Semantics and Pragmatics

Christopher Gauker

Semantics deals with the literal meaning of sentences. Pragmatics deals with what

speakers mean by their utterances of sentences over and above what those sentences

literally mean. However, it is not always clear where to draw the line. Natural languages

contain many expressions that may be thought of both as contributing to literal meaning

and as devices by which speakers signal what they mean. After characterizing the aims

of semantics and pragmatics, this chapter will set out the issues concerning such devices

and will propose a way of dividing the labor between semantics and pragmatics.

Disagreements about the purview of semantics and pragmatics often concern

expressions of which we may say that their interpretation somehow depends on the

context in which they are used. Thus:

• The interpretation of a sentence containing a demonstrative, as in “This is nice”,

depends on a contextually-determined reference of the demonstrative.

• The interpretation of a quantified sentence, such as “Everyone is present”,

depends on a contextually-determined domain of discourse.

• The interpretation of a sentence containing a gradable adjective, as in “Dumbo is

small”, depends on a contextually-determined standard (Kennedy 2007).

• The interpretation of a sentence containing an incomplete predicate, as in “Tipper

is ready”, may depend on a contextually-determined completion.

• The interpretation of a sentence containing a discourse particle such as “too”, as

in “Dennis is having dinner in London tonight too”, may depend on a contextually

determined set of background propositions (Gauker 2008a).

• The interpretation of a sentence employing metonymy, such as “The ham

sandwich wants his check”, depends on a contextually-determined relation of

reference-shifting.

• The interpretation of a conditional sentence, such as “If you turn left at the next

corner, you will see a blue house at the end of the street”, depends on a

contextually-determined range of prospects.

I will call devices such as these the in-between phenomena. Insofar as the truth values of

sentences involving in-between phenomena depend on the contextually-determined

values of certain parameters, explication of the in-between phenomena seems to belong

to semantics. Insofar as the evaluation of utterances of such sentences depends on the

context of utterance, explication of the in-between phenomena seems to belong to

pragmatics.

Toward dividing the labor between semantics and pragmatics, I will urge that we

take care to distinguish between two questions. One question concerns the conditions

under which a sentence is true relative to a context. That question belongs to semantics.

The other question concerns which context pertains to a given act of speech. That

question belongs to pragmatics. So characterized, the issue concerns not just how to

define “semantics” and “pragmatics” but how the truth-conditions of sentences and the
Semantics and Pragmatics 8/4/10 Page 3

settings in which sentences are uttered jointly contribute to a successful linguistic

exchange.

1. SEMANTICS, BRIEFLY

A semantic theory for a language will specify the literal, conventional meaning of

expressions of the language. Typically, it is assumed that the core of such a specification

of meanings is a specification of extensions. The extension of a proper name might be

the thing named. The extension of a verb phrase might be a set of things that the verb

phrase describes. The extension of a whole sentence is a truth value. The extension of a

quantifier might be a function from sets (sets of sets) to truth values. In simple cases, the

extension of a compound expression will be determined by the extensions of its

components and its grammatical structure, and, where this is so, a semantic theory is

expected to specify the way in which this is determined. Thus, the truth value of

“Socrates is snub-nosed” will be truth if and only if the extension of “Socrates”, namely,

Socrates, is a member of the set of objects that is the extension of “is snub-nosed”,

namely, the set of snub-nosed things.

Beyond identifying the extensions of expressions, a semantic theory might be

expected to specify the intensions of expressions. Commonly, the intension of an

expression is identified with a function from possible worlds to extensions. Moreover,

some of the expressions of a language may express operations on intensions. Thus,

“possibly” may be thought of as expressing an operation on intensions that takes an

intension and generates the value truth if and only if that intension is a function from

worlds to truth values that generates truth at some world. Propositions are often
Semantics and Pragmatics 8/4/10 Page 4

identified with the intensions of sentences and are thought of as being, or as determining,

functions from possible worlds to truth values.

Many of the controversies in semantics concern the treatment of specific

linguistic devices within this basic framework. Other controversies have to do with the

basic concepts that I have employed in characterizing this framework. For instance, one

controversy concerns the use of the notion of possible worlds. Furthermore, one may

question the nature of the relation between expressions and their extensions, which we

may call the reference relation. What makes it the case that “Socrates” refers to

Socrates? What makes it the case that “is snub-nosed” describes the objects in the set of

snub-nosed things. Some philosophers have attempted naturalistic reductions of the

reference relation. Others despair of obtaining a naturalistic reduction and consequently

present ideas about how to formulate a semantic theory without our needing a

naturalistically reducible reference relation.

In general, a semantic theory will relativize the truth conditions of sentences to

evaluations of various parameters, such as interpretation of the nonlogical lexical items,

model domain, possible world or world-and-time, or, as we will see, context. Various

notions of validity may be defined by picking out a set of these parameters and, while

holding the values of the other parameters fixed, saying that an argument (a set of

premises and a conclusion) is valid if and only if for each parameter in that set, if the

premises are all true relative to that parameter, then so is the conclusion. For instance,

we may define one very narrow kind of validity by taking the pertinent set to be the set of

all evaluations of all of the parameters (holding fixed only the interpretation of the logical

vocabulary). We may obtain a much broader kind of validity by holding the

interpretation, model domain and world-time fixed and taking the pertinent set to be only

the set of all contexts.

Even theorists who disagree about the basic framework for semantics might

nonetheless agree that a semantic theory has as one of its aims the characterization of

various kinds of validity. That this is so will be important in what follows, because I will

appeal to such logical properties in trying to decide between different approaches to the

in-between phenomena.

2. PRAGMATICS, EVEN MORE BRIEFLY

Suppose, for the moment, that declarative sentences always literally express complete

propositions relative to a context of utterance. Still, what a speaker means by uttering a

sentence may be more than that the proposition it expresses is true. A speaker may utter

the sentence, “There’s a gas station around the corner,” and if that sentence expresses in

context the proposition that there is a gas station around the corner from place d, where d

is the location at which the speaker is speaking, then one thing the speaker may mean by

uttering that sentence is that there is a gas station around the corner from d. But in

addition the speaker may mean that the gas station is open and has gas to sell. Pragmatics

may be conceived as, in part, the study of the way in which speakers manage to mean

more by uttering sentences than what the sentences they utter literally mean relative to

the context in which they are uttered. That, at least, is the conception of pragmatics that

we have inherited from Grice (1989).

Inasmuch as meaning something by an utterance of a sentence is a state of a

speaker’s mind, this way of defining the subject matter of pragmatics assumes that it
Semantics and Pragmatics 8/4/10 Page 6

concerns the speaker’s state of mind. A different way of approaching pragmatics would

be to think of it as the study of the way in which hearers are able to draw conclusions on

the basis of their understanding of the literal meaning of a speaker’s words and their own

knowledge of the state of the environment in which the speaker’s utterance takes place

(Gauker 2001). A hearer who hears a speaker say, “There’s a gas station around the

corner”, may infer from the fact that there is a gas station around the corner and the fact

that he or she is standing at a busy intersection in the middle of the day that probably the

gas station is open and has gas to sell. The speaker may usually intend the hearer to draw

such inferences, but is not obvious that the hearer has to think about what the speaker had

in mind in order to do so.

3. THE LOGIC OF THE IN-BETWEEN PHENOMENA

Part of the challenge posed by the in-between phenomena is that they introduce logical

data that need to be accounted for. We find arguments that are valid in the sense that for

every context in which the premises are true, the conclusion is true, and other arguments

that are invalid in the sense that there is a single context in which the premise is true and

the conclusion is not. So our account of the truth conditions of sentences must explain

how the truth value of a sentence relative to a context depends on the content of the

context in such a way that in terms of that explanation we can explain why some

arguments are valid in this sense and others are not.

Imagine a store that sells nothing but unpainted wooden furniture. In conversing

with a customer, the salesman gestures toward the showroom and declares, “Everything

is made of wood”. In the context in which he speaks, we should regard what he says as
Semantics and Pragmatics 8/4/10 Page 7

true. We would not want to deem it not true, as it would be, for instance, if a plastic chair

were for sale. And yet, the telephone on his desk is not made of wood. The sentence

“The telephone is made of wood” is not true in that context (which is not to say that it is

false). So the argument from “Everything is made of wood” to “The telephone is made

of wood” does not preserve truth-in-a-context. But if there were any context in which

“The telephone is made of wood” were true, then in that context “Something is made of

wood” would be true in that context as well.

For another example, in any context in which “Tipper is ready” is true, “Tipper is

ready for something” will surely be true. And yet, there might be a context in which

“Tipper is ready for something” is true, though “Tipper is ready” is not true. Tipper

might be ready to take her next breath but not be ready for the thing that matters in the

context, such as going to the convention center.

4. APPROACHES TO THE MIDDLE GROUND

4.1. Minimalists

One approach to the in-between phenomena is just to say that, while their contribution to

the proposition that a sentence expresses is independent of the context of utterance, the

propositions expressed are only minimal propositions of a certain sort.

One kind of minimalism is that of Emma Borg (2004, 2007). Borg’s motivation

is the idea that a semantic theory should characterize those aspects of a speaker’s

knowledge of meaning that can be attributed to a cognitive module that deals with

language, while our understanding of contexts cannot be located in any particular

cognitive module. Borg will explain the truth condition of a sentence such as “This is

red” as follows (2004: 192):

(R) If a speaker utters a token of “This is red” and refers with the utterance of “this”
therein to an object x, then that utterance is true if and only if x is red.

Borg’s theory does not assign to the sentence “This is red” a proposition that it expresses,

but on her theory one can still say that an utterance of “This is red” expresses a

proposition, regarding the object referred to by the utterance of “this” therein, that it is

red. Borg is prepared to allow that in a sentence such as “Tipper is ready”, there is a

hidden argument place and a hidden quantification. So she says that “Tipper is ready”

has the same “underlying form” as “Tipper is ready for something” (2007: 350–351).

The minimalism of Cappelen and Lepore (2005) is motivated by the desire to

draw a line. Cappelen and Lepore allow that the proposition that a sentence expresses

may be relative to a context in some ways. A context may be called upon to provide an

interpretation for a limited number of items, such as explicit demonstratives, indexicals,

words and inflections indicating tense, and a few others (2005: 1). But they expect that

the arguments that might persuade us to countenance varieties of context-relativity

beyond those on their list would prove too much, leaving us with no account of what we

understand in understanding the meanings of words. They hold that a speaker who utters

the sentence “There are no French girls” (their example, 2005: 41) might thereby state

that there are no French girls in room 401, but what the speaker’s utterance expresses is

precisely the proposition that there are no French girls. A speaker who utters “The

students are ready” might be interpretable as saying that the students are ready for the
Semantics and Pragmatics 8/4/10 Page 9

exam, but what the speaker’s utterance expresses is precisely the proposition that the

students are ready (2005: 168).

A problem for Borg’s theory is that it is not evident that her truth-conditions for

sentences containing demonstratives can be derived from a recursive definition of truth

for a language. For example, the truth condition of “This is not red” ought to be a

consequence of (R) and a general fact about the truth conditions of negations. But what

is this general fact about the truth conditions of negations? We could try:

(N) If u is an utterance of a sentence s and v is an utterance of the negation of s, then

But this would be wrong for two reasons. First, (N) tells us the truth conditions of a

negation only on the condition that the sentence negated has been uttered, though it is not

in general true that if a negation is uttered then the sentence negated has been uttered.

And second, we cannot assume, as (N) does, that the demonstratives in the utterance of

the sentence negated refer to the same objects as the demonstratives in the utterance of

the negation. In place of (N), we might put:

(N!) If r is the negation of sentence s, then r is true in a context c if and only if s is

But this belongs to a definition of truth-in-a-context, not a definition of truth simpliciter,

whereas it was Borg’s ambition to avoid the relativization of truth to context.

Further, neither variety of minimalism accounts for all of the logical facts that the

in-between phenomena point to. Neither theory identifies any sense in which the

argument from “Everything is made of wood” to “The telephone is made of wood” might
Semantics and Pragmatics 8/4/10 Page 10

fail to be valid. Borg’s theory will not explain in what sense “Tipper is ready for

something” fails to imply “Tipper is ready” because she holds that these two sentences

have precisely the same truth conditions. On Cappelen and Lepore’s theory, what we say

will depend on what we say about the proposition that Tipper is ready, but it seems that

for anything that we might say about that, we will end up either affirming that “Tipper is

ready for something” implies “Tipper is ready” or denying that “Tipper is ready” implies

“Tipper is ready for something”, which seem to be wrong results.

4.2. Partialists

A second point of view on the in-between phenomena is that what they show is that

sentences frequently do not express complete propositions at all and that semantic theory

cannot aspire to anything more than ascribing to sentences a kind of gappy meaning that

a speaker can utilize to convey complete propositions to a hearer, the speaker’s meaning.

This is the sort of view advanced by Recanati (2001, 2004), Carston (1998, 2008) and

Bach (1994, 2005).

The partialists disagree among themselves on several questions. Recanati

disagrees with Carston on certain questions concerning the psychology of interpretation

(2004: 39). Carston (2008) does not understand why Bach thinks that the interpretation

of indexicals, such “I” and “now”, belongs to semantics. And Recanati, in defining a

kind of minimal propositional meaning, sometimes takes into account the hearer’s point

of view (2004: 19); whereas Bach finds no place for such a conception of meaning.

Nonetheless, the partialists all agree in thinking that linguistic convention alone will

usually not get as far as determining for a sentence, or even for a sentence in a context, a
Semantics and Pragmatics 8/4/10 Page 11

full proposition expressed. What semantics assigns to many sentences is only a

proposition radical (Bach) or linguistically encoded meaning (Carston).

Not only do the partialists not accommodate the special logic of the in-between

phenomena, they are altogether silent on questions of logical relations. Inasmuch as

sentences do not express propositions, not even relative to contexts, we cannot define

logical relations between sentences in terms of relations between the propositions

expressed. Carston, for one, holds that only mental representations, not spoken

sentences, are the proper object of truth-conditional semantics and that they are free of

the in-between phenomena; but to say that is not yet to address the question of the logic

of natural language. One can think of various accounts that one might offer on the

partialists’ behalf. If sentences express some kind of gappy meaning, short of a full

proposition, then sentence truth could be relativized to ways of filling the gaps, and

logical relations could be defined in terms of truth relative to ways of filling the gaps.

But as far as I know, none of the partialists has actually addressed the issue in print.

4.3. Indexicalists

An influential model for the treatment of the in-between phenomena has been Kaplan’s

treatment of indexicals such as “I” and “now” (1989). According to Kaplan, the truth of

a sentence must be relativized to both a possible world and a context. A context, for

Kaplan, is an object that assigns to each indexical an appropriate referent. For example,

to “now” it assigns a time. Thus, to each expression we can assign a function from

contexts to intensions, which in turn may be thought as functions from worlds to

extensions. Kaplan called these functions from contexts to intensions characters.

The indexicalist strategy for explicating the in-between phenomena, exhibited in a

series of papers by Stanley (e.g., 2002, 2005), is to treat all of them along the same lines

as indexicals. The in-between phenomena are to be thought of as involving additional,

unspoken lexical items that appear in the deep grammatical structure of a sentence and

receive an interpretation from the context. For example, in Stanley and Szabó 2000,

domains of discourse for quantifiers are supposed to be sets that contexts assign to

variables associated in deep structure with nominal phrases. (I use the term “deep

structure” to refer to mental representations of the grammatical structure of a sentence.

Contemporary conceptions of this deep structure vary.) Similarly, gradable adjectives,

such as “small”, could be handled by letting context assign to them a standard, such as an

average height (Kennedy 2007). King and Stanley even suggest that the way to handle

the context-relativity of conditionals is to let context assign a similarity relation (between

possible worlds) to the words “if” and “then” (2005: 152).

The indexicalist strategy suggests a clear division of labor between semantics and

pragmatics and also accommodates at least some of the logical facts at issue. A semantic

theory for a language will take the form of a recursively specified assignment of a

character, in Kaplan’s sense, to each expression (with contexts now extended as

necessary). It will fall to pragmatics to explain what it takes for a given context to be the

one that pertains to a given utterance of a sentence, that is, to be the context such that we

should evaluate the utterance as true (simpliciter) if and only if the sentence uttered is

true relative to that context in the world in which the utterance occurs. The indexicalist

can even account for some of the special logical facts due to context-relativity, such as

the fact that the argument having “Everything is made of wood” as premise and “The
Semantics and Pragmatics 8/4/10 Page 13

telephone is made of wood” as conclusion is in a sense not valid, by defining the

pertinent kind of validity thus: An argument is valid if and only if for each context if the

premise are true in that context then the conclusion is true in that same context.

However, it is doubtful whether the indexicalist strategy can explain all of the

logical facts pertaining to the in-between phenomena. To see this, consider first a

sentence containing two demonstratives, such as:

(1) This is delicious and this is not delicious.

There is nothing contradictory about this sentence, and if it seems odd on the printed

page, that is only because we cannot observe the gestures that would tell us what different

foodstuffs the two occurrences of “this” denote. To acknowledge that (1) expresses a

truth, we do not have to envision that the context somehow shifts between the evaluation

of the first conjunct and the evaluation of the second. Sentence (1) may be true in a

single context because the two occurrences of “this” may denote different objects in a

single context.

By contrast, each of the following sentences is odd in an additional way:

(2) Every student is happy and some student is not happy.

(3) Tipper is ready and Tipper is not ready.

(4) Dumbo is small, and Mickey is large, and Dumbo is larger than Mickey.

There is an oddity in each of these sentences that our semantic theory should enable us to

characterize (Gauker 2010a). The oddity cannot readily be characterized as merely

pragmatic. We can perhaps imagine situations in which one of these sentences would be

uttered, but even then the situation is either one in which the speaker has spoken poorly
Semantics and Pragmatics 8/4/10 Page 14

though we can see what he or she has in mind or one in which the oddity is deliberately

exploited for special effect (for example, to pose a riddle). The contradictory character of

(2)–(4) is a problem for the indexicalist because, for the indexicalist, (2)–(4) will be

noncontradictory, just as (1) is, since the context-relative elements in these sentences

should be interpretable independently, just as they are in (1).

Incidentally, it has sometimes been supposed that pure indexicals, such as “I” and

“now” differ from demonstratives such as “this” in that the former but not the latter are

governed by strict rules such as that “now” always denotes the time of utterance, and

Kaplan himself may have presumed as much. In light of answering-machine cases and

post-it note cases (Predelli 2003), that doctrine has been thrown in doubt. Consequently,

it is now widely doubted that there is a special logic of indexicals that would ensure that

the sentence “I am here now” has a special kind of validity.

5. A proposal: Contexts versus situations

So far I have followed the common practice of using the term “context” equivocally. I

have treated contexts as parameters relative to which we define the truth of sentences of a

language, and I have spoken of contexts as contexts of utterance, which we may think of

as situations in which an utterance takes place. The first step toward dividing the labor

between semantics and pragmatics in accounting for the in-between phenomena will be to

strictly distinguish between contexts proper and situations. A context will be a certain

formal structure relative to which the truth of sentences will be defined. A situation will

be an arrangement of objects and events such as may include an utterance of a sentence.

Semantics will deal with the definition of truth for sentences relative to contexts.
Semantics and Pragmatics 8/4/10 Page 15

Pragmatics will deal with the conditions under which a given context is the one that

pertains to a given utterance situation.

5.1. Semantics in light of this distinction

Although situations are endlessly complex arrangements of objects and events, contexts

are precisely definable structures. So we have the possibility of precisely defining the

whole set of contexts for a given language. And we have the possibility of defining in a

precise way the conditions under which an arbitrary sentence is true relative to an

arbitrary context. The logical phenomena pertaining to the in-between phenomena can

then be handled by defining an argument to be in a sense valid if and only if it preserves

truth-in-a-context and defining a sentence to be in a sense contradictory if and only if it is

true in no context.

To accommodate context-relative domains of discourse, we can let a context

include a set of objects, the domain (Gauker 1997a). To accommodate the context-

relativity of gradable adjectives, we can let contexts include, for each gradable adjective,

a standard that an object must surpass in order for the adjective to apply. To

accommodate incomplete predicates, we may include a set of propositions and say that,

for example, “Tipper is ready” is true in a context if and only if there is some activity

such that a proposition to the effect that Tipper is ready for that activity belongs to the

contextually-determined set of propositions and there is no activity such that the

proposition that Tipper is not ready for that activity belongs (Gauker 2010b). To

accommodate metonymy we can let contexts include a deferment function that might, for

example, take a ham sandwich as input and yield as output the patron who ordered a ham
Semantics and Pragmatics 8/4/10 Page 16

sandwich. To accommodate the context-relativity of conditionals we might let a context

include (or even be) a set of prospects (themselves conceived as contexts), so that we can

say that an indicative conditional is true in a context if and only if for each prospect in (or

identical to) the context, if the antecedent is true in that prospect, then so is the

consequent (Gauker 2005).

In light of these characterizations of context and the definitions of logical

properties in terms of truth-in-a-context, we can hope to capture the logical data

introduced by the in-between phenomena. We can say that the argument from

“Everything is made of wood” to “The telephone is made of wood” is invalid, because

there is a context c such that every object in the domain of discourse for c is made of

wood but such that the telephone is not in that domain. The sense in which “Tipper is

ready for something” fails to imply “Tipper is ready” is that there is a context in which

the contextually-determined set of propositions contains the proposition that Tipper is

ready for a certain activity, so that “Tipper is ready for something” is true in that context,

but also contains the proposition that Tipper is not ready for a certain other activity, so

that “Tipper is ready” is not true in that context.

To account for the contradictory aspect of (2) we may suppose that the context

specifies a global domain not assigned to any particular indexical. In (2), the truth of

“Every student is happy” in a context demands that the domain for the context include

only happy students, while the truth of “Some student is not happy” in context demands

that the domain include at least one student who is not happy. (More work has to be done

to accommodate the kind of sentences that motivate Stanley, such as “Every student

answered every question” and “Every sailor waved to every sailor”. See Gauker 2010a.)
Semantics and Pragmatics 8/4/10 Page 17

Insofar as a context contains a single contextually-determined set of propositions, we

may find that the conditions under which “Tipper is not ready” is true relative to a

context are dual to the conditions under which “Tipper is ready” is true relative to a

context, so that (3) is true in no context. As for (4), a context may specify a single

standard relative to which both “small” and “big” are evaluated, so that if Dumbo is

larger than Mickey, then no matter our contextually-determined standard for smallness, if

Dumbo satisfies it, then Mickey will not satisfy our contextually-determined standard for

bigness.

5.2. Pragmatics in light of this distinction

We may assume that quite generally an utterance of a declarative sentence is true

(simpliciter) if and only if the sentence uttered is true relative to the context that pertains

to the utterance. In this light, the question for pragmatics becomes: What is it about the

situation in which an utterance takes place that determines that a particular context is the

context that pertains to that utterance?

Many authors writing about the in-between phenomena, including all of those

mentioned in section 4, assume that the shortfall between the semantic properties of the

sentence uttered and what a given utterance of that sentence expresses is made up by the

content of the speaker’s communicative intentions. So long as interpretation is conceived

simply as filling the gaps left open by literal meaning, it may be hard to conceive of any

alternative. But once we conceive of the issue for pragmatics as that of identifying the

context that pertains to a given utterance, alternatives become conceivable. One could

still insist on the primacy of speakers’ intentions, by holding that nothing but the
Semantics and Pragmatics 8/4/10 Page 18

speaker’s intention picks out the context pertinent to a speaker’s utterance. But an

alternative that comes into view is that the pertinent context might be fixed by other

features of a situation, such as these: what is perceptually salient in the immediate

environment, the facts that are relevant to the interlocutors in light of their goals, what it

is true or reasonable to say, and anaphoric connections between context-relative elements

in a sentence and prior discourse (Gauker 2008b).

One problem facing the assumption that speaker intention fixes context is that it is

just not clear what the contents of the pertinent speaker intentions are supposed to be. In

the case of demonstratives, the answer may be clear enough: The intention that

determines that the pertinent context assigns object o to “that” is the intention to draw the

hearer’s attention to object o. But it is far less clear what the content of the intention

would be that determines the membership of the domain of discourse for the pertinent

context. It cannot be, for instance, the intention, for some set D, that the sentence uttered

be evaluated relative to a context for which the domain of discourse is D. That account

of it would be question-begging in that it presumes that we have a prior understanding of

what it is for a certain set to be the domain of discourse for an utterance.

Another basic question for the assumption that speaker intention fixes context is

how a hearer is supposed to recognize the speaker’s intentions in speaking. Normally,

hearers have little access to what a person has in mind apart from an understanding of

what the person says. But, as the in-between phenomena indicate, hearers normally have

to identify the context pertinent to a speaker’s act of speech in order to figure out what

the speaker has said. So normally, hearers will have to have some means of identifying

the pertinent context without independent access to the speaker’s intention. The question
Semantics and Pragmatics 8/4/10 Page 19

then becomes whether speakers have sufficient independent access to the speaker’s

intention or whether the speaker’s intention can determine the pertinent context although

that hearers normally identify the pertinent context on some other basis.

Finally, one has to give some thought to the mental representations that are the

speaker’s underlying intentions. It is not plausible that every aspect of the meaning of an

utterance that is not written out explicitly in the sentence uttered is somehow written out

explicitly in the brain. The meanings of mental representations will depend on

contextually-determined referents and contextually-determined domains of discourse,

much as the meanings of spoken sentences do (Gauker 1997). So to assume that the

speaker’s intention is what determines the context pertinent to a speaker’s utterance just

puts off explaining how something other than speaker’s intention can determine the

context pertinent to a particular token representation.

6. BEYOND THE IN-BETWEEN

There are many linguistic phenomena that challenge the core conception of semantics

described in section 1 although they do not appear to belong to the same family as the in-

between phenomena listed in the introduction. These include anaphoric pronouns, words

that indicate a rhetorical relation between what is being said and prior discourse (such as

“but” and “anyway”), and adjectives whose meanings seems to depend on the nouns they

modify (such “good” in “good actor”). The proposal of section 5.1 does not immediately

suggest a means of accommodating these phenomena in a compositional, or recursive,

semantic theory. But perhaps it gives us reason to hesitate before giving up on the idea
Semantics and Pragmatics 8/4/10 Page 20

that that a compositional or recursive theory of suitable semantic values of some kind

relativized to parameters of some kind might encompass such phenomena as well.

RELATED TOPICS

1.3 Quantifiers and Determiners

1.4 Anaphora

1.7 Adjectives

1.11 Indexicals and Demonstratives

1.12 Indicative Conditionals

1.13 Subjunctive Conditionals

3.9 Philosophy of language for epistemology

4.1 Extensions, Intension, Characters and Beyond

4.3 Presupposition and Implicature

4.7 Pragmatic Enrichment

4.12 Context Sensitivity

5.11 Relevance Theory

5.11 Situation Semantics

REFERENCES

Bach, Kent (1994) “Conversational Impliciture,” Mind and Language 9: 124–162.

Bach, Kent (2005) “Context ex Machina,” in Z. Gendler Szabó (ed.) Semantics vs.

Pragmatics, Oxford: Oxford University Press, pp. 15–44.

Borg, Emma (2004) Minimal Semantics, Oxford: Oxford University Press.

Borg, Emma (2007) “Minimalism versus Contextualism in Semantics,” in G. Preyer and

G. Peter (eds.) Context-Sensitivity and Semantic Minimalism: New Essays on

Semantics and Pragmatics, Oxford: Oxford University Press, pp. 339–359.

Cappelen, Herman and Ernie Lepore (2005) Insensitive Semantics: A Defense of

Semantic Minimalism and Speech Act Pluralism, Oxford: Blackwell Publishing.

Carston, Robyn (1998) “The semantics/pragmatics distinction: a view from relevance

theory,” UCL Working Papers in Linguistics 10: 1–30.

Carston, Robyn (2008) “Linguistic Communication and the Semantics/Pragmatics

Distinction,” Synthese 165: 321–345.

Gauker, Christopher (1997a) “Domain of Discourse,” Mind 106: 1–32.

Gauker, Christopher (1997b) “Universal Instantiation: A Study of the Role of Context in

Logic,” Erkenntnis 46: 185–214.

Gauker, Christopher (2001) “Situated Inference versus Conversational Implicature,”

Noûs 35: 163–189.

Gauker, Christopher (2005) Conditionals in Context, Cambridge, MA: MIT Press.

Gauker, Christopher (2008b) “Against Accommodation: Heim, van der Sandt, and the

Presupposition Projection Problem,” in J. Hawthorne (ed.) Philosophical

Perspectives, 22, Philosophy of Language, Oxford: Blackwell Publishing, pp. 171–

205.

Gauker, Christopher (2008b) “Zero Tolerance for Pragmatics,” Synthese 165: 359–371.

Gauker, Christopher (2010a) “Global Domains vs. Hidden Variables,” The Journal of

Semantics 27: 298–328.

Gauker, Christopher (2010b) “What Tipper is Ready for: A Semantics for Incomplete

Predicates,” Noûs (forthcoming).

Grice, Paul (1989) Studies in the Way of Words, Cambridge, MA: Harvard University

Press.

Kaplan, David (1989) “Demonstratives: An Essay on the Semantics, Logic,

Metaphysics, and Epistemology of Demonstratives and Other Indexicals,” in J.

Almog, J. Perry, and H. Wettstein (eds.) Themes from Kaplan, Oxford, Oxford

University Press, pp. 481–564.

Kennedy, Christopher (2007) “Vagueness and Grammar: The Semantics of Relative and

Absolute Gradable Adjectives,” Linguistics and Philosophy 30: 1–45.

King, Jeffrey C. and Jason Stanley (2005) “Semantics, Pragmatics, Semantic Content,” in

Z. Gendler Szabó (ed.) Semantics versus Pragmatics, Oxford: Oxford University

Press, pp. 111–164.

Predelli, Stefano (2003) Contexts: Meaning, Truth and the Use of Language, Oxford:

Oxford University Press.

Recanati, François (2001) “What is Said,” Synthese 128: 75–91.

Recanati, François (2004) Literal Meaning, Cambridge: Cambridge University Press.

Stanley, Jason (2002) “Making it Articulated,” Mind and Language 17: 149–168.

Stanley, Jason (2005b) “Semantics in Context,” In G. Preyer and G. Peter (eds.)

Contextualism in Philosophy, Oxford: Oxford University Press, pp. 221–253.

Stanley, Jason, and Zoltán Szabó (2000) “On Quantifier Domain Restriction,” Mind and

Language 15: 219–261.