostly focus on action, and
116 Viewpoint: representation and compression

downplay the teller or the emotional response, while early novels rely heavily
on what characters do or say. However, narrative discourse gradually develops
forms like FID, stream of consciousness, unreliable narrators, and all the
varieties of viewpoint compression, all of which enhance narrative access to
minds in action, thus increasing our ‘shared mind’ abilities. Crucially, narrative
discourse has developed to be a means of accessing other minds in ways much
more sophisticated than spoken discourse. The narrative constructions reviewed
in chapters 2 and 3 look like textbook examples of what Tomasello (2008)
would predict – more and more complex linguistic forms providing access to
other minds in an intersubjective context constructed in the text.
Throughout this chapter, I focused on viewpoint-structuring lexical choices;
in Chapter 5, I will consider a variety of referring expressions (pronouns,
proper names, role-descriptors, common nouns) and, specifically, their contri-
bution to the establishment of viewpoint and identity in the context of narrative
discourse.
5 Referential expressions and narrative spaces

Do you want to grow up to be Miss New Jersey just like your mommy?
(American Pastoral, Philip Roth)

This chapter presents an overview of selected patterns in which referential

expressions are used in the narrative. The analyses presented here cannot
exhaust the issues, but they are intended to show that the mental space complex-
ity of narratives creates a number of cross-space links which can then be
manipulated or differently construed through the use of any referential form –
a pronoun, a name, or a role-descriptor. Reference in narratives is thus a matter
of emerging complex referential networks.
There are many specific problems in this area of language use. First of all,
pronouns in the narrative need a specific treatment with respect to the different
nature of deixis in fictional texts.1 In Chapter 3, I have signaled some character-
istics of the first-person versus third-person choices which structure novels in
their entirety, and in Chapter 7 I will return to the issue of the narrative use of
pronouns in speech and thought representation blends. In this chapter, I will
consider two other classes of referential expressions: proper names and role-
descriptors. Section 5.4 will also propose a brief discussion of how the use of
common nouns in constructions such as ‘fictive motion’ (Talmy 1996; Matlock
2004) can profile participants in the narrative space. The chapter ends with an
excursion into contemporary poetry, to add to the understanding of deictic
pronouns in literary contexts.
The crucial way in which my view of reference differs from most ac-
counts is the focus away from anaphoric and cataphoric relations assuming
a certain flow of discourse, and towards personal pronouns as signals of
connectors between mental/narrative spaces.2 This is similar to Rubba’s
(1996) discussion of alternative deictic grounds, where she shows how the
choice of a deictic pronoun may align the speaker with different locales,
cultures, or viewpoints. But my approach also attempts to reveal the speci-
ficity of narrative discourse, which relies heavily on stable and textually
maintained connections across different spaces. The question is directly con-
nected to the broader issue of the cognitive status of identity of characters,

117
118 Referential expressions and narrative spaces

whether in fiction or nonfiction, and the relationship between identity and

viewpoint.

5.1 Compression, decompression, and cross-space mappings

The concept of a narrative space assumes that such a space is structured with a
certain topology, including primarily time, spatial location, and participants.
These participants may, but do not have to be profiled also in other narrative
spaces (a secondary character may appear for the duration of one scene or
chapter only), or they may participate in all the major spaces. If the latter is the
case, the continued identity in all the relevant spaces has to be supported
through linguistic means. As I argued in Chapter 3, the primary difference lies
in the status of the narrator, but identity is also dependent on how a text treats
the very conceptualization of a character as constituting an independent and
coherent whole.
In earlier work I have discussed many instances of the narrative treatment of
reference which requires that a fictional identity be decompressed across differ-
ent narrative spaces, mostly for the purposes of marking differing viewpoints in
the spaces.3 For example, when travel writer Jonathan Raban visits the Alabama
town of Guntersville, the people he meets view him very differently from how
he views himself. They possibly mispronounce his name a bit, so during his stay
in Guntersville he assumes a new identity, that of John Rayburn, which is
construed from the viewpoint of the townspeople.
(1) I liked being John Rayburn. I had fallen into a routine that felt like a good life . . . But
I knew that if I stayed much longer, I would make that one mistake. Scratch John
Rayburn, and he’d confess my own thoughts on politics, books, religion – thoughts
that wouldn’t wash in Guntersville. Late one night . . . I packed Rayburn’s life into
my case. (Hunting Mister Heartbreak, p. 218)

The example suggests a specific kind of decompression of a frame. Raban

definitely preserves his inner persona, and thus his new identity relies on the
way the inhabitants of Guntersville choose to see him. The stay in the town is a
narrative space in which the writer and first-person narrator is construed from a
different viewpoint, while remaining firmly connected to the narrating persona
in the SV-space. Decompression for viewpoint thus does preserve the cross-
space links which allow the reader to follow the story and see the idea of ‘John
Rayburn.’ At the same time, examples like these (and they are very common)
suggest that referential links may be established not only backwards and for-
wards in the flow of discourse, but also across spaces.
Example (1) also uses proper names in a way going beyond the standard
definition of ‘unique reference.’ ‘Rayburn’ is much less of a unique identity, and
much more of a signal of an application of a new frame to the person. I argued
 5.2 Proper names, frame metonymy, and character status 119

elsewhere (Dancygier 2009) that proper names are in fact not ‘unique’ by
definition, but that they make unique reference possible by calling up a very
rich frame. In fact, it is also common to decompress the frame from the actual
person or location and use the proper name to refer to the frame only, as in
the case Everest (in the sense of ‘ultimate achievement’, as in Getting that
award was his Everest), Vietnam (‘a costly and unsuccessful war,’ as in Is Iraq
becoming Bush’s Vietnam?), or Einstein (‘a genius,’ as in So you think you are
an Einstein?). In all of these instances the proper name is used with modifica-
tion, but does not simply become a common noun, as it retains the uniqueness
of the framing.
At the same time, we can argue that ‘Rayburn’ in (1) is also a role that the real
man assumes for the duration of his stay. Role-value mappings (such that the
role of President can be filled by ‘values’ (such as Clinton, Bush, Obama) also
rely heavily on frames (e.g. the role of President emerges out of the frame of a
system of government). In the case of (1) it is not likely that anyone else would
fill the role, but the same man then assumes another identity/role (that of
‘Rainbird’) when he arrives in Seattle. In the next two sections, I will consider
more examples of the use of proper names and role-descriptors in narrative
discourse.

5.2 Proper names, frame metonymy, and the status of a character

Proper names in narrative discourse may have various functions, going beyond
the designation of a character as a participant in the story.4 Examples of such
narrative functions are numerous and varied, and range from meaningful names
such as Squire Allworthy, which do not maintain the pretence of authenticity,
to names associated not only with family history, but also with shared features
of physique and character, as in the case of the Aurelianos and José Arcadios in
Márquez’s novel One Hundred Years of Solitude. Some authors further rely on
proper names in explicitly connecting narrative spaces, as in the case of
Coetzee’s Slow Man, where two women characters (Mariana and Marianna)
are connected not only to each other, but also to the male character’s childhood
memory of the French symbol of revolution, Marianne. The recurring name,
popular across different languages, gives one main female character, Mariana, a
more meaningful role in the story. Names can also signal narrative options and
play with the idea of character identity, as in Nabokov’s Lolita, where proper
names seem to play a special function. First of all, the narrator, Humbert
Humbert, openly admits that his name has been invented, and discusses other
options he has considered, as well as reasons why he rejected them. Other
characters are named to evoke associations and puns of all kinds (Quilty/guilty,
Cue/Q); also, all characters are referred to in multiple ways. To mention only the
various names of the title character of Lolita, Dolores, Dolly, Lo, Lola, or Lolita
120 Referential expressions and narrative spaces

are all varieties of the main character’s name aligned with different viewpoints
(school, family, sex, marriage, et cetera). Clearly, writers do not always treat
proper names as no more than convenient labels to identify characters.
Proper names can also be used without any specific reference, and then they
can signal associated social and ethnic frames. In Philip Roth’s American
Pastoral (to be discussed extensively in the next section) the main character
is a Jew, Seymour Levov, also known by a nickname the Swede, whose ambition
takes him beyond his Jewish roots. The narrator comments on the name choices
Jews have made:
(2) The contradiction of Jews who want to fit in and want to stand out . . . resolved itself
in the triumphant spectacle of this Swede who was actually only another of our
neighborhood Seymours whose forebears had been Solomons and Sauls and who
would themselves beget Stephens who would in turn beget Shawns. (American
Pastoral, p. 20)
The names here, all used in the plural so as to indicate the reference to generations
rather than individuals, all stand for degrees of alignment with the religious
Jewish past: from traditional Jewish names (Solomons and Sauls), to the some-
what Americanized generation which follows (Seymours), and the next era,
where Jewishness is almost erased (Stephens), all the way to becoming indistin-
guishable from any other ethnicity (Shawns). The reliance on frames makes it
possible to outline many years of social and religious change in a just a few
sentences.
Example (1) shows how decompression for viewpoint may be marked by
different proper name choices. In (3), a similar instance of decompression, the
main character in Anne Tyler’s novel Ladder of Years, Delia Grinstead, begins
to see herself as two different people. In this case, the decompression is also
aligned with location: Delia in Baltimore, where she lives with her family, is not
the same person as Miss Grinstead in the town of Bay Borough, where she lives
alone, having left her old identity behind. In this case, however, the decom-
pression is not a jocular reference to the quirky preferences of the locals, but a
genuine sense of disconnect which Delia recognizes and observes with some
astonishment:
(3) Miss Grinstead was Delia – the new Delia. (Ladder of Years, p. 94)
She had noticed that Miss Grinstead was not a very friendly person. The people
involved in her daily routine remained two-dimensional to her . . . . She hadn’t
developed the easy, bantering relationships Delia was accustomed to. (Ladder of
Years, p. 101)

Crucially, Delia remains the name the character and the narrator rely on in
referring to the ‘real’ Delia – to her inner sense of self. Miss Grinstead, on the
other hand, is a person who surprises Delia with her behavior and taste.
 5.3 Role-value mappings as cross-space connectors 121

My final example of proper names functioning in a text comes from a movie,

Murder at 1600:
(4) Will Carla Town be Jack Neil’s Chappaquiddick?
Jack Neil is a fictional US president, and Carla Town is the name of a murdered
White House employee. The investigation reveals some sexual connection to the
president’s family and thus a scandal is possible. The name of ‘Chappaquiddick,’
a little town in Massachussetts, is used here with reference to a tragic accident
in which a young woman drowned and the man who had accompanied her had
left the site of the accident without notifying anyone. The man was Senator Ted
Kennedy, whose political career suffered as a result.
Example (4) also represents a construction, discussed in Dancygier (2009),
which operates primarily on frames, but also uses the genitive form to prompt a
blend in which an experience of an event is used as a frame to then be attributed
to the experience of another person. In the blend in (4), all proper names are
cross-mapped with other, unmentioned names, and their associated frames. The
Jack Neil scenario here is blended with the familiar Ted Kennedy scenario, and
the deaths of two women are also connected in the blend. The point of the
construction is to use familiar frames to construe new situations, but, crucially
to the point being made here, it relies on frames as prompted by proper names.
In (4), the names call up frames, emerging from the story (Carla Town=her
tragic death; Jack Neil=a politician whose career may suffer), or imported via
another input (Chappaquiddick). The blend can then be used in constructing the
movie’s story along the lines of the scenario evoked (which turns out to be
wrong). Proper names are thus used in the narrative not only to refer, but also to
participate in constructions which structure the emergent story itself.
In the next section, I will consider another referential dimension – that of role-
value mappings.

5.3 Role-value mappings as cross-space connectors

Role-value mappings, as briefly mentioned above, connect two entities: a slot in
a socially determined frame and an individual filling that slot. However, as
narrative data suggests, roles may not have to be defined outside of the text, but
may be established by a pattern recurring across different texts (such as ‘rake’,
‘fortune hunter,’ et cetera) or even by a text itself. In what follows, I will discuss
the case of a novel which is structured by frames and associated roles – Philip
Roth’s American Pastoral.
The novel tells the story of Seymour Levov, a New Jersey Jew, and his family.
Seymour is introduced as a high-school friend of Nathan Zuckerman, Roth’s
literary alter ego, and referred to as the Swede. This high-school nickname is
prompted by Seymour’s unusual physique – he is blond and tall, and very
122 Referential expressions and narrative spaces

athletic. The Swede wants to be just an American, not an American Jew, and
starts constructing his life around his central goal – having his own piece of the
American Dream. And indeed the dreams all come true – the Swede makes a lot
of money running his father’s glove factory, moves to an old house in the
beautiful village of Old Rimrock, marries an Irish Catholic girl, Dawn Dwyer,
and has a beloved daughter, Merry. The dream frame, structured by roles such as
‘money-making husband,’ ‘beautiful wife,’ ‘perfect child,’ and ‘respectable
house,’ is complete. In fact, each of the roles in the frame is amply elaborated
in the first part of the novel, entitled ‘Paradise Regained.’
However, other frames are at work too. To Seymour’s family his wife is just a
shiksa – someone from outside of their own frame of a traditional Jewish family.
Dawn is also a former Miss New Jersey, so she is definitely a beauty, but in the
Swede’s eyes her primary role is the ‘wife’ and the ‘mother.’ When a local paper
asks little Merry ‘Do you want to grow up to be Miss New Jersey just like your
mommy?’ the Swede’s dignified answer is ‘My wife isn’t Miss New Jersey – my
wife is her mother.’ He rejects the framing whereby Dawn could be seen as ‘just a
pretty face’ and places her squarely within the system of family roles which lie at
the core of his American Dream. Besides, Dawn is also a working mother, as she
runs a cattle-breeding company known as Arcady Breeders. This completes the
frame with hard, farmlike work to counterbalance the Miss New Jersey image, as
well as the ‘wife of a rich man’ frame. The American Dream has to be hard-
earned, and even their beloved girl participates: Merry shoveled cowshit from the
time she was six. Merry rode tractors. Nobody is allowed to think that the family’s
success and their later tragic undoing emerge from a life of thoughtless luxury.
The frame would not be complete without the ‘house.’ Driving through a
village Seymour spots a house which fits his frame perfectly:
(5) He saw a large stone house with black shutters set on a rise back of some trees. A
little girl was on a swing suspended from a low branch of one of those big trees.
(American Pastoral, p. 189)
When the local paper interviews Dawn (of Arcady Breeders), the article puts all
the crucial elements of the frame together:
(6) Mrs. Levov, the former Miss New Jersey of 1949, loves living in a 170-year-old
home, an environment which she says reflects the values of her family. (American
Pastoral, p. 199)

There is a storybook quality to the image Seymour constructs. This is only

confirmed through the clearly described image of what he means by ‘a happy
American’ – someone like the legendary Johnny Appleseed:
(7) It was one of those kid things you keep in your mind no matter how old you get, but
whom he felt like out in Old Rimrock was Johnny Appleseed . . . Wasn’t a Jew,
wasn’t an Irish Catholic, wasn’t a Protestant Christian – nope, Johnny Appleseed
 5.3 Role-value mappings as cross-space connectors 123

was just a happy American. ‘Who is his wife?’ ‘Dawn. Dawn Appleseed.’ ‘Does he
have a child?’ ‘Sure he has a child. And you know what her name is?’ ‘What?’
‘Merry Appleseed!’ (American Pastoral, p. 316)

The real Johnny Appleseed (John Chapman) was a man who walked the frontier
selling apple seeds and starting nurseries, but the name naturally calls up a frame
of a children’s story familiar to most American parents. The blended proper
names given as answers to Merry’s questions make the whole family participate
in the maintenance of the elaborate ‘happy American family’ frame Seymour
has started.
Merry is the part of the dream which is less than perfect from early on and in
the end ruins the whole of it. As a child, she is a stutterer. As a teenager, she
befriends people with extreme leftist views and is involved in a terrorist attack
on the post office in her village of Old Rimrock. The attack kills four people,
including a local doctor, and Merry disappears. Merry’s crime and disappear-
ance reveal the cracks in the flawless image of Seymour’s American dream, and
so Seymour’s brother, Jerry, talks about the little murderer herself, the monster
daughter and describes the Swede ironically as Just a liberal sweetheart of a
father. The philosopher-king of ordinary life.
The roles of all the family members have shifted. Dawn, for example, realizes
that however much she might prefer to imagine herself as another prematurely
aging reader of Vogue, she is no more than the mother of the Rimrock Bomber.
Their entire life focuses on the shame and guilt they feel, and they now fill role-
slots in a family frame they weren’t even aware of: parents of a child who
committed a crime against their American dream, parents of the Rimrock
Bomber. The narrative form changes gradually to reflect the growing incon-
gruity of the Levovs’ dream and the reality which they still cannot face.
Substantial pieces of the text consist in the elaboration of contrasting frames
which can never be reconciled into anything coherent. As Merry’s crime
destroys the sense of the Levovs’ life, it also forces the story to start reflecting
the Swede’s tormented thoughts:
(8) Five years pass. In vain the Rimrock Bomber’s father waits . . . A bomb goes off in
the Pentagon . . . The bomber leaves a note. “Today we attacked the Pentagon . . .
while US mines and warships are used to block the harbours of the Democratic
Republic of Vietnam”; The Democratic Republic of Vietnam – if I hear that from her
once again, Seymour, I swear, I’ll go out of my mind. It’s their daughter! Merry has
bombed the Pentagon. (American Pastoral, p. 147)

The narrative becomes more and more fragmented in this manner, going from a
narration of Seymour’s psychological state, through the report of the news, and
the prompted memory of Dawn’s words, to the way in which these pieces begin
to fit together in Seymour’s mind. The Rimrock Bomber is no longer a role for
which Merry, and only Merry, is the value, it is now the name of every terrorist
124 Referential expressions and narrative spaces

in the emerging frame which is later referred to as indigenous American berserk.

Seymour is now the father of every youngster out there who has ever used the
proper name Democratic Republic of Vietnam – another rich frame, giving
legitimacy and democratic status to the communist part of the country of
Vietnam which is in the guerrilla war with the US. The cross-frame links now
extend from their own household and the memories of Merry’s adolescence to
every family affected in this way and to every young terrorist who is missing.
Seymour uses all means possible to find Merry. He meets with Rita, a
woman who claims to know Merry’s whereabouts, but the encounter opens a
new set of roles for them. She mockingly describes Merry as the daughter of
the beauty queen and the captain of a football team, sneers at her mother as
Lady Dawn of the Manor, while calling Seymour Mr. Law-abiding New Jersey
Fucking Citizen and the great Swede Levov, all-American capitalist criminal.
These insults, intended to paint Seymour’s family as morally empty and
self-righteously locked in their world of privilege, do not have the desired
effect. But even though the Swede cannot accept Rita’s framing of his life, he
begins to see how it offered social acceptance and respect to his daughter,
the stutterer, who became a soldier of freedom, a pioneer in the great struggle
against repression. These words frame him as a capitalist criminal, but give
Merry the role of which the Rimrock Bomber is only a peripheral fraction. The
same attack which is a senseless crime to some is an act of courage and justice to
others – how could one be surprised that Merry does not wish to come home to
being the Rimrock Bomber?
He does find her at last. She has changed her name to Mary Stolz, and is thus
not seen as his daughter, but she is no longer the Rimrock Bomber or a soldier of
freedom. She has become a Jain – a member of one of the ascetic religions of
India. She has come to reject all that drove her away from her family, but she
does not quite fit the frame of a peaceful monk devoted to meditation. While
devotion to violence has destroyed her family, the devotion to nonviolence
has destroyed Merry herself: What he saw sitting before him was not a daughter,
a woman, or a girl; what he saw, in a scarecrow’s clothes, stick-skinny as a
scarecrow, was the scantiest farmyard emblem of life, a travestied mock-up of a
human being. Not even the most generic framings apply to the creature before
him, and Seymour has to accept having lost his daughter irrevocably – and
losing his American dream by the same token. He sums up his family’s story
against the background of their intended all-American, religion-blind ideal –
They raised a child who was neither Catholic, nor Jew, who was instead first a
stutterer, then a killer, then a Jain. – and thus acknowledges that the American
dream frame he has relied on was just an idealization.
The consistent use of referring expressions to evoke and juxtapose frames can
be seen as the organizing narrative strategy in American Pastoral, in agreement
with the novel’s poignant critique of the frames which structure America’s
 5.3 Role-value mappings as cross-space connectors 125

social and family life. Crucially, the text’s rich array of role-slots and role-
descriptors, not only supports the facts of the story, but also manipulates a
number of viewpoints. Each of the characters is described from different
angles (the Swede, his brother, the press, Merry’s terrorist friends, et cetera),
and as a result an uncomplicated sequential story becomes a multidimensional
narrative of changing and conflicting roles each of the characters plays.
The multiplicity of viewpoint spaces here is sustained through the unusually
broad range of referring expressions, and the changing discourse patterns.
The strategy extends beyond proper names and role-descriptors discussed
above, and also includes relative and adverbial clauses:
(9) The kid in that swing, the kid in that tree. The kid in that tree who was now on the
floor of that room. (American Pastoral, p. 326)

(10) Merry at the afternoon tea, the band playing, before she’d been raped. She had
danced with the headwaiter, his six-year-old child, before she’d killed four people.
Mademoiselle Merry. (American Pastoral, p. 404)

In (9), the phrase the kid in that swing is clearly a role-descriptor. It evokes
Seymour’s first view of his dream house, with a child in the swing. The role is a
part of the American dream frame, which has to include a perfect, happy child.
Merry could fit the role for as long as she was small, but, as Seymour recalls the
scene, she is now Merry-the-scarecrow in a repulsive room in some dilapidated
building. The clause who was now on the floor of that room attempts to describe
the same child that played in the swing, but there are no discernible links
connecting the two. The technique works similarly in (10), where the ‘before’
image of a pampered girl dancing is juxtaposed with the ‘after’ identity of
Merry-the-Rimrock-Bomber. In both examples Merry’s identity is decom-
pressed across the temporal axis, although under normal circumstances the
timeline of one’s life has some coherence and directionality to it. Here, there
seems to be no path leading from the past into the present.
Also the use of distal that in (9) has an interesting effect. Both frames
(the happy house and the awful room) are distant from Seymour’s ‘here-and-
now’, temporally and conceptually, even though now signals narrative anchor-
ing in the part of the story when Seymour has to accept giving up on his
daughter. They both belong elsewhere – in a fantasy or in the corner of one’s
consciousness which had best remain closed. The distance implied by that is in
stark contrast to the use of this in another scene, one which intends to render
Seymour’s raw pain and despair. In (11), he is showing his glove factory to Rita,
in an attempt to convince her that he is not a ‘capitalist criminal’ and to get to
know her enough to reach Merry.
(11) This is called a pique machine, it sews the finest stitch, requires far more skill than
other stitches . . . This is called a polishing machine and this is called a stretcher and
126 Referential expressions and narrative spaces

you are called honey and I am called Daddy and this is called living and the other is
called dying and this is called madness and this is called mourning and this is called
hell, pure hell. (American Pastoral, p. 131)

The fragment starts as a rendering of what Seymour says to Rita as they walk
around the factory. It has all the formal characteristics of direct speech, includ-
ing the subsequent appropriate use of you and I. Most important, all the construc-
tions refer to the names of things (this is called X) rather than using simple
predicative constructions (this is X). The smooth shift from the ‘factory machin-
ery’ discourse to Seymour’s personal feelings suggests that while the initial
clauses can be treated as direct discourse, the rest is a rendering of Seymour’s
thoughts, patterned by the same construction. The combined effect is a blend of
what Seymour says and what he feels. While ostensibly giving Rita a well-
organized factory tour, he is described as tormented by the memory of his daughter
and the pain of her absence, which Rita could cure with one word; the passage thus
represents the trauma of a desperate father at the mercy of an indifferent young
woman who holds the very sense of his life in her hands. The use of proximal this
throughout the fragment blends the truly demonstrative use related to the dis-
course ground (pointing at objects) and the sense of one’s own inner emotional
state – also proximal, and also demonstrative, though in a different way.
The crucial switch from one narrative space to another happens after the word
honey. At some point in the conversation Seymour starts addressing Rita as
honey – she seems so young and her link to Merry frames her as a ‘childhood
friend.’ The choice of a form of address such as honey evokes his family role for
him, so that her being ‘honey’ puts them both in the same frame and forces him to
be construed as ‘Daddy,’ however incongruent it seems in the context of the
conversation. The realization further prompts the memory of the horror of the
bombing and the painful acknowledgement of his own trauma. The entire frag-
ment is an interesting form of Speech and Thought Representation (STR), given
that it effectively blends speech and thought and in fact uses ‘speech’ as a prompt
for the inner thoughts which develop alongside it.5 At the same time, there is a
clear metalinguistic element to the fragment, as the expression is called X is
repeated in most of the sentences in the fragment. The inappropriateness of the
‘guided tour’ framing of the conversation, when in fact they should be talking
about Merry, is what prompts Seymour’s sense of disconnect from the interaction.
The demolition of Seymour’s dream does not end with the disappearance
(and, finally, death) of Merry. His wife has an affair with their neighbor, William
Orcutt III, a member of a very prominent legal family. Dawn thus starts her life
from scratch, back to the Dawn of Hillside Road, Elizabeth, New Jersey – in an
almost forgotten frame of a pretty Catholic girl from New Jersey. Obviously,
Orcutt’s social position is an incentive (as the narrator comments sarcastically
The mother of Merry Levov needs nothing less than William Orcutt III.), and
 5.3 Role-value mappings as cross-space connectors 127

Narrator: Zuckerman
SV-space present

MN-space
NS1 The Levovs NS2 American Dream
NS1 Jewish Johnny Appleseed/businessman
Seymour
Community Dawn Appleseed/Miss New Jersey
Dawn
The Swede Merry Appleseed/the kid on the
Merry
shiksa swing/Mademoiselle Merry
The monster,
daughter
NS3 Media NS4 Rita
father of Rimrock Bomber capitalist criminal
NS5 Jain
mother of Rimrock Bomber Lady Dawn of the Manor
Mary Stolz
the Rimrock Bomber soldier of freedom

NS6 Old Money

Dawn of Hillside Road; William Orcutt III (Mr. America)

Figure 5.1 Narrative spaces and roles in American Pastoral

Dawn hopes to erase the Levov episode of her life and pick up where she left
off – from being Miss New Jersey. Her hopes are well summarized in the
following sentence:
(12) Teamed up with Orcutt she’ll be back on the track. Spring Lake, Atlantic City, now
Mr. America. (American Pastoral, p. 385)
The chain of events started with the beauty pageants in Spring Lake and
Atlantic City will be concluded in the relationship with the most glamorous of
Americans. This final nail in the coffin of the American Dream removes any
doubt as to the role of hard work and earning one’s wealth that Seymour valued
so highly. Earlier, under Orcutt’s influence, Dawn insists on building a new,
modern house and abandoning the historic building which Seymour loved
so much. The disintegration of the frame he thought he recreated through his
life is complete. Other frames have emerged, which he had never predicted,
and the competing conceptualizations of his life are too disparate to be
reconciled.
From the linguistic perspective, American Pastoral is a story of frames and of
the roles they rely on. Major narrative spaces of the story are set up with respect
to the set of roles they profile, and the plot is organized alongside shifting frames
and roles of the main characters. All these spaces contribute to the under-
standing of the story of the Levov family, which represents shifting social
frames, rather than personal stories – as is shown in Figure 5.1. The lives of
128 Referential expressions and narrative spaces

its main characters are presented much less through the description of the events
that affect them, and more conspicuously through the roles they take in the
changing frames. Almost in the spirit of an allegory, the characters have depth
and personality mainly to the degree to which the roles they take on are the true
core of the story being told. Thus the way Dawn and Merry are represented is
almost skeletal, and only Seymour’s character achieves some depth. The sheer
number of role-descriptors, as well as their narrative salience (as in the part
where Seymour is referred to almost exclusively as the father of the Rimrock
Bomber), is well beyond what most narratives do, and they jointly create an
allegorical story of the state of the American Dream. Seymour, however moving
a character he is, is not more than the focal point of the frame he builds his life
around and of its disintegration.

5.4 Common nouns

To complete this overview of referential expressions and viewpoint, I will
consider examples where there are no human participants profiled in the
expressions used, but they are implied through the use of common nouns and
constructions. In example (13), from Tolkien’s trilogy The Lord of the Rings,
several hobbits are travelling through the Shire:
(13) The sun had gone down red behind the hills at their backs, and evening was coming
on before they came back to the road at the end of the long level over which it had
run straight for some miles. At that point it bent left and went down into the
lowlands of the Yale making for Stock; but a lane branched right, winding through
a wood of ancient oak-trees on its way to Woodhall. “That is the way for us,” said
Frodo. (The Fellowship of the Ring, p. 86)

The fragment uses several fictive motion constructions (it had run straight, it . . .
went down, et cetera), as well as descriptions of change. The presence of the
observers is signaled only through the assumption that the sunset and the road
are being taken in by the hobbits. But there are also subtle signals of the
travelers’ decision-making process, as in but a lane branched right, which
seems to report an observation of someone looking for the right road. The
final sentence suggests that the person was Frodo. Such cases would fall under
Langacker’s (1990b) idea of a subjective construal, since there is a clear
indication of someone observing the situation, but in the narrative context the
description becomes in fact a narration of where the travelers went and how
what they saw affected their choices.
Even when motion is not involved, the construal of the observer may be
central to what is being narrated. In J. G. Ballard’s novel Empire of the Sun, a
boy separated from his parents in wartime Shanghai roams the streets in search
of food and help, taking in the horrors of the situation:
 5.5 Personal pronouns, viewpoint, and the narrator 129

(14) Outside the tram station in the Avenue Haig, the hundreds of passengers were
briefly silent as they watched a public beheading. The bodies of a man and woman
in quilted peasant clothes, perhaps pickpockets or Kuomintang spies, lay by the
boarding platform. The Chinese NCOs wiped their boots as the blood ran into the
metal grooves of the steel rails. (Empire of the Sun, p. 40)

The description of the scene is structured as narration, but it assumes the visual
and experiential point of view of the boy, Jim. At the same time, the choice of
descriptive terms can only be attributed to the narrator, and thus the boy’s
viewpoint is compressed with the narrator’s SV-space viewpoint, with the latter
responsible for the choice of expression. Only two paragraphs later, Jim’s visual
viewpoint is acknowledged in the text (Jim watched), but the narrating voice
takes over again in the next sentence:
(15) Jim watched the coolies and peasant women staring at the headless bodies. Already
the press of tram passengers was pushing them aside, submerging this small death.
(Empire of the Sun, p. 40)
These examples suggest that viewpoint of specific story participants may be
embedded in the richness of the images described. In (14) and (15), the brutality
of the scene described is in stark contrast with what one expects a boy to take in
without shock, and the calm mood as well as the rich detail of the description
create a blended viewpoint which is informative and disturbing at the same
time. Without the implied viewpoint of the boy, the description of the scene
would seem less brutal and cold.
The examples in this section suggest that viewpoint of characters is present in
the narration without the intervention of specific constructions representing
characters’ thoughts. The effect is often achieved through direct rendering of
sensory perception (mostly vision), but the maintenance of third-person narra-
tion also supports the blend in which the character and the narrator viewpoints
are both represented.

5.5 Personal pronouns, viewpoint, and the narrator

In Chapter 3, I have postulated an approach whereby the folk theory of a
‘narrator’ reflects our reliance on the basic spoken deictic set-up. In this context,
the claim of the Deictic Shift Theory that the ‘ground’ of the narrative is
textually constructed is true only to a degree.6 I argue that the text does indeed
construct all its narrative spaces, but it also requires a higher-level story-
viewpoint (SV) space, which structures the text-wide viewpoint of the story.
The space may not include information typically expected of the ‘ground,’ such
as temporal or spatial location, but it is positioned as the discourse ground from
which the main narrative space (MN-space) is viewed, which is partly deter-
mined by the choice of tense. In any event, the SV-space profiles a fictional
130 Referential expressions and narrative spaces

narrating subjectivity which may be fleshed out as a character in the MN-space,

or remain deictically bound to the SV-space without taking on an embodied
shape. These choices determine most of the ways in which the narrative unfolds,
but are also crucially tied to pronoun choices.7
For example, the exchange in (16) seems unremarkable until one realizes that
the I of I said in the first line is a man, travel writer Jonathan Raban. He is
visiting a lifestyle specialist, and uses a fictional persona as a test case.
Obviously, Ms Lee can see that the speaker does not fit the description provided,
but the pronoun use structures both the situation in the MN-space of Raban’s
travel, and the narrative viewpoint.
(16) “I have a problem,” I said.
“You’ve come to the right place,” said Ms. Lee.
“I’ve just won the New York Lottery.”
“That’s no problem.”
“But I’m forty, I’m a woman, I work in a deli in the Bronx . . .
“So you want a total makeover, right?” (Hunting Mister Heartbreak, p. 91)

The I of I said is anchored to the discourse space of the entire narrative – its
MN-space (via Raban-traveler) and its SV-space (via Raban-narrator). It is, in as
much as narrative discourse allows it, addressed to the idealized reader. The I of
I have a problem, however, is anchored to the discourse space of the conversa-
tion in Ms. Lee’s office, where it functions as a blend of Raban-traveler’s
physique and voice and Raban-client’s assumed persona, including gender.
While Ms. Lee is aware that the two are impossible to reconcile, the fact that
she responds to the ‘client’ role, rather than the ‘value-for-client’ person, is
perfectly acceptable in the context of a consulting office (and would not be
acceptable later if Ms. Lee and Raban decided to go for a cup of coffee together).
To the reader, both instances of I are cases of Raban-narrator addressing them,
either to report what he does in Ms. Lee’s office, or to report what he says to
Ms. Lee. This effect is possible because the lowest-level speaker’s viewpoint
(that of Raban-client) is compressed all the way up to the highest level of
Raban-narrator viewpoint. Thus the first-person pronouns anchored to different
deictic grounds in different narrative spaces can coexist side by side and jointly
contribute to the same narrative.
Examples like (16) may seem unusual owing to the added complication of
Raban play-acting as a woman, but any first-person narrative has to distinguish
between the viewpoint of the narrator and the viewpoint of the participant, so
that any such text relies on viewpoint compression across the SV-space and the
MN-space. Also, first-person narratives have to rely heavily on third-person
pronouns, so that other participants in the events are talked about too.
First-person narratives occasionally switch to the plural form we, which I
have briefly discussed in earlier work (Dancygier 2004a; see also Margolin
 5.5 Personal pronouns, viewpoint, and the narrator 131

1996). I have argued that the pronoun we does not necessarily include other
subjectivities (whether in the inclusive or exclusive sense), but that instead it
signals access to a broader frame, anchored to the first-person speaker/writer.
This was the case, for example, in example (14) discussed in Chapter 4, We were
shrinking, said/thought by an observer on a ship leaving port. While the
‘shrinking’ refers to the observer’s perceived size of the ship, the we does not
include the writer in any role other than the observer. There is, as a result, no
‘we’ at all, in the sense of plurality, but there is an evocation of the ship and her
passengers, and one person’s point of view on the entire event.
We can also be interestingly used in a third-person narrative, to engage the
narrator in an unusual way. In Haruki Murakami’s After Dark it reappears
throughout the novel, as a token of establishing the visual viewpoint of the
narrative space. The second sentence of the novel is: Through the eyes of a high-
flying night-bird, we take in the scene from mid-air. Then the perspective
gradually zooms in on Tokyo’s entertainment district, until the next section
starts with We are inside a Denny’s. Then, at the beginning of the next section:
(17) The room is dark, but our eyes gradually adjust to the darkness. A woman lies in
bed, asleep. A young, beautiful woman: Mari’s sister, Eri. Eri Asai. We know this
without having been told so by anyone . . . We allow ourselves to become a single
point of view, and we observe her for a time. Perhaps it should be said that we are
peeping in on her. Our viewpoint takes the form of a midair camera that can move
freely around the room. At the moment, the camera is situated directly above the
bed and is focused on her sleeping face. (After Dark, p. 30)

The text then goes on to tell the story, returning briefly to establishing the
viewpoint. It is an explicit description of the point of view which the narrator
wants the reader to take, and, most important, it establishes the visual perspec-
tive only, including the response of the eye to the change of light. The pronoun
we creates an illusion that the narrator and the reader are coordinating the way
they are viewing the situation, or that the narrator gives a narrative guided tour
to the reader. At the same time, the description has a cinematic quality to it
(also prompted by the word camera), so that one can naturally imagine seeing
what happens in the room. Also the idea of gradual zooming in – from the city,
to the district, to the enclosed spaces of restaurants and bedrooms – resembles a
technique often used in film, where it usually plays the role of locating the action
of the movie in a broader spatial context. In terms of the narrative spaces, the
narration seems to flesh out the SV-space viewpoint and at the same time
blend it with the expected reader experience of the text. Similarly to the
cases discussed in Chapter 4, the focus on the visual viewpoint here is best
described as fictive vision (we are invited to ‘see’ the location of the story’s
events), which highlights the experiential aspects of how the narrative is
delivered. Even though most of the text is not different from a typical
132 Referential expressions and narrative spaces

omniscient narrative, these we-passages make the reader part of the epistemic
viewpoint of the story.
The examples above suggest that, along with other formal choices, pronouns
in the narrative do not only serve the establishment of referential links, but are
also viewpoint devices, managing and linking various levels of narrative spaces,
and also engaging the SV-space. In general, they support the concept of view-
point compression, because in the cases analyzed above the primary function of
the pronouns is to maintain viewpoint cross-links leading up to the SV-space.

5.5.1 Mixing person and tense

In the analysis of The Blind Assassin in Chapter 2 I looked at an example of a
fragmented text, where narrative spaces were marked with different tense
and reference choices. However, there are also instances where a continuous
and uninterrupted narrative mixes different narrative forms. Such texts, as, for
example, Günter Grass’s novel The Tin Drum, do not fit neatly into the types of
set-ups discussed in Chapter 3. The fragment in (18), for example, alternates
between past and present narration, also switching between first person and
third person (Oskar) to refer to the same character/narrator.
(18) Oskar couldn’t adjust that easily, however, and looking about for something to take
the place of his ants, shifted his attention to several flat, grayish brown creatures
strolling along the edge of my Kalmuck’s collar. I wanted to catch one and examine
it more closely . . . And to explain my conduct Oskar says . . . (The Tin Drum,
p. 375)
Oskar is the main character and narrator of the story, a 22- year-old man who has
never grown since he was a child. He tells the stormy story of his family and
his own life. The scene in (18) is the scene where Oskar’s father dies, after the
Russians seize Danzig. The fragment starts in third-person past, then switches to
first-person (my Kalmuck, I wanted) mid-sentence, then moves on to a mix of
first and third person (my conduct; Oskar) and present tense (says). These
switches do not seem motivated by any specific pattern of meaning construc-
tion, other than the attempt throughout to tell the story and also represent the
mind of the narrator/Oskar. There is a true decompression signaled through
these switches, such that Oskar attempts to be both a teller of a bigger story and
a character whose thoughts and feelings (past and present) are being reported,
and maintains both roles throughout.
The next scene is the father’s funeral, where a wave of reflection rolls through
Oskar’s mind, until he begins to consider giving up on his decision not to grow.
The question (should I or shouldn’t I?) is repeated a number of times
throughout the scene, with intermittent reports of what Oskar is engaged with
as he ponders the question:
 5.5 Personal pronouns, viewpoint, and the narrator 133

(19) Should I or shouldn’t I? You’re going on twenty-one, Oskar. Should you or

shouldn’t you? You’re an orphan. (The Tin Drum, p. 384)

(20) Then I found a slender cast-iron cross with crumbling ornaments and crusted letters
that spelled Mathilde Kunkel – or Runkel. Then – should I or shouldn’t I – in the
sand between thistles and wild oats – should I – I found – or shouldn’t I – three or
four rusty, flaking metal wreaths. (The Tin Drum, p. 384)

In the course of this narration-plus-internal-monologue blend, Oskar also

addresses himself as you (You’re an orphan). The second person here is part
of the internal monologue, wherein Oskar is trying to convince himself that
becoming an adult is now inevitable. There is thus another decompression, such
that Oskar’s dilemma is the ongoing state of his mind, while he describes other
things he sees and does, and also attempts to turn arguments around in his mind.
The switches are thus in fact attempts to maintain several narrative viewpoints at
the same time, with the viewpoints being structured by various strands of
Oskar’s thoughts. The technique is a remarkable attempt to build the multi-
plicity of mental viewpoints a single character may take. These viewpoints are
further blended with various narrative roles in the text (character/narrator), but
they depend on the setting up of various thought-spaces. Once again, multi-
plicity of viewpoint is a function of the narrative spaces set up, rather than of the
number of story participants, and formal choices locate these viewpoints in the
overall narrative structure.
To conclude, personal pronoun choices are further signals of the narrative
space set-up in any given text. Such choices often depend on the macro-level
narrative space configurations, as described in Chapter 3, but may also reflect
the micro-level maintenance of various viewpoints (as in the case of The Tin
Drum). The examples discussed above show that it is not possible to determine
the type of narration on the basis of the formal choices alone, but that, instead, a
careful analysis of narrative spaces might offer a better understanding of how
these choices are arrived at.

5.5.2 Pronouns and narratorship

Pronoun choices in the narrative may cause one to reconsider the concept of
narratorship, while also questioning the position of the narrator with respect to
the fact/fiction contrast. In Richard Powers’s novel Generosity, the first-person
narrator is construed as a full subjectivity, while not being cross-mapped with
any character. Additionally, the narrating voice is framed as that of a writer, and
so the story is told as a third-person narrative, but occasionally the narrator’s
voice comments on the writing process and on the decisions regarding the
characters’ fate. At the same time, the ‘story’ is given some independence, so
134 Referential expressions and narrative spaces

that the writer may feel restricted in his story-production choices by the current
state of the MN-space and the characters’ dispositions.
The novel introduces its main character, Russell Stone, in the third person:
(21) A man rides backwards in a packed subway car . . . I watch him balance a yellow
legal pad on his toppling book sack. (Generosity, p. 3)

The narrator/writer is construed as observing characters rather than inventing

them, while remaining invisible himself in their shared time and space. He is not
planning his story, but rather lets it be prompted by what is ‘seen’:
(22) Look again: the whole point of heading out anywhere tonight. The blank page is
patient and meaning can wait. (Generosity, p. 3)
He also makes his character a writer and a teacher of creative writing, and
describes him in terms strikingly reminiscent of Powers’s own physique.
Furthermore, he also has access to Stone’s thoughts on writing:
(23) Just beyond South Bend, Stone has an epiphany. He knows why he could never in
his life or anytime thereafter write fiction: he’s crushed under the unbearable burden
of a plot . . . Plot is preposterous . . . Story is antilife. (Generosity, p. 273)

It thus seems possible right from the start that there are cross-links connecting
the writer, Powers, the narrator of Generosity, and the character/writer, Stone.
While tenuous at first, these links are gradually reinforced. In the story, Stone
teaches a writing class, which includes a student, Thassa. He becomes close to
the girl, as her own story, central to the plot, develops. In the final pages of the
novel Thassa attempts to commit suicide, while Stone is with her, trying to help:
(24) He sits down on the floor, shaking, clouded and adrift. And in that instant of
annihilation, art at last overtakes him, and he writes.
He cannot rescind this. (Generosity, pp. 288–9)

Stone thus becomes both a participant in the events of the story, and also the
writer who is responsible for these events:
(25) He can do nothing for her but revise. And he has time to rework entire world
anthologies. In the scene he keeps returning to, all the principals assemble in her
hospital room. (Generosity, p. 290)
Through these lines, the text allows the novel’s blend to gel at last. Stone is the
writer, and the text he writes creates an irreversible set of facts, ones which
constitute the story we are reading. Not only is the character inside the story its
‘outside’ originator at the same time, the story told is also the story he and other
characters live. Life and story become one.
Thassa goes back to her home country, Algeria, where another character, a
filmmaker, visits her. The narrator reports their meeting in a café. Among other
things, Thassa takes out her old writing textbook from her bag:
 5.5 Personal pronouns, viewpoint, and the narrator 135

(26) “it’s not mine,” she says. “Give it to Russell. He will need this.” I will need
much more. Endless, what I’ll need. But I’ll take what is given, and go from
there. (Generosity, p. 295)

In this fragment, the narrator’s presence in the midst of the scene is fully
acknowledged through the use of the first-person pronoun (I will need much
more). At the same time, it becomes clear that the I here is also linked to Russell
Stone, who was a ‘he’ until that moment. The cross-space connections become
relations of identity rather than just analogy. And because of the initial link to
Powers himself, the author is now also a participant in the scene, and, by the
same token, in the story. In the final words of the novel, the writer and the
character share a moment in silence, and then the ‘story’ disappears, back into
the blank page of reality.
(27) She looks across at me . . . She answers in all kinds of generous ways. And for a
little while, before this small shared joy, too, disappears back into fact, we sit and
watch the Atlas go dark. (Generosity, p. 296)

Through the pronominal choices, among others, Powers questions the boundary
between fact and fiction, by gradually redefining the relationships between
implied author and narrator, as well as a character. He shows how fiction
incorporates but also disturbs reality, and he presents the boundaries between
fact and fiction as flexible. He also gives characters an ontological status, such
that, within the space of the book, they can interact with the author/narrator/
character. At the same time, he gives the author only limited control over
characters’ lives, and puts the author at a parallel level of participation.
The blend constructed in Generosity seems to rely not only on the presenta-
tion of the story, but also on an underlying construction of frames which define
books and fiction. The novel seems to treat storytelling as consisting of a blend
of two frames: the ‘book’ frame (with its author, its characters, and its readers)
and the ordinary discourse frame (with the speaker, the addressee, and partici-
pants). In the ‘narrative’ blend, the speaker is integrated with the narrator,
addressee with reader, and participant with character, but the author remains
backgrounded. In Generosity, the emerging blend is more complex: it blends the
author, the narrator, and the character into one, and thus makes the boundaries
between fact and fiction permeable.
Crucially to the point being made here, pronouns and other referential
expressions play a large role in the emergence of the blend constructed in
Generosity. The first-person pronoun is used in the text, but initially it represents
both the narrator and the implied author. The structure of the SV-space thus
profiles two subjectivities, though blended into one. Throughout the text, the
I-narrator is not cross-mapped with any character, but the ‘intrusive’ nature of
this narration does not guide the reader through the process of story construc-
tion. Instead, it seems to follow the author through the process of story-writing.
136 Referential expressions and narrative spaces

The MN-space, while still in the viewpoint scope of the SV-space, is also
partly independent of it, in ways which resemble reality, rather than fiction.
The final blend of I and he is also prompting the blend of reality and fiction,
defined through cross-space participation and epistemic transparency, not
identity.
Throughout the section, I have considered uses of personal pronouns which
represent narrative-space configurations and text-specific construals of the
relationship between the SV-space and other narrative spaces. These pronomi-
nal uses require that we understand textual deixis as more than the function of
the text’s ground, but as a direct representation of the set-up of narrative spaces
and identity. Pronouns here serve as crucial signals of narrative space cross-
mapping, embedding, and compression. They participate in referential blends,
rather than simply in reference relations.
The fact that deixis may reflect complex relationships between language and
the surrounding reality is not new (see Rubba 1996; Hanks 1990). But the
narrative treatment of that ‘reality’ poses further questions both about the nature
of the ground and the nature of referential relationships. In the final section of
this chapter I will return to the issue of deixis once more.

5.6 Deictic I and the construal of subjectivity

Throughout this book, I have been considering the idea of a textually construed
ground which profiles a disembodied construct called the narrator. I have also
shown how the setting up of a mental/narrative space which includes such a
construct affects the cross-space mappings throughout the narrative and the use
of pronouns (and tense) in the text. However, I have not looked in detail at the
interpretive implications of such a set-up. In order to show what a ‘subjectivity’
like that implies, I will look briefly at two examples from poetry. Even though
there are no narrative elements involved, the observations are relevant to my
overall argument.
Poems often use I (profiled or implied) as a reference to the so-called ‘poem’s
persona’ – a subjectivity whose viewpoint structures the content of the poem,
very similarly to the way in which the concept of the narrator structures the
viewpoint of a novel. It is thus assumed that there is a ‘communicator’ behind
the scenes. However, it is possible for a poem to use I to refer to an object,
and the resulting interpretation reveals much about the nature of literary dis-
course. Below, I will consider two poems by Wisława Szymborska (both from
the collection Nothing Twice [1997]). In one of them, “Advertisement,” the I
stands for a tranquilizer drug, in the other, “Archeology,” for the discipline
mentioned in the title. The drug presents itself as capable of handling many
problems.
 5.6 Deictic I and the construal of subjectivity 137

(28) I can take exams or the witness stand.

I mend broken cups with care. All you have to do is take me . . .
I know how to handle misfortune,
how to take bad news.
I can minimize injustice,
lighten up God’s absence
or pick the widow’s veil that suits your face.
The drug promises to alleviate all kinds of emotional pain and distress. We
know that the power of such drugs is to enable us to deal with such situations,
but we also do not feel that taking drugs causes us to relinquish our agency. And
yet this is exactly the effect that the choice of I brings about: the enablement we
expect from drugs is presented as causation and control: the drug, not the
widow, would pick the veil. This slight shift within the broad range of causal
relations creates an effect whereby the drug taker is no longer fully agentive.
What this shift suggests is that the profiling of a speaker (I) also profiles an
agent, capable of controlling emotions and performing various acts. The drug
should be an instrument (an ‘enabler’), but because it can speak in the first
person, it also acquires agentivity and can now cause things to happen. The
further result is the construal wherein the addressee (drug taker) is only a patient
or an experiencer, but has no control. The point of the poem seems clear, but it is
achieved primarily through giving the drug the ability to speak and, conse-
quently, to cause.
In the second poem, “Archeology,” the set-up involves epistemic abilities as
well:
(29) Well, my poor man,
seems we’ve made some progress in my field . . .
I no longer require your stone gods,
your ruins with legible inscriptions.
Show me your whatever
and I’ll tell you who you were . . .
All I need for my ends
is your layer of dirt
and the long gone smell of burning.
The use of the first-person pronoun here concerns our epistemic agency. We
create academic disciplines in order better to understand the world around us,
but we assume that we have control over what kind of knowledge will emerge.
Shifting us into the role of the addressee and giving control of the discourse to
the discipline construes us as mere recipients of the knowledge we think we are
creating. Again, the use of I comes with additional abilities and control, this
time in the epistemic domain. The reader, profiled in the position of the
addressee is thus given the experience of being dominated by constructs
normally considered under his/her control.
138 Referential expressions and narrative spaces

There are several aspects of this construal which call for an explanation. First,
the use of I without the support of an available deictic ground, which would be
the norm in conversation, in fact sets up such a ground and attributes the
construct identified as I as a carrier of all the abilities normally associated
with speakerhood. The nature of the construction is essentially the same in
spoken and written contexts, but in literary texts requires more of an elaboration
of the space functioning as the ground. Primarily, what needs to be profiled is
the addressee, typically aligned with the reader’s participation in the exchange.
But there are also abilities such as agency and epistemic stance which are
associated with the role of the speaker. The relationship between the reader
and the literary I is typically based on having an equal status (as in the intrusive
narrator case); also, the narrator may yield all the ground to Ego-viewpoints and
remain off-stage. But when the first-personhood is attributed to an object, it
becomes clear to what degree that actual literary ground is dependent on the
specific construal created by the text. Thus it is not that I can be used because
some speaker is available in the ground – on the contrary, the ground becomes
available, and is then structured accordingly, because I has been used or
implied.
There is also the problem of making non-human participants speak. The
possibility emerges thanks to the frames evoked by these constructs and the
subsequent blends in which semantic roles such as ‘agent’ or ‘experiencer’
could be attributed to unusual participants. But the frame of, say, drug-taking
already evokes the enabling function, the agent (drug taker) and the instrument
(the drug); at the same time, it profiles the drug taker as a patient or experiencer,
so that the role is available right away. What the blend does is use the frame
evoked by the object (drug), put the object in a context where it obtains the
status of speakerhood, and assign roles differently to participants. But the
meaning emerges primarily out of the availability of the frame.
The kinds of blends described here are not the only ones in which frames
evoked by objects participate in new construals involving a speaker role in a
newly set-up discourse ground. In the next chapter I will look at blends in which
objects are construed as addressees – with a very different effect. I will start with
a final look at the concept of literary deixis and then consider dramatic discourse
as yet another example of a discourse set-up specific to literary texts.
6 Fictional minds and embodiment
in drama and fiction

. . . O happy dagger!
This is thy sheath; there rust and let me die.
(Romeo and Juliet, Shakespeare)

In earlier chapters I have focused on the contribution the language of narratives

makes to various aspects of story construction. I have relied on cognitive
theories of language to reveal the general mechanisms involved, and I tried to
link the discussion to broader issues of embodiment, framing, and intersubjec-
tivity. In this chapter I will consider further aspects of embodiment, as repre-
sented in the narratives intended for the stage. Specifically, I will look at the
linguistic means employed in Early Modern theatre to represent the characters’
minds, and trace some of the subsequent changes in literary style. Overall, I will
argue that literature develops its formal means in order to find more accurate
ways of representing thoughts and emotions.

6.1 Deictic ground in literary discourse

Cognitive grammar and mental spaces theory analysts repeatedly stress the need
to include a more refined discussion of the ‘ground’ and the deictic set-up within
it.1 As the discussion in preceding chapters suggests, literary texts pose further
challenges to what we understand deictic ground to be. Generally, the ground
is claimed to include the presence of the speaker and the hearer, as well as the
location of the exchange, its history and goal, and the specific communicative
goals of the participants. Recent work (Sanders et al. 2009) also expands the
standard set-up into the ‘Basic Communicative Space Network,’ which assumes
not only the presence of the ground, but also epistemic, speech-act, and meta-
linguistic spaces, ready to be filled with contextually available information on
the speaker’s and the hearer’s attitudes and communicative goals.
Generally, the concept of the ground is intended to be more specific than what
is typically referred to by ‘the context,’ since it attempts to reveal the specific
ways in which the structure of the ground affects interpretation. Furthermore,

139
140 Fictional minds and embodiment in drama and fiction

Rubba (1996) used mental spaces theory to expand the idea of deictic ground
and show the effects of speakers switching across different alternative grounds
and adjusting their pronominal choices to the relation between the base
space and the alternative space profiled in the construal. Crucially, the ground
is used by Verhagen (2005) to clarify the ways in which grammatical construc-
tions (which he terms ‘constructions of intersubjectivity’) serve the goal of
structuring not just the content of the interaction, but also the way in which
it addresses the specific communicative relationships between the speaker and
the hearer. For one, what the speaker says may involve other, off-ground
participants (as in He thinks you’re right), or comment in some way on what
the hearer has said earlier in the exchange. But the primary conclusion from
the whole range of mental spaces work on deixis and the ground makes it clear
that the standard default set-up (speaker and hearer, here, now) may yield a
number of construals depending on shifts of perspective or considering multiple
mental spaces at the same time.
Deixis seems to refer to written narratives only partially. Deixis in narrative
is typically talked about as a text-internal phenomenon: for example, Deictic
Shift Theory (Duchan et al. 1995; see also Rauh 1981) explains changes of
person and tense through text-internal concepts. Otherwise, there seem to be
good reasons for not applying deixis in the context of narration. There is no
identifiable ground which the reader shares with the narrator, also the time of
reading does not coincide with the time of writing (the writer may in fact be
long dead before the reader accesses the text), et cetera. Nevertheless, the lack of
an actual shared ground does not prevent us from conceptualizing reading as a
communicative act in which the text plays a role. Depending on the inclination
of the analyst, the reader may be claimed to be responding to an author or a
text, while the reader’s own participation in the act of reading may be seen as
central or tangential. However, nobody is likely to deny the existence of some
kind of ground, such that the reader is prompted into some conceptual or
emotional response by the text, as produced by the author. The communication
is one-way, but communication it is.
In Chapter 3, I have reviewed some of the most common narrative options,
focusing on the construction of the story-viewpoint space. One of the major
points made there was that the SV-space sets up the topology which is projected
over all the narrative spaces of the text and structures the overall narrative
viewpoint established by the text. The set-up of the SV-space may be under-
stood to represent the text’s ground – with the narrator as a kind of speaker
(regardless of whether the narrator is on- or off-stage) and the reader as a kind of
hearer. The more intrusive forms of narration, where the reader (or narratee) is
directly addressed by the narrator, seem to suggest that such a conceptualization
may in fact underlie more experimental contemporary set-ups. Furthermore,
just as the hearer may become engaged with participants not present in the
 6.2 Mental spaces in dramatic narratives 141

ground, who are brought into the discourse through intersubjective construals
such as reported speech and thought, the reader is also given access to story
participants, via the narrative choices made by the narrator.
In the context of such a default set-up we can draw some parallels between
spoken discourse and narrative discourse. The fragments typically described
as narration are thus equivalent to the narrator’s ‘direct speech,’ while repre-
sented speech and thought, in its amazing variety, plays a role similar to
intersubjective constructions in speech. It is not within the scope of this book
to speculate whether colloquial discourse has fewer such constructions than
narrative discourse, or, on the contrary, more of them. But the pervasiveness of
intersubjective construals in speech, as described by Verhagen, may suggest
that various ways of bringing other participants into the ground are common
both in speech and in writing.
The question which remains open is the exact role of so-called direct speech
(or dialogue) in narrative fiction. It is natural to assume that giving unmediated
voice to characters constitutes some kind of stepping back on the part of the
narrator, a temporary yielding of the story viewpoint to a character. But there
are reasons not to accept such an interpretation, since it would mean that
characters’ discourse is beyond the scope of the SV-space, and, more important,
that narrative fiction changes viewpoint levels all the time, without any mech-
anism ensuring the coherence of the whole. We need an explanation which
allows us to see how individual instances of direct speech are incorporated
into higher narrative levels, including the MN-space. Such an explanation in
general terms seems to be offered by the concept of viewpoint compression, as
elaborated throughout this book. But in Chapter 7 I will also look at a number of
cases of direct speech, to suggest that in the narrative it is in fact representative
of a range of various viewpoint phenomena.
In this chapter, though, I want to look at some selected aspects of a narrative
form which, at face value, consists entirely of direct discourse – drama. I will
discuss the ways in which material aspects of theatre are coaligned with its
language,2 and the drama-specific ways of representing fictional minds.

6.2 Mental spaces, physical spaces, and dramatic narratives

Even though one of the goals of a play, like other fictional narratives, is to tell
a story, it is also clear that the means through which the story emerges are
different. Theatre is thus a good experimental ground for testing the assump-
tions proposed in earlier chapters. I will be interested in one major question:
how does a play tell a story? Given the obvious materiality of the storytelling
environment, one can expect the verbal layer of the medium to play a role
which complements the physical materiality of the stage and the embodied
presence of the characters. It is also to be expected that the representation of
142 Fictional minds and embodiment in drama and fiction

the characters’ minds will take a different form. Finally, the engagement of the
audience is different from the response of the reader, given the appeal to the
senses alongside the exposure to language.
A mental-spaces analysis of a theatre event requires many distinctions. The
primary dividing line falls between the in-theatre reality and the out-of-theatre
reality. The significance of the latter manifests itself not only in the contribution
of the author and her/his historical context, but also in that of the director, and
the availability of material objects which are used on the stage to represent the
frames associated with them. And, naturally, the bodies of the actors appearing
on the stage participate in the outside reality as well. All these observations
may seem obvious, but they contribute to the material aspects of theatre and the
storytelling mode thus constructed.
The in-theatre reality is more structured and more complex as well. We have
to distinguish two physical spaces – the stage space and the audience space,
divided by an invisible barrier.3 The categories are somewhat analogous to the
distinction between the story space of a novel and the situation of reading.
However, the audience space is limited spatially to the enclosure of the play-
house, while being physically separate from the out-of-theatre reality, and
only conventionally separated from the stage, which contributes to the holistic
nature of the theatrical experience. At the same time, the audience space can be
thought of in terms of its visual and aural fields – both in its access to on-stage
events and its separation from the out-of-theatre reality. The stage space func-
tions as the physical embodiment of the story space, and thus houses all the
story dimensions: space, time, characters, story events, et cetera; even the events
which take place off-stage (another space!) are somehow made available to the
participants on-stage, via the stage set, a report by a character, the very events
of the play, et cetera. The stage space is further structured with stage-framed
objects such as props, costumes, the sets, the lighting. It is, however, relatively
independent of the text space, which helps construct many of the stage-space
elements, but also adds the whole discourse dimension of the theatre. Crucially,
the text space of the play is structured differently from any other storytelling
text, as it consists in the turns taken by different characters, some of which
represent discourse types different from the ones available in the out-of-theatre
reality, such as a soliloquy, or the chorus. It also divides the story into chapter-
like chunks (acts, scenes) and provides stage directions as a kind of discourse
which mediates between the stage and the text. Stage directions are usually
not numerous and guide basic movements of the characters, such as coming on
or going off the stage. But they do suggest the need for further blending of the
stage and the text in the emergent story.
Each of the spaces is further structured by more conceptually important
distinctions. The stage space, for one, poses interesting questions with respect
to its spatial organization. In spite of its openness, it has gradually come to be
 6.2 Mental spaces in dramatic narratives 143

understood in terms of a bounded space, so that the inner region is allocated

to the representation of the story space. Contemporary theatre often returns to
the traditional set-up and breaks ‘the fourth wall,’ but that does not make the
wall disappear – rather, the concept is being played with. Apart from the in/out
orientation, the stage is also structured by the front/back schema (downstage
versus upstage) – where ‘front,’ or downstage, is closer to the viewers, and thus
also closer to the center of the entire in-theatre space. Finally the up/down axis
also plays a role, as I will argue in section 6.3.
In some cases, the in/out and front/back orientation is used to distinguish
the region off-stage as contributing to the story-blend. The region beyond the
stage is not to be seen, but can signal events happening off-stage. Examples are
numerous. In Chekhov’s play The Cherry Orchard, the final destruction of
the trees, crucial to the story, is signaled only by the sound of wood being
chopped, reaching the characters and the audience from off-stage. This case is
simple, though, since the whole play leads to such a turn of events. In Stoppard’s
Arcadia, sounds reaching us from off-stage are far more important. Arcadia is
fairly unusual in its use of the stage space, since it alternates between the past
(1809) and the contemporary present, never changing or removing the set or
even the props accumulating on the stage. The past and present events take place
in the same schoolroom in Sidley Park, giving a glimpse of the surrounding park
through french windows. The park, however, is the center of attention, as it is
being remodeled in the nineteenth-century subplot and then investigated in the
twentieth-century subplot. Some crucial events talked about throughout the play
are located in the park, including amorous encounters, but also hunts and duels.
As a result, the sound of a popping gun can be associated either with the killing
of a rabbit, or with a duel in which one of the characters can be expected to die.
The sound of a gun coming from off-stage is used repeatedly to suggest the
duel (which in fact never happened) and then the play goes on adding hunting
for ducks and a killing of a rabbit as explanations, thus playing very cleverly
with the viewers’ story-constructing processes and reminding the viewers of
the duel as the central event. In this sense the sound of a popping gun is used as a
means of evocation of frames existing outside the play (danger, killing) and
constructed within the play (the duel, the hunt). But the sound alone cannot
clarify which frame is to be accessed at any given point in the play.
All the frames mentioned here belong to the past space of Arcadia. The play
uses the past and the present to make the viewers solve the mystery and guess
‘what really happened,’ but it also blends them in interesting ways – for
instance, the final scenes in the schoolroom feature the characters from the
past and the present together on the stage, not talking to each other directly, but
engaging both temporal spaces of the narrative. However, before the blend
comes together on the stage and in the viewers’ minds, the sound coming from
off-stage is used to begin the construction. Even though popping guns are heard
144 Fictional minds and embodiment in drama and fiction

throughout the play, they are primarily used to distract, until Scene 4, the last
scene of Act I, which is taking place in the present and ends with the following
stage direction:
(1) The empty room.
The light changes to early morning. From a long way off, there is a pistol shot. A
moment later there is the cry of dozens of crows disturbed from the unseen trees.

Later, the ‘past’ sequence (Act II, Scene 6) starts very similarly:
(2) The room is empty.
A reprise: early morning – a distant pistol shot – the sound of crows.

The directions clearly suggest that the final moment of Scene 4 is a transition
from the present into the past, and then Scene 6 picks up at that very moment.
However, the transition also clarifies that the sound of the gun does not attest
that there was in fact a duel, because a moment after the shot is heard in Scene 6
Septimus Hodge, one of the past characters, comes in with a rabbit he has just
killed. The sound of the shot is thus not simply a profiling of events off-stage,
but a narrative anchor negotiating between the real and the imagined versions
of events at Sidley Park and guiding the viewer. This is just one example of
how Arcadia uses the off-stage space, because the sounds of guns are heard in
every act, further confusing the events narrated. Similarly, through most of the
scenes in Arcadia, there is the sound of a piano next door, though it is not clear
who is playing and whether the music comes from the present or the past. More
narrative anchoring through sound off-stage means more story-construction
effort on the part of the viewer and greater complexity of the narrative. The
off-stage space can thus be seen as a narrative space which complements the
events presented on-stage, while not necessarily repeating the narrative material.
An interesting aspect of the structure of the stage space is the use of eye
contact between an actor and the audience. It is not common, so as to preserve
the boundedness of the stage space. In natural discourse, eye contact triggers
the construal wherein the participants share a discourse space. In a sense, eye
contact is often a correlate of an establishment of a deictic ground in which the
participants can act as the speaker and the hearer. The fact that actors typically
do not make eye contact with the audience is reinforced by the fact that they
rarely speak to the audience. It happens, of course, but when it does, it signals a
deliberate violation of the boundedness of the stage space. Furthermore, eye
contact between an actor and the audience is even more conspicuous in a movie,
especially since then it only goes one way – an actor-as-character speaks to the
viewer, or, more accurately, to the eye of the camera. Every time this happens, as
in some scenes in Woody Allen’s Annie Hall, the viewer is being pulled into the
space of the story and made to feel like a participant, not just a witness. This is
different from the standard construal of the theatrical or movie audience,
 6.2 Mental spaces in dramatic narratives 145

wherein viewers are expected to construct the story in their minds, not to take a
role in it. However, the audience is also an addressee of the entire performance,
and thus the occasional breaches of etiquette which make the viewer feel like a
participant are not in fact exceeding the limits of the entire theatre set-up.
The text of a play is of course different from the text of any other narrative.
Not only is it structured primarily as dialogue, but it also includes varieties of
discourse which are typically not found either in other narratives or in colloquial
speech: a character may be speaking to her/himself, to a material object (a prop),
to a mental image, or to a dead body of another character, et cetera. All these
forms of discourse play a role in story construction and need to be seen in
the context of the narrative goals a play has to fulfill. As a result, construction of
meaning and the emergent story in a theatrical performance require that we
account for all the spaces involved, as represented in Figure 6.1.
The basic division runs through the middle of the diagram, so that the left-hand
side represents all the material aspects of the play, and the right-hand side is
correlated with the contribution of the text. But neither one is a holistic and uniform
concept, as each consists of more spaces, some of them overlapping to a degree.
This combination of the material and the linguistic is at the core of the genre.
In what follows I will use selected examples from Shakespearean plays to
illustrate the functioning of all spaces and to clarify the nature of the interaction
among them. I will argue that a play tells a story through textual and extra-textual
prompts and that the drama story is a blend emerging out of the audience’s
interaction with on-stage prompts and off-stage frames. Theatrical discourse uses
all the spaces available, but it relies crucially on the theatre-specific blends of the
material and the textual.

Stage space
Text space
set
language (form and frames)
props
discourse and constructions
up/down
front/back
time Discourse spaces
Emergent chorus/audience
performance story space character/character
Back-stage space characters character/object
events character/body
...
Reality space
actors Narrative spaces
objects Audience space
narrating
historical facts listening plot construction
seeing
characters’ minds
constructing meaning

Figure 6.1 The mental space schema of a theatrical performance

146 Fictional minds and embodiment in drama and fiction

6.3 Materiality of the stage and fictional minds

Throughout this section, I will be looking at the material aspects of the stage
space and their interaction with the story construction processes. There is an
inherent difficulty in matching the narrative construction taking place in novels
with the parallel processes occurring in a context where every word is spoken
by someone. It is probably clear that there is no natural way to treat dramatic
discourse as an example of spoken discourse, although in contemporary texts
there is more and more convergence. However, in what follows I will be
considering examples from Early Modern plays, hoping to find the core con-
cepts of storytelling in these texts; later, different stylistic means represent the
concepts in the context of a novelistic tradition.
Discussing the materiality of the stage requires distinguishing various instan-
tiations of the phenomenon: the stage itself, the sets, the props, but also the
actors’ bodies, as well as material and embodied actions, such as manipulation
of objects, body posture, gesture, and movement. Some of these elements are
present on the stage as a part of the representation of the storyworld – a character
needs to be played by an actor with matching physique, a chair is usually on
the stage to be sat on, and a sword is used in stage fights. These are all elements
of the materiality of theatre which rely more or less accurately on representation
blends, in which the represented reality (including a fictional reality) is fused
with the representing reality. I have talked at length about representation blends
in literature in earlier chapters, and will not devote more attention to them. I will
focus, instead, on the cases where material objects and bodies on the stage are
used to do more than represent and participate in the story construction and
the emergence of interpretation. The difference between representation and
emergence of meaning is rather basic here. The former consists in the kinds of
blends we interact with every day, while the latter requires a context wherein
the object is used beyond the ordinary. Out of context, an empty chair will
naturally be understood as available for sitting on, and probably not as a prop
in a cabaret performance. But when Liza Minelli sings in Cabaret, she uses a
chair as a dancing prop, never actually sitting on it. That is where the chair
acquires its new meaning.
Similarly in Arcadia. A gun, or a sound of a popping gun, evokes the frame
of using the gun as a weapon – fighting, hunting, or killing.4 But for it to evoke
a specific duel and participate in the construction of the entire story, which
depends on whether the duel happened or not, more than an ordinary evocation
and representation blend is required. The frame used is story-specific and
emerges as the play goes on, rather than being available instantly. The same is
true about material objects in Arcadia. The table in the center of the schoolroom,
which acts as the centerpiece of the set, gradually accumulates many objects,
left there by participants from both the past and the present subplots: an apple, a
 6.3 Materiality of the stage and fictional minds 147

sleeping turtle, a book, et cetera. The objects, such as the apple, have no specific
role to play in the story, but when a character in one subplot takes a leaf off it
and then another character bites it in the other subplot, the apple begins to
work as an anchor, as it connects two inputs of the blend gradually emerging
through the play. Since the anchoring and blending function of objects does
not properly belong to the common meaning of the word ‘prop,’ I will not use
the word in my discussion. By the same token, this approach does not quite
mesh with the discussion of props in literature (see, for instance, Sofer’s
fascinating book, 2003). Similarly, when actors’ bodies are used on the stage
to support the metaphorical meaning of a scene or to foreground crucial
metonymies constructed through language, this is not a specific aspect of acting
in general (and so is not included in Graver’s in-depth review [1997] of options),
but a construal in which various aspects of the theatre environment are used in
meaning synchrony. In the cases discussed below, the body is used to represent
the mind, but this requires that language and embodiment work together in a
stage-specific environment.

6.3.1 Narration on the stage

The role of a narrator involves various tasks: introducing the setting and the
characters, managing the flow of time, revealing or hiding information as the
rules of suspense demand, and, quite often, also telling the reader ‘what
happened.’ Also, one of the major roles of a novel’s narrator is to provide
some access to the minds of the characters, either by giving the reader direct
access to thoughts or by reporting them or mediating between a character’s
thought and the narrator’s own evaluation. None of these tasks can naturally
be performed in a play, and, crucially, it is clear that all the relevant information
cannot be presented on the stage – either because it cannot be staged realisti-
cally, or because it would take too long, or because the information (such as a
character’s mood) cannot easily be expressed in plain words.
However, the need for some rudimentary narrative mode has been present in
theatre since its earliest forms. A classical tragedy chorus performs some of the
narration tasks and comments on the events in the story in a mode best described
as ‘metanarrative’ – commenting on what has happened rather telling the view-
ers what happened. But a chorus is rare and can only frame the narrative in the
most general terms, as the choruslike characters in Romeo and Juliet, who
describe the basic context at the beginning and comment on the events at the end.
What dramatic discourse does, then, is also use characters as tellers – a blend
which is very natural in the context of a play. In Elizabethan drama, characters
report events which have taken place off-stage, often as eyewitnesses, messen-
gers, et cetera, so that typically it is realistic to assume that the character offering
a report had reasons both to witness the events and to come and give an account
148 Fictional minds and embodiment in drama and fiction

of them. However, on occasion such limits of plausibility are crossed. When

Ophelia dies in Hamlet, by drowning in a stream, her death is reported to the
other characters by Gertrude, in beautiful and moving detail. However,
Gertrude never makes it clear how she came to know what happened, and
there is also a nagging question her report prompts – if someone did witness the
death of a young, innocent girl, why wasn’t she saved?5 The question, though
perfectly reasonable, also presupposes its own answer – because what happens
on the stage is not meant to be realistic, because Ophelia couldn’t have drowned
on the stage, and because someone had to tell the story as a result, so that the
viewers know it. Thus what a character says to a character on the stage is not
necessarily a token of actual discourse between two people, but it may in fact
be addressed to the audience. It is of little importance that Gertrude was not
the person who witnessed Ophelia’s death, but in view of the fragment’s
narrative role it is more important that Gertrude’s own viewpoint as a character
is not involved – she is hardly more than a mouthpiece.
There are also events which cannot be plausibly represented by characters. In
such cases the narrating can also be done by ghosts (as in Hamlet) or other
supernatural beings. Thus the witches who tell Macbeth about his fate not only
foreshadow the events of the play, but also seem to reveal Macbeth’s most secret
desires. The witches are undoubtedly adding to the visual appeal of the story,
but they have a role to play, beyond being mysterious and scary, as the viewer
needs to understand the forces driving Macbeth to action. It is not surprising,
then, that in Kurosawa’s film Throne of Blood, an adaptation of the story of
Macbeth, the witches are replaced by an apparition more appropriate to Japanese
aesthetics – a single pale spirit, turning a spinning wheel (a symbol of fate in
numerous cultures). At the same time, the words of the spirit and of other
characters in the movie clearly identify the role of the prophecy as a prompt or
a catalyst, as they attribute secret lust for power to Washizu, Macbeth’s counter-
part in Kurosawa’s adaptation. In a sense, the movie relies on the aspect of
Macbeth’s story which I find crucial – its focus on the character’s mind, not on
the events. All in all, supernatural interventions in both texts have a role to play in
both narrating events and revealing the characters’ dispositions.

6.3.2 The vertical dimension of the stage and representation

of mental states
The stage is conceptualized primarily as a bounded region, structured from the
point of view of the audience as having a front and a back. It is less common,
though, for the vertical direction to start playing a conspicuous role, so it is all
the more surprising to see the up/down schema structuring the primary themes
of a play – in Shakespeare’s Richard II. It is made possible in the context of an
Elizabethan stage, which was divided into a lower and an upper level, so that the
 6.3 Materiality of the stage and fictional minds 149

spatial division along the up/down schema fits naturally there. Crucially for the
approach taken here, the material feature of the stage is exploited throughout
the play and turned into the main motif, linking various aspects of the story
together, but also being reinforced through all available modalities. The vertical
schema emerges in the language of the play, but is also matched by the use of
the upper and lower stage, as well as by the gestures of the characters and the
metonymies relying on the verticality of the actors’ bodies. The synchrony
across all these modes of expression is surprisingly salient, and is then used
metaphorically to structure the complexities of power struggle and personal
degradation which are central to the meaning of the play.
The up/down schema is the basis of many orientational metaphors. First
described in Lakoff and Johnson (1980), these metaphors use the vertical
orientation to structure many polarized concepts, such as ‘status is up/lack of
status is down (He is climbing up the social ladder), ‘well-being is up/lack of
well-being is down’ (We’ve been sinking for a while), ‘happy is up/unhappy is
down’ (Good to see your spirits are up!), and many others. Many of them are
relevant to the play, in which King Richard loses his crown, is humiliated by his
rival, and suffers discouragement, rejection, and pain. Recurrent vertical images
in the play foreshadow events and highlight their meaning in the context of the
story, while also reflecting the king’s feelings of loss and disappointment.
Early on, in Act I, the king’s rival, Bolingbroke, approaches his sovereign
with due respect:
(3) Lord marshal, let me kiss my sovereign’s hand,
And bow my knee before his majesty; (I.3)

Kneeling in front of another person and bending your head down to kiss the
king’s hand are both body postures signifying submission and respect. Richard
receives the gesture kindly and offers to abandon his high position in return:
(4) We will descend and fold him in our arms. (I.3)

This simple exchange of courtesies has some significance, though. Here, so early
in the story, Bolingbroke still maintains the pretence of being the loyal subject,
and this is signaled through body posture and its metonymic function. On the
other hand, the king’s willingness to ‘descend’ not only shows friendship, but
also foreshadows his imminent downfall as partly caused by his softness. But the
Duke of York is not fooled by Bolingbroke’s gesture and comments later:
(5) Show me thy humble heart, and not thy knee,
Whose duty is deceivable and false. (II.3)

His response not only underlines the intended meaning of the body posture,
but also highlights another body-part metonymy, wherein the heart is the locus
of one’s feelings.
150 Fictional minds and embodiment in drama and fiction

Later in the play, in Act III, Richard knows he is losing the battle for his
crown, but reassures his men, evoking the up/down schema again: Look not to
the ground, Ye favourites of a king; are we not high? (III.2). In the next scene
Richard awaits Bolingbroke’s arrival on the walls of the castle. He is thus still
‘high,’ occupying the upper level of the stage and his high position, but he
knows the war is lost. He engages in a long reflection on the coming events:
(6) What must the king do now? Must he submit?
The king shall do it: must he be deposed?
...
Or I’ll be buried in the king’s highway,
Some way of common trade, where subjects’ feet
May hourly trample on their sovereign’s head;
For on my heart they tread now whilst I live;
And buried once, why not upon my head? (III.3)

The vocabulary of this monologue reinforces the downfall Richard expects

(submit, be deposed). Richard also constructs a spatial image of the reversal
of his fortunes. He should be towering over his subjects, but his burial might
mean that they will trample on him, and his power of control. In effect, to
describe his humiliation, Richard creates more levels of up-and-down structure,
by imagining himself buried under the feet of his subjects. The body-part
metonymies are very meaningful here: the head, as the center of control, and
the heart, as the center of feeling, have both fallen so low that the subjects feet
are above. The power of the scene comes directly from the broad figurative
exploitation of metonymy, and the projecting of much of it onto the bodies of
the participants.
Finally, Northumberland arrives as Bolingbroke’s emissary, and asks Richard
to come down and meet Bolingbroke in the court. In the context of his actually
being ‘up,’ both on the castle wall and on the stage, the request has a very
straightforward spatial sense. But Richard’s words immediately expand the
meaning of the scene into the metaphorical sense of ‘losing power’:
(7) Down, down I come; . . .
In the base court? Base court, where kings grow base . . .
In the base court? Come down? Down, court! Down, king! . . .

As in the preceding scenes, the wording describing descent to the lower court
is exploited to prompt the metaphorical reading. Crucially, the structure of the
stage and the motion of the actor’s body are also co-opted into the rich construal
of the fate awaiting Richard.
Later in the scene, Richard is waiting in the court as Bolingbroke arrives. The
king knows the meaning of this, but Bolingbroke upholds the pretences.
Because of all the metaphorical load of the previous scenes the viewer will
not be deceived by Bolingbroke’s politeness as he kneels again. Richard as well
 6.3 Materiality of the stage and fictional minds 151

understands the empty gesture, as he comments Fair cousin, you debase your
princely knee . . . Then he lifts Bolingbroke and makes him stand up, but,
contrary to his words, this is naturally understood not as courtesy, but as an
admission of defeat. Then he continues:

(8) Up, cousin, up; your heart is up, I know.

Thus high at least, although your knee be low. (III.3)

The contrast between Richard’s earlier Down, down I come and Up, cousin, up
is in each case built into the bodies of the actors and their motion, but the
metaphorical sense prevails again. Perhaps the most interesting part of the line
is the expression Thus high at least – stage directions suggest that Richard
touches his head as he says this, but even without them it is entirely clear that
his head, the locus of the crown and the center of control, must be in focus here
(especially since both the ‘heart’ and the ‘knee’ metonymies have been
exploited in the same sentence). The play is again building its meaning into
the actors’ bodies.
Throughout the play, the same metaphors (good is up; status is up;
happy is up; . . .) and metonymies (head for center of control;
heart for feeling; knee for submission; . . .) are profiled through a
complex blend of linguistic and material means, including the body, motion,
gesture, structure of the stage, and language. Crucially, meaning is prompted
through various combinations of linguistic and embodied means. In the case of
Down, down I come the actor is announcing and/or performing motion down-
wards and speaking at the same time. Such a description of action to be taken
is less likely to happen in spontaneous discourse, but it may occur when the
consequences of the action are highlighted, as when a parent warns I’m coming
in so that the teenager is given one last chance to start behaving. But in the play
the words and embodied actions may be used to prompt for independent
meanings, as when Richard touches his head while talking about his opponent’s
heart. Only his gesture suggests the crown, while the language remains a more
typical instantiation of an orientational metaphor of status. In Cienki’s discus-
sion of metaphoric gestures (2008), embodied motion typically represents the
source domain of the metaphor (so up/down motion could signal well-being or
social status), but Richard’s gesture, being metonymic, points to the head to
signal the crown, even though in the scene Richard is its bearer, in the political
sense, or literally wearing it. In other words, the gesture makes sense in the
context where Richard is thought of in terms of his role, the king, rather than in
reference to his personal tragedy. The various embodied and linguistic means
thus prompt various instances of a rich blend relying on the body posture,
metonymic meanings of body parts, social and political context, et cetera. It is
because of this rich blend that Richard’s deposition is given its political,
152 Fictional minds and embodiment in drama and fiction

emotional, and moral complexity. We can refer to this overall mapping as Body/
Mind blend.
The Body/Mind blend prompts for up/down metaphors throughout the play,6
giving unity and poetic force to the story. Furthermore, the means of story
construction here support the earlier discussion of a variety of roles theatrical
discourse can play. When Richard is describing his emotions and commenting
on his fate, he is speaking in the presence of his courtiers, and thus, ostensibly, to
them. But the lengthy monologue does not really function as ‘direct speech,’ but
provides a view of Richard’s understanding of the events, and thus addresses
the audience first of all. There is perhaps no need to distinguish in every case
between words addressed to characters and words addressed to the audience,
since, obviously, the audience is the ultimate addressee in any case and thus
the boundedness of the stage space is naturally permeable to anything that
can be seen or heard. But, linguistically speaking, the discourse that profiles a
genuine interaction between characters treats the audience as primarily an
‘overhearer’, and is thus different in terms of both linguistic form (such as
style, length of the turn, et cetera) and content.

6.3.3 Material objects and the human mind on the stage

Shakespearean drama has often been described as creating the ground for
literature to represent and shape the human mind. The source of the phenom-
enon is often sought in the power of poetry and the complexity of the characters
and their plights. I want to argue, however, that the effectiveness of Shakespearean
drama is partly due to his truly ingenious use of the materiality of the stage and
the construction of complex blends which represent thoughts and emotions in
creative ways.
I will look in more detail at the tomb scene in Romeo and Juliet (V.3). Romeo
arrives to find Juliet lying in her tomb, and assumes she is dead. He is not aware
that she is only in a drug-induced deep sleep, and thus he succumbs to despair.
He kills himself, drinking potent poison, and falls dead. A moment later, Juliet
awakes, and seeing her husband dead, kills herself with his dagger. The events
of the scene are dramatic and emotionally loaded, but the crucial aspect of what
we see is the representation of the feelings of the two young lovers. How can
this be achieved?
One of the interesting features of the scene is that both Romeo and Juliet
are, in turn, the only characters speaking. That is not unusual, and most
Shakespearean soliloquies are based on a similar idea, but the discourse of
this scene is still surprisingly animated and uses a number of constructions
resembling spoken discourse. Both Romeo and Juliet are justifying their deci-
sion to take their own lives, but their words sound much less abstract than, say,
Hamlet’s reflection on the reasons to refrain from suicide in the famous To
 6.3 Materiality of the stage and fictional minds 153

be or not to be soliloquy. The crucial difference, it seems, is that Hamlet talks

about death as an abstraction, while Romeo and Juliet perform acts of suicide,
which requires that they both talk about death, but also perform the acts of
killing themselves. Hamlet’s proverbial inaction seems to be reflected also in
the discourse the character engages in. For comparison, each of the lovers in
turn addresses the body of the other, they also speak to objects (poison vial,
dagger) and to their own body parts. None of these ‘addresses’ can seriously be
taken to be genuine vocatives, but they do serve a function which needs to be
described in terms of both linguistic constructions and embodied cognition.
Also, the use of framing and metaphor here is again, as in the Richard II scenes,
closely intertwined with the material aspects of the scene. Overall, I will argue
that the scene represents the characters’ minds by relying on material objects,
language, and actors’ bodies in a very complex way.
When Romeo finds Juliet dead, he sits by her and looks, surprised to see her
unchanged. He delves into a complex metaphor wherein death is an adversary in a
love contest, trying to take Juliet away from her Romeo. The metaphor of death
as an adversary is common in colloquial language, while the only one
perhaps even more common is death is departure, as in She passed away,
She left us, et cetera. Thus we might expect Romeo to talk about leaving the world
of the living, going away, following Juliet, but he does not. Instead, he says:
(9) For fear of that, I still will stay with thee
And never from this palace of dim night
Depart again: here, here will I remain.

The decision to remain ‘here’ makes perfect sense if we consider that the scene
is taking place in a tomb, where only the dead dwell. But at the same time,
Romeo’s decision to die is portrayed as less dramatic, almost obvious – what
could be more natural than a husband remaining by his wife’s side? Moreover,
this metaphorical reversal also reinforces the setting of the scene, and would
not make much sense without it. Again, the text is relying on the materiality
and framing of the stage and thus foregrounding its metaphorical meaning
through what might seem a plain description of a spatial location. As in all
the best uses of metaphor, the literal and the figurative are becoming the same.
But Romeo is also aware that dying means losing consciousness, and thus not
being able to delight in the presence of his love:
(10) Eyes, look your last!
Arms, take your last embrace! and, lips, O you
The doors of breath, seal with a righteous kiss
A dateless bargain to engrossing death!

These faux addresses to the body (eyes, arms, lips) prompt a different series
of blends. They evoke the body parts and the sensory and emotional frames
154 Fictional minds and embodiment in drama and fiction

associated with them, so that the viewers can appreciate Romeo looking at his
wife for the last time, embracing and kissing her in a gesture of parting, even if
they cannot see the lovers properly. The way these linguistic forms work is quite
complex. From the formal perspective, the constructions consist of two parts – a
vocative (O, eyes) and an imperative (look your last). But the entities evoked by
the vocatives have no intentionality, and thus cannot be implored to act in any
way, which raises questions about the meaning of the imperative here. In fact,
imperative form in English is often used not to impose actions on others (as
would be the case in standard instances such as Bob, open the door, please!),
but to profile imagined future events (Just open the door and you’re finished)
(see Dancygier and Sweetser 2005 for more discussion). The door-opening
might happen in either of the two instances exemplified, but the verb form itself
does not determine whether the speaker desires it or not. Similarly, in (10),
Romeo’s desires are not part of the construal, even though he is describing
actions his own body might take. But the goal, in the context of the stage, is not
to describe what he wants or orders, but what he will imminently do.
The ‘address’ and the ensuing imperative (look your last, take your last
embrace), evoke the actions Romeo is taking as he is speaking, along with
their emotional meaning – which is that of parting. In effect, the language is
describing what Romeo is doing, and framing his actions emotionally in an
unambiguous way. It is thus a performative blend similar to the ones described
by Sweetser (2000), such as performing a hunt-dance before the actual expedi-
tion, as a way of increasing the chances of a successful hunt. Crucially, the
information which needs to reach the viewer is reinforced through poetic
evocation of the actor’s body and its actions.
A moment later, Romeo ‘addresses’ the poison vial in his hand and drinks the
potion:

(11) Come, bitter conduct! Come, unsavoury guide!

Thou desperate pilot, now at once run on
The dashing rocks thy sea-sick weary bark!
Here’s to my love! – O true apothecary!
Thy drugs are quick. – Thus with a kiss I die.

As in the cases above, the address and the imperative describe (rather than
demand) the imminent act of drinking poison, while announcing Romeo’s
performance of it. Also, the scene is typically played with Romeo’s hand
holding the vial raised in front of or above his face and his eyes looking up at
it – incidentally, this also allows the viewers to recognize the vial for what it is
and see Romeo’s face well. But there is an additional element – the terms used to
refer to the vial. The nouns are conduct, guide, and pilot – all of them metaphori-
cally referring to the expectation of what the poison will do, rather than what it
is, thus describing a change of state (death) conceptualized as transfer from one
 6.3 Materiality of the stage and fictional minds 155

location to another. The initial metaphor of ‘staying here’ is substituted with the
standard departure metaphor for death.
At the same time, the adjectives – bitter, unsavoury, desperate – gradually
move away from the expected sensation of taste to Romeo’s emotional state.
The emotional attributes of the young man committing suicide are transferred
onto the instrument he uses in achieving his goal. Also, in a brief evocation
of the apothecary who provided the poison, Romeo announces that the drug is
taking its effect quickly. And so he dies, having also talked his way through the
process in very moving detail.
Remarkably, this part of the scene takes significantly more time on the stage
than it would be likely to take in real life, even including all the parting gestures.
The story events are actually crisp and fast-moving, but the pace of the play is
different. But good storytelling does not have to be realistic in matching stage
time and story time. In fact, as in all narratives, narrative time is allocated based
on the emotional load and overall importance, and not some realistic rendering
of time flow. The play does the same, but cannot just extend the time by adding
words or slowing down the movements of the actors’ bodies: the engagement
of the actors through discourse and gesture helps in animating the stage time
devoted to an emotionally loaded event.
A moment after Romeo dies, Friar Laurence arrives, too late. Juliet wakes
and sees Romeo dead, but she is immediately urged to run and seek shelter
elsewhere. Her response matches the lines earlier spoken by Romeo, as she is
clearly determined to stay:
(12) Go, get thee hence, for I will not away.

Again, her words match both the literal meaning of her refusal to escape and
the metaphorical mapping death is a staying in a location, set up
earlier by Romeo’s words. Then a pageboy warns her about a noise, and she
makes her decision quickly:
(13) – then I’ll be brief – O happy dagger!
This is thy sheath; there rust and let me die.

Again, her words mimic the pattern established earlier in Romeo’s words, as she
speaks to the instrument of her own death. The crucial difference is that while
Romeo used the adjective desperate, she describes the dagger as happy – the use
of the adjective signals her more willing acceptance of death, but can also be
read as framing the dagger not so much as an instrument of death, but as an
embodiment of the lost lover, welcoming the intimacy.
The use of language throughout the scene suggests its reliance on blends
which allow the characters to reveal their thoughts and feelings without stating
them, and while engaging the material objects in unusual linguistic patterns. In the
blends, feelings are attributed to objects which only humans can experience, and
156 Fictional minds and embodiment in drama and fiction

Discourse frame Killing frame

speaker killer
addressee victim
instrument

On-stage frame
speaker/killer
speaker/victim
addressee/instrument

Figure 6.2 Acting/Speaking blend

prompted to act in ways which match human actions. These are indeed blends,
specific to the stage space, where things appear in order to participate in the
construal of the story.7
The cases of the poison vial and the dagger are similar. With the exception of
their being used with adjectives suggesting different emotions, they both suggest
death, and they frame people both as users of instruments of death (killers) and
as victims. But on the stage, these props are also framed as addressees, and
characters about to use them as speakers. As a result, the roles in the action of
killing are blended with the roles in the discourse on the stage. Speaking and
acting become indistinguishable aspects of the performance.
As Figure 6.2 shows, the blend involves two frames. One is the frame of
killing, metonymically evoked by the objects (poison vial, dagger); the frame
profiles the roles of ‘killer,’ ‘victim,’ and ‘instrument.’ The other input is the
general discourse frame, profiling a speaker and an addressee. In the on-stage
frame, a blend of the other two, the speaker is both the killer and the victim, while
the instrument is an addressee. In the emergent meaning, speech and action are
blended, in that the characters’ words evoke the act of committing suicide, which
is taking place in the same narrative space, albeit moments after the discourse
ends – so that the viewers can absorb the frames and apply them to the events on
the stage. I’ll refer to this kind of blend as the Acting/Speaking blend.
The address to the body parts, quoted in (10), prompts a very similar blend
(see Figure 6.3). The frames evoked, in the context of the relationship between
the characters, are connected with affectionate behavior: eyes evoke admiration
(as they consistently do in Shakespearean sonnets), arms intimacy, and lips
passion. There is no ‘acting’ in the proper sense, but there is showing one’s
feelings of love and parting. In the on-stage blend the speaker is the experiencer
of the feelings, the addressee is the body part associated with the feeling,
and the other lover is an object of emotion. The nature of the acts profiled is
 6.3 Materiality of the stage and fictional minds 157

Discourse frame Emotion frame

speaker experiencer
addressee associated body part
object of emotion

On-stage frame
speaker/experiencer
addressee/associated body part
addressee/object of emotion

Figure 6.3 Feeling/Speaking blend

different, as the role prompted is not ‘agent’ but rather ‘experiencer,’ so in this
variety we can talk about a Feeling/Speaking blend.
Both blends work similarly: material objects in the stage space metonymi-
cally evoke frames. These frames profile roles which are not directly available
on the stage, and then the speaking character is profiled in one (or more) of the
roles. The use of the blends allows characters to talk their way through impor-
tant acts on the stage which, if they were to be simply performed, would seem
more like pantomime. The discourse of the drama is a crucial element of the
functioning of the stage space, but often needs to be materially supported
through objects brought onto the stage in order to evoke frames necessary in
processing the events. Ultimately, these blends also make better use of stage
time, as they give more temporal and discursive salience to the crucial elements
of the scene.
The blends connecting acts on stage and feelings of characters involve props
and the actors’ bodies. None of the blends above are intended to represent what
characters think, but we might expect to find such blends as well. It seems that
they are represented by the cases where Romeo speaks to the dead bodies, lying
there in the tomb. He addresses the body of Paris (“O, give me thy hand . . . I’ll
bury thee in a triumphant grave”), of Tybalt, whom he has killed (“Tybalt, liest
thou there in thy bloody sheet?”), and, of course, of Juliet (“Ah, dear Juliet, why
art thou yet so fair?”). Romeo finds himself in a tomb, surrounded by the bodies
of people who have died in the recent events, so it is natural that he would reflect
on their fate. As in the other parts of the scene, he is choosing addressees who
cannot answer, but their role is that of objects of reflection. The fact that the
vocatives may play such different roles in the constructions does not make the
construction any less consistent in its performative sense – when a single
character on the stage addresses any material aspect of the surroundings, be it
158 Fictional minds and embodiment in drama and fiction

a material object, a dead companion, or his own body, the on-stage meaning is
bringing the object to the attention of the audience and describing the charac-
ter’s thoughts, feelings, or actions focused on that object. Thus Romeo buries
Paris as he talks about it (Acting/Speaking blend), while both Tybalt and Juliet are
addressed as objects of his reflection, in a variety of the blends established above –
Thinking/Speaking blend, wherein ‘speaking to’ stands for ‘thinking about.’
It seems useful to distinguish the three kinds of blends, in spite of their
similarity, because they are achieved through different discourse patterns. They
all profile aspects of the characters’ actions, emotional responses, and mental
processes, and they are all marked by faux forms of address. However, the cases
differ not only with respect to the choice of the faux addressee, but also in the
aspects of the mental and emotional functioning to be represented by various
aspects of the body. As in the case of Richard II, discourse is structured by the
Body/Mind blend. The specific blends profiling actions, emotions, and thoughts
are all instantiations of this larger blend. The kinds of constructions we find in
these cases, as well as the kinds of blends, distinguish the discourse of drama
from other forms of literary discourse, since it relies heavily on the availability
of material objects and on the presence of actors’ bodies. It also agrees with
the story-constructing needs of the viewers, who have to rely on all the cognitive
modalities the theatrical performance appeals to: hearing, vision, processing
language, understanding gesture, and evocation of frames.
The Body/Mind blend, in its various theatrical instantiations, is not as unique
to the theatre as the specific forms of discourse are. The ‘Mind Is a Body’
metaphors are very common, and sometimes utilize very complex domains. For
example, the domain of ‘food’ is naturally co-opted into the construal of
thought, as in thinking is chewing, swallowing is accepting an
idea, digesting is understanding, et cetera. However, there is an
important difference between such uses and the blends just described. The on-
stage blends do not rely on the knowledge of the body and its functioning – on
the contrary, they rely on the direct evocation of embodied experience, as
relayed through the body of the actor. The fact that Romeo talks about the
‘last embrace’ is only a part of the impact – the actor playing Romeo embraces
the actress whose body represents Juliet. That is why the language surrounding
such acts can be unclear or unusual, or rely heavily on metaphor (recall the
conduct, guide, and pilot terms, all describing poison), since the blend relies in
the same degree on the performance of the acts in question.
The props and the actors’ bodies thus function as ‘material anchors’ in the
sense introduced by Hutchins and as ‘narrative anchors’ (see Chapter 2). The
anchoring effect is clear in Romeo and Juliet, where props play such an
important role, but there are objects used in other plays which are central to
the entire narrative construction – enough to mention Desdemona’s handker-
chief in Othello (cf. Yachnin 2002). The handkerchief is a token of the bond
 6.3 Materiality of the stage and fictional minds 159

between Othello and Desdemona, and thus its appearance in the hands of
another man has to be understood as betrayal. In fact, it does not seem to matter
much what kind of object plays the role of such a token – it could have been a
ring, for one. What matters is that the object can be framed as a gift, as something
one can carry around at all times, and which is at the same time easy to take
away (as the actual events in the play demand) or to lose (as Iago’s plot construes
it). The choice of the handkerchief is possibly the best narrative choice, as it fits
all aspects of the description perfectly and is easily seen from a distance.
The actors’ bodies are also used as anchors in all plays that use cross-dressing
(Two Gentlemen of Verona, Twelfth Night, et cetera), as they prompt identity
blends which inevitably play with the femininity and masculinity of characters.
In the context of Elizabethan theatre, the blends are further complicated by the
fact that boy actors may thus represent women who pretend to be men – a
fascinating issue for blending theory, though beyond the scope of this chapter.
The blends shown are also interesting in that they can be seen as precursors of
the numerous constructions of represented speech and thought, used abundantly
since the nineteenth century. Some of the forms, such as free indirect discourse,
can be claimed to be specific to literary discourse, and thus the above discussion
of the Body/Mind blend does not make these blends so exceptional. The
representation of the characters’ minds may be simple – they can just tell the
audience how they feel, via a confession to other characters or in a soliloquy. But
in the scene described, feeling is only a part of what needs to be appreciated, as
there is action as well – and action so crucial to the story that there can be no
misunderstanding (what if the viewers missed the fact that what Romeo drinks is
poison?). Moreover, it is infinitely more true to the genre to create constructions
which also use the very bread-and-butter of theatre – the materiality of the stage,
of props, and of the actors’ bodies. Shakespeare has clearly mastered the skill.
The linguistic constructions he uses are clearly a part of the Early Modern
poetic repertoire, as can be seen in the final scene of Marlowe’s Dr Faustus. The
main character is on the stage alone, being dragged to hell by frightening
creatures. The monologue is rather long, and in the course of it Faustus is
using a substantial number of constructions strikingly resembling the vocative-
cum-imperative pattern of the Romeo and Juliet scene. He speaks to himself
(Ah Faustus, Now hast thou but one bare hour to live), to God, to Christ, and to
Lucifer, to mountains, to stars, to Earth (Earth, gape), to his body and soul
(Now, body, turn to air . . . O soul, be changed into little water drops . . .), to
serpents and adders, and finally pleads Ugly hell gape not, come not Lucifer.
The scene allows Faustus to display a range of emotions (desire for a reprieve,
fear, guilt, regret), but it also allows the viewer to follow what happens to
Faustus and appreciate the bodily pain of being dragged to hell by fierce
serpents. The Body/Mind blend constructions are the primary means in the
scene of allowing the audience to grasp what Faustus’s body and mind go
160 Fictional minds and embodiment in drama and fiction

through as he dies. The fact that all the vocatives call up frames of imagined
situations or otherworldly occurrences explains the lack of material support on
the stage, but the constructions remain the same – they represent Faustus’s
thoughts and feelings.

6.3.4 Ghosts and other supernatural occurrences on the stage

The preceding section mentions the witches in Macbeth as a good example of
how the stage uses embodiment to represent mental constructs. The witches
take a material form, but they seem to be more of an unacknowledged voice in
Macbeth’s head than the sources of knowledge of the future. They may at the
same time be modeled after a tragedy convention when an oracle or a sphinx
predicts future events which seem very unlikely, so that the play can then show
the workings of fate. But considering the rest of the play, Macbeth is not simply
manipulated by fate, and the role of the witches may be closer to revealing to
him what he does not wish to acknowledge. For comparison, Oedipus does all
he can to avoid his prophesied destiny, but fails because in the classical tragedy
world fate rules a person’s life entirely. Macbeth could not avoid becoming
thane of Cawdor, but the fulfillment of the rest of the prophecy is his own doing.
The witches may be telling him what will happen, promising ultimate power,
but they may also be helping him to realize what his desires are.
Another example of a meaningful supernatural intervention is the Birnam
wood. In this case, Macbeth reads the conditional prophecy as a promise of
impunity, assuming nature would not allow the forest to move. So when the
‘forest’ arrives, Macbeth is supposed to receive this as a realization that his
crimes violated the natural order in a way that cannot be forgiven. Macbeth is a
play full of supernatural interventions and mysterious apparitions, but all these
instances are carefully constructed to contribute to our understanding of what
is going on in the characters’ minds, rather than in the story alone. From this
perspective, it seems possible to talk about Macbeth as a play about the mind.
Not all of the subjective construals in Macbeth involve material occurrences
on the stage. In the famous scene in Act II where Macbeth speaks to the dagger
‘in his mind’s eye’ (“Is this a dagger which I see before me, The handle toward
my hand?”) the viewers hear about the dagger but cannot see it – indeed, how
could they, if Macbeth himself seems unsure about the nature of this vision? It
is possible, of course, that the difficulty in giving a real dagger its ghostly
appearance has prevented the writing of the visible dagger into the play,
especially since Macbeth also draws out a real dagger, an instrument of the
murder he has planned. But the entire monologue, preparing the viewer for
Duncan’s murder, is another example of the faux address, as described above,
and thus uses a verbally constructed vision of the material object as a catalyst
through which Macbeth’s unspoken thoughts could be represented. The
 6.3 Materiality of the stage and fictional minds 161

underlying materiality of the immaterial dream dagger plays the same role it did
in Juliet’s suicide scene.
In fact, the monologue constructs two strands of thought, not one, which was
clearly shown in the movie directed by Polański. Macbeth and the viewers
both see the dagger, eerily floating in the air, to the accompaniment of ghostly
sounds. However, the dagger is visible only when Macbeth is directly address-
ing it in his thoughts, and in these fragments the text is heard as a voiceover:
(14) I see thee yet, in form as palpable
As this which now I draw.
Thou marshall’st me the way that I was going;
And such an instrument I was to use. (II.2)

But when Macbeth reflects more soberly upon the strange appearance in his
mind, the dagger is no longer visible, and the actor turns his head away from the
apparition and begins to speak:
(15) Mine eyes are made the fools o’ the other senses,
Or else worth all the rest.

The visual choices made by the director help underscore the claims made above:
the discourse of the play needs to represent the hidden inner thoughts of
characters, and often relies on material objects to serve as addressees in linguis-
tic constructions specializing in thought representation. It may also represent
various levels of inner reflection, through choices involving all the material
aspects of the stage/movie environment – the actor’s body, the actor’s voice, and
material objects.
Returning now to the question of what is or is not visible, there is a significant
difference between things only a character can see (and thus has to describe to the
viewers), those that the audience can see as well, and those that other characters
can see too. The witches, for one, appear to Macbeth and to Banquo, but
apparently not just because the fate of both of them is revealed to each one
and to the audience. The way I am reading the scene, Banquo is there so that
Macbeth’s pride in the high position he is promised can be hurt by the prediction
that Banquo’s descendants will inherit the throne after him. As all the stories of
this kind convince us, the ambition to be king is nothing compared to wanting
your children to continue being kings – and the witches put Banquo in the way of
that desire, creating a conflict between the two characters. Interestingly, Banquo
himself is not tempted to take action to help fulfill the prophecy – otherwise,
what would stop him from killing Macbeth? But the logic of the play is focused
on Macbeth and the conflict between his hidden desires and his sense of morality.
This might support my view of Banquo’s role in the play.
The whole scene with the witches is crucial for the viewers to appreciate the
conflicting desires. But they don’t have to see the dagger, especially if it is to be
162 Fictional minds and embodiment in drama and fiction

clear that it is in Macbeth’s imagination. Once it is mentioned, it metonymically

evokes thoughts of killing, and Macbeth’s words clarify the rest. Another such
instance is the blood spot on Lady Macbeth’s hand, which stands metonymi-
cally for the murders she committed and thus also for her guilt, because, like
the aftermath of the crime, it does not go away. What is more, the hand is
commonly used metonymically to represent willful participation (as in have a
hand in something, give someone a hand), and thus the spot also points to Lady
Macbeth as the instigator of the killings – a memory she wants to erase. She tries
to wash off the blood, addressing it in the now familiar construction (“Out,
damned spot! out, I say!”), and thus we know she is haunted by remorse and
horror of her crimes, but also afraid of being discovered – in her hallucinatory
state, she is not aware that others cannot see the spot. Both the dagger and the
spot are products of the characters’ imaginations, and thus are visible only to
them. Other characters cannot see them and the audience does not have to,
because the discourse patterns profile their presence with sufficient salience. As
in the other cases, speaking to an object, even an imagined one, is used to
prompt for blends which profile the speakers as experiencers of emotions and
thoughts (wish to kill, fear, guilt, et cetera).
Ghosts on the stage are interesting blends. The ghost is assumed to be a
disembodied spirit of a person who is now dead, and yet it has to appear on the
stage in some embodied form, so that it can be appreciated as a ghost and yet
seen by the audience, and possibly the characters. Turner (2004) discusses the
use of death and ghosts in Henry the Sixth, Part One, and in Hamlet, relying to
some degree on the concept of ‘ghost physics,’ as proposed by Talmy (2000).
Wetmore points out (2008) that ghosts appear quite often in Early Modern
revenge tragedies and compares their function with the role ghosts play in
Japanese nō plays. As he argues, Early Modern ghosts are not really agents in
the play: they come to call for revenge, reveal crimes and their perpetrators, and
sometimes also witness the acts of revenge to ascertain that their bidding has
been done. Japanese nō ghosts, by comparison, come to tell their whole story
and often act as avengers themselves – they can be narrators, but also have full
agency in the play. This fits the description of the ghost of Old Hamlet. First of
all, he, like all the ghosts considered by Wetmore, is visible to everyone – he is
thus much more of a participant than just an instigator of revenge. He tells the
part of the story which took place before the events on the stage start develop-
ing, so he does some crucial narrating. Thus this ghost connects the past space
of the story with its present, and provides motivation for the plot as we watch it.
There is also an interesting difference in how the bodies of the ghosts are
represented. Western ghosts are eerie but not necessarily scary, while, as Wetmore
suggests, Japanese ghosts come in full embodied shape, either the way they
looked when they lived, or the way they looked when they died (and thus possibly
disfigured). But they also can appear in more than one form. As Surma points out
 6.3 Materiality of the stage and fictional minds 163

(2009), a nō ghost can truly exist in both the otherworld and the real world, and
can choose the form it appears in, depending on the role the scene plays in the
story. However, due to the highly stylized nature of the poetics of nō, the choices
of how a ghost will appear can also guide the viewers. For example, in the play
Atsumori, the ghost of the main character first appears in the shape of a simple
man, a grass-cutter. The grass-cutter carries a flute, which was in Atsumori’s hand
when he died, so the flute is an anchor to the slain youth. Even though ghosts have
to be played by masked actors to be distinguished from mortals, the grass-cutter is
unmasked. However, his appearance will still be recognized by savvy viewers of
nō as being only a disguise because he would be played by a high-ranking (shite)
actor. This same actor would also play the ghost of Atsumori in his true warrior
form, which appears in the second half of the play. As this example shows, nō
uses various anchors, including the actor’s body, the actor’s position in the world
of theatre, the mask, and material objects to project some of the information
about the role being played. Crucially, the mask defines a character as a certain
type (say, a noble youth, as in the case of Atsumori), but does not allow emotions
to be expressed through the actor’s face – all is done through words and actions.
Ghosts in nō are clearly participants in both the present and the past of the
play – they exist as if in both worlds, and are fully embodied, and thus visible to
all. By comparison, the ghosts in Early Modern revenge tragedies have very
little agency, but they also link the present of the play with its past – it would
thus be accurate to claim that one of the roles of all vengeful ghosts is to
construct a past narrative space which serves as the viewpoint space for the
events of the play’s present. Moreover, as Surma points out (2009), ghosts in nō
also serve as bridges between the real and the unreal, which coexist in some
sense. They come back from the past to tell their story, but can also be agents in
the story’s present. Later, in kabuki theatre, the presence of a ghost on the stage
can be signaled by fireflies, but otherwise a ghost may look like an earthly
human being. Alternately, it may wear make-up in colors which signal death (an
equivalent of a nō mask), which enhances the material aspects of the play’s
meaning.
While the vengeful ghosts such as Old Hamlet may be more common in Early
Modern drama, they are not the only kind. One good example of the different
type of apparition is Banquo’s ghost, coming to sit at Macbeth’s table during a
feast. This ghost comes from a very recent past, as Banquo has just been
murdered, but he does not tell his story to anyone and does not seek revenge.
He appears to Macbeth, and Macbeth alone, so that he is visible to the audience
but not to other guests at the table. Macbeth shows signs of fear and distress,
but nobody knows why. Banquo’s ghost is thus a material form similar to Lady
Macbeth’s blood spot – an embodied token of Macbeth’s guilt. One can ask,
then, why he is the first ghost to appear in the play – there have been earlier
murders and the victims could have appeared to haunt the killers. If the ghosts
164 Fictional minds and embodiment in drama and fiction

were kinds of projections of feelings from the characters’ minds, there would be
no reason to wait till Banquo dies. But Banquo’s ghost appears towards the end
of the play, when it is time for everyone to start wondering what is bothering
Macbeth. He appears at the banquet, not in Macbeth’s sleep (as the blood spot)
and not in the privacy of his being alone (as the witches). I argue that Banquo
needs to appear not just so the viewers can appreciate Macbeth’s distress, but
so that his friends and courtiers can. Furthermore, Banquo takes the seat at the
table which has been reserved for Macbeth, another visual and material means
of reminding Macbeth that his royal seat will not be his for long. Literally and
metaphorically, Banquo takes the royal seat away from Macbeth.
As the examples suggest, ghosts may play a variety of roles in the play, but
what they share is an embodied form which can occupy a specific position in
the narrative, including prompting access to past narrative spaces, never repre-
sented on the stage. They are also perfect means of evoking revealing behavior
of other characters, and prompting discourse patterns which allow characters to
freely express their thoughts. Alongside other material objects and presences
on the stage, they are catalysts for on-stage demonstrations of hidden thoughts
and feelings.
Overall, discourse of the play performs most of the functions later identified
in novels, albeit in stage-specific forms. There is clearly a place for narration
on the stage, as well as for direct and indirect speech and thought representation.
In other words, when characters speak on the stage, they are not necessarily
speaking only to other characters within the story space – they also speak to the
audience, and sometimes speak to the audience via material objects and bodies
as purported addressees. Material aspects of theatrical space are exploited to
profile subjective construals beyond the characters’ words and play a central
role in prompting story construction processes. The reliance on materiality has
a clear advantage over relying on discourse alone, since supporting psycho-
logical components of the story with speech and objects conveys information to
the audience through multiple channels. Also, stage time is allocated to events
based on their emotional impact, not verisimilitude. The meaning of the play
emerges as a blend prompted by textual and material means, and this close
correlation between text and materiality is genre-specific.

6.4 From dramatic narratives to novelistic narratives

As section 6.1 proposes, literature can be seen a special case of a deictic set-up,
where the reader is receiving communication via the text and can rely on the
authority of a narrator. At the same time, it is also a complex intersubjective
construction, wherein the reader is given access to the lives and minds of charac-
ters. In this sense, a literary text is also a site of joint attention (Tomasello 1999,
2008; Oakley 2009; Tobin 2006; 2010), so that the effect of reading is achieved
 6.4 From dramatic narratives to novelistic narratives 165

through the text or the narrator drawing the reader’s attention to the aspects of
narrated situations. Similar claims can be read from Bakhtin’s concept of dialogic
imagination. All of these concepts support the idea that literature, being essentially
a linguistic form, relies in its specific ways on the broad conceptualizations
governing any exchanges of utterances.
In the preceding section I looked at the discourse of drama. Ostensibly, it is
the closest to the deictic set-up wherein there needs to be a shared ground and
one has to see something happen or hear it said to receive the message. But, as
I argued above, even such a presumably unambiguous set-up may be used in
interesting ways to communicate mental and emotional states without them
being directly verbalized. Even if the linguistic constructions suggest otherwise,
however, the pretence of the ‘you have to hear it or see it’ communication is
retained in the structure of discourse, assigned to different characters.
In prose fiction, the pretence of the deictic ground does not disappear too
quickly. One of the telling examples of the transition is La Celestina, a 1499
Spanish ‘novel in dialogue’, by Fernando de Riojas. La Celestina is considered
a literary achievement (some claim it is the first Western novel) mostly because
of its interesting characters and complex plot, but it is remarkable from the point
of view of the topic of this section in that it is a bizarre mixture of drama and
prose. The text is divided into acts, though there are sixteen of them, more in the
spirit of contemporary chapters. Characters speak as they usually do in a play,
but their turns can be pages long and include very good examples of narration
and report. While much of what is interesting about the text is its novelistic
spirit, what is said has to be said by one character to another.
The first-hand access to information is also what seems to underlie another
early narrative form – the epistolary novel, such as Richardson’s Clarissa.
While much later than La Celestina (it is dated 1748), the novel is also infinitely
more complex in its representation of the characters’ feelings and intentions.
Its central narrative form is the strand of letters between innocent Clarissa and
Lovelace, the rake of the story. While what she says is to be taken as a direct
representation of the heroine’s thoughts and feelings, his text is meant to
deceive Clarissa into accepting his advances. Zunshine (2006) offers a detailed
analysis of how these exchanges can be construed from the point of view of the
‘theory of mind,’ but it is equally important for the understanding of the novel
to note that the very format of the epistolary work assumes that the reader also
gets first-hand access to the character’s words, and processes them as originat-
ing in the mind of the character. The epistolary formula backgrounds narratorial
choices and downplays the role of the SV-space in the novel, offering instead
an illusion of direct communication of thoughts and feelings, a variety of Ego-
viewpoint. Of course, the letters also have to contain all the narrative elements
(so that the reader finds out what the plot was), but a text like Clarissa seems to
represent two needs, very salient at the time: the need to tell the story, but also
166 Fictional minds and embodiment in drama and fiction

the need to represent the characters’ thoughts and emotions. And what could
be more reliable than the voices of the characters themselves? As it turns out,
Lovelace deceives Clarissa, though not the reader, but the narrative has now
clearly established its important goal – representation of characters’ minds.
The need to see the mind at work is further elaborated in the so-called it-
narratives – stories told from the point of view of a material object or a pet (a
lapdog, a guinea coin, a coach, et cetera).8 As Lupton suggests (2006), this
ephemeral literary form, very popular in the eighteenth century and then gone,
reflects the emerging interest in consciousness as the topic of literary produc-
tion. From the point of view of the model explored here, it-narratives constitute
a very interesting viewpoint set-up – with an object or animal taking the role of
the character, so that the frequent changes in its ownership and location can be
used to profile more narrative spaces. However, what it-narratives add to the
discussion in this chapter is the observation that material objects can naturally
be used as anchors to narrative spaces and can have human perceptions and
emotions projected onto them. The connection between Juliet’s dagger and the
story of a coin seems tenuous, but it relies on a cognitively salient pattern of
projection of mental phenomena onto non-human participants on the narrative
‘stage.’
The eighteenth century is also when the intrusive mode of storytelling is
common. Again, it seems that the presence of the on-stage narrator derives from
the expectation that a novel is a kind of witness report, even though the illusion
of the narrator as actually being able to know all the information reported is
not being seriously upheld. However, it seems possible to argue that drama
starts a narrating model wherein all the relevant information has a deictically
identifiable source. That source is a direct, on-stage participant, who can be
narrating, commenting, or genuinely involved in a conversation with another
participant, but the information is provided first-hand. The access to partici-
pants’ minds is not postulated beyond what the participants themselves talk
about, but attempts are made to use material aspects of the narrative space
(whether dramatic or novelistic) to relay the information on mental states which
are not being communicated directly. Clearly, narratives are naturally interested
in the minds of characters, but the stylistic means of making mental states
accessible to the reader are initially more reliant on material anchors. However,
we need to acknowledge that the techniques in all cases involve the emergence
of specialized linguistic constructions. These constructions develop, especially
with the emergence of modernist techniques, but the path to interior mono-
logues or free indirect discourse starts with the vocative-cum-imperative forms
and intrusive narration.
Even after the novel is well developed, authors may resort to theatre-like
forms, as in the case of Melville’s classic Moby-Dick, published in 1851. The
narration of Moby-Dick is truly multidimensional, with frequent excursions into
 6.4 From dramatic narratives to novelistic narratives 167

nonfiction, longish treatiselike fragments, embedded shorter stories, and even a

part which looks like it has been taken out of a play. The style is so varied and
original that it would require a longer study to do it justice, but I want to mention
just two fragments in which representation of the characters’ minds takes a truly
innovative form. One of the turning points in the novel is the scene in which
Captain Ahab announces that the Pequod’s expedition has a special goal – that
of killing the white whale (rather than just finding any whales to hunt). He
communicates his passion to the crew and receives their pledge of support. The
goal is still controversial and dangerous, though, and the reader can expect that
the men on board may have different emotional reactions to what has happened.
At this point the narrator, Ishmael, ostensibly just one of the crew but also the
novel’s primary voice, steps back to let others speak. It is narratologically not an
easy task, since the novel is essentially the first-person narration of an eyewit-
ness. Melville thus resorts to a theatrical setting for the duration of four short
chapters, named Sunset, Dusk, First Night-Watch and Midnight, Forecastle.
These chapters are thus overtly covering one evening, right after Ahab’s pas-
sionate announcement. In Sunset, the first-person voice is given to Ahab, in
Dusk it is given to first mate Starbuck, and in First Night-Watch to second mate
Stubb. Finally, Midnight, Forecastle represents all the mixed voices of the crew,
in the form of a play. All four chapters contain ‘stage directions’, including
rather traditional ones, such as (Aside). Essentially, the first three of these
chapters are internal monologues of the main officers of the Pequod, reflecting
on the situation. The chapters represent reflective thought only, and each of the
characters represents a different emotional attitude.
The viewpoint of the narrative here is almost modernist in its complete shift
to Ego-viewpoint in each case and in the way in which the narrator’s role is
reduced to the ‘stage directions.’ However, these stylistic choices are clearly
not just narratorial experiments – they have a very specific role to play in the
development of the plot. Elsewhere in the novel, mental states or participants
other than the narrator are not central to the telling of the story, but the dramatic
scene in which the crew come to share Ahab’s obsessive goal requires some
stepping back and taking in of the consequences of the emotional shift. When
the ‘faux stage directions’ return at the end of the novel, it is again to announce
that a character (Ahab, or Pip) is either left alone on the ‘stage’ or speaks to
himself (and the reader, but not other characters) – again, the formula is a kind
of monologue. The theatricality of this narrative solution is somewhat artificial,
but it effectively takes the floor away from an on-stage narrator and gives full
voice to a character for an extended turn, which could not have been handled
with dialogue alone. Crucially to my argument, the need to represent thoughts
and feelings is what prompts this narrative innovation, and the most natural
solution involves giving unmediated, first-hand voice to the character. The
‘seams’ are showing, but the goal has been achieved. The role of the hearer in
168 Fictional minds and embodiment in drama and fiction

the theatre has been blended with that of the reader of a novel, and the goal has
not changed.
In this section, I have focused on the role of supernaturally embodied
narrative constructs in the framing of the minds of characters. The next section
briefly considers the recent emergence of texts which rely on the body in more
natural terms, while also focusing on the mind.

6.5 Fictional minds, bodies, and brains

While earlier literary narratives use bodies and materiality as inputs into mind-
constructing blends, some recent texts display different tendencies. On the
one hand, bodies are directly used metaphorically as representations of the
characters’ minds; on the other, there are numerous texts featuring characters
with various disorders related to unusual or deficient brain activity. The latter
tendency is particularly interesting, as it suggests that literary production
remains sensitive to the issue of the representation of minds and uses recent
knowledge in neuroscience as inspiration.
Rushdie’s Midnight’s Children is an excellent example of how bodies are
used in contemporary texts. Some of the main characters are described as
having particular and much exaggerated bodily features, which the text aligns
with their psychological states. The main character, Saleem, has an unusually
large nose, always congested and runny. But this very feature of his anatomy
is what makes it possible for him to communicate telepathically with other
children born on the cusp of India’s independence; once his nasal passages are
clear, he loses his gift. His mother, Amina, hides her continued love for her first
husband, but when she starts seeing him again in secret, she also starts suffering
from awful corns on her feet; naturally, the corns go away when she breaks the
relationship. Also, Saleem’s counterpart and enemy, Shiva, is distinguished
by his unnaturally large and protruding knees, which are a mark of his brutal
strength. Other characters are also marked by their physique – Naseem’s warts,
Padma’s body hair, Jamila’s voice, et cetera. While all these narrative choices
can easily be subsumed under the novel’s magic realist style, these bodily
features become central to the narrative as metonymic for the characters’ special
roles and gifts.
Examples of what I refer to as ‘embodied characterization’ are common in
other contemporary texts. The main character in Coetzee’s Slow Man has his leg
amputated after an accident, but his limited ability to move is also aligned with
his mental state, which prevents him from moving beyond the tragedy that
befell him and trying a different life. Also, the main character of Günter Grass’s
novel The Tin Drum, Oskar Matzerath, refuses to go on growing at the age of
three, and remains physically, though not mentally, a child for many years. This
embodied refusal to participate in the society in which he finds himself qualifies
 6.5 Fictional minds, bodies, and brains 169

the novel as a form of magic realism, but also foregrounds its primary themes
throughout the text. The heavy-bodied character of E. Annie Proulx’s The
Shipping News is also not too agile mentally; the two clueless lovers in The
Blind Assassin are cross-linked to another pair of lovers – she mute, he blind, et
cetera.
We can also note a recent emergence of numerous novels using characters with
various mental and neurological disorders. To mention just a few, there were main
characters suffering from Tourette’s syndrome, Asperger’s syndrome, Capgrass
syndrome, Huntingdon disease, and de Clerambaut syndrome. All these texts,
which I have dubbed ‘neuro-literature,’ inspect the human soul (and build good
stories) through the functioning of the brain, rather than via psychoanalytically
defined inner conflicts or postcolonially confused identities. They try to reach into
the very embodied core of the human mind and human emotions.
I consider this narrative tendency to be the new form of the literature of
embodiment, informed by current advances in neuroscience and the flood of
information on the brain pouring off the pages of books for the general public
and popular magazines. Literary production, I argue, has strived to give readers
access to the minds of characters since its inception. The means have relied on
the body and the material world all along, but have gradually adapted both
linguistically and psychologically. At the linguistic end of the spectrum, the old
faux imperatives were substituted with more and more intricate constructions
of speech and thought representation, including achievements such as the
stream-of-consciousness techniques or the form of free indirect discourse. At
the same time, the body regained its space in our attention, as a crucial means
of shaping our mental life and giving it an expression. Also, the general knowl-
edge of what our mental and emotional life depends on has led to literary
solutions exploring the consequences of what neuroscience has taught us.
To sum up, many of the changing literary conventions I have looked at occur
as improvements in the representation of fictional minds. These improvements
rely on our folk theories of how other minds can be accessed, such that are
available at any given time. Early on, human thoughts are primarily accessed
through embodied actions and self-reflexive linguistic expressions, but, as time
goes by, providing ‘access’ to other minds becomes an important function of
literature, one which can be identical with plot construction. The fictionality of
the minds portrayed creates the kind of focus and time compression which real
interactions do not provide, and thus literature gradually develops linguistic and
narrative constructions which allow story-time to be allocated in ways which
help represent the minds, rather than events.
From the narrative perspective, direct textual access to the minds of charac-
ters has become a norm; intrusive narration, with its overt assumption of shared
deictic ground, is now mostly used in literature for children and humorous texts.
It is assumed that some linguistically negotiated access to the mind is possible,
170 Fictional minds and embodiment in drama and fiction

even a direct access, as modeled after modernist texts, especially stream-of-

consciousness ones. Interestingly, these narrative forms, which one could refer
to as ‘direct thought,’ are often linguistically indistinguishable from ‘direct
speech.’ Literary style seems to be redefining the role of constructions such
as direct speech into tools of close representation of thought. In Chapter 7, I
will consider the emerging constructional peculiarities of the speech/thought
contrast.
7 Speech and thought in the narrative

Can Rita and I blow up the post office?

(American Pastoral, Philip Roth)

In this chapter I will consider a range of literary language forms in which the
formal distinction between speech and thought is used in different constructions
specific to narrative discourse. Narrative discourse is typically treated as a dis-
course genre independent of colloquial usage, although, as I argued in Chapter 6,
there are conceptual similarities, even if accompanied by formal differences.
Further distinctions within the broad category typically address the difference
between generic narrative purposes (pure narration) and the representation of
discourse attributed to specific characters (Speech and Thought Representation, or
STR); the latter are in turn talked about as direct discourse (as in He said, ‘I like
it’), indirect discourse (He said that he liked it) and free indirect discourse (He
liked it). The literature on these constructions is extremely rich, with much of the
work recently brought together in Vandelanotte (2009), but since the issue is
beyond the scope of my present concerns, I will only discuss selected instances of
uses pertinent to the issue of viewpoint and story-construction.
Above, I only used the term ‘discourse,’ without distinguishing speech from
thought.1 In fact, I will be arguing below that not only is the distinction not
salient enough, it is also misleading. As I have shown in Chapter 6, linguistic
forms specific to speech have been co-opted into the representation of thoughts
and feelings quite early on. In fact, these forms themselves suggest that the
nature of all literary forms is such that, being based in the language resources,
they need to rely at least partially on available constructions to express mean-
ings which are not the core of the constructions’ semantics. In drama, indeed,
speech, and more specifically speech profiling a potential addressee, is a common
choice in representation of thought and feeling. Languages rarely (if at all) use
constructions which formally distinguish speech from thought, since what mat-
ters most is the attribution of content to a specific character and thought is often
communicated through speech.
The crucial point is not whether any given content was communicated
(speech) or not (thought), but how to distinguish the literary representation of

171
172 Speech and thought in the narrative

communicated or communicable content from the state of a character’s con-

sciousness. For example, an actual perception of warmth and sunlight, regis-
tered by someone’s consciousness, may never need to be verbalized, but in a
novel its representation may take any number of different forms, including a
description of the weather (It was sunny and warm), a description of a charac-
ter’s bodily experience (She felt the warmth of the sun on her face), and an STR
construction (How nice and warm, she thought/said, turning her face towards
the sun). Any of these choices (and many others) would perform the function
of letting the reader know what the character experienced, but the issue of
STR only becomes relevant when the perception is communicated (speech) or
consciously formulated in the way in which it could be communicated (what
counts as ‘thought’ in STR).2
In Chapter 4, I discussed narrative forms which use the pretence of visual
perception to represent conceptualizations. In this chapter, I will look at a
related phenomenon, wherein speechlike constructions in novels clearly play
the role of the representation of the characters’ minds, without being mediated
through the viewpoint of a narrator. In the final section, I will compare them
with selected examples of more typical uses of STR.

7.1 Types of discourse spaces in the narrative

Relying on the concept of mental spaces and their special case, narrative spaces,
offers some possibilities for a treatment of narrative discourse which does not
rely entirely on the opposition between narration and STR. It is based on the
assumption elaborated earlier in this book that narrative spaces constitute a
multileveled network, with each space marked by a viewpoint (such as an SV-
space, the MN-space, or an Ego-Viewpoint space). When a narrative space is
marked with a viewpoint of a character, all the structure of the space is accessed
that way, regardless of the linguistic realization of that space (whether it is
structured more like speech, like thought, or like narration). As we have seen in
Chapter 6, a narratorial experiment such as Melville’s theatrical chapters may
blur these categories and blend them into one. Each of the three Ego-Viewpoint
chapters in Moby-Dick is at the same time an instance of speech (since it is
framed in theatrical terms), of thought (because it represents a character’s
reflection on the events), and, to some degree, of narration (because the view-
point is compressed up to the SV-space, and participates in the construction of
the story of the hunt for Moby-Dick).
Such instances are common in novels. It matters much less whether a given
space is structured as a specific discourse type, but it does matter which view-
point is represented and how the space participates in the overall story-
constructing blend. In (1), Anne Tyler’s character Delia is preparing herself
mentally for a job interview:
 7.1 Types of discourse spaces in the narrative 173

(1) Good afternoon, she would say. I’m here to ask if . . .

No, not ask. Ask was too tentative. (Ladder of Years, p. 94)

The italicized fragments in the first line participate in several narrative spaces.
As speech, they are lines in an imagined future conversation Delia is planning.
As thought, they belong to the space of Delia’s plan for the interview the next
day. And technically speaking, along with the rest of the fragment, they are also
parts in the narratorial rendering of Delia’s thoughts about the interview.
However, the narrator’s intrusion is limited to she would say, which also reports
the content of Delia’s plan as she rehearses it to herself, and is thus itself an
example of free indirect thought, along with the second line of (1). What is
more, ask in the second line of (1) belongs to a metalinguistic space in Delia’s
thought (as reported by the narrator), and is a quote from the imaginary
conversation in Delia’s mind. It seems that the crucial distinction is between
Delia’s reflective space and the imagined conversation space, and not in making
the (difficult) decision as to the allocation of Good afternoon to speech or
thought category – in fact, it is both, depending on the space whose viewpoint
we are considering. Also, ask represents two different ‘thought’ spaces – Delia’s
plan and Delia’s metalinguistic evaluation of the word choice. Finally, temporal
and referential compression is what obliterates the categories (as well as allow-
ing for a consistent choice of tense and reference), so that in fact the distinctions
are interpretive only. The viewpoint really matters here, but the speech/thought
distinction not as much.
Examples of speech in imagined conversations are not uncommon, and
participate in very interesting construals. The fragment in (2), from Eggers’s
novel, is a part of a complex narrative space network. The main character, Dave,
is driving to meet a girl while also worrying about the fact that he has left
his little brother, Toph, with a babysitter. He imagines a scenario wherein the
babysitter is a murderer, and kills the boy, whereupon Dave is accused of
parental negligence:
(2) There will be a trial, a show trial –
How did you meet this man, this babysitter?
We found a posting, on a bulletin board.
And how long did your interview of him take?
Ten, twenty minutes. (AHWOSG, p. 126)

The conversation (between the prosecutor and Dave?) is imagined as a deposi-

tion in the trial in which Dave (not the babysitter!) is the defendant. The entire
court scenario is a narrative space embedded in the train of thought starting with
the imagined murder. It is again both direct speech and free indirect thought,
depending on which viewpoint is adopted. But primarily, what is being elabo-
rated is the ‘emotional’ space of Dave’s paralyzing though unreasonable guilt,
woven into every bit of the narrative he constructs.
174 Speech and thought in the narrative

The approach which distinguishes narration from various forms of STR has
much to contribute to our understanding of viewpoint phenomena. Later in
this chapter I will consider some examples to show how the framework being
developed here treats the important grammatical effects of STR (tense choices,
pronominal choices, et cetera). First, however, STR phenomena must be seen in
the larger context of a fictional narrative, including the way reported content is
embedded in higher spaces and the complexity of such embedding/embedded
relations in the context of the entire narrative structure.

7.2 Speaking for thinking

It is generally true that narratives use similar forms to represent speech and
thought, especially in the ‘reported’ or ‘embedded’ cases. What is less talked
about, however, is the forms which are appropriate to the representation of
‘direct thought,’ as opposed to ‘direct speech.’ The shared understanding seems
to be that the term ‘thought’ refers to the ideas in a character’s mind which are
not structured in a dialogic form of exchange of utterances, and are thus not
addressed to anyone in particular. As I will show below, however, direct thought
can rely on the dialogic structure of conversation, profiling various viewpoints
as discourse participants. I will refer to this narrative technique as speaking for
thinking.
The question of a potential addressee is crucial to this discussion. In the
examples analyzed in Chapter 6, thought can take dialogic form because the
text profiles a material or imaginary addressee – hence the constructions in
which material objects, bodies, or mental images are being spoken to so that
the speaker’s thoughts and feelings can be expressed and actions identified. But
in the context of a novel, dialogic structure typically constitutes only a portion
of the text, and then the speaker and the addressee are characters, situated in
the context of the same ground, as emerging from the story being constructed.
Moreover, the reader may be an off-stage addressee of all this, but becomes an
on-stage addressee only in the case of intrusive narration.
However, it is possible to treat much of what would qualify as ‘direct
thought’ as a one-way dialogue similar to narration, though with an important
difference. When the narrator is on-stage, the viewpoint of the SV-space is
directly accessible to the reader. But when a character’s thoughts are commu-
nicated, they are relayed through an off-stage narrator. In an example like (3),
from Roth’s American Pastoral (discussed in some detail in Chapter 5),
discourse structured similarly to conversational discourse is presented through
the filter of the SV-space viewpoint and construed as though the Ego-Viewpoint
here were a part of a conversational structure in which the person asking
the question also answers it – an effect which is possible also in colloquial
discourse.
 7.2 Speaking for thinking 175

(3) Had there ever been a happier child? A less destructive child? A little signorina any
more loved by her mother and father? No. (American Pastoral, p. 229)

The questions are rhetorical, as is in fact the answer, since the point of the
fragment is to express the amazement at the discord between the pampered
child of the past and the teenage terrorist she became. The text in (3) reflects a
shocked and pained reasoning of the father, as reported by a narrator. The
narrative space structure of the novel as a whole includes a number of levels
of embedding – from the on-stage narrator, Nathan Zuckerman, who runs into
his high-school friend, through Zuckerman’s attempts to find out more about
the Swede’s life, down to his complete identification with the character whose
story he is telling, where (3) belongs. The example thus represents plausible
thoughts of the Swede, as they are imagined and reported by Nathan. Structured
as speech, they are clearly thoughts; formed as questions, they are expressions
of shock and disbelief. But Zuckerman’s narration, which dominates the entire
novel, is indistinguishable from an off-stage narration with a zero-narrator, that
is, a narration where the viewpoint of the character is the only one represented.
Zuckerman quickly moves away from acknowledging what people tell him in
conversations, and yields the ground to his imagined experience of the Swede’s
life. While much of the text represents the Swede’s thoughts and experiences,
mostly from the third-person perspective, constructions of represented thought
are rare and fragments like (3) are used instead. They are a signal of viewpoint
compression, prompted by the constructions typical of direct speech, and
rendering the raw experience of the character. What seems to dominate in
such cases is the ‘directness,’ miming first-hand access to the thoughts, rather
than marking any narratorial presence. But the possibility of such a choice of
expression is based in the mechanism of viewpoint compression, not STR.
A similar degree of compression, across different spaces, can be found in (4).
Having shifted the viewpoint into the space of the Swede’s mind, Zuckerman
recreates the man’s memories of his daughter’s childhood. They are brought
in, as in (3), to underscore the desperate state the family is in after Merry’s
disappearance. The context of (4) is the Swede’s reflection on Rita, who presents
herself as Merry’s terrorist friend and torments the Swede with veiled threats and
false promises. Thinking of Merry’s friends, he remembers Merry and her child-
hood friend Patti, and the space is elaborated through questions of the kind Merry
could have asked at the time. Again, there is a pretense of direct speech, while on
a higher narrative level these are the Swede’s thoughts, imagined by Nathan.

(4) Mom, can Patti stay for dinner? Mom, can Patti wear my boots? Mom, can you drive
me and Patti to the village? (American Pastoral, p. 321)

Later in the fragment, thoughts about Rita are set against the memory of Merry’s
childhood.
176 Speech and thought in the narrative

Narrator: Zuckerman
SV-space
present

MN-space: The Levovs

Roles: Mother, Father, Daughter, Daughter’s Friend

NSI – Merry’s childhood; past NS1 – the Rimrock Bomber; present

Daughter – Merry / child Daughter – Merry / terrorist
Daughter’s Friend – Patti Daughter’s Friend – Rita
Childhood activities Terrorist activities

DS 1 / 2 / 3/... Imagined DS

Speaker: Merry Speaker: Merry

Asking permission to do things with Patti Asking permission to do things with Rita

Figure 7.1 Discourse blend (examples 4 and 5)

(5) Can Rita stay for dinner? Can Rita wear my boots? Can you drive me and Rita to the
village? . . . Can Rita and I blow up the post office? (American Pastoral, p. 321)

The blend in (5) builds on the generic role of ‘Merry’s friend’ (see Figure 7.1).
The ‘Patti’ input in (4) belongs to the past of the story, and explicitly evokes an
addressee (Mom). The ‘Rita’ sequence of questions builds on the frame estab-
lished in (4) and substitutes a young terrorist in the ‘friend’ role, which changes
the nature of what the girls would now want to do. The blend belongs to the
present of the story, and is crucially built on disanalogy – the innocent child-
hood activities are supplanted by dangerous acts of terrorism, and yet, in the
Swede’s mind, Merry is still his beloved child, seeking the parent’s approval of
her friend. The lack of form of address in (5) suggests that the Swede may be
imagining himself as an addressee of these requests, which underlines his
dramatic situation of a father who has to accept the daughter’s friend if he
wants to continue accepting the daughter.
The ‘speech’ pattern in (5) is thus not identifiable as speech at any level. It is,
instead, a representation of a mental state, that of having to deal with irrecon-
cilable frames. Crucially, Roth’s narrative choices completely blend actual facts
with imagined experience, and minimize the role of the narrator. In American
Pastoral, Zuckerman is not just a narrator, he is more like a medium who gives
voice to the Swede’s story. Viewpoint compression is the primary narrative
technique here.
What is also interesting about such examples is that the fragments which
mimic conversation may not only be imagined, but also generic – it is hard to
say whether a sequence such as (4) was ever uttered and whether it is this
particular occasion that is remembered. It could also be intended as frame
 7.2 Speaking for thinking 177

metonymic – so that it evokes the frame of ‘childhood friend activities,’ which

the discourse in (4) is representative of. Similar examples of quotations used as
‘selective depictions’ have been discussed in Clark and Gerrig (1990) to refer
to instances where a particular stretch of discourse is selected to play the role of
a depiction, not a verbatim quotation. What one might add to this observation
is the point that such ‘depictions’ cannot play their role unless the selected
quote has a frame-metonymic function, so that it evokes a frame in which the
particular expression played a role, to use that frame in the construal of the
situation.3
There seem to be numerous examples of direct speech expressions used
metonymically, for example in Eggers’s novel. Example (6) comes from a
fragment when Dave agonizes over the impression strangers have of Toph’s
happiness. He is frustrated by a school photo, in which, in his view, Toph looks
sad, and this makes him remember ‘the Phone Voice Problem’ – Toph’s habit of
answering the phone in a weak and insecure voice.

(6) And of course people wonder. What’s wrong with Toph? they ask me when he hands
me the phone. And always I have to be cavalier. That’s just Toph! ha, ha. (AHWOSG,
p. 322)

People can be expected to wonder about the happiness of a small boy being
raised by a rather colorful character of a brother, but we are aware that the
question reported in (6), as well as the answer, do not have to be quotes of
speech fragments uttered on any specific occasion. Being generic, they evoke a
frame wherein the ‘Phone Voice’ is indeed a ‘Problem,’ even though it may
appear problematic only from the point of view of Dave’s exaggerated fears.
As (7) shows, the discourse then continues into another generic pattern, of
Dave’s attempts to change Toph’s habits. At the same time, the description of
what the voice ‘sounds like’ is vivid and detailed and soon develops into what is
clearly Dave’s childhood memory of being frightened by an overpowering parent:

(7) So I’ve been imploring him to sound normal. Please sound normal, Toph, you are
normal, we are normal so just sound normal please can’t you? Don’t sound like I’ve
been beating you, like you’re in the bathroom hiding from me, because I’ve been there,
have hidden from parents before . . . and I have hidden there, and have seen, darkening
the white slit of light under the door to this closet, his shoes. (AHWOSG, p. 322/3)

It is not quite relevant to ask if this is speech or thought. It starts as a

metonymic speech pattern, but ends up being Dave’s personal reflection on
his own childhood, one which he probably never included in his actual
conversations with Toph. There is a shared viewpoint, of Dave fearing being
thought abusive, and building the fear on his own childhood experiences. That
viewpoint space participates well in the overall narrative, and thus the status of
specific mock-utterances remains unresolved.
178 Speech and thought in the narrative

The problem of the school photo which triggered the fragment on the Phone
Voice Problem continues when Dave looks at the picture set again:
(8) All these pleading Tophs, Eleven Tophs saying Look at my sad life, you people, you
viewers of junior high pictures! . . . Save me from him, because every night before dinner
he’s asleep on the couch and so dead to the world, and when he can’t get up he tugs on
my shirt and begs, he makes me cook for us and then later . . . (AHWOSG, p. 325)

What the pictures are ‘saying’ is presumably a fair description of Dave’s routine
behavior, not always in keeping with a ‘perfect father’ figure. The viewpoint
compression here is quite complex: Dave’s guilty awareness of his behavior is
the highest viewpoint here, but is attributed to Toph instead. Dave would not
want to accuse himself of bad habits, as this would imply an attempt to improve,
but he attributes the accusatory thoughts to Toph, while also worrying that
other people will become aware of them too, and blame Dave. In the end, the
confusion of viewpoints here means that Dave wants Toph to look happy, so that
people don’t assume he is unhappy, so that in turn Dave can feel good in his
parental/brotherly role. The fact that Toph is communicating unhappiness
through the picture is an interesting construction – viewers may interpret the
represented reality of the picture in this way, so the picture is ‘saying’ that it is
so. In this representation blend, the viewpoint of the potential-viewer-cum-
addressee is the one that Dave constructs, and then tries to respond to. But all
these scenarios are playing out in Dave’s mind only, and are properly a part of
his thoughts.
As these examples suggest, networks of spaces in narratives may be structured
by present and past events, memories, imagined scenarios, frame-metonymic
prompts, representation blends, roles, et cetera. All these spaces are subject to
viewpoint compression, so that we can see the relevance of any specific space
set-up in the text to the story as a whole. But whether these spaces are linguis-
tically structured as speech or not is not correlated with their place in the network
and their viewpoint role.

7.3 Levels of embedding in thought representation

The examples above signal the fact that a network of narrative spaces may
involve a number of levels of embedding, rather than a two-level structure
profiling the narrator and the character. The depth of embeddings can then
affect the way in which various forms of discourse will be interpreted. At the
same time, thought-spaces may be set up not only from the point of view of
the narrator (when they are reported as actually occurring in the character’s
mind), but also from the viewpoint of other characters – so that we learn what
characters think about each other’s thoughts. One such example is presented
in (9), which comes from Philip Roth’s novel The Human Stain. The main
 7.3 Levels of embedding in thought representation 179

characters there are Coleman, a retired college professor, and Faunia, his lover, a
much younger woman with a very tragic and difficult past. Given how much
she has been through, she can be expected to ruminate on her fate, and Coleman
attributes just such thoughts to her:
(9) He thinks she is thinking about how long it has all gone on, the mother, the stepfather,
the escape from the stepfather, the places in the South, the places in the North . . . and
maybe she is. Maybe she is even if, alone now on the grass while the boys are
smoking and cleaning up from lunch, she thinks she is thinking about crows. She
thinks about crows a lot of the time . . . and today they are cawing all over campus,
and so instead of thinking of what she is thinking the way Coleman thinks she is
thinking it, she is thinking about the crow that used to hang around the store in Seeley
Falls. (The Human Stain, p. 164)

What the fragment makes clear is that thought does not have to verbalize clearly
the actual issues considered; it may meander through different topics, while
at the same time, deep down, it always revolves around the problem. How,
then, are such thoughts to be reported? Roth’s narrative choices, again relayed
through Zuckerman, are similar to those in American Pastoral, where the
lowest-level Ego-viewpoint is compressed up to the narrative level. In (10),
the train of Faunia’s thoughts is recreated:
(10) That crow’s voice. She remembers it at all hours, day or night, awake, sleeping, or
insomniac. Had a strange voice. Not like the voice of other crows probably because
it hadn’t been raised with other crows. Right after the fire, I used to go and visit that
crow at the Audubon Society, and whenever the visit was over and I would turn to
leave, it would call me back with this voice. (The Human Stain, p. 165)

The narrative gradually goes deeper into Faunia’s thoughts. Starting with the
present-tense third-person narration of the nature of Faunia’s thoughts (She
remembers it), which follows the pattern of the rest of the narrative, through free
indirect discourse past-tense report of the content of her thoughts (Had a
strange voice.), to the sudden viewpoint shift when the text takes a first-person
past-tense form, representing Faunia’s thoughts without the mediation of the
narrator. Not only does this narrative choice open a lengthy internal monologue,
but it also affects the entire structure of the novel, gradually compressing out
all the levels set up earlier, and separating Faunia’s thoughts from the SV-space.
It is thus in fact a shift down into the consciousness of one of the characters.
Faunia’s thoughts revolve around the nature of crows, which she admires
and seems to understand. She feels so close to the crow she once went to visit
that she feels fully identified with it:
(11) (. . .) why can’t I be a crow locked up in this body? Yeah, and where is the doctor
who is going to do what they do to let me out? Where do I go to get the surgery that
will let me be what I am? . . .
I am a crow. I know it. I know it! (The Human Stain, p. 169)
180 Speech and thought in the narrative

SV-space Narrator
‘Nathan Zuckerman’

he Coleman thinks

Faunia thinks
she (about her Faunia thinks
unhappy life) she (about crows)

Faunia thinks
about her life by I I’m a crow
blending it with the
life of a crow

Figure 7.2 The ‘crow’ blend (examples 9, 10, and 11)

Overall, the fragment sets up a blend of narrative spaces which not only gives
the reader access to Faunia’s thoughts but also elaborates on the mutual relation-
ship between the main characters (see Figure 7.2). Two ‘thought’ spaces have
been set up for Faunia – one in which Coleman imagines her depressed about
her past, and the other wherein she is indeed thinking about it, but in a different
way (instead of thinking of what she is thinking the way Coleman thinks she
is thinking it, she is thinking about the crow). She rethinks her life through her
bond with crows and finally imagines herself being one of them, even if locked
into a human body. Without acknowledging it, the narrator is constructing
a blend in which different concepts in the minds of two different characters
can yield a similar meaning, even though Coleman clearly underestimates
the complexity and depth of Faunia’s trauma. It is achieved through the addi-
tional level of mental space structure – he ‘thinks’ she is thinking, while she ‘is
thinking.’ This narrative choice manipulates viewpoint in interesting ways,
and also allows characters’ thoughts to be expressed and work jointly in the
narrative.
Roth’s narrative technique is worth a pause. In the two novels mentioned
here, he starts with an on-stage narrator, the writer Zuckerman, whom his
faithful readers know from earlier novels. As the narrator hears more from or
about the characters, the narration becomes gradually off-stage, as the story
being constructed comes into focus. In both cases, the texts go into lengthy
internal monologues, until there is a complete shift of focus to the character’s
Ego-viewpoint. But a careful reader remembers that all of the text is embedded
under an SV-space in which Roth’s alter ego, Zuckerman, is in fact weaving it
together from the facts structuring the SV-space (e.g. his role of a high-school
 7.3 Levels of embedding in thought representation 181

friend of the main character), the stories heard, and finally those he constructs
himself. It may not be Roth’s intention, but the successive descent into the
character’s mind seems to be a perfect description of how fiction might emerge.
The narrative role of direct speech can be somewhat complicated in the
context of the network even when the expressions represent what has been
said. Example (12), from Anne Tyler’s novel A Patchwork Planet, describes
a scene where the main character and first-person narrator, Barnaby, comes
home to find a message from his parents:
(12) My answering machine was blinking . . . I leaned over and pressed the button.
“Barnaby,” my mother said, “this is your mom and dad.”
What a thrill.
“We just wanted to say Happy New Year, sweetie . . . Call us sometime, why
don’t you? Bye.”
Click.
I flopped back on my bed and looked up at the ceiling. (A Patchwork Planet,
p. 43)

All the direct speech in this example is the message, and so represents what
Barnaby hears in the MN-space, and only secondarily what his mother has
said earlier, when she called. The viewpoint overall is thus that of Barnaby as
the addressee of the message, which is further reinforced with his inner com-
ments (What a thrill.) and further things he hears (Click.) Thus, even though it
is a genuine example of direct speech, its embedding in the temporally later
space of Barnaby’s return home contextualizes it differently; for one, if Barnaby
had been home when his mother called, he would not have such a natural
opportunity to comment sarcastically on his parents’ calling. In the fragment,
all the viewpoints profiled in the phone call are embedded in the space of
Barnaby listening.
Space embedding, even when there are clearly two different sources of what
is being said, does not have to take a marked constructional form. In the movie
The Hunt for Red October the main character, Jack Ryan, talks to the National
Security Advisor, Jeffrey Pelt, asking for his help. Pelt answers:
(13) I am a politician, which means that I’m a cheat and a liar, and when I’m not
kissing babies I’m stealing their lollipops. But it also means that I keep my options
open.

It is clear that not all that Pelt says represents his viewpoint, and that he is
evoking another discourse space wherein speakers other than himself describe
politicians in negative terms. In spite of the fact that the ‘cheat and liar’ part of
the utterance is naturally seen as echoic of some contextually available belief,
this description would not suffice for two reasons: one, there is no indication of
such an utterance being available in the context of the movie, and two, an echoic
utterance is naturally described as embedded in the current discourse, even if
182 Speech and thought in the narrative

there are no constructional signals present. Thus, in spite of the fact that (13)
represents an entry in a conversation, the example blends two viewpoints.
Throughout the utterance, Pelt is referring to himself as I, without adding any
hedges such as people think (I’m a cheat and a liar), but it is clear that other
people’s thought-space is being evoked. The use of the first-person pronoun is
thus a result of the compression of the ‘he must be a cheat and a liar’ viewpoint
into the higher discourse space of the conversation, in which Pelt is the speaker,
and thus refers to himself as I.
The fact that (13) comes from a fictional dialogue does not invalidate the
point. In colloquial discourse speakers often say similar things. When one
reveals a secret but wants to remain anonymous, it is natural to say something
like You never heard it from me, which puts the official ‘line’ into the mouth of
the addressee, thus preventing him from identifying the source. Such cases also
rely on blended discourse, as the full utterance would probably include an
embedding phrase such as You mustn’t tell anyone (that you heard it from me).
Embedding of discourse without profiling a complete STR construction is in
fact more common than is generally thought. Recent work by Pascual (2002,
2006a, 2006b, 2008) describes a class of discourse blends which she refers to as
‘fictive interaction.’ The description applies to many formal constructions in
which a discourse fragment that functions as a token of a discourse situation is
incorporated into another utterance. Some of Pascual’s (2006a) examples are
given in (14):

(14) a. You need to go in with the attitude that yes, I can do this . . . (allnurses.com
forum)
b. You’ll learn the winning attitude of YES, I CAN DO IT! (personal training
website)
c. Develop a “Yes, I can do it” attitude. (“Twice – Exceptional Students”
Newsletter)

What is crucial to these constructions is that a discourse fragment begins to

play a meaningful role of its own, and can contribute to another utterance by
occupying various syntactic positions (a clause in 14a, an NP in 14b and an
adjective in 14c). What fictive interaction shares with examples like (13) is the
possibility of adding a second viewpoint to the viewpoint of the broader utter-
ance, but what is different is the pronominal choice. In (13), the compression of
the embedded space into the embedding one requires that the referential features
and the ground of the higher space are preserved. In all the instances in (14), the
pronoun I is preserved, while the utterance as a whole uses second person you
(also implied in the imperative in [14c]). Switching to you in any of the ‘Yes, I
can do it’ expressions would cause confusion or unacceptability, which sug-
gests that the ‘token’ nature of the phrase requires that any person assuming the
attitude will blend her/himself with the I, because the I is in fact not identifying
 7.4 Viewpoint compression and constructional phenomena 183

any speaker in any specific discourse ground. Example (13) is crucially different
in this respect, as the speaker blends his identifiable critics’ views into his own
discourse ground.
The examples throughout this section suggest that there are various possible
embedding configurations in representations of thought, and that they may yield
a variety of different constructions. While these constructions may not be
well established as stable grammatical patterns, they all rely on viewpoint
blends. They sometimes suppress the overt expression of embedding, but they
also maintain a constant connection to the SV-space of the narrative. Viewpoint
compression thus offers an explanation of a range of narrative choices intended
to negotiate different viewpoints.

7.4 Viewpoint compression and constructional compositionality

The preceding section looked primarily at the expressions which represented
thoughts without profiling any STR constructions. In most cases, the actual
linguistic form was modeled after direct speech, even though the actual inter-
pretation involved numerous levels of embedding. In what follows I will look
at selected constructional phenomena related to STR. However, the range of
constructions specific to STR is defined differently by different analysts, and
thus I will not discuss such proposals here or attempt to view each from the
perspective of my framework. Instead, I will use some very standard examples
of STR to clarify further the mechanism of viewpoint compression, as it
participates in textual representation of speech and thought.
Example (15) comes from Anne Tyler’s novel Ladder of Years (see Figure 7.3).
The main character, Delia, is being interviewed by Mr. Pomfret, her future employer:
(15) Mr. Pomfret didn’t mention references. His sole concern was the nature of her past
duties. Had she typed, had she filed, taken shorthand? . . . He said she would start
tomorrow; her hours were nine to five. Sorry the pay was just minimum wage, he
said . . . Also she was expected to brew the coffee; he hoped that wasn’t a problem.
Of course it wasn’t, Delia said brusquely, and she rose and terminated the interview.
(Ladder of Years, p. 95)

The fragment represents some salient forms of represented speech: indirect

speech (He said she would start tomorrow) and free indirect speech (Had she
typed, had she filed, taken shorthand?; her hours were nine to five.). It also
features some less typical solutions, wherein free indirect utterances are accom-
panied by an additional he said (Sorry the pay was just minimum wage, he said).
All these forms are marked with some standard shifts in grammatical forms,
resulting from the shift away from the represented speaker’s and hearer’s deictic
ground and an alignment with that of the narrator. Consequently, the pronoun
referring to Mr. Pomfret as the speaker is shifted into he, while the you with
184 Speech and thought in the narrative

SV-space Narrator: 0
present

MN-space

Character: Delia (she) (he)

past (she)

Discourse space 1

Speaker: Mr. Pomfret (‘I’)

Addressee: Delia (you)

Figure 7.3 Third-person report (example 15)

which Mr. Pomfret addresses Delia changes into she; also tense forms are
shifted from present to past. These changes (even if not always equally con-
sistent) have been described in the relevant literature.4
Within the framework advocated here, the same facts can be accounted for
naturally. The viewpoint of the SV-space in the case of (15) is the present, with
an off-stage narrator, which accounts for all the characters in the narrative being
referred to via third-person pronouns. Two of the characters in the past MN-
space are Mr. Pomfret (he) and Delia (she); in (15) they are in conversation, and
thus in a separate discourse space within the MN-space. In a discourse space,
one expects its specific ground to be profiled, and thus the participants would
refer to themselves and each other according to the rules: I for the speaker and
you for the hearer. But there are two options available from here onwards. Either
the discourse space is a direct speech space, and then does not need to be
incorporated into the network other than through stable cross-mappings main-
taining the identity of characters, places, et cetera, or it is to be linked with the
higher space (via embedding, reporting, STR, et cetera). If the discourse space is
integrated into higher narrative spaces, it needs to be coordinated with them
with respect to viewpoint established by the SV-space. Thus if direct speech
(i.e. independence of the discourse space) is maintained, the I/you forms remain
restricted to the discourse ground of the conversation, and cannot be projected
upwards. But if the discourse space becomes a part of the blend, it has to start
using the basic deictics (pronouns and tense) as they are used throughout the
network of the MN-space (he/she). The discourse remains a part of the MN-
space in either case, but its degree of incorporation is different: from establish-
ing cross-space mappings, to some degree of blending.
The topology of the MN-space in (15) is embedded in the SV-space with no
on-stage narrator. It is possible to argue that a zero-narrator (off-stage narrator),
 7.4 Viewpoint compression and constructional phenomena 185

such as the one in Ladder of Years, cannot be cross-mapped to any of the

characters and never requires the use of the form I. In order that the consistency
of reference is preserved throughout the network, the unspoken I remains
reserved for the position of the narrator, but there is no character in the lower
space whose viewpoint could be compressed up to the SV-space, thus justifying
the use of I in the MN-space. Discourse spaces in which direct speech is used
are different, since they rely on their own ground, even if the characters from
the discourse space are cross-mapped with the characters in the MN-space.
Thus, even though Delia and Mr. Pomfret are profiled as characters in both
the MN-space and the discourse space, and these spaces are cross-linked, they
would use I/you only if their conversation were represented as direct speech, in
an unincorporated discourse space. In (15), the discourse space is fully inte-
grated into the MN-space, and thus the forms shift to he/she.
Now, what about the form of STR known as indirect speech? Example (15)
has one instance of it, He said she would start tomorrow. How is it different
from the examples of free indirect speech in the fragment, such as Had she
typed, had she filed, taken shorthand? and from the free indirect/indirect hybrid
Sorry the pay was just minimum wage, he said? It seems that the he said
expression in both cases participates in the MN-space, not the discourse
space, since it belongs to the descriptions of actions performed rather than
representing discourse of a character. The clause representing the words of a
character, whether fully incorporated syntactically (she would start tomorrow)
or not thus participates in the discourse space/MN-space blend. Both cases
preserve much of the discourse style of the discourse space: there is the temporal
deictic tomorrow, which is future with respect to the discourse space ground,
not the MN-ground, and the emotive marker Sorry. Such local gaps in full
viewpoint coordination signal the degree to which the blend is integrated, but
do not change the nature of the process.
To sum up, the MN-space has discourse spaces embedded in it (alongside or
as a part of other narrative spaces). These discourse spaces are embedded in the
MN-space but may be marked with varying degrees of integration. They may
not be integrated at all, and then the form of choice is direct speech, with the
I/you pronominal choice. They may be embedded via appropriate MN-space
expressions (he said) and thus integrated to a high degree. Or there may be a
more loose connection between the spaces, such that the pronouns and tense are
typically integrated, but there is no MN-expression in which the reported
sentences are directly embedded, and the expressive flavor of the discourse
space is preserved.
Theoretically, nothing prevents a given discourse space from being incorpo-
rated into the MN-space in any one of the three ways – example (15) uses only
indirect and free indirect speech, but if the conversation were longer, one could
imagine different parts reported differently. Contrary to what much of the
186 Speech and thought in the narrative

discussion of STR suggests, discourse spaces are embedded in the MN-space

as spaces, not as individual sentences. It is not surprising that the choice of the
STR construction may vary from sentence to sentence, because a construction-
ally consistent report (say, indirect speech only) would be stylistically unac-
ceptable. But there needs to be minimal consistency as well – any form other
than direct speech needs to be blended to some degree into the MN-space; at a
minimum, the temporal topology and reference need to be coordinated. This
does not prevent shifts (as in the case of [10] above), but it precludes random
changes from one sentence to another.
The story network functions differently when the story’s SV-space profiles an
on-stage first-person narrator. Such a narrator is often cross-linked to a story
participant who describes the events he or she participates in. Discourse spaces
within the MN-space do not necessarily involve the narrator/character, and
these behave roughly as in the off-stage narrator set-up. But when the narra-
tor/character (N/CH) does participate in the discourse space, then in the instan-
ces where N/CH is the speaker the pronoun I is used. However, the I is then
aligned not just with the discourse space in question, but also with all the
narrative spaces in which N/CH is a participant. In (16), for example, Dave
(in Eggers’s novel again) reports talking to his brother about his day at school
and then quotes what he says to the parents of another child at school. He is
consistently using I to refer to himself throughout:

(16) “What happened today?” I ask.

“Today Matthew told me that he hopes that you and Beth are in a plane and
that the plane crashes and that you both die just like Mom and Dad.”
“They didn’t die in a plane crash.”
“That’s what I said.”
Sometimes I call the parents of Toph’s classmates.
“Yeah, that’s what he said,” I say.
“It’s hard enough, you know,” I say.
“No, he’s okay,” I continue, pouring it on this incompetent moron who has
raised a twisted boy. “I just don’t know why Matthew would say that. I mean, why
do you suppose your son wants Beth and me to die in a plane crash?” (AHWOSG,
p. 89)

There are two (or more) discourse spaces here: one is the conversation between
Dave and his brother, the other is the generic discourse space of conversations
with parents in which he tries to make them feel guilty about their children.
Whether Dave is the speaker in one space or the other, or whether he narrates
(I ask, I say, I call, I continue), the pronoun I consistently links all his narrative
space roles across the contexts. There is a cross-space mapping linking the
narrator, the character in the MN-space, and the speaker in the discourse space,
so that, via Access Principle,5 any one of these instances can use the pronoun I.
Interestingly, this makes for a much more integrated narrative, since the
 7.4 Viewpoint compression and constructional phenomena 187

presence of the first-person voice throughout connects the emerging blend very
tightly via a shared viewpoint.
Example (17) shows another instance, this time of Dave’s internal mono-
logue, scattered as usual, but largely consisting of questions he is considering.
(17) There is so much to do . . . I have to get a résumé together, and we have to find a new
place to live when the sublet ends, and how will Toph get to school if I get an early
job? Will Beth pull her weight, will she be too busy, will we kill each other? . . .
How much should I/ can I/ will I burden Kristen? . . . Should I lighten my hair?
(AHWOSG, p. 66)

What is central to the narrative space set-up here is that Dave is considering
questions about his own possible future, so the pronoun I in each of the
questions (such as Should I lighten my hair?) means that he is the person asking
the question and also its addressee; this blended speaker/addressee viewpoint is
further compressed up to his role as a narrator, making the narration very close
to the core of the main character’s mind.
Crucially, this narrative space set-up is different from an STR form recently
identified and described by Vandelanotte (2004a, 2004b, 2009) under the name
of DIST (Distancing Indirect Speech and Thought). Vandelanotte’s description
of DIST compares it to FIST (Free Indirect Speech and Thought) and notes
several differences. One directly relevant to my discussion here is the issue of
the meaning of the pronoun I. FIST would involve uses like the one in (17), but
DIST allows the first-person pronoun to identify the addressee, not the speaker
in the discourse space being reported, but that addressee also has to be what
Vandelanotte refers to as the ‘current speaker’ – which in the context of first-
person narration means the narrator. The DIST use of I is well illustrated in (18),
which is a fragment from Henry Miller’s non-fiction account of his visit to
Crete, The Colossus of Maroussi. He talks there to a Greek intellectual whom he
calls Mr. Tsoutsou, since he has forgotten his real name:
(18) He started off on Aragon – had I read Le Paysan de Paris? Did I remember the
Passage Jouffroy in Paris? What did I think of St. Jean Perse? Or Nadja of Breton?
Had I been to Knossus yet? I ought to stay a few weeks at least – he would take me
over the island from one end to another. He was a very hale and hearty fellow and
when he understood that I liked to eat and drink he beamed most approvingly. (The
Colossus of Maroussi, pp. 115–116)

Miller, as the first-person narrator, is reporting a conversation in which

Mr. Tsoutsou asks a lot of questions. All the reported questions use the pronoun
I to refer to Miller as the addressee (you in the discourse space), while also
marking him as the N/CH of the narrative. In this case, then, the pronoun choice
is aligned entirely with the SV-space and the MN-space, but not with the
discourse space – this seems to be a narrative space rendering of the nature of
DIST. The same applies to the sentence I ought to stay a few weeks at least, as it
188 Speech and thought in the narrative

SV-space Narrator: ‘I’

present

MN-space

Narrator/character ‘I’
past

Discourse space 1
Speaker (‘I’)

Addressee/narrator
(‘you’)

Figure 7.4 First-person report (example 18)

clearly renders the Greek’s suggestion, not Miller’s own determination. But in the
final sentence, he understood that I liked to eat and drink (indirect speech), the
text uses I in a more common reported sense, as referring to Miller’s own
appreciation of his enjoyment of earthly pleasures, albeit acknowledged by Mr.
Tsoutsou (see Figure 7.4).
We can now compare the role of pronouns in the use of questions in the three
examples above. Had she typed? is an example of free indirect discourse, or
FIST, and the pronoun she substitutes you (as a marker of the addressee) while
aligning it with the MN-space framed by an off-stage narrator. Should I lighten
my hair? is also an example of FIST, and the I connects to the first-person on-
stage narrator-cum-character, as a speaker and as an addressee. Finally, Had I
been to Knossus yet? is an example of DIST, where I links to the roles of the
addressee and narrator, but not the actual speaker. Each of the pronouns is thus a
prompt for a complex blend, linking the viewpoints of the relevant narrative
spaces, and, perhaps primarily, situating the utterance with respect to the view-
point of the SV-space. It seems an oversimplification to treat such grammatical
choices as sentential only, as they appear to be the means narrative discourse
relies on in achieving story-wide coherence and in prompting the processes of
narrative space integration. To put it simply, such pronominal choices partic-
ipate in overall processes of story construction.
It is thus not the case that all of the aspects of pronoun interpretation in the
narrative are a part of the meaning of pronouns. On the contrary, I argue that
a pronoun like I is indeed no more than a pointer to the current speaker (also in
the textual sense of the current narrator), seen against the available ground. The
complexities of reference emerging in the examples throughout this chapter
result from the narrative space network, not the pronoun itself, but there are
specific regularities in the network. For example, an off-stage narrator in an
 7.4 Viewpoint compression and constructional phenomena 189

SV-space cannot be referred to as I, but once the narrator comes on-stage (in
intrusive narration or first-person narration) the first-person pronoun can be
used. Moreover, if there is a character cross-mapped with the narrator, that
character will also be an I in various narrative spaces, as long as there is a cross-
space connection coming through a network of blended spaces.
The coordination of grammatical (and especially deictic) forms in the narra-
tive thus happens via three processes, but the same processes also explain the
emergence of the story. First, lower-level spaces are embedded in the higher-level
spaces. This applies to forms such as indirect discourse, where the embeddings
are at the sentential level, and thus coordinate pronouns and tense to the required
degree. But it also applies to larger spaces, such as flashbacks, spaces prompted
by different temporal and spatial topology, subplots, et cetera. Grammatical
coordination may be less strict then, but some is required: for example the
embedded novel in The Blind Assassin is coordinated with the rest of the novel
by strict avoidance of proper names and complete reliance on pronouns, but the
narrative tense is present, even though we learn early on that the space elaborates
a sub-story which resides in the past of the narration.
Second, there is blending. Clusters of spaces are gradually integrated into
larger chunks of the story, until the complete story emerges (this does not
preclude lower-level, sentential blends such as FIST). This process is coordi-
nated with respect to the SV-space, whose constant viewpoint allows temporal
progression to emerge and also maintains identity cross-links (so that a charac-
ter referred to as Alex in one of the spaces of The Blind Assassin is finally cross-
mapped with he in the novel and the ‘blind assassin’ role in the pulp fiction
novel, of which he is also the author). The process of reading constantly
increases the number of cross-space connectors, and thus the process of the
emergence of the story (through projection and integration) progresses.
Finally, there is the mechanism of viewpoint compression. I have argued
throughout the book that compression of viewpoint is what allows spaces to be
integrated by their shared viewpoint. Even though their viewpoints may be
different (let’s say, because of a different temporal ground or different narrator),
one of the spaces may be selected so that its viewpoint is the main narrative
viewpoint for a part of the novel or the whole (see the relevant discussion
in chapters 2 and 4). Also, as I argued elsewhere (Dancygier 2005a), additional
spaces may be set up when a space is decompressed into more spaces to mark
different viewpoints within it, and then the viewpoint of one of the decom-
pressed spaces is compressed with the main narrative viewpoint, substituting
it for as long as the space remains active. Similar phenomena can be observed
when decompression is not involved. For example, in example (15) above, the
part of the main narrative space being elaborated is the interview. The report
includes primarily what Mr. Pomfret said, and Delia’s concluding words, but
none of this is done through direct speech. The interview space thus contains
190 Speech and thought in the narrative

another space – the report of Mr. Pomfret’s words. The report is first introduced
through the narrator’s summary of how Mr. Pomfret conducted the interview,
then sets up a number of sentential-level STR spaces, detailing his words. But
when a sentence starts with He said she would start tomorrow, it explicitly sets
up the space of ‘Mr. Pomfret speaking,’ where the viewpoint is clearly that of
the narrator (SV-space). The ensuing content of his utterance is embedded in
the ‘speaking space.’ The next sentence, her hours were nine to five, is a blend
of the reporting space and the reported space, but the viewpoint is that of the
lower space, so the words of the character become the narrative viewpoint for
the duration. Viewpoint compression thus foregrounds the words of the char-
acter while not abandoning the narrative viewpoint of the fragment. Viewpoint
compression here is the mechanism which contributes to the integration of the
FIST blend, and to the coherence of the story as a whole.
Another issue is the distinction between two cases: embedding and blending.
Mental space embedding, such as indirect discourse, involves some phenomena
such as pronoun or tense shift. But this is primarily the result of a phenomenon
widely talked about in mental space theory, that of projection. A higher,
embedding space (such as He said) projects its tense and reference pattern
into the lower, embedded spaces (she would start tomorrow, rather than you will
start tomorrow), very much like in conditional constructions, where the if-space
(If I were you) projects its stance, marked by past tense, and its referential
pattern, into the lower, main-clause space (I would marry him). Blended spaces,
for comparison, do not necessarily maintain the hierarchical organization or the
syntactic dependence, while also allowing the tense and reference to be adjusted
to the participating spaces. As a result, a free indirect utterance such as Had she
typed? marks many of the same shifts, but does not maintain the embedding
structure. However, as I suggested above, the viewpoint of the higher spaces
still dictates the grammatical choices. The fact that these embedding and
blending processes occur is the reason why lengthy narratives can be read as
coherent. Contemporary narratives break some of these rules and experiment
with much more fragmented structure, but they can be appreciated for what they
are precisely because we have a tacit understanding of how viewpoint com-
pression smoothes out narrative space boundaries.
Discussions of STR often consider reasons why adverbs of time are some-
times excluded from the viewpoint adjustment, as in He said she would start
tomorrow above, where tomorrow is future with respect to the time of the
conversation and not the time of the report. This is in fact similar to what has
been labeled the ‘was-now’ construction, recently put in the construction-
grammar perspective in Nikiforidou (2010). I argued earlier that such not-
fully-integrated uses of adverbs should be treated as signals of the maintenance
of their spaces in the network. My interpretation is slightly different from
Nikiforidou’s, in that I do not assume that this is a constructional pattern
 7.4 Viewpoint compression and constructional phenomena 191

resulting from coercion. Nikiforidou discusses the resulting blend of spaces

elsewhere, but my point is that this is much less a signal of a blend than a signal
of incomplete integration which serves a specific purpose. Tomorrow is not only
the day after the conversation between Mr. Pomfret and Delia, it is also the next
day from the point of view of the story as a whole. Deictically, tomorrow is
situated identically with respect to the interview space and the SV-space, which
has a narrator in the present and tells the story in the past.
The ‘was-now’ construction is found in the narration proper as well as in the
STR constructions. Since the former case seems more interesting from the point
of view of the framework being developed here, I will consider some narration
examples below. Example (19) repeats example (6) from Chapter 3; it is a
fragment from Fielding’s Tom Jones; example (20) is another classic – the first
paragraphs of James Joyce’s “Eveline”:

(19) He now lived, for the most part, retired in the country, with one sister, for whom he
had a very tender affection. This lady was now somewhat past the age of thirty.
(Tom Jones, p. 17)

Example (19) comes after two initial paragraphs in which Squire Allworthy is
introduced, along with his story, and thus the now functions as a kind of
resumptive pronoun, announcing that the excursion into the pre-story past has
ended and the in-story past begins. It is thus a means to organize the readers’
understanding of where along the axis of time narrative spaces belong. The
resulting construal highlights the current temporal viewpoint of the narrative,
setting it against a broader temporal space. Interestingly, while being stylisti-
cally salient, the now in the fragment is not indispensable to the processing of
the narrative.
The function of now is more explicitly resumptive in (20). The story starts
with the image of Eveline sitting at the window, looking outside. The next part
describes her childhood memories, prompted by the current situation – in this
part many past adverbs are used, so that the elaboration of the narrative space of
Eveline’s childhood is maintained. However, the last sentence of the paragraph
starts with now and then frames the imminent events which have prompted the
memories – Eveline’s planned departure for America.

(20) She sat at the window watching the evening invade the avenue. Her head was
leaned against the window curtains and in her nostrils was the odour of dusty
cretonne. She was tired . . .
. . . One time there used to be a field there . . . Still they seemed to have been rather
happy then. Her father was not so bad then; and besides, her mother was alive. That
was a long time ago; she and her brothers and sisters were all grown up; her mother
was dead. Tizzie Dunn was dead, too, and the Waters had gone back to England.
Everything changes. Now she was going to go away like the others, to leave her
home. (Dubliners, p. 25)
192 Speech and thought in the narrative

Being used at the beginning of the sentence, now is here also a discourse marker
announcing a sudden shift of viewpoint. Both (19) and (20) show that the
adverb is not simply an unexpected reference to the present in the middle of a
past story, to be blended into the past background, but that it is a specialized
viewpoint marker, signaling the return to the narrative space which is being
currently elaborated. So if it is a temporal marker at all, it picks out a specific
narrative space as the current ground of the narration. It is also worth noting
that in (20), unlike (19), now is a necessary addition to the text, since without it
the paragraph-ending sentence would be difficult to process. Clearly, there is a
shift of viewpoint from one space to another happening here, and the sentence-
initial use of now marks that shift.
Example (20) supports the claim that narratives operate through the setting up
of narrative spaces and their layering in terms of their embeddedness and
subsequent blending. There is much textual signaling of the different spaces,
but the use of now considered here shows that viewpoint shifts may also be
marked through specific means. This use of now does not mark an independent
meaning, specific to the adverb’s narrative function exemplified here, because
the new meaning emerges at the intersection of the deictic meaning of the
adverb (a pointer to the current space) and the structure of the narrative network
(so that the past space can be selected as the current space to be elaborated).
Both (19) and (20) do not have narrators who are also characters. Example
(21), for comparison, uses now in the first-person narration of The Blind
Assassin. In the fragment, Iris, the narrator and character, describes looking at
old notebooks, hidden by her sister to prevent the truth about her life from being
revealed:

(21) I riffed through the other notebooks. History was blank, except for the photograph
Laura had glued into it – herself and Alex Thomas at the button factory picnic, both
of them now colored light yellow, with my detached blue hand crawling towards
them across the lawn. (The Blind Assassin, p. 627)

The use of now here is not clearly resumptive, but it also signals the temporal
structure of what is being described. The adverb comes within a description of
the photograph Iris sees in the notebook, in which a hand at its edge has been
tinted blue. But the blue belongs to the time when Laura colored the photo and
put it in her notebook, while the yellowish shade of the photo overall (we can
assume it was black-and-white) has emerged over the years that have passed.
So the ‘yellow’ here is the color observed years later, when Iris is looking at it.
The fact that Laura glued the photo in the notebook is the pre-past of the story,
while the yellow tinge belongs to the now of the scene narrated – when Iris finds
the notebooks.
Now thus emerges as a marker of time, but not the present time. Instead, it
highlights the narrative space currently in focus, especially when the context
 7.4 Viewpoint compression and constructional phenomena 193

highlights some earlier events. The now marks a return to the temporal space
currently elaborated, and in some cases this use may further be resumptive.
Throughout the chapter I have looked at the way in which the lowest-level
discourse spaces profiling speech or thought are represented in the text. While
there are clearly various constructional phenomena involved, it is not the
case that traditional categories such as direct, indirect, or free indirect discourse
naturally refer to classes of sentences falling under a cohesive linguistic
description. Rather, the history of research into STR shows intensive negotia-
tion of the boundaries of each category, of new categories needed, or of the
actual status of adverbs and expressive elements. In this section, I have not
attempted either to summarize or to revise the existing categories. Instead,
I have proposed an approach wherein individual formal signals, such as a
pronoun, a temporal adverb, a verb form, or a syntactic pattern, contribute their
meaning to the overall construal, but allow the construal to emerge as a result of
a blending process reconciling all the components of the expression. Thus a
pronoun I points deictically to the speaker situated against a ground, but if there
are several grounds in the network, I may prompt a chain of projection and
compression in which different spaces will be involved, and the designated
participant may appear not to be the speaker in the standard sense. This approach
goes beyond Rubba’s discussion of alternate grounds (1996), because the matter
of reference in the narrative may require involving the network of spaces, rather
than different or competing spaces.
The discussion of STR often yields new categories – such as DIST, proposed
by Vandelanotte. There is indeed a crucial difference in its design, the shift
towards the deictic centre of the current speaker (rather than represented
speaker). But the difference between DIST and FIST is primarily in how deictic
forms such as pronouns are aligned with the two mental spaces in the required
network (current versus represented). Perhaps this is a criterion which could be
further studied profitably – in fact Dancygier and Vandelanotte (2009) discuss
other uses of a similar deictic alignment in a range of colloquial and literary
uses, under the term distanced discourse. Clearly, alignment with the current
speaker is a compositional feature of more than one construction, though it also
reliably distinguishes two forms of STR.
In a completely different context, the vocative-cum-imperative construction
discussed in Chapter 6 builds on the same mechanism of constructional com-
positionality – the imperative evokes a desirable though not predictable future
occurrence, and in a typical context it involves a speech act of ordering or
requesting, with its appropriate felicity conditions. But this is, so to speak, the
‘strong’ form of the imperative, which is absent from uses such as Open the
door and I’ll shoot! (cf. Dancygier and Sweetser 2005), where the action of
opening the door is not only not ordered or desired, but outright discouraged.
When Dr Faustus urges the Earth to ‘gape’ and swallow him so he can escape
194 Speech and thought in the narrative

hell, he is not ordering, but is imagining the desirable state of affairs occurring.
The most salient use of a form (such as the ‘order’ for the ‘imperative’) is not
necessarily the most accurate. In fact, it is usually better to have a minimalist
description of a form’s meaning, and then find out how this minimal meaning
gets enhanced in actual constructions.
The narrative is too complex a language form to be naturally analyzable in
terms of a set of fully determined constructions. Throughout this book I have
attempted to focus on selected mechanisms responsible for the emergence of
the forms and meanings readers pull out of texts. But, as I tried to show, all
those meanings are functions of specific kinds of narrative space networks, and
of phenomena such as embedding, projection, and, perhaps first of all, view-
point compression. These mechanisms, however, do not yield specific con-
structions, but rather propose interpretive strategies which make sense of the
amazing variety of narrative choices available.
8 Stories in the mind

Life is a trace in the ocean of stories.

The analyses and theoretical constructs proposed throughout this book apply
to some of the more common questions raised by fictional narratives, but there
is a host of more specific problems which can gradually be addressed. In this
chapter, I will sum up some of the main directions the examples considered
suggest and outline the general conclusions emerging.

8.1 The linguistics of literature

Stories are linguistic artifacts, but they require a language theory which allows
the analyst to consider questions of meaning and form in some correlation, and
deal with patterns of emergent interpretations. The primary interest of cognitive
linguistics is the elucidation of form–meaning mappings and the uncovering
of underlying cognitive concepts, and thus it seemed worthwhile to test it in a
broad literary perspective. In the narrative context, where forms are less natu-
rally in focus and meanings are of primary interest to readers and analysts alike,
cognitive linguistics opens new avenues of investigation, with respect to both
form and meaning.
The nature of narrative meaning is a further question I tried to address. I have
argued that such meaning emerges in the reader’s mind and cannot be attributed
to the textual prompts alone, but relies also on the frames available, on the
reader’s disposition, on the cultural context, et cetera. Thus, while the text
serves as a prompt, it does not limit the ways in which the prompts will be
interpreted. What matters most is the range of processes which may yield
different readings, and not the readings themselves. This approach appears to
question some assumptions analysts of literature may hold dear – either the
expectation that linguistic tools can predict the interpretive result, or the belief
that such a result is primarily a matter of influences beyond the text. However,
no such claims are being made. While interpretive acts are motivated by the
linguistic processes I tried to describe, they are not predictable. Also, the frames
available in the reader’s mind participate in the construal alongside the text and
195
196 Stories in the mind

come into play at any level. It is the interaction of the linguistically motivated
aspects of interpretation with cultural or historical frames that may yield a
specific reader’s specific reading. But the nature of that interaction is at the
core of the process and needs to be formulated.
Throughout the book, I talked about macro-level phenomena and micro-level
phenomena, and relied on two mechanisms that explain the cross-level coher-
ence. The first mechanism is blending – a convenient term to represent pro-
cesses which yield emergent meanings. The second mechanism, central to my
argument, is viewpoint compression. While blending gracefully explains the
fact that narrative meaning emerges through selection of conceptual material
and its integration into new, higher-level structures, viewpoint compression was
argued to be the main driving force behind these processes, playing a major role
in the actual textual choices.
I defined narrative viewpoint independently of the tradition which assumes
a viewing Ego, and linked it to the concept of a narrative space, as a partial
construct structured by a certain topology. A narrative space affords various
viewpoint dimensions – temporal, spatial, emotional, et cetera – but it also relies
on the presence of participants and their dispositions. Viewpoint can thus be
aligned with one of those participants, but may shift to time or space, or to an
unprofiled, imagined participant, or to another narrative space. Those viewpoint
shifts, crucial to the complexity of fictional narratives, are managed across
narrative levels through compression – a process which may incorporate a
viewpoint of a lower space into the higher space, and gradually reach the highest
narrative level, wherein the story as a whole is construed as being narrated.
The approach tested some assumptions I had made – both about the role of
the text and about the role of frames in the processing of texts. It was also a test
of the applicability of blending in the analysis of longer texts. I will briefly
review the conclusions below.

8.1.1 The sentence versus the text

The main linguistic problem I tried to address is the ways in which the lowest-
level grammatical and lexical choices participate in the meaning of the text in
its entirety. Naturally, not all such sentence-level forms contribute significantly
to how a narrative will be read, but there are at least two instances discussed
here where the choices matter throughout the text as a whole. One such example
is the choice of the structure of the SV-space and its relation to MN-space.
Grammatical options such as pronouns or tense reduce or expand the distance
between the locus of the narrating and the actual storyworld. The degree of
distance, in turn, affects the epistemic impact of the text and, consequently, the
ways in which what is narrated is processed. First-person narratives usually
mark less distance, but also limit the range of viewpoints available and leave
 8.1 The linguistics of literature 197

significantly more unsaid, or require that it be said differently. For example, the
first-person narrator of Atwood’s The Blind Assassin yields a very personal
tone of the narration, but the text also needs to construct other major narrative
spaces to profile epistemic viewpoints which are indispensable in the grasping
of the story. Similarly, in Raban’s travel writing, frequent decompressions of
identities and concepts, marked by sometimes unusual textual devices, serve
the purposes of viewpoint management at the level of the story. At the same
time, Woolf’s multi-Ego narrative plays with many perspectives all the time, but
creates no pretense of a consistent epistemic viewpoint and requires that all
the work of piecing the story together is done in the reader’s mind. In all specific
cases, lower-level narrative choices are responsible for the kind of interpretive
work the reader does.
Also, there is the question of lexical choices or minute reactivations of images
and scenes, which I discussed under the rubric of narrative anchors. While my
most complex example is that of anchors in The Blind Assassin, there is a fair
degree of anchoring and reactivation patterns in all narratives. Quite often, the
anchors are material anchors written into the story (such as the photograph in
The Blind Assassin), so there is a clear correlation between evocation patterns
identified in cultural artifacts and those developed in narratives. One such
example, not mentioned earlier in the text, is Melville’s Moby-Dick, where
unusually numerous objects contribute to the understanding of what motivates
the characters and open more general avenues of interpretation (let me mention
only the doubloon, Queequeg’s coffin, the Pequod’s tiller, or Ahab’s leg). In
each such case the object built into the story is endowed with story-specific
meaning, which is reactivated every time the object is mentioned and contrib-
utes to story construction.
As I argued throughout the book, these, and other, lowest-level choices play a
role in the understanding of the text in its entirety and achieve their narrative
power through the processes of viewpoint compression. The compressions
reduce and remove the distance across various narrative spaces, thus contribu-
ting to the compact and well-integrated structure which finally emerges as the
story. Viewpoint compression, as I argue, is the central mechanism leading to
story coherence, as the emergent construct gradually reaches its completeness.
The process of reading is thus not simply linear and does not rely primarily on
the accumulation of information. It is a multidimensional process, reaching
across narrative spaces in different directions. In the responses to my earlier
work I often heard colleagues sounding uneasy about the idea that a sentence
may have a narrative status comparable to a fragment, or even an independent
subplot, as they can both be treated as narrative spaces. But there is no reason
to assume that compression does not work across sentences, and is only valid
when more major narrative fragments are involved. True, when the text chooses
to separate narrative chunks in terms of viewpoint, as in The Blind Assassin,
198 Stories in the mind

viewpoint compressions are more visible, but the process is equally valid in
Eggers’s text, where numerous (but very short) escapes into imaginary scenar-
ios also require compression to yield the general impact of the obsessive and
frustrated narration. The narrative choices may be more or less obvious, but the
process remains the same.
This approach made my discussion much less focused on the distinctions
which narratology relies on. Much ink was spilled to elucidate the distinctions
between narration (as the mode in which the narrator ‘tells’ the events of the
story) and various kinds of speech and thought representation. In my discussion
of texts, I have assumed that these distinctions are primarily available at the
lower levels of narration, since ultimately the entire text is in the scope of the
SV-space and thus needs to be compressed to contribute to the narrative as a
whole. I have also not attempted to rely on the idea of a ‘narrative event’ and
assumed that a fantasy in the narrator’s mind is no different in its narrative status
from a description of a character’s behavior: either one might be compressed
into the story to add to the understanding of the character’s plight, and whether
it was worded as a narrator’s description or a direct insight into the mind is a
matter independent of the story-emergence mechanisms. The mechanisms were
my main focus.

8.1.2 Constructions in the narrative

My approach relies heavily on the concept of a construction, but it uses the term
in a sense which seems more appropriate in the context of the narrative.
Narratives naturally rely on the constructions available in the colloquial context
(such as the ditransitive construction or the caused-motion construction), just
as they rely on the entire linguistic repertoire of the language in question.
However, they also create constructions specific to the genre, where linguistic
resources are deployed in a less strictly patterned way. I have referred to this
phenomenon as constructional compositionality, but the narrative-specific ways
in which linguistic forms are used may call for a discussion of constructions
supporting specific features of the genre.
One example of a construction which seems to be appearing primarily in
dramatic narratives is the ‘vocative-cum-imperative’ construction in Early
Modern drama. Both the vocative and the imperative are language forms available
in English, but their function in the cases like Come, bitter conduct! is adjusted to
the meaning of the scene and strictly dependent on the actions performed on stage.
There are apparently many such constructions worth discussing. For example,
I have not devoted much attention to the use of tense in the narrative, but there is
clearly a need to elucidate what the present tense in narratives means, beyond
being compressed to the SV-space to yield a less distanced narration. It definitely
relies on the meanings available in colloquial discourse, but because it is not
 8.1 The linguistics of literature 199

clearly correlated with a standard deictic ground, there are also differences in
meaning. Questions can also be asked about the narrative use of demonstratives,
articles, modals, and many other linguistic forms. There is much similarity in
usage, but there are also differences.
As I have shown, constructions may be specific to the linguistic ramifications
of the genre. The ‘vocative-cum-imperative’ has been used in drama, while
interesting uses of deictic pronouns can be found in poetry. Various types of
narrative modes are available in novels and short stories, and some of them may
be present in epic poems, but the deployment of constructional resources may
be genre-specific and thus opens the door to an enquiry of the textual choices a
genre makes available. These questions connect with the recent research on
spoken and written genres. For example, studies of conceptual metaphor in
discourse (cf. Semino 2008, 2010) suggest that different types of texts might
call for a different use of conceptual metaphor. As Semino shows, scientific
discourse may rely on elaborating an improbable source domain in order to
allow for a new construal of a complex target. Also, a recent study by Moder
(2010) shows that genres such as radio interviews may require constructions
which rely on available resources (such as similes or copulative constructions)
but yield a result geared into the context where the listeners have no access to
visual contact with the guest in the studio. Narrative study confirms that specific
discourse contexts use constructional resources of English in new ways.
What seems to be required, then, is an understanding of how the particular
type of contact between the recipient and the text may alter prototypical uses of
linguistic forms. At the same time, the constructional and compositional
approach would add to the linguistic study of genres, literary and non-literary.
Possibly the most interesting constructional aspect of narratives is the use of
deixis. My analysis has not fully covered the tense dimension of the use of
deixis, but I have assumed that the SV-space constitutes a correlate of the deictic
ground, so that the communication occurring in that context, even if written and
directed at an unspecified addressee, assumes a default temporal dimension.
Thus the SV-space time, like the time of any exchange, is present, and the story
can either be distanced from it, which means told in the past tense, or com-
pressed to it, which means that the choice is present tense. But the past itself
is also an interesting choice, in that, on the one hand, it supports the default SV
set-up and, on the other hand, it puts events in specific relations with each other,
such as precedence.
My analysis has been more specific in its treatment of pronouns, but these
questions go far beyond the default deictic set-up. In various texts, pronouns
have widely varying functions. They may open links to other narrative spaces
(as in The Blind Assassin) or support decompressions for viewpoint (as in
Raban’s work), or signal STR blends and other more unusual blends when a
viewpoint of one space is compressed with that of another. They may also, as in
200 Stories in the mind

the case of Szymborska’s poems, signal unusual deictic set-ups in which partici-
pant roles are assigned in non-standard ways. What is systematic about all these
options is that the choice of a deictic pronoun like I or you calls up a deictic centre
(whether previously set up or not) and assigns communicative roles to the
participants, even inanimate ones. Furthermore, once such a set-up is established,
pronouns may serve functions which have much less to do with the flow of
discourse and the resulting anaphoric or cataphoric relations, but instead connect
various available spaces ‘at various angles.’ The range of pronoun use in the
narrative is extremely broad and should be investigated further.
The instability of constructional phenomena in the narrative is also correlated
with the issue of narration versus speech and thought representation. As I
argued, it is difficult to distinguish types of STR either as independent con-
structions or as instances of narration, speech, or thought, as the constructional
correlates are not systematic enough. Also, any of the specific choices may have
limited bearing on the overall interpretation: in the case of the reports included
in the first paragraphs of The Blind Assassin, the specific words and thoughts
of different participants are a stylistically interesting way of structuring two
viewpoints, but there is a gap between the meanings specific to the constructions
used and to the overall narrative effect.
Furthermore, the fluidity of STR categories and their links to narration clearly
supports the idea of compositionality, and poses questions which need inves-
tigation. Naturally, we can investigate the role of pronouns and tense in the
management of viewpoint, and consider contextually interesting uses of
adverbs from the point of view of their alignment with viewpoint spaces, but
sentence types and complex sentences can also be viewed as compositionally
interesting. If the form of a question can represent a character’s or on-stage
narrator’s thought, then we need an explanation of how this effect emerges
and what low-level aspects of questionhood had to be tweaked for the use to
emerge. Similarly, while Verhagen (2005) has pointed out the intersubjective
nature of reported speech and thought constructions, many questions still
remain about intersubjective construals emerging out of different reporting
choices. Practically any aspect of grammar has its interesting correlates in the
narrative context.

8.2 The storyworld reality

The question of the status of the worlds structured through narrative discourse
was another important recurring issue. I was arguing against the assumption
that we need to surgically remove narratives from the scope of our ordinary
assumptions about the world and language. Narrative worlds constructed lin-
guistically (or visually, as in film) may depart from the reality of our experience
in many ways, but they do not depart from it more than our everyday
 8.3 Blending and narrative analysis 201

imagination might allow. As almost any paper on blending or metaphor shows,

our ability to create imaginative construals of situations is not as limited as the
strict division between reality and fiction assumes. My approach is thus closer
to the work on text worlds than on possible worlds, since the former is much
more of a discourse theory than the latter.
I have also tried to show the ways in which reality participates in narrative
construals. First, any text relies on our reality-based knowledge of frames: if we
read about someone making a phone call, we understand it because we have an
easily accessible frame of devices making communication over long distances
possible. Frames and patterns of frame evocation constitute a reality-based
conceptual network which underlies any construal a narrative may come up
with. Even a science-fiction story has to rely to the same degree on accessible
frames to understand how the imagined world questions the frames observed
in reality (knowing gravity in real life is crucial to understanding what absence
of gravity might mean). Second, the historical context of the text or the influence
of the author’s experience should be treated as an integral part of the emerging
meaning of the narrative, rather than as separate from it. Narratives could not
be written coherently without being immersed in some cultural, social, or
individual knowledge and experience. The reader may apply different frames,
but the very reliance on frames is what makes narrative interpretation possible.
Another dimension is the narrative rendering of characters’ experience of the
world. I have talked about fictive experience, which might mean that narrative
events and other construals are represented in the text as if they were seen or
heard by someone, but there are more interesting patterns to explore if we
consider various ways in which narratives, whether viewed as produced or
received, rely on the experiential and embodied construals of experience. I have
focused for a while on fictional minds and the representation of mental states,
but questions of fictive experience reach far beyond these categories.

8.3 Blending and narrative analysis

In order to describe the fluid nature of narrative meaning construction I have
selected a framework which seemed best suited to an analysis of meaning
emergence – conceptual integration. The analysis suggests two conclusions.
One, blending can indeed be applied in the analysis of more complex texts.
And two, narratives may expand our understanding of how blending works. But
the question of further implications of my theoretical choice remain difficult
to answer. For example, researchers have asked an important question – what is
the psychological reality of blending? Opinions vary. The proponents of blend-
ing (Fauconnier and Turner 2002) make strong claims about the role of double-
scope blending in the evolution of modern human beings. Followers in many
disciplines, from language through music, art, and performance to literature,
202 Stories in the mind

find blending a very helpful tool in reaching beyond structure-based theories.

And yet it is possibly too early for experimental work in neuroscience or
psychology which would tell us if blending is ‘real’ and can naturally be
distinguished from other paths of conceptualization.1
However, as an interpretive tool, blending has become possibly the single
most useful theory of meaning. It allows one to capture the so far unattainable
scope of phenomena in a coherent way, and it is the first theory of meaning
which is not encumbered by being restricted to this or that area of enquiry. It has
uncovered hosts of observations on human processing of prompts and the
economy of our meaning construction mechanisms. Blending may not become
scientifically satisfying for a while, but it is very satisfying already from the
point of view of language analysis. As regards stylistics and literary criticism,
the complaints are (often ironically) different – couldn’t we say the same
without blending? isn’t it too scientific to be useful to me? It is hard to imagine
a theory which could satisfy all its critics, especially if they come from different
academic cultures, with their specific expectations. But perhaps the fact that
these divergent groups of researchers consider the pros and cons seriously is a
signal that it is appealing enough to warrant more effort. I hope to have shown
that there are merits to the blending framework, whichever academic culture
one comes from.
The examples analyzed throughout this book are useful in putting these
questions in context. First, they show that meaning construction processes
work in similar ways through different levels of language structure and that
there is a natural way to connect sentence-level phenomena with textual-level
processing. Second, blending allows one to incorporate complex cultural
frames into the picture. Also, it does not restrict interpretive efforts unreason-
ably. Although it does require that more attention be paid to what the text does,
it opens itself to any culturally or historically based set of questions. And it may
still reach its experimental phase. Many of the concepts emerging out of
cognitive linguistics research, such as embodiment or the metaphorical nature
of abstract thought, have prompted experimental work which supports the
validity of these research tools.2 Thus, while blending is so useful because it
allows one to capture processes rather than static, complete conceptual struc-
tures, it is also difficult to test for, for the very same reason.
Perhaps the material presented in this book might at least explain why it is so
hard to do. Narratives evoke a number of complex research questions: the role
of memory mechanisms which link language forms to representations of com-
plex situations; the nature of linguistically prompted emotional responses; our
embodied sense of how actions signal mental processes; the nature of mental
processes involved in the construction, maintenance, and development of con-
structs called storyworlds; reliance on one’s own experience with various
cultural artifacts, including narratives; the dynamic nature of text processing;
 8.4 A bridge to the truth 203

the ways in which linguistic expressions can negotiate various forms of access
to other people’s minds; and the evolutionary processes that led to the emer-
gence of all these abilities. No less. So until and unless methods in psychology
and cognitive science can help in answering these questions, blending will have
to remain an interpretive tool and continue to inspire more difficult questions.
It has achieved these goals when applied to the material in this book.

8.4 A bridge to the truth

In the motto to Chapter 1 of the book I gave a verse from Arabian Nights which
claims that a tale is a bridge that leads to the truth. Perhaps “the truth” is an
exaggeration, but I found the quotation appropriate to this project, because it
highlights the fact that whatever understanding a reader might acquire, it is not
contained ‘in’ the story, but can only be arrived at through the interaction with
it. The understanding of the narrative as a kind of container into which a reader
reaches to pull out the story seems to be an inheritance of the Conduit Metaphor,
as described in Reddy (1979). It is a complex metaphor which conceptualizes
linguistic expressions as containers into which meanings are packaged in order
to be sent to the receiver, who unpacks them. The primary fallacy of this
construal is that meaning is stable, bounded and possible to transmit intact.
The conduit fallacy explains some of the unrealistic language-based assump-
tions about the meaning of fiction being contained in the language of the text,
ready to be unpacked. Apparently, the fallacy has also led to a broad rejection of
the view that the text matters in reading, beyond the basic meanings of words
and the accessibility of syntactic forms. The rejection, in turn, has prompted the
belief that meaning is so unstable that an analyst cannot presume to find it,
though she or he may uncover some important elements, including those which
the author may have never intended to include (this would go, among others,
for the variety of interpretations intent on finding a feminist, racist, or other
politically defined angle in the text). These approaches to narrative meaning
tend not to look at the form of the narrative as a relevant criterion, and thus are
also unrealistic, though in a different way.
The approach I have argued for suggests that narrative meaning, as any other
linguistic meaning, is prompted by linguistic forms, but emerges in the process
of interpretation, guided by general linguistic as well as specifically narrative
processes. Included in this approach is the idea that a narrative communicative
act is in some ways like any other communicative act – the author, like the
speaker, constructs the text, and the reader, like a hearer, interprets it, based on
the textual prompts, available contextual ground, and general knowledge, and
in an intersubjective context where other viewpoints are accessible. Much of
this book was devoted to processes leading to the interpretation.
204 Stories in the mind

But the nature of the process, the length of narratives, and the vast variety of
background knowledge authors and readers bring into the exchange are jointly
responsible for the fact that readers of the same text may not arrive at the same
reading. They may (though do not have to) uncover the sequence of events, but
that alone does not constitute the content of the story. The story defined here
is a much richer construct, in which dispersed identities are established and
mental states constitute important narrative material. Representation of minds
is as central to the narrative as representation of events, and it is possibly the
aspect of texts which yields most innovative narrative techniques. Sequence of
events is the result of the characters’ motivations and causal chains thus created,
not the source of narrative coherence.
My approach stresses the dynamic nature of the process, which is elegantly
captured in the motto to this chapter: Life is a trace in the ocean of stories. Lives
are depicted in narratives, but the depiction chooses a path through the maze
of events affecting the main characters. The path may meander in strange ways,
but that is also a part of the process, since finding coherence means constructing
an emergent structure that puts some events in focus and backgrounds others.
The story moves through events surrounding it, and emerges out of the links
the reader’s mind adds. It is not built into the ocean of meaning we float in: it
has to emerge as we move through the waves.
 Notes to the text

C H A P T E R 1 L A N G U A G E A N D L I T E R A RY N A R R AT I V E S
1. For more discussion of this phenomenon, see Fleischman 1989; Dancygier and
Sweetser 2005.
2. Cf. Fauconnier 1994 [1985], 1997; Fauconnier and Sweetser 1996; Fauconnier and
Turner 1994, 1996, 1998a, 1998b, 2002; Turner and Fauconnier (1995, 1999); Coulson
2001; Coulson and Oakley (2000, 2005). Also, check for more sources at markturner.
org/blending.html.
3. The work in cognitive stylistics has already made a significant contribution to our
understanding of the ways in which literature and cognition are linked (see Tsur 1992,
2003; Stockwell 2002; Gavins and Steen 2003; Bex, Burke, and Stockwell 2000;
Burke 2005; Freeman 2000, 2005; Semino and Culpeper 2002), and much new work
is being done in that context.
4. There is indeed a significant amount of linguistic work within blending, too much
perhaps to be sufficiently represented here. The relevant papers will be referenced as
the need arises.
5. One volume on cognition and the arts, which holds articles on visual art, music,
architecture, and many general questions of aesthetics, is Turner (2006b); work by
theatre and performance scholars has been given some focus in Hart and McConachie
(2006) and Cook (2007, 2010); work specifically devoted to music has been done
primarily in numerous articles and books by Zbikowski (2002, 2006 are very repre-
sentative examples); film was taken up in Rubba (2009).
6. The initial outline of the view of Literary Darwinism can be found in Carroll (1995,
2005), and Gottschall and Wilson (2005). Also, a special issue of Style (Style 2008)
contains not only a programmatic article by Carroll, but also a number of (mostly
critical) responses to the idea, with a concluding rejoinder by Carroll.
7. Some of the discussion on literature and ‘evolution’ is fleshed out in the papers in
Gottschall and Wilson (2005).
8. Cf. Mar and Oatley (2008); Djikic, et al. (2009).
9. The concept of embodiment in the analysis of linguistic meaning is now one of the central
issues of cognitive linguistics. Starting with the earliest work by Johnson (1987) on
image schemas, and Lakoff and Johnson’s introduction of metaphor theory (1980, 1999),
the idea that linguistic meaning emerges out of patterns of embodied experience has
become the central claim in cognitive linguistics and has been applied to a broader and
broader range of phenomena. Recent publications, such as Johnson (2007) and the papers
included in Hampe (2005) and Frank et al. (2008), move the issues further into cognitive
science, and into an exchange with work in philosophy, such as Gallagher (2005).

205
206 Notes to pages 25–64

10. Much of this work is reported in the papers gathered in Pecher and Zwaan (2005b).
The relevant grounding research is also interestingly reported in the introduction
(Pecher and Zwaan 2005a).
11. Any discussion of the framework is beyond the scope of this work, but a very clear
introduction can be found in Richerson and Boyd (2005).
12. Work in evolutionary psychology has been publicized mainly through articles by
John Tooby, Leda Cosmides, and their collaborators (cf. Barkow et al. 1992).
Forthcoming broader volumes, such as Cosmides and Tooby (in press) and Tooby
and Cosmides (in press) will soon become the definitive sources.

C H A P T E R 2 B L E N D I N G , N A R R AT I V E S PA C E S , A N D T H E
E M E R G E N T S TO RY
1. The work on mental spaces is now too broad to review here. Many interesting papers
can be found in Fauconnier and Sweetser (1996); there is also interesting work on
tense and aspect (Cutrer 1994; Fauconnier 1997), conditionals and coordinate con-
structions (Dancygier and Sweetser 2005), or constructions of intersubjectivity,
including negation, adverbial conjunctions, complementation, etc. (Verhagen 2005).
It might be worth noting that much of the work on mental spaces has gradually
merged with research on blending.
2. Some examples are: Oakley’s paper (1998) on Art Spiegelman’s Maus, Rubba’s
(2009) paper on Mulholland Drive, Sweetser’s paper on Cyrano de Bergerac
(2006), Turner’s work on compression and representation in cartoons (2006a), and
Dancygier’s (2006) discussion of the movie Annie Hall.
3. I first introduced narrative spaces in Dancygier (2005a), then developed the concept in
Dancygier (2007) and (2008a); the concept owes much of its structure to the dis-
cussion by Sanders and Redeker (1996) of mental spaces in journalistic discourse.
4. The discussion of the narrative in terms of possible worlds can be found, among
others, in Doležel (1976, 1998); Maitre (1983); Pavel (1986); Ronen (1994); and
Ryan (1991). For an excellent, broad review of the framework, see Semino (2009).
5. Some of the more useful discussion can be found in Chatman (1978); Rimmon-Kenan
(1983); Toolan (2001); Porter Abbott (2007).

C H A P T E R 3 S TO R I E S A N D T H E I R T E L L E R S
1. The concept of the narrator has been discussed in much of the work in narratology and
stylistics. Some general terminological distinctions can be found in Booth 1961;
Chatman 1978; Toolan 2001; Rimmon-Kenan 1983; Herman 2007.
2. For a discussion of the views of authorship in Coetzee’s recent novels, see Dancygier
2010; also, some brief discussion of Powers’s Generosity, in Chapter 4.
3. The role of evaluative aspects of stance in the construction of narrative viewpoint is
beyond the scope of this project. See Harding (2004; 2007) for interesting examples
and analyses.
4. Since my focus throughout the book is on the elucidation of the viewpoint profiled in
a narrative space, I will not discuss the various construals referring to the postulated
recipient of the text, whether that subjectivity is termed ‘narratee,’ ‘implied reader,’ or
the like. These concepts can naturally be built into the framework, but are not my
focus here.
 Notes to pages 69–118 207

5. I am thus rejecting Banfield’s (1987) concept of an ‘empty’ deictic center in the

narrative. On the contrary, I assume that every narrative act assumes a space wherein
there is a source of viewpoint, which can then take the form of a narrator. The on-
stage/off-stage distinction is meant to take care of the differences in the degree of
profiling of the narrating subjectivity.
6. Now in past narratives can have other functions, to be discussed in Chapter 7.
7. Example (10) is earlier used in Chapter 2 on page 38.
8. Mental space embedding was best described in Sweetser (1996), and then developed
in Dancygier (2002). The crucial point to be restated here is that the embedded space
inherits the viewpoint, especially epistemic viewpoint, of the mother spaces. Both
papers argue for this effect of embedding by looking at verb forms in conditionals, but
in the context of the narrative the structure inherited by the embedded space may
include much of the space topology.

C H A P T E R 4 V I E W P O I N T: R E P R E S E N TAT I O N
AND COMPRESSION
1. Focalization is the central concept in narratological enquiry, and it is not possible to
give it justice here. The best introduction to the concept and its revisions can be found
in Rimmon-Kenan 1983; Bal 1985; Van Peer and Chatman 2001; Herman 2002.
2. More discussion of ‘speech’ and ‘thought’ will be offered in Chapter 7.
3. ‘Represented speech and thought’ is yet another case of a concept which is central to
narrative enquiry, and has thus inspired an impressive body of research. The most
pertinent texts are Genette 1980; Leech and Short 1981; Banfield 1982; Bal 1985;
Fludernik 1993; recent revisions, based on different methodologies, add much to the
discussion (consider, for example, Sanders and Redeker 1996; Semino and Short
2004; Vandelanotte 2005 and 2009). For a good general review, consult Toolan 2001.
4. Cf. Lanser 1981; Rimmon-Kenan 1983; Ronen 1994; Prince 2001; Van Peer and
Chatman 2001.
5. The idea of fictive motion was introduced in Talmy (1996), later developed in Talmy
(2000); Matlock (2004); Langacker (2005). The ‘fictive’ idea has also been extended
to ‘fictive interaction,’ introduced in Pascual (2002, 2006a, 2006b, 2008).
6. Some discussion of forms of STR and some of their constructional features will be
included in Chapter 7. The discussion here is restricted to the possibility that STR
phenomena are essentially an instance of the more general pattern of viewpoint
compression.

CHAPTER 5 REFERENTIAL EXPRESSIONS AND

N A R R AT I V E S PA C E S
1. For a general treatment of deixis, see papers in Rauh (1981). Wales (1966) is an
excellent review of the issues posed by pronouns in general. The cognitively informed
view of deixis can be found in Fillmore (1997).
2. It is not possible to review such approaches here, but let me briefly mention the
framework developed in Van Hoek (1997), where referential relations are treated in
terms of ‘reference points,’ a concept earlier established in cognitive grammar.
3. See Dancygier 2004a, 2005a, 2008b.
208 Notes to pages 119–172

4. In this discussion I am primarily interested in the role proper names play in story
construction. They do, however, perform important discourse-organizing functions as
well. For example, Van Vliet (2008) discusses the contrast between the choice of a
pronoun or a proper name as important in signaling the episodic structure of narra-
tives. Also, Vandelanotte (2004b, 2009) observes the role of proper names in signal-
ing relations of accessibility relevant to the selection of speech and thought
representation constructions.
5. See Chapter 7 for more discussion of this narrative technique.
6. Deictic Shift Theory has been proposed in the collection of papers edited by Duchan
et al. (1995); see especially Galbraith (1995) and Wiebe (1955). It proposes a treat-
ment of the specificity of textual deixis.
7. This chapter considers examples of pronoun use with the exception of constructions
of represented speech and thought. These will be discussed in Chapter 7.

CHAPTER 6 FICTIONAL MINDS AND EMBODIMENT IN

DRAMA AND FICTION
1. For a discussion of the approaches to the ‘ground,’ see Langacker (1990); Coulson
and Oakley (2005b); papers included in the volume edited by Oakley and Hougaard
(2008); and Verhagen (2005).
2. For a very basic introduction to the linguistics of drama, see Sanger (2001).
3. Theatre theorists include further divisions of the on-stage space into locus and platea,
concepts emerging from medieval theatre and continued in Early Modern drama (see,
among others, Weimann 2000); an in-depth discussion of these concepts is beyond the
scope of this chapter, but it is interesting to note that even the earliest theatrical forms
require different functional allocations of space.
4. I am relying here on the concept of a frame, as it is better matched with my overall
framework. However, broader discussion of how objects carry meaning can be found
in semiotics literature, e.g. Elam (1980).
5. Thanks to my colleague Patsy Badir for bringing this example to my attention.
6. Beyond the scenes discussed here, the Body/Mind blend and the vertical schema
metaphors are used in almost every scene, including the central ‘seize the crown’
scene, where Richard and Bolingbroke are holding the crown together and Richard is
describing his emotions through the metaphor of the bucket filled with tears, moving
down with the weight of the sorrow. Further discussion of the scene would go beyond
the scope of this chapter.
7. For more discussion in the context of theatre criticism, see Harris and Korda (2002).
8. Blackwell’s collection of essays (2007) explores it-narratives in some detail.

C H A P T E R 7 S P E E C H A N D T H O U G H T I N T H E N A R R AT I V E
1. There is a rich tradition (Leech and Short 1981; Semino and Short 2004) distinguish-
ing a number of types of STR, including constructions specialized in thought or in
speech (or indeed, writing) representation.
2. Some authors (cf. Brinton 1980; Adamson 1995) distinguish reported perception as a
separate STR category, but the question remains open: how do we distinguish
these STR modes in conceptual terms? In Chapter 4 I argued for a category of
 Notes to pages 177–202 209

‘representation of conceptualization,’ rather than thought, and these comments apply

here as well.
3. The discussion of quotations used as ‘selective depictions’ has further been advanced
through Pascual’s concept of ‘fictive interaction’ (2002, 2006a), where stretches of
discourse play the role of a ‘token’ of a given ‘type’ of an interactive move.
4. A review of the literature on STR cannot be undertaken here. Let me just refer the
reader to the most comprehensive volumes, such as Banfield (1982); Leech and Short
(1981); Fludernik (1993); Semino and Short (2004); and Vandelanotte (2009).
5. Fauconnier (1994 [1985]) uses the term Access Principle to describe a regularity in
language use which depends on cross-connections between spaces. If two spaces are
connected, speakers can choose an expression from one of the spaces to refer to an
element in the other spaces. For example, if the book called Poetics, an element in a
space of books, is cross-linked to the name Aristotle in the space of authors, one can
talk about reading Aristotle without confusion.

CHAPTER 8 STORIES IN THE MIND

1. The search for the final experimental confirmation of the reality of blending still
continues (see Gibbs 2000). Furthermore, there have been dismissive remarks also
from the cognitive linguistics camp (as in an interview with George Lakoff reported in
Sanchez Garcia 2003), suggesting that blending is not interesting precisely because it
is present everywhere and it is in fact not more than neural binding – an unremarkable
phenomenon. (The framework which Lakoff supports instead has become known as
Neural Theory of Language [cf. Feldman 2006; Feldman and Narayanan 2003;
Gallese and Lakoff 2005; Lakoff 2008].) If that, however, turned out to be true,
then, as Mark Johnson commented in another interview (Pires de Oliveira and de
Souza Bittencourt 2008), the role of blending is not diminished, because, unremark-
able or not, neural binding is a hard fact.
2. See, for example, Spivey (2007), or papers in Pecher and Zwaan (2005b).
 References

Adamson, Sylvia. 1995. From empathetic deixis to empathetic narrative: stylisation and
(de)subjectivisation as processes of language change. In Dieter Stein and
Susan Wright (eds.) Subjectivity and Subjectivisation: Linguistic Perspectives.
Cambridge University Press, 195–224.
Ariel, Mira. 1990. Accessing Noun Phrase Antecedents. London: Routledge.
Bal, Mieke. 1985. Narratology: Introduction to the Theory of Narrative. University of
Toronto Press.
Banfield, Ann. 1982. Unspeakable Sentences, Narration and Representation of the
Language of Fiction. Boston: Routledge and Kegan Paul.
1987. Describing the unobserved: events grouped around an empty centre. In Nigel Fabb,
Derek Attridge, Alan Durant, and Colin MacCabe (eds.) The Linguistics of Writing:
Arguments between Language and Literature. Manchester University Press, 265–285.
Barkow, Jerome, Leda Cosmides, and John Tooby (eds.). 1992. The Adapted Mind:
Evolutionary Psychology and the Generation of Culture. New York: Oxford
University Press.
Baron-Cohen, Simon. 1995. Mindblindness: An Essay on Autism and Theory of Mind.
Cambridge, Mass.: MIT Press.
Barsalou, Laurence W. 1999. Perceptual symbol systems. Behavioral and Brain Sciences
22: 577–660.
Bex, Tony, Michael Burke, and Peter Stockwell. 2000. Contextualized Stylistics: In
Honour of Peter Verdonk. Amsterdam/Atlanta: Rodopi.
Blackwell, Mark (ed.). 2007. The Secret Life of Things: Animals, Objects, and It-
Narratives in Eighteenth Century England. Lewisburg: Bucknell University Press.
Booth, Wayne C. 1961. The Rhetoric of Fiction. Chicago University Press.
Boyd, Brian. 2005. Evolutionary theories of art. In Gottschall and Wilson (eds.),
149–178.
2006. Fiction and theory of mind. Philosophy and Literature 30(2): 517–581.
2009. On the Origin of Stories: Evolution, Cognition, and Fiction. Cambridge, Mass.:
Belknap Press of Harvard University Press.
Bradley, Raymond and Norman Swartz. 1979. Possible Worlds: An Introduction to Logic
and Its Philosophy. Oxford: Basil Blackwell.
Brinton, Laurel. 1980. Represented perception: a study in narrative style. Poetics 9:
363–381.
Bruner, Jerome. 1991. The narrative construction of reality. Critical Inquiry 18(1): 1–21.
Burke, Michael. 2005. How cognition can augment stylistic analysis. European Journal
of English Studies 9(2): 185–195.

210
 References 211

Canning, Patricia. 2008.“The bodie and the letters both”: “blending” the rules of early
modern religion. Language and Literature 17: 187–203.
Carroll, Joseph. 1995. Evolution and Literary Theory. Columbia: University of Missouri
Press.
2005. Literary Darwinism: Evolution, Human Nature, and Literature. Evanston, Ill.:
Northwestern University Press.
Chatman, Seymour. 1978. Story and Discourse: Narrative Structure in Fiction and Film.
Ithaca: Cornell University Press.
1990. Coming to Terms: The Rhetoric of Narrative Fiction and Film. Ithaca: Cornell
University Press.
Cienki, Alan. 2008. Why study metaphor and gesture? In Cienki and Müller (eds.), 5–25.
Cienki, Alan and Cornelia Müller (eds.). 2008. Metaphor and Gesture. Amsterdam: John
Benjamins Publishing Company.
Clark, Herbert H. and Richard J. Gerrig. 1990. Quotations as demonstrations. Language
66(4): 764–805.
Cook, Amy. 2007. Interplay: The method and potential of a cognitive scientific approach
to theatre. Theatre Journal 59: 579–594.
2010. Shakespearean Neuroplay: Reinvigorating the Study of Dramatic Texts and
Performance through Cognitive Science. New York: Palgrave Macmillan.
Cook, Guy. 1994. Discourse and Literature. Oxford University Press.
Copland, Sarah. 2008. Reading in the blend: collaborative conceptual blending in the
Silent Traveller narratives. Narrative 16(2): 140–162.
Cosmides, Leda and John Tooby (in press). Universal Minds: Explaining the New Science of
Evolutionary Psychology (Darwinism Today Series). London: Weidenfeld & Nicolson.
Coulson, Seana. 2001. Semantic Leaps: Frame Shifting and Conceptual Blending in
Meaning Construction. Cambridge University Press.
Coulson, Seana and Todd Oakley (eds.). 2000. Conceptual blending (special issue).
Cognitive Linguistics 11(3/4).
(eds.). 2005a. Conceptual blending (special issue). Journal of Pragmatics 37.
2005b. Blending and coded meaning: literal and figurative meaning in cognitive
semantics. Journal of Pragmatics 37: 1,510–1,536.
Currie, Gregory. 1995. Image and Mind: Film, Philosophy, and Cognitive Science.
Cambridge University Press.
2004. Arts and Minds. New York: Oxford University Press.
2007. Framing narratives. In Hutto (ed.), 17–42.
Cutrer, Michelle. 1994. Time and tense in narratives and everyday language. PhD
dissertation. University of California, San Diego.
Damasio, Antonio. 1999. The Feeling of What Happens: Body and Emotions in the
Making of Consciousness. New York: Harcourt Brace & Company.
2003. Looking for Spinoza: Joy, Sorrow, and the Feeling Brain. Orlando, Fla.:
Houghton.
Dancygier, Barbara. 1993. Interpreting conditionals: time, knowledge, and causation.
Journal of Pragmatics 19: 403–434.
1998. Conditionals and Prediction: Time, Knowledge, and Causation in Conditional
Constructions. Cambridge University Press.
2002. Mental space embeddings, counterfactuality, and the use of unless. English
Language and Linguistics 6(2): 347–377.
212 References

2004a. Identity and perspective: the Jekyll-and-Hyde effect in narrative discourse. In

Michael Achard and Suzanne Kemmer (eds.) Language, Culture, and Mind.
Stanford: Center for the Study of Language and Information, 363–375.
2004b. Visual viewpoint, narrative viewpoint, and mental spaces in narrative fiction.
In Augusto Soares da Silva, Amadeus Torres, and Miguel Gonçalves (eds.) Cultura
e cognicão: estudos de linguistica cognitiva (2 vols.) Coimbra: Livraria Almedina,
vol. 2, 347–362.
2005a. Blending and narrative viewpoint: Jonathan Raban’s travels through mental
spaces. Language and Literature 14(2): 99–127.
2005b. Constructional compositionality and blending: the case of Polish SLVF con-
structions. In the Proceedings of the Thirtieth Annual Meeting of the Berkeley
Linguistics Society. Berkeley, Calif., 456–467.
2006. Preface: what can blending do for you? Language and Literature 15(1): 5–15.
2007. Narrative anchors and the processes of story construction: the case of Margaret
Atwood’s The Blind Assassin. Style 41(2): 133–152.
2008a. The text and the story: levels of blending in fictional narratives. In Oakley and
Hougaard (eds.), 51–78.
2008b. Personal pronouns, blending, and narrative viewpoint. In Andrea Tyler,
Yiyoung Kim, and Mari Takada (eds.) Language in the Context of Use: Cognitive
and Functional Approaches to Language and Language Learning. Berlin: Mouton
de Gruyter, 167–182.
2008c. Review of Arie Verhagen’s Constructions of Intersubjectivity: Discourse,
Syntax, and Cognition (Oxford University Press 2005), Linguistics 46(3): 658–667.
2009. Genitives and proper names in constructional blends. In Vyvyan Evans and
Stephanie Pourcel (eds.) New Directions in Cognitive Linguistics. Amsterdam and
Philadelphia: John Benjamins Publishing Company, 161–184.
2010. Close encounters: the author and the character in Elizabeth Costello, Slow Man
and Diary of a Bad Year. In Graham Bradshaw and Michael Neill (eds.) J. M.
Coetzee’s Austerities, Farnham, UK: Ashgate Publishers, 231–252.
in press a. Language, cognition, and literature. In Mark Collard and Edward Slingerland
(eds.) Creating Consilience. Oxford University Press, 409–423.
in press b. Narrative time, sequence, and memory: a blending analysis. In Ralf Schneider
and Marcus Hartner (eds.) Blending and the Study of Narrative. Berlin/New York:
Mouton de Gruyter.
Dancygier, Barbara and Eve Sweetser. 2005. Mental Spaces in Grammar: Conditional
Constructions. Cambridge University Press.
Dancygier, Barbara and Lieven Vandelanotte. 2009. Judging distances: Mental spaces, dis-
tance, and viewpoint in literary discourse. In Geert Brône and Joeren Vandaele (eds.)
Cognitive Poetics: Goals, Gains, and Gaps. Berlin: Mouton de Gruyter, 319–370.
Deacon, Terrence. 1997. The Symbolic Species: The Coevolution of Language and the
Brain. New York: W. W. Norton.
2006. The aesthetic faculty. In Turner (ed.), 21–56.
Decety, Jean and Thierry Chaminade. 2003. Neural correlates of feeling sympathy.
Neuropsychologia 41: 127–128.
Dennett, Daniel. 1997. Consciousness Explained. New York: Little, Brown.
Djikic, Maja, Keith Oatley, Sara Zoeterman, and Jordan B. Peterson. 2009. On being moved
by art: how reading fiction transforms the self. Creativity Research Journal, 21: 24–29.
 References 213

Doležel, Lubomir. 1976. Narrative modalities. Journal of Literary Semantics 5: 5–14.

1998. Heterocosmica: Fiction and Possible Worlds. Baltimore and London: Johns
Hopkins University Press.
Donald, Merlin. 1991. Origins of the Modern Mind: Three Stages in the Coevolution of
Culture and Cognition. Cambridge, Mass.: Harvard University Press.
2001. A Mind So Rare: An Essay on Faculty Psychology. Cambridge, Mass.: MIT
Press.
2006. Art and cognitive evolution. In Turner (ed.), 3–20.
Duchan, Judith, Gail Bruder, and Lynn Hewitt (eds.). 1995. Deixis in Narrative:
A Cognitive Science Perspective. Hillsdale: Lawrence Erlbaum Associates.
Dutton, Denis. 2009. The Art Instinct: Beauty, Pleasure, and Human Evolution. New York:
Bloomsbury Press.
Elam, Keir. 1980. The Semiotics of Theatre and Drama. London/New York: Methuen.
Emmott, Catherine. 1997. Narrative Comprehension: A Discourse Perspective. Oxford:
Clarendon Press.
2002. “Split-selves” in fiction and medical “life stories”: cognitive linguistic theory
and narrative practice. In Semino and Culpeper (eds.), 153–182.
2003. Reading for pleasure: a cognitive poetic analysis of “twists in the tale” and other
plot reversals in narrative texts. In Gavins and Steen (eds.), 145–160.
Emmott, Catherine, Anthony J. Sanford, and Eugene J. Dawydiak. 2007. Stylistics meets
cognitive science: studying style in fiction and reader’s attention from an interdis-
ciplinary perspective. Style 41(2): 204–227.
Evans, Brad. 2007. Introduction: rethinking the disciplinary confluence of anthropology
and literary studies. Criticism 49(4): 429–445.
Evans, Vyvyan. 2004. The Structure of Time: Language, Meaning and Temporal
Cognition. Amsterdam: John Benjamins.
Fauconnier, Gilles. 1994 [1985]. Mental Spaces: Aspects of Meaning Construction in
Natural Language. Cambridge University Press.
1997. Mappings in Thought and Language. Cambridge University Press.
Fauconnier, Gilles and Eve Sweetser (eds.) 1996. Spaces, Worlds, and Grammar.
University of Chicago Press.
Fauconnier, Gilles and Mark Turner. 1994. Conceptual Projection and Middle Spaces.
San Diego: University of California, Department of Cognitive Science Technical
Report 9401.
1996. Blending as a central process of grammar. In Adele Goldberg (ed.) Conceptual
Structure, Discourse, and Language. Stanford: Center for the Study of Language
and Information, 113–130.
1998a. Conceptual integration networks. Cognitive Science 2(22): 133–187.
1998b. Principles of conceptual integration. In Jean-Pierre Koenig (ed.) Discourse
and Cognition: Bridging the Gap. Stanford: Center for the Study of Language and
Information, 269–283.
2002. The Way We Think: Conceptual Blending and the Mind’s Hidden Complexities.
New York: Basic Books.
Feldman, Jerome. 2006. From Molecule to Metaphor: The Neural Theory of Language.
Cambridge, Mass.: MIT Press.
Feldman, Jerome and Srini Narayanan. 2003. Embodied meaning in a neural theory of
language. Brain and Language 89: 385–392.
214 References

Fillmore, Charles J. 1985. Frames and the semantics of understanding. Quaderni di

Semantica 6: 222–254.
1990. Epistemic stance and grammatical form in English conditional sentences. CLS
26: 137–162.
1997. Lectures on Deixis. Stanford, Calif.: CSLI Publications.
2006. Frame semantics. In Dirk Geererts (ed.) Cognitive Linguistics: Basic Readings.
Berlin: Mouton de Gruyter, 373–400.
Fillmore, Charles J. and Paul Kay. 1999. Grammatical constructions and linguistic
generalizations: The What’s X doing Y? construction. Language 75(1): 1–33.
Fillmore, Charles J., Paul Kay, and Mary Catherine O’Connor. 1988. Regularity and
idiomaticity in grammatical constructions: the case of let alone. Language 63(3):
501–538.
Flaherty, Alice W. 2004. The Midnight Disease: The Drive to Write, Writer’s Block, and
the Creative Brain. Boston/New York: Houghton Mifflin Company.
Fleischman, Suzanne. 1989. Temporal distance: a basic linguistic metaphor. Studies in
Language 13(1): 1–50.
1990. Tense and Narrativity: From Medieval Performance to Modern Fiction.
London: Routledge.
Fludernik, Monika. 1993. The Fictions of Language and the Languages of Fiction.
London and New York: Routledge.
1994. Second-person narrative as a test case for narratology: the limits of realism. Style
28(3): 445–480.
1997. Towards a “Natural” Narratology. London and New York: Routledge.
2003. Natural narratology and cognitive parameters. In Herman (ed.), 243–270.
Fodor, Jerry A. 1983. The Modularity of Mind: An Essay on Faculty Psychology.
Cambridge, Mass.: MIT Press.
Frank, Roslyn M., Rene Dirven, Tom Ziemke, and Enrique Bernardez (eds.). 2008. Body,
Language, and Mind (2 vols.), vol. 2: Sociocultural Situatedness. Berlin / New York:
Mouton de Gruyter.
Freeman, Margaret H. 2000. Poetry and the scope of metaphor: toward a cognitive theory
of literature. In Antonio Barcelona (ed.) Metaphor and Metonymy at the
Crossroads: a Cognitive Perspective. Berlin: Mouton de Gruyter, 253–281.
2005. The poem as complex blend: conceptual mappings of metaphor in Sylvia Plath’s
“The Applicant.” Language and Literature 14(1): 25–44.
Galbraith, Mary. 1995. Deictic shift theory and the poetics of involvement in narrative. In
Duchan, Bruder, and Hewitt (eds.), 19–59.
Gallagher, Shaun. 2005. How the Body Shapes the Mind. New York: Oxford University
Press.
Gallagher, Shaun and Daniel D. Hutto. 2008. Understanding others through primary
interaction. In Zlatev et al. (eds.),17–38.
Gallese, Vittorio, and George Lakoff. 2005. The brain’s concepts: the role of the sensori-
motor system in conceptual knowledge. Cognitive Neuropsychology 22: 455–479.
Gavins, Joanna. 2003. Too much blague? an exploration of the text worlds of Donald
Barthelme’s Snow White. In Gavins and Steen (eds.), 129–144.
2007. Text World Theory: An Introduction. Edinburgh University Press.
Gavins, Joanna and Gerard Steen (eds.) 2003. Cognitive Poetics in Practice. London and
New York: Routledge.
 References 215

Genette, Gérard. 1980. Narrative Discourse: An Essay in Method. Translated by

J. E. Lewin. Ithaca: Cornell University Press.
Gerrig, Richard. 1993. Experiencing Narrative Worlds. New Haven: Yale University Press.
Gerrig, Richard and Giovanni Egidi. 2003. Cognitive psychological foundations of
narrative experiences. In Herman (ed.), 33–55.
Gibbs, Raymond W. Jr. 1994. The Poetics of Mind: Figurative Thought, Language, and
Understanding. New York: Cambridge University Press (also translated into Italian,
Korean, Spanish, and Japanese).
1999. Intentions in the Experience of Meaning. New York: Cambridge University
Press.
2000. Making good psychology out of blending theory. Cognitive Linguistics 11(3/4):
347–358.
2005a. Embodiment and Cognitive Science. Cambridge University Press.
2005b. The psychological status of image schemas. In Hampe (ed.), 113–136.
Gibbs, Raymond W. Jr. and Teenie Matlock. 1999. Psycholinguistics and mental repre-
sentations. Cognitive Linguistics 10(3): 263–270.
Goldberg, Adele E. 1995. Constructions: A Construction Grammar Approach to
Argument Structure. University of Chicago Press.
Gottschall, Jonathan. 2008a. The Rape of Troy: Evolution, Violence and the World of
Homer. Cambridge University Press.
2008b. Literature, Science, and a New Humanities. New York: Palgrave Macmillan.
Gottschall, Jonathan and David Sloan Wilson (eds.). 2005. The Literary Animal: Evolution
and the Nature of Narrative. Evanston, Ill.: Northwestern University Press.
Graver, David. 1997. The actor’s bodies. Text and Performance Quarterly 17(3):
221–235.
Haiman, John. 1980. The iconicity of grammar: isomorphism and motivation. Language
56: 515–540.
Hamilton, Craig. 2002. Conceptual integration in Christine de Pizan’s City of Ladies. In
Semino and Culpeper (eds.), 1–22.
Hampe, Beate (ed.). 2005. From Perception to Meaning: Image Schemas in Cognitive
Linguistics. Berlin: Mouton de Gruyter.
Hanks, William F. 1990. Referential Practice: Language and Lived Space among the
Maya. Chicago University Press.
Harding, Jennifer Riddle. 2004. Simple regrets: counterfactuals and the dialogic mind.
PhD dissertation, University of Maryland.
2007. Evaluative stance and counterfactuals in language and literature. Language and
Literature 16(3): 263–280.
Harnad, S. 1990. The symbol grounding problem. Physica 42: 335–346.
Harris, Jonathan and Natasha Korda (eds.). 2002. Staged Properties in Early Modern
English Drama. Cambridge University Press.
Hart, Elizabeth and Bruce McConachie (eds.). 2006. Performance and Cognition:
Theatre Studies and the Cognitive Turn. New York: Routledge.
Herman, David. 2002. Story Logic: Problems and Possibilities of Narrative. Lincoln,
Nebr., and London: University of Nebraska Press.
2003a. Stories as a tool for thinking. In Herman (ed.), 163–193.
(ed.) 2003b. Narrative Theory and the Cognitive Sciences. Stanford: Center for the
Study of Language and Information.
216 References

2006. Genette meets Vygotsky: narrative embedding and distributed intelligence.

Language and Literature 15(4): 357–380.
(ed.) 2007. The Cambridge Companion to Narrative. Cambridge University Press.
Hiraga, Masako. 2005. Metaphor and Iconicity. New York: Palgrave Macmillan.
Hogan, Patrick Colm. 2003. The Mind and Its Stories: Narrative Universals and Human
Emotion. Cambridge University Press.
Humphrey, Robert. 1954. Stream of Consciousness in the Modern Novel. Berkeley:
University of California Press.
Hutchins, Edwin. 1995. Cognition in the Wild. Cambridge, Mass.: MIT Press.
2005. Material anchors for conceptual blends. Journal of Pragmatics 37:
1,555–1,577.
Hutto, Daniel D. (ed.). 2007a. Narrative and Understanding Persons. Cambridge
University Press.
2007b. The Narrative Practice Hypothesis: origins and applications of folk psychol-
ogy. In Hutto (ed.), 43–68.
Iser, Wolfgang. 1974. The Implied Reader: Patterns of Communication in Prose Fiction
from Bunyan to Beckett. Baltimore: Johns Hopkins University Press.
Jackendoff, Ray. 1975. On belief contexts. Linguistic Inquiry 6(1).
Jackson, Tony E. 2005. Explanation, interpretation, and close reading: the progress of
cognitive poetics. Poetics Today 26(3): 519–532.
Johnson, Mark. 1987. The Body in the Mind: The Bodily Basis of Meaning, Imagination,
and Reason. Chicago University Press.
2007. The Meaning of the Body: Aesthetics of Human Understanding. Chicago
University Press.
Klein, Kitty. 2003. Narrative construction, cognitive processing, and health. In Herman
(ed.), 56–84.
Labov, William. 1972. Language in the Inner City. University of Philadelphia Press.
Labov, William and Joshua Waletzky. 1967. Narrative analysis: oral versions of personal
experience. In June Helm (ed.) Essays on the Verbal and Visual Arts. Seattle:
University of Washington Press, 12–44.
Lakoff, George. 1996. Sorry, I’m not myself today: the metaphor system for conceptu-
alizing self. In Fauconnier and Sweetser (eds.), 91–123.
2008. The Neural Theory of Metaphor. In Raymond Gibbs (ed.) Cambridge Handbook
of Metaphor and Thought. New York: Cambridge University Press, 17–38.
Lakoff, George and Mark Johnson. 1980. Metaphors We Live By. University of Chicago
Press.
1999. Philosophy in the Flesh: The Embodied Mind and Its Challenge to Western
Thought. New York: Basic Books.
Lakoff, George and Mark Turner. 1989. More than Cool Reason: A Field Guide to Poetic
Metaphor. University of Chicago Press.
Langacker, Ronald W. 1990a. Concept, Image, Symbol: The Cognitive Basis of Grammar.
Berlin: Mouton de Gruyter.
1990b. Subjectification. Cognitive Linguistics 1: 5–38.
1991 [1987]. Foundations of Cognitive Grammar, 2 vols. Stanford University Press.
2001. Discourse in cognitive grammar. Cognitive Linguistics 12: 143–188.
2005. Dynamicity, fictivity, and scanning: The imaginative basis of logic and linguis-
tic meaning. In Pecher and Zwaan (eds.), 164–198.
 References 217

Lanser, Susan Sniader. 1981. The Narrative Act: Point of View in Fiction. Princeton
University Press.
Leech, Geoffrey N. and Michael H. Short. 1981. Style in Fiction: A Linguistic
Introduction to English Fictional Prose. London: Longman.
Lehrer, Adrienne and Eva Feder Kittay (eds.). 1992. Frames, Fields, and Contrasts: New
Essays in Semantic and Lexical Organization. Hillsdale, NJ: L. Erlbaum Associates.
Lermitte, Janet. 2010. Faith and art, frames and anchors: Annie Dillard’s Holy the Firm.
Paper presented at the PALA Conference in Genoa, Italy.
Lupton, Christina. 2006. The knowing book: authors, it-narratives, and objectification in
the eighteenth century. Novel 39(3): 402–420.
Lyons, John. 1977. Semantics. Cambridge University Press.
Maitre, Doreen. 1983. Literature and Possible Worlds. London: Middlesex Polytechnic
Press.
Mar, Raymond A. and Keith Oatley. 2008. The function of fiction is the abstraction and
simulation of social experience. Perspectives on Psychological Science 3: 173–192.
Margolin, Uri. 1996. Telling our story: on “we” literary narratives. Language and
Literature 5(2): 115–133.
2003. Cognitive science, the thinking mind, and literary narrative. In Herman (ed.),
271–294.
Matlock, Teenie. 2004. Fictive motion as cognitive simulation. Memory and Cognition
32: 1,389–1,400.
McGregor, William B. 1997. Semiotic Grammar. Oxford: Clarendon Press.
McNeill, David (ed.). 2000. Language and Gesture. Cambridge University Press.
Michaelis, L. 2003. Word meaning, sentence meaning, and syntactic meaning. In
Hubert Cuyckens, Rene Dirven, and John Taylor (eds.) Cognitive Approaches to
Lexical Semantics. Berlin/New York: Mouton de Gruyter, 163–210.
Moder, Carol. 2010. Two puzzle pieces: fitting discourse context and constructions into
Cognitive Metaphor Theory. English Text Construction 3(2): 294–320.
Moschovakis, Nicholas. 2006. Topicality and conceptual blending: Titus Andronicus and
the case of William Hacket. College Literature 33(1): 127–149.
Moss, John and Tobi Kozakiewich (eds.). 2006. Margaret Atwood: The Open Eye.
University of Ottawa Press.
Murakami, Haruki and Philip Gabriel. 2008. What I Talk about When I Talk about
Running. New York: Alfred Knopf/Random House.
Nikiforidou, Kiki. 2010. Viewpoint and construction grammar: The case of past+now.
Language and Literature 19(3): 265–284.
Noh, Eun-Ju. 2000. Metarepresentation: A Relevance-Theory Approach. Amsterdam:
John Benjamins.
Nuñez, Raphael E. and Eve Sweetser. 2006. With the future behind them: convergent
evidence from Aymara language and gesture in the crosslinguistic comparison of
spatial construals of time. Cognitive Science 30: 401–450.
Oakley, Todd. 1998. Conceptual blending, narrative discourse, and rhetoric. Cognitive
Linguistics 9(4): 321–360.
2009. From Attention to Meaning: Explorations in Semiotics, Linguistics, and
Rhetoric. Berne: Peter Lang AG.
Oakley, Todd and Anders Hougaard. 2008. Mental Spaces in Discourse and Interaction.
Amsterdam and Philadelphia: John Benjamins.
218 References

Palmer, Alan. 2003. The mind beyond the skin. In Herman (ed.), 322–348.
2004. Fictional Minds. Lincoln, Nebr., and London: University of Nebraska Press.
Panther, Klaus-Uwe. 2006. Metonymy as a usage event. In Gitte Kristiansen,
Michel Achard, Rene Dirven, and Francisco Ruiz de Mendoza Ibañez (eds.)
Cognitive Linguistics: Current Applications and Future Perspectives. Berlin and
New York: Mouton de Gruyter, 147–186.
Pascual, Esther. 2002. Imaginary Trialogues: Conceptual Blending and Fictive
Interaction in Criminal Courts. Utrecht: LOT.
2006a. Fictive interaction within the sentence. Cognitive Linguistics 17(2): 245–267.
2006b. Questions in legal monologues: fictive interaction as argumentative strategy in
a murder trial. Text & Talk 26(3): 383–402.
2008. Fictive interaction blends in everyday life and courtroom settings. In
Anders Hougaard and Todd Oakley (eds.) Mental Spaces in Discourse and
Interaction. Amsterdam: John Benjamins, 79–107.
Pavel, Thomas G. 1986. Fictional Worlds. Cambridge, Mass., and London: Harvard
University Press.
Pecher, Diane and Rolf Zwaan. 2005a. Introduction. In Pecher and Zwaan (eds.), 1–8.
(eds.) 2005b. Grounding Cognition: The Role of Perception and Action in Memory,
Language, and Thinking. Cambridge University Press.
Perrakis, Phyllis. 2006. Negotiating with the looking glass: Atwood, her protagonists,
and the journey to the dead. In Moss and Kozakiewich (eds.), 349–359.
Pinker, Steven. 1994. The Language Instinct. New York: William Morrow and Company.
2002. The Blank Slate: The Modern Denial of Human Nature. New York: Penguin.
Pires de Oliveira, Roberta, and Robson de Souza Bittencourt. 2008. An interview with
Mark Johnson and Tim Rohrer: from neurons to sociocultural situatedness. In Frank
et al. (eds.), 21–52.
Porter Abbott H. 2007. Story, plot, and narration. In Herman (ed.), 39–52.
Pratt, Mary Louise. 1977. Toward a Speech Act Theory of Literary Discourse.
Bloomington: Indiana University Press.
Prince, Gerald. 2001. A point of view on point of view or refocusing focalization. In Van
Peer and Chatman (eds.), 43–64.
Prinz, Jesse J. 2002. Furnishing the Mind: Concepts and Their Perceptual Basis.
Cambridge, Mass.: MIT Press.
2004. Gut Reactions: A Perceptual Theory of Emotion. New York: Oxford University Press.
2005. Passionate thoughts: the emotional embodiment of moral concepts. In Pecher
and Zwaan (eds.), 93–114.
Propp, Vladimir. 1968 [1928]. Morphology of the Folktale. Translated by Laurence
Scott. Austin: University of Texas Press.
Rauh, Gisa. 1981. Aspects of deixis. In Gisa Rauh (ed.) Essays on Deixis. Tübingen:
Gunter Narr Verlag, 9–60.
Reddy, Michael. 1979. The conduit metaphor – a case of frame conflict in our language
about language. In Andrew Ortony (ed.) Metaphor and Thought. Cambridge
University Press, 284–297.
Richerson, Peter J. and Robert Boyd. 2005. Not by Genes Alone: How Culture
Transforms Human Evolution. Chicago University Press.
Rimmon-Kenan, Shlomith. 1983. Narrative Fiction: Contemporary Poetics. London:
Methuen.
 References 219

Ritchie, David. 2004. Lost in “conceptual space”: metaphors of conceptual integration.

Metaphor and Symbol 19(1): 31–50.
Rohrer, Tim. 2005. Mimesis, artistic inspiration, and the blends we live by. Journal of
Pragmatics 37: 1,686–1,716.
Ronen, Ruth. 1994. Possible Worlds in Literary Theory. Cambridge University Press.
Roy, Wendy. 2006. The body of/as evidence: Margaret Atwood, The Blind Assassin, and
the feminist literary mystery. In Moss and Kozakiewich (eds.), 361–371.
Rubba, Jo. 1996. Alternate grounds in the interpretation of deictic expressions. In
Fauconnier and Sweetser (eds.), 227–261.
2009. The dream as a blend in David Lynch’s Mulholland Drive. In Vyvyan Evans,
and Stephanie Pourcel (eds.) New Directions in Cognitive Linguistics. Amsterdam
and Philadelphia: John Benjamins Publishing Company, 465–498.
Ryan, Marie-Laure. 1991. Possible Worlds, Artificial Intelligence, and Narrative Theory.
Bloomington and Indianapolis: Indiana University Press.
2007. Toward a definition of narrative. In Herman (ed.), 22–38.
Sanchez Garcia, Jesus. 2003. Language and cognition: George Lakoff on some internal
and external complexities. An interview. In Francisco J. Ruiz de Mendoza (ed.),
Annual Review of Cognitive Linguistics 1: 233–267.
Sanders, José and Gisela Redeker. 1996. Perspective and representation in speech and
thought in narrative discourse. In Fauconnier and Sweetser (eds.), 290–317.
Sanders, Ted, José Sanders and Eve Sweetser. 2009. Causality, cognition, and communica-
tion: a mental space analysis of subjectivity in causal connectives. In Ted Sanders and
Eve Sweetser (eds.) Causal Categories in Discourse and Cognition. Berlin/New York:
Mouton de Gruyter, 19–60.
Sanger, Keith. 2001. The Language of Drama. London: Routledge.
Scalise Sugiyama, Michelle. 2005. Reverse-engineering narrative: evidence of special
design. In Gottschall and Wilson (eds.), 177–195.
Scalise Sugiyama, Michelle and Lawrence S. Sugiyama (in press). “‘Once the child is
lost he dies’”: Monster stories vis-à-vis the problem of errant children. In
Mark Collard and Edward Slingerland (eds.) Creating Consilience. Oxford
University Press, 363–371.
Schnepf, Jennifer. 2006. Role-value mappings and narrative structure in Atwood’s The
Blind Assassin. Paper presented at the Literature and Cognitive Science Conference,
Storrs, Conn., April 6–9, 2006.
Semino, Elena. 1997. Language and World Creation in Poems and Other Texts. London:
Longman.
2002. A cognitive stylistic approach to mind style in narrative fiction. In Semino and
Culpeper (eds.), 95–122.
2003. Discourse worlds and mental spaces. In Gavins and Steen (eds.), 83–98.
2006. Blending and characters’ mental functioning in Virginia Woolf’s Lappin and
Lapinova. Language and Literature 15(1): 55–72.
2007. Mind style 25 years on. Style 41(2): 153–203.
2008. Metaphor in Discourse. Cambridge University Press.
2009. Text worlds. In Geert Brône and Joeren Vandaele (eds.) Cognitive Poetics:
Goals, Gains, and Gaps. Berlin/New York: Mouton de Gruyter, 33–72.
2010. Unrealistic scenarios, metaphorical blends, and rhetorical strategies across
genres. English Text Construction 3(2): 250–274.
220 References

Semino, Elena and Jonathan Culpeper (eds.). 2002. Cognitive Stylistics: Language and
Cognition in Text Analysis. Amsterdam: John Benjamins.
Semino, Elena and Michael Short. 2004. Corpus Stylistics: Speech, Writing, and
Thought Presentation in a Corpus of English Writing. London: Routledge.
Sharifian, Farzad. 2008. Distributed, emergent cultural cognition, conceptualization and
language. In Frank et al. (eds.), 109–136.
Short, Mick. 2007. Thought presentation twenty-five years on. Style 41(2): 227–240.
Simpson, Paul. 1993. Language, Ideology, and Point of View. London and New York:
Routledge.
Slingerland, Edward. 2005. Conceptual blending, somatic marking, and normativity: a
case example from ancient Chinese. Cognitive Linguistics 16(3): 557–584.
2006. What Science Offers the Humanities: Integrating Body and Culture. Cambridge
and New York: Cambridge University Press.
Sofer, Andrew. 2003. The Stage Life of Props. Ann Arbor: University of Michigan
Press.
Spivey, Michael J. 2007. The Continuity of Mind. New York: Oxford University Press.
Spivey, Michael J., Daniel C. Richardson, and Monica Gonzalez-Marquez. 2005. On the
spatial and image-schematic underpinnings of real-time language processing. In
Pecher and Zwaan (eds.), 246–281.
Stockwell, Peter. 2002. Cognitive Poetics: An Introduction. London: Routledge.
Style. 2008. Style, 42 (2/3), special issue on literary Darwinism.
Surma, Szymon. 2009. The significance of ghosts in nō and kabuki: an analysis of stories
and forms. Unpublished manuscript.
Sweetser, Eve. 1990. From Etymology to Pragmatics. Cambridge University Press.
1996. Mental spaces and the grammar of conditional constructions. In Fauconnier and
Sweetser (eds.), 318–333.
1999. Compositionality and blending: semantic composition in a cognitively realistic
framework. In Theo Janssen and Gisela Redeker (eds.) Cognitive Linguistics:
Foundations, Scope, and Methodology. Berlin: Mouton de Gruyter, 129–162.
2000. Blended spaces and performativity. Cognitive Linguistics 11(3–4): 305–334.
2004. “The suburbs of your good pleasure”: cognition, culture, and the bases of
metaphoric structure. The Shakespearean International Yearbook 4, 24–55.
2006. Whose rhyme is whose reason? Sound and sense in Cyrano de Bergerac.
Language and Literature 15(1): 29–54.
Talmy, Leonard. 1988. Force dynamics in language and thought. Cognitive Science 12:
49–100.
1996. Fictive motion in language and “ception”. In Paul Bloom, Mary A. Peterson,
Lynn Nadel, and Merrill F. Garrett (eds.) Language and Space. Cambridge, Mass.:
MIT Press, 211–276.
2000. Towards a Cognitive Semantics (vols. 1 & 2). Cambridge, Mass.: MIT Press.
Tobin, Vera. 2006. Ways of reading Sherlock Holmes. Language and Literature 16(1)
73–90.
2010. Joint attention, To the Lighthouse, and modernist representations of intersub-
jectivity. English Text Construction 3(2): 185–202.
Tomasello, Michael. 1999. The Cultural Origins of Human Cognition. Cambridge,
Mass.: Harvard University Press.
2008. Origins of Human Communication. Cambridge, Mass.: MIT Press.
 References 221

Tooby, John and Leda Cosmides (in press). Evolutionary Psychology: Foundational
Papers. Cambridge, Mass.: MIT Press.
Toolan, Michael. 2001. Narrative: a Critical Linguistic Introduction. (2nd edn.) London
and New York: Routledge.
Tsur, Reuven. 1992. Toward a Theory of Cognitive Poetics. Amsterdam: Elsevier.
2003. On the Shore of Nothingness: A Study in Cognitive Poetics. Charlottesville:
Imprint Academic.
Turner, M. 1991. Reading Minds: The Study of English in the Age of Cognitive Science.
Princeton University Press.
1996. The Literary Mind: The Origins of Thought and Language. New York: Oxford
University Press.
2003. Double-scope stories. In Herman (ed.), 117–142.
2004. The ghost of anyone’s father. The Shakespearean International Yearbook 4:
72–97.
2006a. Compression and representation. Language and Literature 15(1): 17–29.
(ed.) 2006b. The Artful Mind: Cognitive Science and the Riddle of Human Creativity.
Oxford University Press.
Turner, Mark and Gilles Fauconnier. 1995. Conceptual integration and formal expres-
sion. Metaphor and Symbolic Activity 10(3): 183–203.
1999. A mechanism of creativity. Poetics Today 20(3): 397–418.
Vandelanotte, Lieven. 2002. But forced to qualify: distancing speech and thought
representation as a symptom of uninformedness in Larkin. Leuvense Bijdragen
(Leuven Contributions in Linguistics and Philology) 91(3–4): 383–426.
2004a. Deixis and grounding in speech and thought representation. Journal of
Pragmatics 36(3): 489–520.
2004b. From representational to scopal “distancing indirect speech or thought”: a cline
of subjectification. Text 24(4): 547–585.
2005. Types of speech and thought representation in English: syntagmatic structure,
deixis and expressivity, semantics. PhD dissertation. University of Leuven.
2009. Speech and Thought Representation in English: A Cognitive-Functional
Approach. Berlin: Mouton de Gruyter.
2010. “Where am I, lurking in what place of vantage?” The discourse of distance in
John Banville’s fiction. English Text Construction 3(2) 203–225.
Van Hoek, Karen. 1997. Anaphora and Conceptual Structure. University of Chicago
Press.
Van Oort, Richard. 2003. Cognitive science and the problem of representation. Poetics
Today 24: 237–295.
2007. The culture of criticism. Criticism 49(4): 459–479.
Van Peer, Willie and Seymour Chatman (eds.). 2001. New Perspectives on Narrative
Perspective. Albany: State University of New York Press.
Van Vliet, Sarah. 2008. Proper nouns and pronouns: the production of referential
expressions in narrative discourse. PhD dissertation. Publications of the
Netherlands Graduate School of Linguistics.
Vargas Llosa, Mario. 1997/2003 [1997]. Letters to a Young Novelist. Translated by
Natasha Wimmer. New York: Picador Books.
Verhagen, Arie. 2005. Constructions of Intersubjectivity: Discourse, Syntax, and
Cognition. Oxford University Press.
222 References

2008. Intersubjectivity and the architecture of the language system. In Zlatev et al.,
(eds.), 307–332.
Wales, Katie. 1996. Personal Pronouns in Present Day English. Cambridge University
Press.
Weber, Jean Jacques. 2004. A new paradigm for literary studies, or: the teething troubles
of cognitive poetics. Style 38(4): 515–523.
Weimann, Robert. 2000. Author’s Pen and Actor’s Voice: Playing and Writing in
Shakespeare’s Theatre. Cambridge University Press.
Werth, Paul. 1999. Text Worlds: Representing Conceptual Space in Discourse. London:
Longman.
Wetmore, Kevin J., Jr. 2008. “Avenge me!”: ghosts in English Renaissance and Kabuki
revenge dramas. In Wetmore (ed.) Revenge Drama in European Renaissance and
Japanese Theatre: From Hamlet to Madame Butterfly. New York: Palgrave
Macmillan, 75–90.
Wiebe, Janyce. 1995. References in narrative text. In Duchan et al. (eds.), 263–285.
Wilson, Sharon. 2006. Blindness and survival in Margaret Atwood’s major novels. In
C. A. Howells (ed.) The Cambridge Companion to Margaret Atwood. Cambridge
University Press, 176–190.
Yachnin, Paul. 2002. Wonder-effects: Othello’s handkerchief. In Harris and Korda (eds.),
316–334.
Zbikowski, Lawrence. 2002. Conceptualizing Music. New York: Oxford University Press.
2006. The cognitive tango: music, mind, and brain. In Turner (ed.), 115–132.
Zlatev, Jordan. 2008. The co-evolution of intersubjectivity and bodily mimesis. In Zlatev
et al., (eds.), 215–244.
Zlatev, Jordan, Timothy P. Racine, Chris Sinha, and Esa Itkonen (eds.). 2008. The Shared
Mind: Perspectives on Intersubjectivity. Amsterdam and Philadelphia: John
Benjamins Publishing Company.
Zunshine, Lisa. 2006. Why We Read Fiction: Theory of Mind and the Novel. Columbus:
Ohio State University Press.
 Literary works cited

Amis, Martin. 1991. Time’s Arrow. New York: Vintage International.

Atwood, Margaret. 2000. The Blind Assassin. Toronto: McClelland and Stewart.
Austen, Jane. 2009. Pride and Prejudice. London: Penguin Classics.
Ballard, J. G. 1984. Empire of the Sun. New York: Simon and Schuster.
1994/1995. Rushing to Paradise. London: Flamingo.
2008. The Miracles of Life. London: Fourth Estate.
Bulgakov, Mikhail. 1966/1997. The Master and Margarita. Translated by Richard Pevear
and Larissa Volokhonsky. London: Penguin Books.
Calvino, Italo. 1979/1981. If on a Winter’s Night a Traveler. Translated by William Weaver.
Toronto, Canada: Lester & Orpen Dennys.
Chatwin, Bruce. 1988. Utz. New York: Penguin Books.
Coetzee, J. M. 2003. Elizabeth Costello. London: Secker and Warburg.
2005. Slow Man. London: Secker and Warburg.
2007. Diary of a Bad Year. London: Harvill Secker.
Dillard, Annie. 1977/1988. Holy the Firm. New York: Harper Perennial.
Eggers, Dave. 2000. A Heartbreaking Work of Staggering Genius. Toronto: Vintage
Canada.
Fallaci, Oriana. 1979/1980. The Man. Translated by William Weaver. New York: Simon
and Schuster.
Fielding, Henry. 2000. The History of Tom Jones. Cologne: Köneman.
Fitzgerald, F. Scott. 2004 [1925]. The Great Gatsby. New York: Scribner.
García Márquez, Gabriel. 2006 [1970]. One Hundred Years of Solitude. Translated by
Gregory Rabassa. New York: Harper Perennial.
Garland, Alex. 1999. The Tesseract. New York: Riverhead Books.
Grass, Günter. 2009. The Tin Drum. Translated by Breon Mitchell. Boston/New York:
Houghton Mifflin Harcourt.
James, Henry. 1963 [1881]. The Portrait of a Lady. London: Penguin Books.
Joyce, James. 2000. Dubliners. Oxford University Press.
Marlowe, Christopher. 2007. Dr Faustus (edited by Michael Keefer). Toronto:
Broadview Editions.
McEwan, Ian. 2001/2002. Atonement. Toronto: Vintage Canada.
Melville, Herman. 1998 [1851]. Moby-Dick. New York: New American Library.
Miller, Henry. 1941/1958 [1941]. The Colossus of Maroussi. New York: New Directions
Books.
Murakami, Haruki. 2007/2008. After Dark. Translated by Jay Rubin. Toronto: Vintage
Canada
Nabokov, Vladimir. 1991[1955]. Lolita (annotated edition). New York: Vintage Books.

223
224 Literary works cited

Paretsky, Sara. 1999. Hard Time. New York: Dell Publishing.

Philip, Neil. 1994. The Arabian Nights. New York: Orchard Books.
Potocki, Jan. 1995. The Manuscript Found in Saragossa. Translated by Ian Maclean.
London: Penguin Books.
Powers, Richard. 2009. Generosity: An Enhancement. New York: Farrar, Straus and
Giroux.
Proulx, Annie. 1993. The Shipping News. New York: Scribner.
1985. Foreign Land. London: Picador Books.
1986. Coasting. London: Picador/Pan Books.
1987 [1979]. Arabia through the Looking Glass. London: Picador Books.
1991. Hunting Mister Heartbreak. New York: Harper Collins.
Raban, Jonathan. 1998 [1981]. Old Glory: An American Voyage. New York: Vintage
Books.
2000 [1999]. Passage to Juneau. New York: Vintage Books.
Richardson, Samuel. 1747/1985. Clarissa. London: Penguin Books.
de Riojas, Fernando. 1962. The Celestina. Translated by Lesley Byrd Simpson. Berkeley
and Los Angeles: University of California Press.
Roth, Philip. 1997/1998. American Pastoral. New York: Vintage International.
2000. The Human Stain. New York: Vintage International.
Rushdie, Salman. 1980. Midnight’s Children. New York: Penguin Books.
Shakespeare, William. 1994. The Complete Works. The Shakespeare Head Press Oxford
Edition. New York: Barnes & Noble Books. (Macbeth, Romeo and Juliet, Richard
II, Othello)
Shields, Carol. 1997. Larry’s Party. Viking Penguin.
Stephenson, Neal. 1992/1993. Snow Crash. New York: Bantam Books.
Stoppard, Tom. 1993. Arcadia. London/New York: Faber and Faber.
Szymborska, Wisława. 1997. Nothing Twice: Selected Poems / Nic Dwa Razy: Wybór
Wierszy. Translated by Stanisław Barańczak and Clare Cavanagh. Cracow:
Wydawnictwo Literackie.
Theroux, Paul. 1978. The Picture Palace. New York: Penguin Books.
Tolkien, J. R. R. 1954. The Fellowship of the Ring. Boston: Houghton Mifflin Company.
Tyler, Anne. 1995. Ladder of Years. New York: Fawcett Columbine.
1998. A Patchwork Planet. Toronto: Penguin Books.
Woolf, Virginia. 1953 [1925]. Mrs. Dalloway. New York: Harcourt Brace Jovanovich
Publisher.
 Index

Acting/Speaking blend, 156, 158 decompression for viewpoint, 100–102, 118,

agency, 137, 163 120, 133
Atwood, Margaret, 10, 40, 50 default communicative context, 19–21, 31, 65,
authorship, 48, 58, 79, 82, 83, 84, 135 129, 139–141
deictic ground, 20, 31, 64, 68, 76, 117, 136,
Bakhtin, Mikhail, 165 139–140, 165, 183, 199
Ballard, J. G., 22, 23, 128 Deictic Shift Theory, 129, 140
blending, 7, 31–35, 38, 56, 189, 201 deixis, See deictic ground
compression, 18, 59, 70, 93 Dennett, Daniel, 20
optimality constraints, 57 Dillard, Annie, 43
selection, 58 discourse blends, 182
vital relations, 56 DIST, See speech and thought representation
vs embedding, 190 distanced discourse, 193
Body/Mind blend, 152, 158, 159 distributed cognition, 14, 18, 109
Booth, Wayne, 74, 87 Donald, Merlin, 14, 16, 28
Boyd, Brian, 21 drama, 162–163
Bulgakov, Mikhail, 65, 81, 83, 86 and blending, 155–159
and mental spaces, 141–145
Calvino, Italo, 76 and supernatural phenomena on stage,
Canning, Patricia, 9 160–164
Carroll, Joseph, 27 and traditional Japanese theatre, 162–163
Chatman, Seymour, 53 body posture, 149–152
Chatwin, Bruce, 60 discourse of, 152–160
Chekhov, Anton, 143 narration, 147–148
Cienki, Alan, 151 spatial construals, 142–144, 148–152
Clark, Herbert, 177 vs early novels, 164–168
Coetzee, J. M., 79, 80, 119, 168 dual voice, 88, 111
cognitive linguistics, 6, 31, 195
conceptual distance, 72–73, 74, 75, 99 Eggers, Dave, 38, 60, 62, 73, 74, 173,
conceptual integration, See blending 177, 186
constructional compositionality, 8, 10, 33, embodiment, 25, 26, 29, 113, 151, 158,
75–76, 193, 198 168–169
constructions, 20, 88, 105, 121, 140, 166, emergence of literature, 16–19
194, 198 Emmott, Catherine, 30, 51, 54
vocative-cum-imperative, 153–160, 198 epistemic stance, 22, 55, 89, 90
cross-input projection, 40, 42, 43, 44, 46, 52 epistemic viewpoint, 48, 59, 61, 62, 65, 70,
Currie, Gregory, 29, 115 85, 132
evidentiality, 17, 22
Damasio, Antonio, 14 evocation, 44, 47, 49, 99, 124, 143, 146
de Riojas, Fernando, 165 evolutionary psychology, 11, 13, 28, 109
Deacon, Terrence, 7, 14, 24, 28, 50 eye contact, 144

225
226 Index

fact vs fiction, 22–23 Literary Darwinists, 13

Fallaci, Oriana, 76 literary interpretation, 9–11
Fauconnier, Gilles, 31, 33, 201 literature and art, 14–16
Faulkner, William, 74 Lupton, Christina, 166
Feeling/Speaking blend, 157
fictive interaction, 182 Margolin, Uri, 130
fictive motion, 103, 117, 128 Marlowe, Christopher, 159
fictive vision, 102–103, 104, 107, 131, 201 material anchors, 158, 163, 166
Fielding, Henry, 64, 191 material objects, 136–138, 146–147, 152–161,
Fillmore, Charles, 32 166, 197
FIST, See speech and thought representation McEwan, Ian, 84
flashbacks, 6, 37, 54, 63, 100, 189 meaning construction, 4–6, 10, 25, 35, 145, 202
Fludernik, Monika, 6, 78, 88 Melville, Herman, 166, 172, 197
focalization, 87, 108 mental spaces, 32, 35, 36, 39, 94, 141–145
frame evocation, 46, 157 and theatre, 141–145
frame metonymy, 33, 50, 177 metaphor, 149, 151, 153, 155, 158, 164
frames, 32–33, 35, 43, 48, 50, 51, 52, 54, 101, metaphors of time, 55
120, 121, 135, 138, 143, 156, 201 metonymy, 149, 150, 151, 162, 168
micro/macro viewpoint, 62–63, 87, 133
Gallagher, Shaun, 113 Miller, Henry, 187
García Márquez, Gabriel, 119 mind style, See narrator, unreliable
Garland, Alex, 40, 107 mirror neurons, 113
Gavins, Joanna, 36 Moder, Carol, 199
Genette, Gérard, 60, 87, 91 Monterroso, Augusto, 37
Gerrig, Richard, 18, 54, 177 Murakami, Haruki, 21, 131
gesture, 149, 151, 155, 158 mythic past, 17
Gibbs, Raymond, 59
Grass, Günter, 132, 168 Nabokov, Vladimir, 119
Graver, David, 147 narratee, 68, 76
Ground, 19 narration, 53, 63, 64, 69, 73, 75, 111, 167,
grounding problem, 24, 25 191, 200
narrative anchors, 42–50, 82, 144, 147, 158,
Herman, David, 6, 15, 23, 24, 27, 53, 163, 166, 197
54, 109 narrative and cognition, 23–27
Hutchins, Edwin, 158 narrative and evolution, 27–28
Hutto, Daniel, 113, 114 Narrative Practice Hypothesis, 114, 115
narrative spaces, 36–37, 38, 40–42, 51, 53, 59,
image schemas, 24, 26, 27, 28 61, 63, 73, 74, 76, 81, 83, 84, 86, 93, 127,
imperative, 154, 193 136, 140, 164, 172, 173, 188, 192
intentional stance, 20 and meta-representation, 108–112
internal monologue, 67, 133, 167, 179 embedding, 84–86, 175, 178–183, 190
intersubjectivity, 112–116, 140, 141, types and relations, 38, 63–67
164, 200 narrative transport, 18
Iser, Wolfgang, 58 narrative uptake, 59
it-narratives, 166 narrative viewpoint, 49, 61, 63, 67, 81, 84, 86,
88, 93, 115, 196
James, Henry, 66 experiential viewpoint, 103–106, 111, 128
Johnson, Mark, 55, 149 visual viewpoint, 91–94, 128, 131, 161
joint attention, 164 narrator
Joyce, James, 191 as character, 60, 62, 63, 71, 78, 132, 135, 186
as epistemic viewpoint, 61, 64–70, 136
Labov, William, 56 intrusive, 64, 68
Lakoff, George, 55, 100, 149 multiple narrators, 80–84
Lanser, Susan, 115 off-stage, 66, 69, 174, 188
Leech, Geoffrey, 103 omniscient, 64, 66
 Index 227

on-stage, 65, 166, 167, 174, 180, 186, 189 Semino, Elena, 74, 103, 199
types of, 60–67 sentences as spaces, 37–39
unreliable, 61, 73 sequence of events, 37, 51, 53, 54–56
narratorship, 21, 31, 48, 58, 59, 74, 80, 82, Shakespeare, William, 148, 152
86, 133 Shared Mind theory, 112–115
natural narratology, 6 Shields, Carol, 96
Nikiforidou, Kiki, 190 Short, Mick, 103
now, 69, 93, 125, 190–193 Slingerland, Edward, 12
Sofer, Andrew, 147
Oakley, Todd, 164 speaking for thinking, 174–178
Oatley, Keith, 21 speech and thought representation, 70, 88–91,
Odyssey, 54 108, 110, 111, 159, 164, 169,
omniscience, 59, 61, 64, 68–70 171–194, 200
oral stories, 17–19, 56, 114 constructional phenomena, 183–194
direct discourse, 88, 126, 141, 174, 181
Palmer, Alan, 21, 29, 115 Distancing Indirect Speech and Thought
Paretsky, Sara, 110 (DIST), 187
Pascual, Esther, 182 free indirect discourse, 69, 88, 107, 111,
performative blend, 154 183, 185
person, See tense and person choices Free Indirect Speech and Thought
personal pronouns, 31, 51, 52, 64, 130, (FIST), 187
133–136, 182, 184–189 indirect discourse, 183, 185
Pinker, Steven, 11, 12, 13, 27 speech vs thought, 88, 103, 126, 171–178
poetry, 136–138 Spivey, Michael, 27, 35
possible worlds, 36 split-self metaphors, 100
Potocki, Jan, 80, 84, 86 Stoppard, Tom, 143
Powers, Richard, 133 stories as tools for thinking, 15, 23
Prinz, Jesse, 25 story construction, 37, 38, 39, 42, 50, 51, 53,
proper names, 34, 48, 52, 118–121 56–57, 68, 86, 91, 143, 152, 164,
props, 147, 157 188, 197
Proulx, E. Annie, 169 stream-of-consciousness, 70, 103, 169
structuralist narratology, 6
Raban, Jonathan, 91, 96, 118, 199 Sweetser, Eve, 8, 9, 10, 33, 55, 154, 193
reader response, 58 symbolic representation, 25, 26, 27, 50
Reddy, Michael, 203 Szymborska, Wisława, 136, 200
referential expressions, 42, 51–53, 117–128
representation blend, 45–46, 92–96, 101, Talmy, Leonard, 117, 162
102, 178 telling vs showing, 87
representational stance, 20, 50 tense and person choices, 8, 41, 63, 67, 68–79,
Richardson, Samuel, 165 85, 132–133, 190
Rimmon-Kenan, Shlomith, 87, 90 text vs story, 53–54
role-descriptors, 119, 121–128 text world theory, 36, 51
role-value mappings, See role-descriptors theory of mind, 21, 28, 29, 58, 109,
Roth, Philip, 120, 178, 180 112, 165
Rubba, Johanna, 117, 139, 193 Theroux, Paul, 94
Rushdie, Salman, 168 Thinking/Speaking blend, 158
time vs tense, 75
Scalise Sugiyama, Michelle, 17, 18 Tobin, Vera, 164
scene vs summary, 90 Tolkien, J. R. R., 128
science vs humanities, 12–13 Tomasello, Michael, 17, 20, 29, 114,
semantic roles 116, 164
agent, 138, 157 Toolan, Michael, 87
experiencer, 138, 156, 162 Turner, Mark, 9, 24, 26, 27, 28, 31, 162, 201
instrument, 138 Tyler, Anne, 66, 69, 120, 172,
patient, 138 181, 183
228 Index

van Oort, Richard, 13, 24, 26, 28 we, 53, 104, 130–132
Vandelanotte, Lieven, 74, 88, 103, 171, Werth, Paul, 36, 51
187, 193 Wetmore, Kevin, 162
Vargas Llosa, Mario, 22 Woolf, Virginia, 64, 67, 69, 75
Verhagen, Arie, 19, 114, 140, 141, 200
vertical integration, 12 Yachnin, Paul, 158
viewpoint and focus, 39 you narratives, 53, 76–79, 133
viewpoint compression, 40, 70, 96–100,
106–108, 110, 112, 141, 175, 176, 178, Zlatev, Jordan, 112
183, 189, 196 Zunshine, Lisa, 21, 28, 58, 109,
visual viewpoint, See narrative viewpoint 112, 165