Pragmatics 1: meaning and context

Recall the definitions of semantics and pragmatics at the beginning

of this chapter. Pragmatics concerns both the relationship

between context of use and sentence meaning, and the

relationships among sentence meaning, context of use, and

speaker’s meaning. In this section we will focus on those aspects

of pragmatics having to do with how the context of use

contributes to semantic meaning. Indexicality, context-

dependency, and anaphora Indexical are words whose semantic

meanings depend in a direct way on the context of use. Some

simple examples are I, you, here, and now. The pronoun I refers to

the person who says it (except in cases like direct quotes, where it

refers to the person being quoted); the word here refers to the

place where it is spoken, and so forth. Indexicality is similar to the

phenomenon of vagueness, discussed earlier. Recall that

adjectives like old can be vague, in that what counts as old

depends on what is being talked about. “What is being talked

about” is an aspect of context of use. Another term, deixis, is also

used for cases in which meaning depends on context of use, and

the terms deixis and indexicality are often used interchangeably.

The “classic” indexicals (like I, you, here, and now) are strictly tied

to the context of use; their reference depends solely on the

context of use. Other words are more flexible. They sometimes get

their meaning from the context of use, but not always. An

example is she. I could say She is cute in a situation in which we

are both looking at a baby girl. In that case, she would refer to this

girl by virtue of the context of use. However, the reference of she

can be determined in other ways. In (34a), she can get its meaning

from the previous sentence, and in (34b) it can get its meaning

within the sentence itself, from no girl in this class: (34) a. Do you

see that baby girl over there? She is cute. b. No girl in this class

doubts that she is cute. When a word or phrase picks up its

meaning from some other piece of language nearby, the

relationship between the two is called anaphora. A word which

gets its meaning in this way is an anaphor, and the piece of

language which gives the anaphor its meaning is its antecedent. In

(34a), the antecedent of she is that baby girl over there.

Box 4.4

Discourse anaphora There are several kinds of anaphora. Some

types of anaphora are semantic in nature, while others are purely

pragmatic. The relationship between the quantifier no girl in this

class and she in (34b) is an example of purely semantic anaphora;

the semantic nature of this anaphora is clear because the

antecedent of she doesn’t in any sense refer to a particular girl.

Therefore, we can’t say that the context of use tells us what she

refers to. Rather, you might say that the meaning of she is derived

from the (quantificational) meaning of its antecedent. Example

(34a) illustrates a kind of pragmatic anaphora called discourse

anaphora. Discourse anaphora occurs when the anaphor and

antecedent are in different sentences. It only works with certain

antecedents.

While the antecedents in (i)–(iii) allow discourse anaphora, (iv)–(v)

do not:

(i) This baby girl was born on January 1. She is cute.

(ii) Maria was born on January 1. She is cute.

(iii) A baby girl lives next door to me. She is cute.

(iv) No baby girl lives next door to me. ??She is cute.

(v) Most baby girls are born with hair. ??She is cute.

The first three sentences are acceptable because the antecedents

in these cases provide a reference for the anaphor she. In

contrast, no baby girl and most baby girls (in (iv) and (v)) don’t

provide a reference for she, since they are quantifiers, and this is

why anaphora doesn’t work in these examples.

Presupposition

The sentence John stopped crying at noon only makes sense if it is

assumed that John was crying just before noon. If I say this to you,

and you know that John had not been crying at all, you would feel

that what I say is out of place or that you’d been misled. We say

that the sentence John stopped crying at noon presupposes that

John was crying just before noon. Many words, phrases, and

structures create presuppositions. Here are some examples of

sentences followed by their presuppositions (in parentheses):

(35)

a. Each of the boys in the room is nice. (There are some boys in the
room.)

b. That pig is fat. (That is a pig.)

c. John is crying again. (John has cried before.)

d. It is Bob who stole the chocolate. (Somebody stole the

chocolate.)

Presuppositions are not just anything which a speaker happens to

take for granted. Presupposition occurs when a speaker’s choice

of words shows that he or she is taking something for granted. For

example, part of the meaning of the word again is that someone

who uses it indicates that he/she is taking for granted that

whatever is being talked about happened before. For this reason,

we say that (35c) presupposes that John has cried before. (The

speaker of (35c) might take other things for granted, such as that

you should comfort people who are crying, but you can’t
determine this just by looking at (35c) itself. Therefore, we

wouldn’t say that (35c) presupposes, in the linguist’s sense, that

one should comfort crying people.)

Presuppositions are often understood in terms of the notion of

common ground, as discussed by the philosopher Robert Stalnaker

(Stalnaker 1974, 1978). The common ground is the set of

propositions which the participants in a conversation mutually

assume. (They don’t have to actually believe these propositions,

but they at least act as if they do.) In any normal conversation,

many things are implicitly in the common ground – that the sun

comes up every morning, that the speaker and the hearer are

alive, that things fall down when you release them in the air, and

so forth. Other propositions may get into the common ground

because they are explicitly stated. For example, if I say to you

“I’m hungry” (and you think I’m being sincere), henceforth the

proposition that I’m hungry will be part of the common ground;

that is, we will both assume it to be true (until something changes

– for example, I eat something). The common ground is a major

part of the context of use, and helps us make explicit the role of

presupposition in the use of sentences like John stopped crying at

noon: the sentence is only appropriate if the common ground

already contains the information that John was crying just before

noon or if that information can be added without arousing

controversy. In television shows about lawyers and detectives,

presupposition is often turned to the benefit of a crafty

investigator. Suppose a detective suspects X of committing a

murder in New Jersey, but has no clear evidence that X has

traveled there. The detective might casually ask, “How was your

trip to New Jersey?” This sentence presupposes that the suspect

has been there, and so any answer at all will provide the crucial

evidence that X has recently been to New Jersey.

Pragmatics 2: meaning and the intention to communicate

Indexicality and presupposition are aspects of pragmatics which

mostly have to do with the relationship between context of use

and semantic meaning. In this section, we’ll investigate the other

major subdomain of pragmatics, the relationships among semantic

meaning, context of use, and speaker’s meaning.

The Gricean view of meaning

Semantics views meaning from the compositional perspective: the

meaning of a sentence is built up from the meanings of its parts.

The smallest parts get their meanings from the lexicon, and then

these meanings get put together according to rules which pay

attention to the grammatical structure of the sentence.

However, not all aspects of meaning can be explained by this

compositional “bottom-up” approach, and a complementary

“topdown” view of meaning has focused on the intentions of

language users. More precisely, when A says something to B, A

intends for B to be affected in a certain way. If A says “It’s raining,”

for example, A may intend for B to believe that it’s raining (and

perhaps to open an umbrella or come inside). This perspective

helps us understand many aspects of speaker’s meaning

Implicature

The idea that meaning is based in the intentions of speakers is

most clearly revealed in H. P. Grice’s theory of conversational

implicature (Grice 1957, 1975). Very often, when someone says

something, he or she doesn’t mean exactly what the words

literally mean. That is, the (speaker’s) meaning differs from the

(semantic) meaning. For example, the semantic meaning of

“There’s a bear sneaking up behind you!” doesn’t involve the

concept of warning; it just reports a fact. However, it’s quite likely

that a warning is part of what the speaker means. This “extra

meaning” which goes beyond what the words literally say is an

implicature of the sentence. Grice explained how speaker’s

meaning can be determined in such cases by positing a

Cooperative Principle that all speakers and hearers assume when

speaking to each other: Cooperative Principle: speaker’s meaning

can be calculated on the basis of semantic meaning and the

assumption that speakers are behaving rationally and

cooperatively. Grice broke this general principle into four

conversational maxims to explain what rationality and

cooperativeness are: The maxim of Quality: make your

contribution one that is true rather than false. The maxim of

Quantity: provide the information that is required for the

purposes of the conversation, but no more. The maxim of

Relevance: make your contributions relevant. The maxim of

Manner: be clear and orderly in your talk. These maxims are not

rules to be followed in the sense that traffic laws are. Rather, they
are assumptions which we use to try to make sense of what

people say. That is, we assume that people follow the four maxims

when they talk, and this helps us figure out what they mean.

Consider (36), for example:

(36) There are three students in the class: Mary, Bob, and Jill.
A: Which students passed the exam?

B: Mary and Bob.

In this conversation, in addition to concluding that Mary and Bob

passed the exam, A is likely to infer that Jill didn’t. However, B

never said that Jill didn’t pass the exam, so why would A infer this?

By assuming that B is following the four maxims, A can figure that

B gave as much true information as was required and relevant

(maxims of quality, quantity, and relevance). Since it would be

relevant to say that Jill passed if she actually had passed, A can

infer that B didn’t include Jill in the list of people who passed

because B doesn’t think that Jill passed (so long as other

assumptions hold, such as that B knows Jill is in the class).

Moreover, B knows that A would figure this way, and so said

“Mary and Bob” didn’t pass becomes part of the speaker’s

meaning of B’s utterance. That is, B uses the Cooperative Principle

and maxims to implicate that Jill didn’t pass. With the

understanding that A would conclude that Jill didn’t pass. In this

way, the idea that Jill didn’t pass becomes part of the speaker’s

meaning of B’s utterance. That

is,

B uses the Cooperative Principle and maxims to implicate that Jill

didn’t
pass.

Another example of implicature was hinted at earlier in

example (12): (12) Elvis Presley made a peanut butter

sandwich and sat down beside the pool.

This sentence seems to mean that Elvis made the peanut

butter sandwich before going to the pool. This ‘before’

meaning is not part of the semantic meaning of and (as given

by truth conditions); it is an implicature. According to Grice’s

maxim of Manner, we should present information in an

orderly way, and in most cases that includes mentioning

events in the order in which they occurred. Therefore, a

hearer can conclude that the speaker means to say that Elvis

made the sandwich before sitting down by the pool.

Because the Gricean maxims are not rigid rules, like rules

of law, but are rather flexible assumptions about how

speakers behave, they can be broken, or flouted, to

implicate further meanings. Flouting a maxim occurs

when a speaker uses language in a way which appears, in

an obvious way, to violate a maxim. For example, if you

ask me whether I think your new shirt is attractive, and I

say “It was probably inexpensive,” my reply seems to

violate the maxim of relevance – I didn’t answer your

question. However, because you assume that, despite

appearances, I am conforming to relevance, you try to

figure out how what I said could be relevant. Since my

utterance avoided answering your question by

mentioning a reason why you might have bought an

unattractive shirt, you will infer that I don’t like the shirt.

This inference can become an implicature of the

sentence, that is, part of my speaker’s meaning.

Speech acts
Another important figure in the development of pragmatics is John
Austin

(Austin 1962). He pointed out that when people use language,

they are performing a kind of action. He called these actions

speech acts. It’s easy to see the “act” nature of language when

a minister says, “I now pronounce you husband and wife” in a

wedding ceremony. By virtue of this sentence being said by an

appropriate person, the engaged couple becomes a married

couple. Most speech acts are not so “official,” but they all rely

on the speaker using an utterance to signal his/her intention

to accomplish some action and the hearer inferring that action

from the utterance. When people make bets and threats and

promises, offer congratulations and apologies, or issue orders

or challenges, they are using language to accomplish actions.

Consider the contrast between (37a) and (37b):

(37) a. I promise to visit tomorrow.

b. She promised to visit tomorrow.

(37a) performs the act of promising: if you say it, you’ve

promised something. If you don’t visit the next day, the person

you said it to can complain that you broke your promise.

Example (37b) simply reports a promise by somebody else, you

haven’t promised anything yourself.

Sentences which perform actions, like (37a), are known as

performatives, while other sentences, (37b), are called

constatives. A good test of whether a sentence is a

performative is whether you can insert the word hereby

before the verb: “I hereby promise/challenge/bet. . .” make

sense, but “I hereby walk/see/like. . .” do not. As Austin

pointed out, however, even constatives perform actions of

some sort; (37b) performs the action of reporting her

promise. Thus, the distinction between performatives and

constatives may not be as important as the idea that all

sentences can be used to perform actions of various sorts.

In trying to understand the various types of acts that sentences

may perform,

Austin proposed three “levels” of speech act:

Locutionary acts: grammar-internal actions like articulating a

certain sound, using a certain morpheme, referring to a

particular person.

(These are the “acts” which fall under phonetics, phonology,

morphology, syntax, and semantics. They are usually not of much

interest to people studying pragmatics.)

Illocutionary acts: actions of communication like

asserting a fact, asking a question, requesting an

action, making a promise, or giving a warning.

Perlocutionary acts: actions which go beyond communication,

like annoying, frightening, or tricking someone by what you

tell them.

(38) There’s a bear sneaking up behind you!

At the locutionary level, A utters the word there and refers to

the addressee with the word you (among many other

locutionary acts).

At the illocutionary level A asserts a fact (that there’s a bear

sneaking up behind B) and warns B that he or she is in danger. At

the perlocutionary level, A frightens B and causes B to run away.

Linguists often speak of the illocutionary force of a sentence. The

illocutionary force is the type of communicative intention that

the speaker has. For example, (38) has the illocutionary force of

warning, while (37a) has the force of promising and (37b) has the

force of reporting or asserting. The context in which the sentence

is uttered is crucial in interpreting the illocutionary force of a

speech act; if a loan shark to whom you owe money says “I

promise to visit tomorrow,” the speech act intended may be a

threat (disguised as a promise).

Pragmatics summary

Pragmatics is fundamentally about how the context of use

contributes to meaning, both semantic meaning and speaker’s

meaning. The core topics of pragmatics are indexicality,

presupposition, implicature, and speech acts, but in reality there is

no limit to the ways in which context can influence meaning.

Situations can even develop which allow words to mean things

they never meant before. For example, several families are having

dinner together, and two of the teenagers are, unbeknownst no

anyone

else, dating. They each separately make an excuse to leave the

dinner to their parents, expressing a wish to go work on their

chemistry assignment, and they have an enjoyable time

together. After this, they start to say things like “Don’t you need

to work on your chemistry homework?” to indicate a desire to

sneak off together – a new pragmatic meaning for sentences of

that kind.