SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Literal Meaning and Objectivity
  5. Crary's Arguments Against Literal Meaning
  6. The Illocutionary Force Argument
  7. Open-Ended Variation in Truth Conditions
  8. Concluding Remarks
  9. References

Alice Crary has recently developed a radical reading of J. L. Austin's philosophy of language. The central contention of Crary's reading is that Austin gives convincing reasons to reject the idea that sentences have context-invariant literal meaning. While I am in sympathy with Crary about the continuing importance of Austin's work, and I think Crary's reading is deep and interesting, I do not think literal sentence meaning is one of Austin's targets, and the arguments that Crary attributes to Austin or finds Austinian in spirit do not provide convincing reasons to reject literal sentence meaning. In this paper, I challenge Crary's reading of Austin and defend the idea of literal sentence meaning.

Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Literal Meaning and Objectivity
  5. Crary's Arguments Against Literal Meaning
  6. The Illocutionary Force Argument
  7. Open-Ended Variation in Truth Conditions
  8. Concluding Remarks
  9. References

J. L. Austin's work is emerging from a period of relative neglect to occupy a prominent position in contemporary theories of language. François Recanati and Charles Travis cite J. L. Austin as an inspiration for their radical contextualist criticisms of truth conditional semantics, and Austin's work is also credited as a source of a more modest epistemological contextualism.1 Even though contextualists are inspired by Austin and draw examples of context sensitivity from his work, they have not spent much time on Austin exegesis.2 In contrast, Alice Crary has recently developed an extended and radical reading of Austin (Crary 2002; 2007). The central contention of Crary's reading is that Austin gives convincing reasons to reject the idea that sentences have context-invariant literal meaning.3 While I am in sympathy with Crary about the continuing importance of Austin's work, and I think Crary's reading is deep and interesting, I do not think literal sentence meaning is one of Austin's targets, and the arguments that Crary attributes to Austin or finds Austinian in spirit do not provide convincing reasons to reject literal sentence meaning. In this paper, I challenge Crary's reading of Austin and defend the idea of literal sentence meaning.

Literal Meaning and Objectivity

  1. Top of page
  2. Abstract
  3. Introduction
  4. Literal Meaning and Objectivity
  5. Crary's Arguments Against Literal Meaning
  6. The Illocutionary Force Argument
  7. Open-Ended Variation in Truth Conditions
  8. Concluding Remarks
  9. References

Crary's discussion of Austin and literal meaning is part of a wider attack she mounts against a metaphysical assumption that she calls the ‘abstraction requirement’. The abstraction requirement holds that for a concept to count as objective, it must be independent of our ‘subjective endowments’, because (according to advocates of the abstraction requirement) ‘the subjective (i.e., perceptual and affective) endowments we draw on in thinking and talking about how things are have an essential tendency to distort our view of reality … it is therefore only by abstracting from such endowments that we can assure ourselves of having our minds around more than mere appearance’ (Crary 2007: 20).4

Crary aims to replace the conception of objectivity that meets the ‘abstraction requirement’ with a broader conception of objectivity that makes room for the idea that a wide range of concepts (including moral concepts) count as objective even though they involve subjective capacities and responses. Crary is interested in the concept of literal meaning because it seems to ‘encode’ the abstraction requirement:

To speak of literal sentence meaning is … to assume both that there must be rules for [identifying the contribution that context-sensitive expressions make to the meaning of a sentence] (e.g., rules for assigning values, in particular circumstances, to the parameters that indexicals pick out) and, further, that is must be possible to represent our mastery of these rules, when combined with our knowledge both of the meanings of the words that compose a sentence and of other rules of the language, as a kind of algorithm for generating the meaning of the sentence as uttered on particular occasions. Now, if we assume that we can arrive at a grasp of the meaning of a sentence by thus applying a kind of algorithm, and if we also assume (as advocates of the idea of literal sentence-meaning generally do) that our ability to apply such an algorithm is independent of our possession of any particular subjective endowments, it will appear to be possible to grasp the meaning of a sentence and to determine whether it accurately represents the world in a manner that satisfies an abstraction requirement. (ibid.: 50)

Many semantic theorists do believe that our knowledge of meaning can be represented as ‘a kind of algorithm for generating the meaning of the sentence as uttered on particular occasions’, in that they think that the meaning of sentences is built up out of the meaning of their parts by repeated application of simple operations like function application.5 And, insofar as determining the meaning of a sentence consists in something akin to performing simple mathematical operations, it is possible for something without ‘our particular subjective endowments’—a computer, say—to do so.

But as far as I know, no one thinks that ‘determining whether [a sentence] accurately represents the world’ is algorithmic. Even if one knows the referents of all context-sensitive expressions in a sentence, determining whether the sentence accurately represents the world may require a great deal of additional ingenuity. For example, if I point to an airliner passing far overhead and say ‘That airplane is carrying more than 100 passengers’, you may be able to compositionally determine the meaning of the sentence, but even if you are able to figure out what airplane I'm referring to, determining whether what I have said accurately represents the world would require a substantial amount of additional work—it would require somehow finding out how many passengers were on board that particular flight. That is clearly beyond anything knowledge of literal meaning could accomplish. And so I find it strange to attribute the view that ‘whether [a sentence] accurately represents the world’ can be determined ‘independent[ly] of our possession of any particular subjective endowments’ to ‘advocates of literal sentence meaning’.6 It is unlikely that there are any ‘advocates of the idea of literal sentence meaning’, if literal sentence meaning is characterized as it is by Crary in the passage quoted above.

But if Crary's formulation of the concept of literal sentence meaning is modified so as to remove the requirement that grasp of literal meaning enables one to determine whether the sentence accurately represents the world, it does pick out a conception of literal meaning that some contemporary philosophers subscribe to. That modified conception of literal meaning is that grasp of the meanings of words and the rules by which they are combined into complex expressions (including sentences) enables one to know what has to be the case in order for the sentence to be true. That is, grasping the literal meaning of the sentence enables one to know the sentence's truth-condition. Recanati calls this view ‘Literalism’:

[T]he meaning of the expression provides a rule which, given the context, enables the interpreter to determine the content [the truth condition, when the expression is a sentence] of the expression in that context. The content thus determined in context by the conventional meaning of words is their literal content. (Recanati 2004: 3)

Recanati says that Literalism is ‘the dominant position’ in theorizing about the relation between semantics and pragmatics.7

One might think that Literalism, with its commitment to the idea of assigning sentences truth conditions, is at odds with Austin's view that it is statements, and not sentences, that are the bearers of truth value. Austin observes that sentences, considered apart from the circumstances in which they are used, cannot be classified as true or false:

If you take a bunch of sentences … there can be no question of sorting them out into those that are true and those that are false; for …the question of truth and falsehood does not turn only on what a sentence is, nor yet on what it means, but on, speaking very broadly, the circumstances in which it is uttered. (Austin 1962: 110–11)

But that idea, and the idea that it is only particular uses of sentences that have truth values, is compatible with the assignment of truth conditions to sentences. A sentence like ‘Today is fine’ might be assigned the truth condition An utterance u of ‘Today is fine’ is true just in case the day on which u is uttered is fine, which makes it manifest that it isn't the sentence that is true or false, but utterances of the sentence in particular circumstances.8

Does Austin, according to Crary, give an argument against literal meaning understood in the sense of the idea that sentences can be assigned truth conditions algorithmically? That would be an interesting—indeed radical—argument. Or does Austin merely repudiate the dubious idea that whether a sentence accurately represents the way the world is can be determined algorithmically, without relying on our ‘subjective endowments’? It seems that Crary wants to attribute the more radical argument to Austin.

Crary's Arguments Against Literal Meaning

  1. Top of page
  2. Abstract
  3. Introduction
  4. Literal Meaning and Objectivity
  5. Crary's Arguments Against Literal Meaning
  6. The Illocutionary Force Argument
  7. Open-Ended Variation in Truth Conditions
  8. Concluding Remarks
  9. References

Crary rejects two influential, opposed readings of Austin advanced by John Searle and Jacques Derrida (Derrida 1988, Searle 1968). Crary argues that both Searle and Derrida assume that ‘the notion of objective truth stands or falls with the idea of literal sentence meaning’ (Crary 2007: 73). According to Crary, Searle (in his early interpretations of Austin) takes this tie between literal sentence meaning and objective truth to ‘speak in favor of searching for an interpretation of Austin's doctrine of speech acts that is consistent with the idea of such [literal] meaning’, while Derrida moves in the other direction, arguing that even though Austin presents himself as ‘receptive’ to the idea of objective truth, preserving a commitment to objective truth is in tension with Austin's attack on literal meaning, ‘and that charity speaks for reading Austin as rejecting the notion of objective truth outright’. Crary aims to undermine both readings of Austin by rejecting the shared assumption: One can reject literal sentence meaning while retaining a commitment to objective truth (Crary 2002: 73–4, 2007: 72–3).

Crary gives two arguments that are supposed to show that we have reason to give up the idea of literal sentence meaning. The first argument claims that it is only once one has identified the illocutionary act performed in the utterance of a sentence that one can identify the literal meaning of the sentence uttered. That would make it impossible to assign literal meanings (algorithmically determined truth conditions) to sentences prior to observing their use in particular utterances, thereby blocking the standard role that literal meaning plays in explanations of our ability to understand novel sentences of the language. The second argument is a version of radical contextualist arguments targeting the idea that the meaning of a sentence determines a truth condition independently of a non-systematizable contribution from the context of utterance. An essential part of this second argument involves the attempt to show, on a priori grounds, that attempts to systematize the contribution that the context of utterance makes to the truth-conditional content of utterances must fail. I will argue that these arguments do not succeed in problematizing the idea of literal meaning and that it is doubtful that they capture Austin's attitude towards the idea of literal meaning.

The Illocutionary Force Argument

  1. Top of page
  2. Abstract
  3. Introduction
  4. Literal Meaning and Objectivity
  5. Crary's Arguments Against Literal Meaning
  6. The Illocutionary Force Argument
  7. Open-Ended Variation in Truth Conditions
  8. Concluding Remarks
  9. References

Crary sets up the first argument she attributes to Austin by introducing an example. She says,

Let me offer one illustration of the kinds of considerations that make Austin suspicious of the idea that sentences have ‘literal’ or ‘conventional’ meanings that they carry with them into every context of their use. Consider the sentence ‘The car is red.’ If any sentence has a clear ‘literal meaning’, this sentence seems like a good candidate. It appears to have a constant and unvarying descriptive meaning—one which its indexical elements (along with contextual clues) will pin down to the particular contexts in which it is used. Imagine the following situation:

Two people, A and B are sitting at a table that has on it a checker set and a small green toy car. A picks up the green car and says: ‘The car is red.’ B is confused. Nonetheless, because she takes herself to understand (what she thinks of as) the ‘literal meaning’ of A's sentence, she thinks she can determine (by appealing to this ‘literal meaning’ along with a few contextual clues) what factual statement A is making. Suppose that B now responds: ‘Of course it's not red. It's green.’ A laughs and rejoins: ‘That's not what I meant. This checker set is missing one of its red pieces. I was suggesting that we use this little car here as one of the red pieces!’ (Crary 2002: 68–9)

Crary reads the example as showing that it is only once the illocutionary act performed in the utterance of the sentence has been understood is it possible to determine whether the meaning of ‘The car is red’ is the same when it is used to make a statement as it is when it is used to establish a rule (as it does in the example). Crary says that what the example ‘teaches’ is that

it is only once B has understood how A's sentence is to be taken that she is in a position to make such a decision [a decision about the relation between the meaning of the sentence as used in a statement and the sentence used in the creation of a rule] at all … So it is not in virtue of her grasp of (what she thinks of as) the literal meaning of A's sentence that B understands what A is saying. This means that B's appeal to the literal meaning of A's sentence does not here play the role that it is often called on to play in philosophy. (ibid.: 69)

Crary doesn't say explicitly what the ‘role that [literal meaning] is often called on to play in philosophy’ is, but it seems clear from the context of her argument that she has in mind the idea that sentences that have different illocutionary forces can share the same ‘core’ of literal meaning. According to this picture, a semantic theory is a theory of this literal meaning ‘core’, and a ‘theory of force’ supplements the theory of literal meaning. The theory of force would give rules that determine how a sentence with a single ‘core’ literal meaning can be used with a variety of different illocutionary forces.9 On this standard picture, (disambiguated) sentence types have literal meanings which can be studied in isolation from the illocutionary forces those sentences express when uttered. The standard justification for this division of labor is that it is in the service of an explanation of speakers’ ability to learn a language and understand novel utterances (employing familiar words) of the language. Only if the expressions of the language have a meaning that does not vary from context to context can an explanation be given of how a speaker can understand novel utterances in the language.

With that understanding of the ‘role that [literal meaning] is often called on to play in philosophy’, how is the argument Crary finds in Austin supposed to work? The basic thought is that person B in Crary's example can't determine whether the literal meaning of ‘This car is red’ is the same when it is used as a description of a fact and when it is used as the creation of a rule until she has understood the illocutionary force of both utterances. If that were the case, then the project of explaining how speakers understand novel utterances by explaining how speakers first determine the semantic value of the literal meaning ‘core’ and then combine that with features of the context of utterance to determine the illocutionary force with which the sentence was uttered would not be possible. One would not be able to grasp the literal meaning ‘core’ without also grasping the illocutionary force, so grasp of force could not be explained in terms of a prior grasp of the ‘core’ plus some supplementary rules for the determination of force.

But I don't see why the example needs to be understood in the way that Crary understands it. A defender of the traditional role of literal meaning would say that person B successfully grasps the literal meaning of ‘This car is red’ in Crary's example, but that she misjudges the relevant contextual clues that indicate what illocutionary force the utterance of the sentence has, wrongly taking the utterance to be an assertion when it is in fact instituting a rule. Crary doesn't provide any reasons for thinking that this way of understanding the example is mistaken, so the standard view of the relation between literal meaning and illocutionary force remains unshaken.

Whether or not Crary's first argument is successful in challenging the received view of literal meaning, is it an argument that Austin would accept? I don't think it is. Though Crary identifies Austin's ‘locutionary act’ with ‘grammatically correct and meaningful sentences’ (2007: 62) and attributes to Austin the view that identifying the locutionary act performed by an utterance requires ‘an appreciation’ of the illocutionary act that the utterance is used to perform, Austin nowhere explicitly commits himself to the idea that identifying the locutionary act performed by an utterance requires an appreciation of the illocutionary act performed by that utterance as well.10 And in fact, Austin makes some remarks that seem incompatible with the idea that identifying the locutionary act performed by an utterance requires ‘an appreciation’ of the illocutionary act the utterance is used to perform. For example, Austin says that it ‘might be perfectly possible … to make entirely plain “what we were saying” ’ (that is, what phonetic, phatic and rhetic acts—which together make up the locutionary act—were performed when the sentence was uttered) when we utter a sentence, without indicating what illocutionary act was performed by the utterance of the sentence:

it might be perfectly possible, with regard to an utterance, say ‘It is going to charge’, to make entirely plain ‘what we were saying’ in issuing the utterance, in all the senses so far distinguished [phatic, rhetic and phonetic acts], and yet not at all to have cleared up whether or not in issuing the utterance I was performing the act of warning or not. It may be perfectly clear what I mean by ‘It is going to charge’ or ‘Shut the door’, but not clear whether it is meant as a statement or warning, &c. (Austin 1975: 98)

If it is possible to make plain what was said (in the sense of the components of the locutionary act) by the utterance of a sentence without indicating what illocutionary act was performed by the utterance, then one does not need to appreciate what illocutionary act was performed by the utterance of a sentence before one can grasp ‘what is said’ (that is, the phatic and rhetic acts, the components of the locutionary act performed) by the utterance of the sentence.11

Moreover, when Austin illustrates the difference between locutions, illocutions and perlocutions, the way he identifies locutionary acts appears to be independent of any appreciation of the illocutionary acts it is used to perform:

Act (A) or Locution

He said to me ‘Shoot her!’ meaning by ‘shoot’ shoot and referring by ‘her’ to her.

Act (B) or Illocution

He urged (or advised, ordered, &c.) me to shoot her (ibid.: 101–2).

Austin therefore appears committed to the possibility of identifying the locutionary act performed by an utterance without an appreciation of the illocutionary act performed.12

Open-Ended Variation in Truth Conditions

  1. Top of page
  2. Abstract
  3. Introduction
  4. Literal Meaning and Objectivity
  5. Crary's Arguments Against Literal Meaning
  6. The Illocutionary Force Argument
  7. Open-Ended Variation in Truth Conditions
  8. Concluding Remarks
  9. References

Crary presents a second argument that ‘might lead us to share Austin's hostility to the idea of literal sentence-meaning. [Crary's] goal in [explaining the argument] is to suggest … that we should follow in his footsteps in rejecting this idea …’ (Crary 2007: 74). About this argument, Crary says the following:

One of Austin's methods for making this point involves showing us how, even when its indexical elements are held constant, a given sentence may be used to express thoughts that not only differ but, more specifically, differ in truth values. (ibid.: 75)

Is it correct to attribute this method (which, as Crary acknowledges, is most prominent in the work of Charles Travis) to Austin?13 In support of her claim about Austin, Crary cites Austin's remarks about the sentences ‘France is hexagonal’, ‘Lord Raglan won the battle of Alma’ and ‘All snow geese migrate to Labrador’. Crary says that in these cases, Austin presents

sentences with a simple descriptive content (e.g., ‘France is hexagonal’) and observ[es] that they can be used to express thoughts that not only differ but, moreover, differ in their truth-values. (Thus ‘France is hexagonal’ may say something true when uttered by a topranking general and something false when uttered by a geometer …). (ibid.: 68)

But this characterization of Austin's method isn't exactly right.14 In his discussion of ‘France is hexagonal’, and ‘Lord Raglan won the battle of Alma’, Austin does not claim that the truth value of the thoughts expressed by sentences changes in different contexts. Instead, of ‘France is hexagonal’, he says ‘It is a rough description; it is not a true or a false one’ (Austin 1975: 143), and of ‘Lord Raglan won the battle of Alma’, he says ‘As “France is hexagonal” is rough, so “Lord Raglan won the battle of Alma” is exaggerated and suitable to some contexts and not to others; it would be pointless to insist on its truth or falsity’ (ibid.: 144). Austin is not saying that these sentences can be used to express thoughts that differ in their truth values depending on changes in the context of utterance. He is saying that it is pointless to insist on their truth and falsity.

But it might be possible to make a case for Crary's claim that Austin expresses commitment to the view that variations in truth value of a sentence are possible even when indexical elements of the sentence are held constant by first pointing out that Austin does say that ‘intents and purposes’ are important for assessing an utterance's truth or falsity:

Suppose that we confront ‘France is hexagonal’ with the facts, in this case, I suppose, with France, is it true or false? Well, if you like, up to a point; of course I can see what you mean by saying that it is true for certain intents and purposes. It is good enough for a top-ranking general, perhaps, but not for a geographer …

[I]n the case of stating true or falsely, just as much as in the case of advising well or badly, the intents and purposes of the utterance and its context are important; what is judged true in a school book may not be so judged in a work of historical research. (ibid.: 143)

And one might then argue that in this passage Austin is endorsing something very much like the contemporary contextualist commitment to the idea that the truth value of an uttered sentence can vary depending on a variety of features of context that go beyond those needed to fix the value of indexical elements in the sentence.

Whether or not the strategy can be traced to Austin, how is commitment to the idea that sentences can have different truth-values in different contexts, even when the values of all the indexical elements of the sentence are held fixed, supposed to convince us to reject literal sentence meaning?

Crary begins her argument by asking us to consider the sentence ‘This vase is red’. She says that even once the ‘indexical elements [of the sentence] are pinned down to a particular setting, it can be used to express different thoughts’ (with different truth conditions). Then she introduces the following two scenarios in which the sentence is uttered:

  1. A and B are walking through a pottery studio, and A stops in front of a shelf that holds a set of finished items that have been painted with a green glaze. A is explaining to B that, while the other potters in the studio use a more standard gray clay, she herself works with red clay found in parts of the American Southwest. Pointing to a vase she made that, like the other items, is glazed green, A says: ‘This vase is red’.
  2. A is in the pottery studio when B, a painter, calls to ask whether she has any red pieces that could be used for color in a still life painting. Standing in front of the shelf that holds her vase, A says ‘yes’ and then adds, ‘This vase is red’. (Crary 2007: 75–76)

After describing the two scenarios in which the sentence ‘This vase is red’ is uttered, Crary then says: ‘It seems natural to describe the example comprised of these two scenarios as illustrating that the sentence “The vase is red” can be used to say two distinct things—one of them (viz., [1]) true, the other (viz., [2]) false’ (Crary 2007: 76).

While I agree that it does seem natural to describe the cases that way, we might, with Austin, equally naturally say that the first utterance is ‘suitable’, while the second is ‘unsuitable’, but resist describing the case in terms of a difference in truth value. Or, again, equally naturally, we might think that both utterances are true, since there is some sense in which the vase is red in both cases, but in the second context, the utterance is misleading (Sainsbury 2001). Or we might think that both utterances are false, since normally, when you describe a vase as colored, you are talking about what color it is glazed. So there are at least four different options with regard to how to describe the truth values of the utterances in Crary's example, and in terms of intuitive plausibility and ‘naturalness’, none is obviously preferable to any other.

Even if Crary's intuitions were the most plausible response to the case she describes, it wouldn't constitute a convincing argument against the idea of literal meaning that she wants to reject without a response to the various ways that advocates of literal meaning have proposed to explain intuitive variations in the truth value of what is said by an utterance in different contexts.

Suppose that the truth value of utterances of ‘This vase is red’ changes in the scenarios as Crary describes. How might a defender of literal sentence meaning accommodate that fact within her position? There are several live options for explaining the contextual variation displayed by color adjectives: One might argue that the intuitive difference in truth-value is best explained in terms of our intuitions tracking not a change in a semantic feature of the uttered sentence, but a pragmatically conveyed proposition, generated by an expectation that the vase is red ‘in the normal way’ (ibid.). Another approach Crary focuses on that might be invoked to accommodate changes in intuitive truth value introduces the idea of parameters that receive different values depending on changes in the context of utterance. Zoltán Gendler Szabó has proposed such an account, arguing that there is a hidden ‘part’ variable at the level of logical form in color adjectives (Szabó 2000, 2001).

Szabó discusses a similar example (from Travis 1997) involving leaves that are painted green (but are red underneath), but I have modified his discussion so it explicitly applies to Crary's glazed vase example:

An object is [red] if some contextually specifiable (and presumably sufficiently large) part of it is [red]. The logical form of [‘red'] is [‘(red(C, P))(x)’], where ‘C’ is a class standing for a comparison class and ‘P’ is a variable standing for a certain part of the object. It seems to me that the case of the [glazed vase] fits this pattern. If one is sorting [vases] for [use in a still life], what matters is the color of the outside, if one is trying to identify the [type of clay used], what matters is what we find under the [glaze]. This suggests that the context-dependency that appears in [Crary's] example is a relatively easily characterizable kind: it is a matter of different contextually specified values for the variable ‘P’. (Szabó 2001: 137–8) 15

On Szabó's analysis, in the context where the conversation concerns what kind of clay the vase is made out of, an utterance of ‘This vase is red’ has the content:

(red(C, under the glaze))(This vase)

whereas in the context where the conversation concerns whether the vase is suitable to use in a still life, an utterance of the same sentence has a different content:

(red(C, on the surface))(This vase)

The difference in the truth value of the two utterances is then explained in terms of the different values assigned to the part variable in the two contexts.16

Crary says of this kind of ‘parameter'-based approach to explaining variation in intuitive truth-value in different contexts that it seems like a promising way to reconcile context dependence with the idea of literal meaning that she wants to reject. But she argues that the parameter-based approach to reconciling context dependence and literal meaning cannot succeed.

Her argument begins with a description of an empirical investigation of how linguistic expressions interact with their contexts of utterance:

We might survey a large and and variegated set of cases in which ‘This vase is red’ is used and formulate a set of rules for the semantic contribution of ‘red’ by appeal to these cases; we might then hypothesize that our rules capture the semantic contribution that ‘red’ makes to the sentence ‘This vase is red’ on every occasion of its use; and, finally, we might repeat the procedure for sentences in which ‘red’ is predicated, not of vases, but of different kinds of objects. (Crary 2007: 79–80)

That is, we observe how a sentence is actually used and try to determine whether there are any regular ways in which the truth conditions of utterances containing the word ‘red’ are affected by context. If there are, we formulate a set of rules that capture those regularities and propose that those rules capture the semantic contribution made by ‘red’ to the truth conditions of utterances in which the word figures. That seems like a reasonable description of how to conduct an empirical investigation into the meaning of ‘red’.

But Crary is skeptical about the possibility of successfully conducting such an investigation. Her skepticism is motivated by an argument that ‘exploits and expands on the main line of reasoning in Austin's How to Do Things with Words and is appropriately Austinian in spirit’, but which also ‘receives reinforcement’ from passages in Wittgenstein's Philosophical Investigations that ‘target the assumption that applications of words are, in Wittgenstein's words, “everywhere bounded by rules” ’ (ibid.: 80–1).17

Crary's ‘Austinian-Wittgensteinian’ argument against literal sentence meaning goes as follows. Suppose that, after examining ‘a large and variegated set of cases’ in which ‘red’ is predicated of different objects in different contexts, we have formulated a set of rules that, as far as we can tell, capture the semantic contribution that ‘red’ makes to the truth conditions of utterances in which it occurs. Crary raises a question about what we're entitled to if we are in such a position:

Would we at this point be justified in crediting ourselves with a complete set of rules for the semantic contribution of ‘red’ to speech acts in which it figures? The answer is clearly no. For, however complete our set of rules strikes us as being, we cannot foreclose the possibility that on some further employment ‘red’ will play a role that our rules do not cover. (ibid.: 80)

I agree with Crary that we are not justified in crediting ourselves with a ‘complete set of rules’ governing the semantic contribution that ‘red’ makes to the truth conditions of utterances in different contexts, and for precisely the reason that she describes: It can always turn out that there is some ‘further employment’ of ‘red’ that our rules do not cover. But defeasibility of that kind is a completely ordinary feature of any empirical investigation—it's always possible that our attempts to explain our observations about the expansion of gases or the movement of planets or the heritability of traits might turn out to be mistaken.

Crary does not view this aspect of the investigation into the meaning of ‘red’ as a harmless feature of all forms of empirical inquiry, however. Instead, she sees it as a fatal flaw of investigations of meaning:

Moreover, should this happen [that we discover that there is some employment of ‘red’ that our rules do not cover], it would be illegitimate to protest that it simply shows that we have not yet generated rules constitutive of all of the parameters that we operate with in using ‘red’. Given that we cannot at any point say with authority that we have generated all these rules, we seem obliged to conclude that, in our efforts to generate them, we are relying on a form of appreciation of what is said by uses of sentences involving ‘red’ that is not even implicitly informed by such rules. Further, given that we are thus forced to recognize that, in our efforts to grasp what this and other sentences say on different occasions, we invariably rely on forms of linguistic appreciation that resist formulation in the kinds of rules in question, we seem obliged to conclude that our talk of literal sentence-meanings is no better than empty and that we should therefore reject as bankrupt the thought of grasping the meaning of a sentence in a manner independent of sensitivities characteristic of use as language-users. (ibid.)

Because Crary's argument is so compressed, I will attempt a more explicit reconstruction of it, sticking as closely to her language as possible:

  1. We cannot say with authority that we have generated all the semantic rules for an expression e.
  2. If we cannot say with authority that we have generated all the semantic rules for e, then in our efforts to formulate the semantic rules for e, we are relying on a form of appreciation of the meaning of e that is not even implicitly informed by such rules.
  3. And if, in our efforts to formulate the semantic rules for e, we are relying on a form of appreciation of the meaning of e that is not even implicitly informed by such rules, then our talk of literal sentence-meanings is no better than empty.
  4. So (from 1, 2 and 3) our talk of literal sentence-meanings is no better than empty.

What should we make of Crary's ‘Austinian-Wittgensteinian’ argument against literal sentence meaning? Premise 1 in Crary's argument is compelling because, as discussed above, it is merely a statement of the defeasibility of semantic theory, which is a feature of all forms of empirical inquiry. The rest of the argument, however, is more problematic.

As stated, premise 2 is hard to understand because both the antecedent and the consequent involve negations. The contrapositive is easier to comprehend:

2′. If, in our efforts to formulate the semantic rules for e, we are relying on a form of appreciation of the meaning of e that is implicitly informed by such rules, then we can say with authority that we have generated all the semantic rules for e.

So formulated, premise 2′ is easier to understand, but it is also easier see what would falsify it: To do so, we need to find a situation in which we are relying on an appreciation of the meaning of an expression that is implicitly informed by semantic rules, but in which we cannot say with authority that we have generated (i.e., represented) all the semantic rules for that expression. Here is a situation that seems to meet those two requirements: Suppose I have learned the rules of chess, including the meaning of technical chess expressions like ‘castling’, by playing the game, asking questions about the rules, and so on. My appreciation of the meaning of ‘castling’ is thus implicitly informed by the semantic rules for ‘castling’. But I'm aware that the semantic rules for ‘castling’ are complex, and if I try to explicitly formulate the rules I suspect that I will always leave something out (even if I in fact do not). At any given point in such a situation, I cannot say with authority that I have generated all the semantic rules informing my grasp of the meaning of ‘castling’, but my appreciation of the meaning of ‘castling’ is still implicitly informed by those rules.

The upshot of the example is that we may not have authoritative knowledge that we have represented the complete set of semantic rules governing our grasp of the meaning of an expression, even when our grasp of the meaning of that expression is clearly implicitly informed by those semantic rules. That shows that premise 2 in the ‘Austinian-Wittgensteinian’ argument against literal sentence meaning is false (the consequent of the conditional can be false when the antecedent is true). Crary's second, a priori, argument targeting literal sentence meaning should therefore be resisted.

Concluding Remarks

  1. Top of page
  2. Abstract
  3. Introduction
  4. Literal Meaning and Objectivity
  5. Crary's Arguments Against Literal Meaning
  6. The Illocutionary Force Argument
  7. Open-Ended Variation in Truth Conditions
  8. Concluding Remarks
  9. References

Austin's targets, including the ‘descriptive fallacy’ and the ‘true-false’ fetish, arise from failures to appreciate the variety of what language can be used to do. Accepting the idea that declarative sentences of a language have literal meanings in Crary's sense—that is, that they have truth conditional contents that are algorithmically derivable from the content of their parts—need not obscure that variety. It might be the case that literal meaning is rule-governed and insensitive to contextual variation in a way that illocutionary acts are not. Determining whether or not that's the case is, to borrow a phrase from Austin, ‘a matter of prolonged fieldwork’ (Austin 1975: 149).18

Crary's reading of Austin is salutary because it encourages close attention to Austin's project and treats his work as making a contribution to our understanding of language that is of continuing significance. But literal meaning is not one of Austin's targets, and the arguments Crary develops, whether or not they are Austin's, do not succeed in threatening it.19

Notes
  1. 1

    See, for example, Recanati 2004: 146, where he says that the most radical version of contextualism (‘Meaning Eliminativism’) is ‘very much in the spirit of Austin and Wittgenstein’, and see Travis 2008: 2–7 for an account of the influence of Austin on Travis's views about ‘occasion-sensitivity’. For the claim that Austin's work is an inspiration for epistemological contextualism, see Brady and Pritchard 2005: 162 and Ludlow 2005: 12.

  2. 2

    See Travis 2011 for a recent exception.

  3. 3

    ‘Literal meaning’ is not a phrase that Austin frequently uses. There is only one use of the phrase in Austin's published work, in the essay ‘Pretending’, when he uses it to describe the meaning of ‘Praetendere’ (Austin 1979a: 260). (My attention was drawn to this passage in Austin by Bauer 2011). The only other place where Austin uses the word ‘literal’ is in How to Do Things With Words, when he says that ‘insinuating’ is a ‘non-literal [use] of language’ (Austin 1975: 122).

  4. 4

    Wright 2001: 3 calls this an ‘ultra-objective’ requirement.

  5. 5

    See, for example, Fodor and Lepore 2007: 678.

  6. 6

    As pointed out by an anonymous reviewer, the truth value of complex sentences conjoined with truth-functional connectives can be generated algorithmically out of the truth-values of simple sentences, once those are known.

  7. 7

    There are a variety of competing views on how to understand the notion of literal meaning. See Borg 2004, Recanati 2004, and Stojanovic 2006 for discussion.

  8. 8

    An alternate approach would be to understand literal meaning as a function from contexts to contents, along the lines of Kaplan 1989.

  9. 9

    Dummett articulates this conception of the relation between the ‘core’ of a sentence's meaning (reference and sense) and a theory of force (with the caveat that Dummett is talking about verification conditions rather than truth conditions): ‘[A systematic theory of meaning involves] a central part giving the theory of sense and reference (here conceived of as an inductive specification, for each sentence, of the method of its verification), and a supplementary part giving a uniform means of deriving, from that feature of any sentence determined by the central part, every aspect of its use’ (Dummett 1993: 40–1).

  10. 10

    Crary says that Austin is ‘drawn towards’ his view that ‘to perform a locutionary act is in general … also and eo ipso to perform an illocutionary act’ by his ‘thought that there is no such thing as identifying the meaning of a combination of words (or no such thing as identifying the “locutionary act” performed when a combination of words is uttered) independently of an appreciation of how those words are being used to say something to someone on a particular occasion (or: independently of an appreciation of their “illocutionary force”)’ (Crary 2007: 63).

  11. 11

    For a similar conclusion see Garner 1972: 207.

  12. 12

    There is also tension between Crary's statement that Austin is ‘suspicious of the idea that sentences have “literal” or “conventional” meanings that they carry with them into every context of their use’ (2002: 68) and a central element of Austin's theory of truth, namely the idea that there are ‘descriptive conventions’ that correlate sentences with types of situations in the world (Austin 1979b: 121–2). Thanks to an anonymous reviewer for pointing out this additional tension.

  13. 13

    For examples of the method, see, for example, Travis 1996; 1997.

  14. 14

    An anonymous reviewer points out that Austin differs from contemporary contextualists in that he does not embrace the notion of the thought or proposition expressed by an utterance. That should make one suspicious of Crary's attribution to Austin of the contextualist claim that a given sentence, even with its indexical elements held fixed, can express different thoughts in different contexts.

  15. 15

    See also Szabó 2000: 112.

  16. 16

    Szabó's analysis is presented here because it is a clear example of a ‘parameter’ based account of color adjectives. But there are convincing arguments that Szabó's analysis of color adjectives is inadequate: See Hansen 2011, Kennedy and McNally 2010, and Rothschild and Segal 2009 for discussion.

  17. 17

    The Wittgenstein remark is in Wittgenstein 1958: §84.

  18. 18

    James Higginbotham, 1988: 29, puts the idea as follows: ‘The rules of semantics, since they relate forms to meanings, are independent of context; but they come alive only in use. Disentanglement is therefore necessary: we can arrive at what is context-independent about meaning only by judicious comparison of context-bound cases’.

  19. 19

    I would like to thank Zed Adams, Ted Cohen, Alice Crary, Michael Kremer, Chauncey Maher and an anonymous reviewer for the European Journal of Philosophy for their very helpful comments on earlier versions of this paper. The research leading to these results has received funding from the European Research Council under the European Community's Seventh Framework Programme (FP7/2007-2013)/ERC grant agreement no 229 441—CCC.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Literal Meaning and Objectivity
  5. Crary's Arguments Against Literal Meaning
  6. The Illocutionary Force Argument
  7. Open-Ended Variation in Truth Conditions
  8. Concluding Remarks
  9. References
  • Austin, J. L. (1962), Sense and Sensibilia. Oxford: Oxford University Press.
  • Austin, J. L. (1975), How to Do Things with Words. Oxford: Oxford University Press.
  • Austin, J. L. (1979a), ‘Pretending’, in Philosophical Papers. Oxford: Oxford University Press.
  • Austin, J. L. (1979b), ‘Truth’, in Philosophical Papers. Oxford: Oxford University Press.
  • Bauer, N. (2011), ‘What Is to Be Done with Austin?’ in M. Gustafsson and R. Sørli (eds), New Essays on the Philosophy of J. L. Austin. Oxford: Oxford University Press.
  • Borg, E. (2004), Minimal Semantics. Oxford: Oxford University Press.
  • Brady, M. and Pritchard, D. (2005), ‘Epistemological Contextualism: Problems and Prospects’, The Philosophical Quarterly, 55: 161171.
  • Crary, A. (2002), ‘The Happy Truth: J. L. Austin's How to Do Things with Words ’, Inquiry, 45: 5980.
  • Crary, A. (2007), Beyond Moral Judgment. Cambridge, MA: Harvard University Press.
  • Derrida, J. (1988), Limited Inc. Evanston, IL: Northwestern University Press.
  • Dummett, M. (1993), The Seas of Language. Oxford: Oxford University Press.
  • Fodor, J. and Lepore, E. (2007), ‘Brandom Beleaguered’, Philosophy and Phenomenological Research, 74: 677691.
  • Garner, R. T. (1972), ‘On Saying What Is True’, Noûs, 6: 201224.
  • Hansen, N. (2011), ‘Color Adjectives and Radical Contextualism’, Linguistics and Philosophy , 34: 201221.
  • Higginbotham, J. (1988), ‘Contexts, Models, and Meanings: A Note on the Data of Semantics’, in R. M. Kempson (ed.), Mental Representations: The Interface Between Language and Reality. Cambridge: Cambridge University Press.
  • Kaplan, D. (1989), ‘Demonstratives: An Essay on the Semantics, Logic, Metaphysics and Epistemology of Demonstratives and Other Indexicals’, in J. Perry, J. Almog and H. Wetstein (eds), Themes from Kaplan. Oxford: Oxford University Press.
  • Kennedy, C. and McNally, L. (2010), ‘Color, Context, and Compositionality’, Synthese, 174: 7998.
  • Ludlow, P. (2005), ‘Contextualism and the New Linguistic Turn in Epistemology’, in G. Preyer and G. Peter (eds), Contextualism in Philosophy: Knowledge, Meaning and Truth. Oxford: Oxford University Press.
  • Recanati, F. (2004), Literal Meaning. Cambridge: Cambridge University Press.
  • Rothschild, D. and Segal, G. (2009), ‘Indexical Predicates’, Mind & Language, 24: 467493.
  • Sainsbury, R. M. (2001), ‘Two Ways to Smoke a Cigarette’, Ratio, 14: 386406.
  • Searle, J. (1968), ‘Austin on Locutionary and Illocutionary Acts’, The Philosophical Review, 77: 405424.
  • Stojanovic, I. (2006), ‘What Is Said, Linguistic Meaning, and Referential Expressions’, Philosophy Compass, 1: 373397.
  • Szabó, Z. G. (2000), Problems of Compositionality. New York: Garland.
  • Szabó, Z. G. (2001), ‘Adjectives in Context’, in I. Kenesei and R. M. Harnish (eds), Perspectives on Semantics, Pragmatics, and Discourse: A Festschrift for Ferenc Kiefer. Amsterdam: John Benjamins.
  • Travis, C. (1996), ‘Meaning's Role in Truth’, Mind, 105: 451466.
  • Travis, C. (1997), ‘Pragmatics’, in B. Hale and C. Wright (eds), A Companion to the Philosophy of Language. Oxford: Blackwell.
  • Travis, C. (2008), ‘On Constraints of Generality’, in Occasion Sensitivity. Oxford: Oxford University Press.
  • Travis, C. (2011), ‘Truth and Merit’, in Objectivity and the Parochial. Oxford: Oxford University Press.
  • Wittgenstein, L. (1958), Philosophical Investigations, 2nd ed. Oxford: Blackwell.
  • Wright, C. (2001), Rails to Infinity: Essays on Themes from Wittgenstein's Philosophical Investigations. Cambridge, MA: Harvard University Press.