Edinburgh University Press

Chapter Title: Doing things with words

Book Title: An Introduction to English Semantics and Pragmatics

Book Author(s): Patrick Griffiths and Chris Cummins
Published by: Edinburgh University Press. (2017)
Stable URL: http://www.jstor.org/stable/10.3366/j.ctt1g09x36.16

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted
digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about
JSTOR, please contact [email protected].

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
http://about.jstor.org/terms

Edinburgh University Press is collaborating with JSTOR to digitize, preserve and extend access to An
Introduction to English Semantics and Pragmatics

This content downloaded from 128.184.220.23 on Wed, 24 May 2017 00:40:33 UTC
All use subject to http://about.jstor.org/terms
 11  Doing things with words

Overview
What is the point of communicating? That may seem like a strange
question to ask at this stage of a book. But most of the chapters so far
have focused primarily on utterances that make statements of fact.
Stating – passing on facts that will be news to our addressees – is cer-
tainly an important function of language, but it is not the only one.
Much of what we do in language is not reducible to the act of stating: we
exchange greetings, we ask questions, we issue requests, we apologise,
we complain, we warn, we thank, and much else besides.
J. L. Austin was one of the first researchers to focus attention on this
question. His 1955 lecture series, published posthumously in 1962 under
the title How to Do Things with Words, made the point that – in contrast
to the assumptions of the then-­prevailing positivist view – relatively
few of the things uttered in natural language could actually be said to
be “true” or “false”. This, however, is not to say that the utterances are
not meaningful, rather that we need to analyse them in a different way,
specifically in such a way as to reflect their force as social actions.
In this chapter, we will consider briefly what kinds of social action
can be performed through the medium of language, and how speakers
and hearers manage to perform and recognise these actions.

11.1  Speech acts

Here we will follow Searle (1975, 1979) in using the term speech acts to
describe the basic units of linguistic interaction: that is to say, the things
that we can accomplish through the use of language. (11.1) lists a small
sample of speech acts, along with examples of sentences that could be
uttered in order to bring them about.

141

This content downloaded from 128.184.220.23 on Wed, 24 May 2017 00:40:33 UTC
All use subject to http://about.jstor.org/terms

GRIFFITHS & CUMMIN PRINT.indd 141 14/10/2016 08:05

 142 an introduction to english semantics and pr agmatics

(11.1) a. statement: “I lived in Edinburgh for five years.”

b. order: “Pay this bill immediately.”
c. question: “Where are you from?”
d. prohibition: “No right turn.”
e. greeting: “Hello.”
f. invitation: “Help yourself.”
g. felicitation: “Happy New Year!”
h. apology: “I’m terribly sorry.”
It has proved difficult to develop an exhaustive list of possible speech
acts: Austin (1962) reckoned that there could be several hundred dis-
tinct entries on such a list. Austin zoomed in particularly on a category
of items that he called performative, which appear to have the prop-
erty to perform social acts merely by virtue of being uttered, under the
appropriate contextual conditions. The examples in (11.2) are of this
type.

(11.2) a. I name this ship the Queen Elizabeth.

b. I now pronounce you husband and wife.
c. I apologise for my lateness.

The speaker of (11.2a) performs the action of naming the ship by

saying these words. Of course, in practice, saying these words will only
have the effect if the speaker is authorised to perform the action. The
same is true of (11.2b). In the case of (11.2c), unlike the other examples,
the action will be successfully performed by any (sincere) utterance of
the words, as there are no social conventions governing who may and
may not apologise for their own lateness.
A proposed diagnostic for performative utterances is that the word
hereby can be inserted before the verb. It would be coherent to say “I
hereby apologise” but not, for instance, to say “I hereby sing”. By this
definition, there are many performative or potentially performative
verbs in English, including apologise, complain, protest, resign, object, declare,
open, close, vote, propose, and so on. It is not entirely clear whether these
automatically constitute distinct acts, or whether they could naturally
be grouped into distinct classes. Various researchers have attempted
to do this on more or less principled grounds; in recent years, work in
computational linguistics has paid attention to this issue, as it is relevant
for building and training artificial dialogue systems.
It is also not clear whether the kinds of speech act that we
might naturally group together are best understood as the same sort
of thing. Take the case of questions. In the following section, we will
see that distinguishing questions from other kinds of speech act is

This content downloaded from 128.184.220.23 on Wed, 24 May 2017 00:40:33 UTC
All use subject to http://about.jstor.org/terms

GRIFFITHS & CUMMIN PRINT.indd 142 14/10/2016 08:05

 doing things with words 143

not entirely straightforward, but let’s suppose that we can do this.

Then we can consider (11.3) as showing various different examples
of questions.
(11.3) a. What’s your name?
b. Is today Tuesday?
c. Today’s Tuesday, isn’t it?
We can think of all these as information-­seeking questions: the speaker
is performing the social action of trying to obtain information from the
hearer. Yet there are systematic differences between the three exam-
ples. (a) is a wh-­question, and admits arbitrarily many possible answers.
(b) is essentially a yes/no question, also called a “polar” question, and
admits only the answer “yes” or “no” (although the hearer could opt out
of answering by saying “I don’t know”). (c) is also a yes/no question but
seems to encode the fact that it has a preferred response, namely “yes”.
This kind of question is sometimes called a “check-­question”. Although
we could characterise what the speaker of any of these sentences was
doing as “asking a question”, it’s not obvious that these three utterances
are really performing the same social action. For instance, if we were to
categorise them by the kind of response that they elicit from a cooperative
hearer, we might come to the conclusion that they were different things.
And looking at the effect on the hearer would seem to be a reasonable
way of trying to determine what kind of social action is being performed.
We can characterise speech acts at several different levels of organi-
sation. Austin proposed a distinction between locutionary, illocution-
ary and perlocutionary acts. In this scheme, approximately speaking,
locution covers what was actually said, illocution what was meant and
perlocution the effect that the utterance had. Suppose we are at dinner
and you ask me to pass the salt by saying the words “Could you pass the
salt?”, and as a result of this I do so. We could report this in several differ-
ent ways. As a locutionary act, we would describe it as follows: “You said
‘Could you pass the salt?’” As an illocutionary act, we would describe it
as: “You asked me to pass the salt”. As a perlocutionary act, we would
describe it as: “You got me to pass the salt”. Here we are focusing on the
illocutionary level, but it is sometimes useful to know what is and is not
considered to fall within the jurisdiction of speech act theory.

11.2  Sentence types and other indications

For speech acts to be effective, hearers have to be able to recognise them
as such. Other conditions also have to hold – for instance, hearers have
to be willing to comply with them – but a failure to recognise a speech

This content downloaded from 128.184.220.23 on Wed, 24 May 2017 00:40:33 UTC
All use subject to http://about.jstor.org/terms

GRIFFITHS & CUMMIN PRINT.indd 143 14/10/2016 08:05

 144 an introduction to english semantics and pr agmatics

act as, say, a request will probably prevent it from achieving its intended
outcome. The problem of speech act recognition turns out to be a diffi-
cult one, despite the availability of various indications or cues of speech
act type, as we shall see.

11.2.1  Syntactic cues and indirect speech acts

One major category of cues to speech act type is syntactic. The
speech acts in (11.1a–c) were placed at the head of the list because
they ­correspond with the three major sentence types of English (and
many other languages). These sentence types are declarative, com-
monly associated with statements; interrogative, commonly associated
with questions; and imperative, commonly associated with orders. In
English these sentence types are clearly distinguished syntactically, as
the example sentences for (11.1a–c) illustrate.
Nevertheless, there are two limitations to attempting to identify
speech act types on the basis of syntactic cues. As discussed above, there
are potentially many more distinct speech act types than just these three,
and it is clear that these other types cannot always be ­distinguished just
by syntactic considerations. The sentences I promise you that . . ., I warn
you that . . . and I advise you that . . . all use the same syntactic frame and
are distinguished just by the lexical items involved. But even if we just
restrict ourselves to statements, questions and orders (and we assume
that questions represent a single speech act type), it turns out that the
mapping from sentence type to speech act type is somewhat inconsist-
ent. (11.4) illustrates this.
(11.4) a. I’d be grateful if you would close the window.
b. Could you close the window?
c. Close the window.
Formally, (11.4a) is a declarative sentence: it expresses an aspect of
the speaker’s state of mind; (11.4b) is an interrogative; and (11.4c) is an
imperative. Yet all three appear to be used to perform the same speech
act, in this case a request. When a sentence type is used to perform
a speech act that is not customarily associated with it, we have what
is called an indirect speech act (see Searle 1975). By this definition,
(11.4a) and (11.4b) are examples of indirect speech acts. They are by no
means unusual: Levinson (1983) argues that the majority of speech acts
are in fact indirect. Certainly, we might think that (11.4c) represents an
unusually brusque way of trying to bring about the effect of getting the
hearer to close the window, and would only be appropriate if the action
was a matter of urgency.

This content downloaded from 128.184.220.23 on Wed, 24 May 2017 00:40:33 UTC
All use subject to http://about.jstor.org/terms

GRIFFITHS & CUMMIN PRINT.indd 144 14/10/2016 08:05

 doing things with words 145

There are several accounts of how we could get from the “literal”
speech act associated with a sentence to the indirect speech act that it
performs. Consider the interpretation of (11.4b) as a request, rather than
as a question enquiring into the hearer’s capability to close the window.
Gordon and Lakoff (1971) propose that, in such a case, the speaker
already knows that the answer to the question is “yes”, and the hearer
knows this: consequently, the hearer can infer that the purpose of the
utterance is not simply to ask a question, and has to arrive at a different
interpretation. Searle (1975) offers a slightly more detailed account: he
argues that asking the question (11.4b) signals that the answer (known
to be “yes”) is relevant to the speaker. The hearer can infer that the
reason it is relevant whether the hearer can close the window is that the
speaker wants the hearer to do so. For that reason, the question can be
reinterpreted as an indirect request.
We can construct some kind of rationale for various other cases of
indirect speech act interpretation, initially acting as though the literal
interpretation is intended, and subsequently performing additional rea-
soning when the literal interpretation leads us into apparent absurdity.
However, since the 1970s, the consensus opinion seems to have moved
away from the idea that we actually do this – perhaps the syntactic cues
are not quite as privileged as this kind of account would suggest.

11.2.2  Lexical cues

In the case of performative verbs, it is clear that the presence of particu-
lar lexical items signals the presence of particular speech acts (leaving
aside the question of whether every different performative verb signals
a different speech act, as touched upon earlier). These are not entirely
reliable signals: we can use performative verbs in non-­performative
ways. For instance, we can use them to report the performative actions
of others: if I say She named the ship, that is clearly a statement rather than
an act of naming in itself. Nevertheless, when we encounter a performa-
tive verb in the present tense and in a sentence in which the subject is
the speaker, we can be reasonably confident about the speech act being
performed.
In other cases, we might find particular non-­performative lexical
items to be useful indicators of the kind of speech act that is being
­performed. Consider the Could you . . . of (11.4b). In terms of its semantic
meaning, this appears to suggest that the sentence is going to ask about
the hearer’s capability to do something (perhaps politely or hypotheti-
cally, given the use of could instead of can). However, experience sug-
gests that Could you . . . is an idiomatic way of making a polite request,

This content downloaded from 128.184.220.23 on Wed, 24 May 2017 00:40:33 UTC
All use subject to http://about.jstor.org/terms

GRIFFITHS & CUMMIN PRINT.indd 145 14/10/2016 08:05

 146 an introduction to english semantics and pr agmatics

just as Why not . . . is an idiomatic way of making a suggestion. This could
be because the reasoning outlined in the previous subsection has been
somehow routinised by hearers, or it could be that through encoun-
tering many instances of Could you . . . being used in requests we have
simply learned that that is a potential use for these words. In either case,
a competent user of English is entitled to think it somewhat likely that
an utterance beginning Could you . . . will constitute an indirect request.
There are, of course, even more obvious lexical markers of particu-
lar kinds of speech act that are also not performative verbs. Sorry as
a marker of apology is perhaps the most obvious; hello is maybe even
more emphatic as a marker of greeting, and its other uses in dialogue
are rather marginal. Although there are ways of greeting without saying
hello, the use of this particular lexical item is reasonably reliable as an
indicator that what is taking place is an act of greeting. In the middle of
an established interaction, saying “hello” is blatantly anomalous.
We might even be able to use smaller pieces of semantic informa-
tion than that to help us identify the speech act being performed. Some
speech acts, for example, relate predominantly to the speaker: if I protest,
or resign, for instance, I do so largely without the involvement of the
hearer. Even an apology doesn’t have to be clearly directed towards a
particular recipient. However, for other speech acts, such as requesting
or advising, it is clear that the purpose of the act is to effect some kind of
change on the mental state of a particular hearer. We might conjecture,
therefore, that speech acts involving first-­person pronouns (I, we) are
likely to be of one of the former types, while those involving second-­
person pronouns (you) are a little more likely to be of one of the latter
types.

11.2.3  Discourse cues

One of the reasons that the utterance “Hello” is likely to be perceived as
anomalous in the middle of a conversation is that it is customarily used
to perform the speech act of greeting, but that speech act is only really
appropriate at the beginning of a conversation. More broadly, we have
access to a potentially rich set of knowledge about how interactions
take shape, including the kinds of social actions that are likely to be per-
formed during them, and the interdependencies that can exist between
those social actions.
One example that we’ve already discussed in this book is the fact that
there tends to be a close dependency between questions and answers.
There is clearly a logical dependency here – something cannot be an
answer unless it corresponds to a question – but there is also a rich

This content downloaded from 128.184.220.23 on Wed, 24 May 2017 00:40:33 UTC
All use subject to http://about.jstor.org/terms

GRIFFITHS & CUMMIN PRINT.indd 146 14/10/2016 08:05

 doing things with words 147

expectation that questions are almost immediately followed by answers.

As discussed in Chapter 8, we will naturally try to interpret something
as an answer if it is said in response to a question – (11.5) presents
another example of this. Moreover, if there is no answer forthcoming,
we will readily interpret silence as communicatively meaningful: even a
short silence in response to A’s utterance in (11.5) could be interpreted
as indicating a wish on B’s part to decline the invitation.
(11.5) A: Would you like to go for a coffee with me?
B: I have to finish this assignment.
In this case, in speech act terms, A’s utterance can be understood as an
invitation, and the specific expectation here is that B’s response will
either be to accept or decline the invitation. Similarly, if A’s utter-
ance constitutes an offer, we will expect B either to accept or decline
the offer. And, just as in the general question–answer case, we can use
this expectation to help us understand B’s utterance: even if it doesn’t
superficially appear to serve the purpose of responding to A’s utter-
ance, we will try to enrich it pragmatically by appeal to our back-
ground assumptions in order to make it interpretable as an appropriate
response.
This idea was developed by Schegloff and Sacks (1973). They used
the term adjacency pair to describe the relationship between two
conversational turns that are highly likely to happen sequentially, such
as question–answer. When we encounter the first part of an adjacency
pair being performed as a speech act, we can quite strongly expect that
the next speech act will be the second part of the adjacency pair. Clark
(2004) argues that this kind of relation can also exist between wordless
communicative acts such as gestures, raising the issue of whether we
can have things like question–answer sequences playing out in purely
gestural communication. However, we won’t delve further into that
possibility here (see Clark 2012 for more discussion).
We can also use higher-­level information about the status of the
discourse to tell us something about the kind of speech act that is being
performed. The fact that greetings go at the beginning of conversations
is potentially helpful in recognising something as a greeting. Unlike
many languages, English doesn’t have an all-­purpose expression that
serves both to open and close conversations – although ciao has been
borrowed in from Italian and seems to fulfil this dual function in
English too. Importantly, assuming that the opening and the closing
of a conversation are essentially different social acts, it is nevertheless
easy to tell which one is being performed by a particular utterance
of “ciao” – if there isn’t currently a conversation going on between

This content downloaded from 128.184.220.23 on Wed, 24 May 2017 00:40:33 UTC
All use subject to http://about.jstor.org/terms

GRIFFITHS & CUMMIN PRINT.indd 147 14/10/2016 08:05

 148 an introduction to english semantics and pr agmatics

speaker and hearer, it’s a greeting, and if there is, it’s rather likely to be
a ­conversation ender or “leave-­taking” device.
But there are many other pieces of knowledge available to us about
the possible actions at a given stage in a conversation, and it seems
likely that we use this to constrain our guesses as to the speech act
being performed by a speaker. For instance, if someone says their own
name, they could be introducing themselves, but this is only a possible
interpretation if there are people present who do not know them (or
who might not know them). If a customer goes into a restaurant and
says “table for two”, that’s analysable as a request, but if the waiter says
it to the customer first, it’s analysable as something more like a check-­
question. There is not space here to scratch the surface of all the knowl-
edge we have that is potentially relevant to the identification of speech
acts, but it might be helpful to be able to identify some broad classes of
relevant knowledge, as that offers us some basis for thinking about the
processes that are going on under the surface. That will be the topic of
the ­following, final subsection of this book.

11.2.4  Integrating the information

How do we take these various sources of information about speech
acts – along with other potential sources of information not discussed
above, such as prosody, gesture, and so on – and use them to determine
which speech act is being performed? As we discussed earlier, a tradi-
tional approach holds that we initially read the speech act directly off
the syntax, and then perform pragmatic enrichments to correct this if
it gives us some kind of anomalous outcome. But, as we’ve also seen,
this idea is problematic: it only accounts for a relatively small portion
of the possible speech acts in the system (those associated with distinct
sentence types). Also, on psycholinguistic grounds, this is a somewhat
implausible approach, as it suggests that we have to perform a lot of
pragmatic reasoning in order to understand the purpose of utterances,
and yet in practice we are able to recognise indirect speech acts very
rapidly. If someone says “Could you pass the salt?”, a cooperative inter-
locutor passes the salt, rather than whiling away seconds thinking about
how to explain this apparent question as some other kind of utterance.
There are at least a couple of other broad approaches to this problem.
The plan-­based approach, initially laid out by Perrault and Allen (1980)
and elaborated in much subsequent work, argues that what hearers are
doing is trying to recognise the goals of the speaker and to cooperate
in their achievement. From this perspective, the speaker is seen as a
rational agent trying to bring about certain changes in the world that

This content downloaded from 128.184.220.23 on Wed, 24 May 2017 00:40:33 UTC
All use subject to http://about.jstor.org/terms

GRIFFITHS & CUMMIN PRINT.indd 148 14/10/2016 08:05

 doing things with words 149

they find desirable – which sounds rather grandiose, but in practice

could be something as simple as causing a currently open window to
be closed. The speaker then constructs a plan to achieve these changes,
which may involve performing speech acts. The hearer is able to draw
inferences about what the speaker seems to be doing at a given point
in time, and use these to derive conclusions about what the speaker
wants to be the case; but also the hearer can use their understanding of
the speaker’s likely goals to draw conclusions about what the speaker
is trying to do. This approach is appealingly general, as an account
of interaction, but when applied in detail to even relatively simple
­examples it seems to involve a lot of work.
The dominant approach in the computational literature over the past
thirty years or so has been the alternative “cue-­based” approach. The
idea here is that we simply consider all the evidence that could bear
upon the question of what speech act is being performed (syntax, words,
context, etc.), weigh this evidence probabilistically, and thus arrive at a
guess as to the speech act that is most likely to be taking place. In prac-
tice, it seems rather unlikely that we actually use all this information,
and it’s not clear how we represent some of it (how do we keep track
of what is likely in a given conversation, for instance?). However, this
approach has proven fruitful for small-­scale artificial dialogue systems.
Ultimately, the question of which kind of approach is better is a psy-
chological (or in some cases a practical) one, and discussing this problem
in more detail takes us some way beyond the scope of this book. But as
far as the description of speech acts in English is concerned, this is still
an informative debate. We seem to have good evidence that speech acts
are signalled by a range of linguistic devices, doubtless including syntax,
but with syntax not playing such a central role as suspected by Searle
and colleagues in the 1970s. It also seems that the recognition of speech
acts depends on context, but here we are not talking just about the lin-
guistic context but also the cultural context; and it seems that encyclo-
pedic knowledge about how the world works plays an important part in
this kind of discourse pragmatics. If we take the treatment of speech acts
to be a part of linguistic semantics and pragmatics, it becomes increas-
ingly difficult to draw a clear boundary between linguistic knowledge
and world knowledge.

Summary
The idea that we can achieve social actions through the use of words is
perhaps the essential reason why linguistic communication is so crucial
for our species. Many distinct actions can be performed in this way. In

This content downloaded from 128.184.220.23 on Wed, 24 May 2017 00:40:33 UTC
All use subject to http://about.jstor.org/terms

GRIFFITHS & CUMMIN PRINT.indd 149 14/10/2016 08:05

 150 an introduction to english semantics and pr agmatics

one broad category of sentences, called performatives, the verb explicitly

identifies the social action that the speaker is performing (such as promise in
sentences beginning “I hereby promise . . .”). However, in most cases, the
intended purpose, or illocutionary force, of the sentence is not explicitly
marked and must be inferred. Syntactic features of the sentence, such as
the sentence type, are a helpful cue to this, but many speech acts are indi-
rect, in the sense that their purpose does not correspond to the “typical”
purpose for the sentence type. Although theories differ as to how we take
into account all the information that might be of use to us, it seems clear
that in practice we draw upon many distinct linguistic and non-­linguistic
factors when we perform the task of recognising the purpose of an utter-
ance (a task we perform almost every time we use language).

Exercises
1. Using the notions of speech acts and presupposition, give a brief
description of the wording of this notice seen in a bus: “Thank you
for not smoking. MAXIMUM FINE £100.” (In the same frame there
was a picture of a cigarette with a slash through it, inside a red circle
that indicates mandatory prohibition.)
2. For each of the following, name the kind of direct speech act that
would “normally” be associated with the sentence, given its sentence
type, and say what indirect speech act the example would probably
be used to perform.

(a) Can’t you stop talking?

(b) Help yourself to milk and sugar.
(c) Have you heard, our team’s leading 18 to 15?
(d) You have my sympathy.
(e) Don’t imagine that entailment and implicature are the same
thing.
(f) Accept my profound condolences.
(g) Have I ever let you down?
(h) I recommend that you keep a copy of the letter.

3. What kind of function does the politeness formula “I should let you
go” tend to serve in a conversation? How can we explain this effect?

Recommendations for reading

Lycan (2000) and Saeed (2015) offer good introductory treatments of
speech acts. Searle (1975) is worth reading as a more detailed treatment

This content downloaded from 128.184.220.23 on Wed, 24 May 2017 00:40:33 UTC
All use subject to http://about.jstor.org/terms

GRIFFITHS & CUMMIN PRINT.indd 150 14/10/2016 08:05

 doing things with words 151

of how indirect speech acts can be calculated from a literal starting

point. Chapter 5 of Levinson (1983) is also a good guide to the problem
of speech act identification, and the limitations of Searle’s approach.
Jurafsky (2004) gives a lucid account of the competing approaches to the
recognition of speech acts from a computational perspective.

This content downloaded from 128.184.220.23 on Wed, 24 May 2017 00:40:33 UTC
All use subject to http://about.jstor.org/terms

GRIFFITHS & CUMMIN PRINT.indd 151 14/10/2016 08:05