Philosophy of Language

Curtis Brown

Review for the Final Exam

Here is a summary of some of the material we have covered since the midterm exam. You should be familiar with, and able to write about, the issues, terminology, and arguments mentioned or briefly summarized below:

I. Pragmatics

1. Semantic Pragmatics (Lycan, chapter 11)

Semantic pragmatics is concerned with features of linguistic meaning that are dependent on the context of utterance. It is thus intermediate, in a sense, between semantics and "pragmatic pragmatics."

semantics deals with linguistic expressions rather than with particular utterances of those expressions. So, in a sense, linguistic meaning is independent of particular contexts of utterance, since it is a feature of linguistic expressions that they carry with them, so to speak, in any context in which they are uttered.

However, there are some linguistic expressions which have a kind of built-in context dependence. The most obvious examples are "deictic" or "indexical" expressions such as "I," "here," and "now." A singular term that is or contains an indexical expression requires a context in order to determine what its reference is, and a sentence that contains an indexical expression requires a context in order to determine what proposition it expresses.

"Semantic Pragmatics" is concerned with the way in which the content of indexical expressions depends on particular contexts. As we discussed in class, a common way to understand indexicals like "I" and "here" is as follows. The extension or referent of "I" (on a particular occasion of use) is an individual. The intension is a function from possible worlds to extensions. In the case of indexicals, this is a constant function: "I" refers, with respect to every possible world, to the same individual. But, unlike non-indexical expressions, the linguistic meaning of "I" is not to be identified with its intension. Rather, its linguistic meaning may be thought of as a function from contexts of utterance to intensions. In any context of utterance, the intension of "I" is a constant function from possible worlds to the agent of the context. (Usually the "agent" of the context is the person who, in that context, is thinking or saying the word "I.")

2. Pragmatic Pragmatics (Lycan, chapter 12; Austin, How to Do Things with Words)

A general overview of the book follows. I also have somewhat more detailed summaries of sections 1-3 and 4-7.

A. Relation of Speech Act Theory to Verificationism

Like the verificationists, Austin identifies utterances that look like statements but aren't. The verificationists focused on two such kinds of utterances: (a) "metaphysics," i.e. things that look like statements but are not analytic and can't be verified or disconfirmed; (b) utterances whose purpose is something other than stating facts about the world, such as ethical language, which according to Ayer was used not to state facts but to express emotions.

Austin focuses attention on a third category: (c) performative utterances, such as "I christen this ship the Santa Maria," or "I promise I will pay you back." These don't state a fact, but rather accomplish something -- the first does not state that I am christening a ship, but rather actually christens it; the second does not state that I am promising something, but rather actually promises it.

B. Performatives vs. Constatives

You should be able to describe Austin's distinction between performatives and constatives. Such a description should include the idea that a constative utterance is a "saying" whereas a performative utterance is a "doing"; the idea that the characteristic flaw of constatives is being false, whereas the characteristic flaw of performatives is being infelicitous. (You should be able to give examples of ways of being infelicitous.)

A number of things must be true in order for a performative utterance to fully succeed. You should be able to illustrate the main kinds of infelicities:

A. there must be a conventional procedure with a conventional effect, where the procedure includes uttering certain words, and in the particular case the people and circumstances must be appropriate for invoking the procedure. ("I hereby divorce you" doesn't work because there is no such conventional procedure for divorcing someone. A christening doesn't work if the wrong person does it or the wrong boat is present, even though there is a conventional procedure for christening. --an aside: often the conventions in question may not be explicitly written down anywhere, and there may be disagreement over exactly what they require: compare the current controversies over issuing marriage licenses to same-sex couples.)

B. The procedure must be executed by all participants correctly and completely. (E.g. if the marriage license isn't signed, the participants are not legally married. I think.)

Γ. Where the procedure is designed for use by people with certain thoughts and feelings, who will later act in certain ways, they must have the appropriate thoughts and feelings, intend to act in the appropriate ways, and actually follow up by acting in those ways. (E.g. the practice of promising is designed to be used by people who intend to keep their promises, and who actually do go on to act in accordance with the promise. The practice of apologizing is designed to be used by people who feel sorry for what they have done. And so on.)

Whereas failing to meet conditions A or B results in a failure of the performative act to succeed at all (a "Misfire"), failing to meet Γ results instead in the act being performed, but the procedure has been abused. (If I promise to pay you back but don't intend to, then I really have promised, and have the moral obligations associated with promising, but the promise was not sincere.)

C. Problems with the Performative/Constative Distinction. 

You should also be able to describe Austin's growing doubts about the performative-constative distinction (notably the points that (a) constatives are subject to something very like infelicities, and (b) many performatives, especially things like judging, advising, etc., are intimately linked to propositions that can be true or false.

D. A More Comprehensive Speech-Act Theory

And you should be able to describe the more comprehensive view that Austin eventually replaces the performative-constative distinction with, including the following features:

E. Kinds of Illocutionary Act Verbs

(A slightly more detailed diagram of this material is here.)

3. Implicative Relations (Lycan, chapter 13; Grice, "Logic and Conversation")

Up to this point, our discussion of pragmatics had to do primarily with individual speech acts. Grice takes up the issue of how speech acts fit together into longer stretches of communication, in particular, into conversations.

Maxims of conversation:

Conversational Implicature: what happens when a maxim is violated? In some cases this may be simply because the conversationalist is incompetent. More often, however, it's a way of communicating something other than what is literally said. The maxims may be deliberately violated ("flouted") as an indirect means of communication. 

One problem, noted by Lycan: the reasoning involved in conversational implicatures involves a negative phase (the utterance violates a maxim, so the speaker must be trying to communicate something other than the literal meaning of the utterance) and a positive phase, in which one attempts to determine what the speaker is trying to communicate. It's a lot clearer how the reasoning in the negative phase works than how the reasoning in the positive phase works.

II. Metaphor 

(Lycan, chapter 14) (I have a somewhat more detailed summary here.)

Davidson: no such thing as "metaphorical meaning," only literal meaning. Metaphors cause people to have interesting thoughts, but the metaphor in no sense means the thoughts it provokes; they are just causal consequences of someone hearing the metaphor. A consequence would appear to be that we can't speak of metaphors being true or false, or of someone misinterpreting a metaphor.

Naive Simile view: a metaphor is just a shorthand way to express a simile. "Juliet is the sun" as short for "Juliet is like the sun." Objections: (1) this seems to understate the "tension" involved in hearing a metaphor. Similes aren't "anomalous or puzzling," so if metaphors are just short for similes, they shouldn't seem anomalous or puzzling either. But they do. (2) Doesn't explain much: just shifts the burden of explanation to the simile. Everything is like everything else in some respect or other, so similes seem just as vacuously true as metaphors often are obviously false. Typically a use of a simile seems to be an attempt to say something more specific than merely that A is like B in some respect or other. So now we need an account of the meaning of a simile! (3) The similarity between A and B may itself not be literal. "Susan is a block of ice." In what respect is Susan like a block of ice? Well, they are both cold. But wait . . . Susan is not literally cold!

Figurative Simile Theory. Metaphors abbreviate similes in which the resemblances are figurative rather than literal. Objections: (1) but then we still need an account of figurative meaning! Otherwise the account seems just circular. (metaphor  = simile in which the resemblances are metaphorical . . . this doesn't seem to get us very far!) (2) seems to have the result that virtually any sentence is ambiguous between a literal and a metaphorical reading. (3) some metaphors are too complex to be understood as similes. [My example: Nabokov: "Either the drizzle had stopped or Fialta had got so used to it that she herself did not know whether she was breathing moist air or warm rain." what exactly is the simile? Fialta is like a woman, and the drizzle is like her breath, and . . . somehow this seems way off. Nabokov again, from the same story ("Spring in Fialta"): "Occasionally, in the middle of a conversation her name would be mentioned, and she would run down the steps of a chance sentence, without turning her head." Again, it's hard to know even where to start in converting this into a simile.]

Pragmatic Theory: metaphorical meaning is not a kind of semantic meaning but rather a kind of pragmatic meaning. That is, the meaning is not (entirely) determined by the literal meanings of the words, but rather is a matter of what a speaker is using those words to communicate on a particular occasion of use. Metaphor on this account is a matter of speaker meaning, not linguistic meaning. And it's pragmatic because it attaches not to a sentence per se, but to the use of a sentence in a particular context of utterance.

III. Syntax

We didn't read anything on this topic, but I raised a few issues in class about Chomsky's views on syntax.

1. Rationalism/Empiricism. One reason Chomsky's work is interesting for philosophy is its relation to the debate between rationalists and empiricists over innate ideas. Rationalists like Descartes, Spinoza, and Leibniz thought that there were innate ideas, ideas which we have and use but which we did not and could not learn from experience. (For Descartes, the related ideas of God and infinity are like this.) For empiricists like Locke, on the other hand, all our ideas derive from experience. For Locke, the infant mind was like a blank slate, or a dark room, or an unmarked piece of wax. Experience wrote ideas on the slate, or shone them into the dark room, or imprinted them on the wax. Although Chomsky doesn't address ideas in the sense of the early modern philosophers, his view about grammar has a clear affinity with rationalism. Chomsky holds that much of our knowledge of grammar is not acquired from experience, but rather is innate, present from birth (although like other innate abilities, possibly not activated or triggered until sometime later; similarly, in some sense the newborn "knows how to walk," but the knowledge isn't triggered until an appropriate stage of physical development is reached).

2. Nature of grammar. Grammar or syntax of course is the main focus of Chomsky's scientific work; we barely scratched the surface of this. But a couple of important points stand out.

A. The recursive nature of language. Chomsky argues that the only way finite capacities can allow us to determine the grammaticality of an infinite number of sentences is if our knowledge of grammar consists of recursive rules. (Roughly, rules are recursive if they can be reapplied the results of an earlier application. For instance, a sentence can consist of a noun phrase followed by a verb phrase, but a sentence may also consist of a sentence followed by a conjunction followed by a sentence. So once we get a sentence or two, perhaps by the NP + VP rule, we can combine them by putting a conjunction between them. But the result is also a sentence, so we can use the same rule over again to produce a longer sentence. And there's no limit to the number of times we can do this.)

B. Universal Grammar. This is said to characterize the initial state one is in prior to learning a particular language. Then language acquisition is the transition from this initial state to a final state which is the knowledge of a particular natural language. For the empiricists, the initial state has to be empty, and all of grammar has to be acquired in learning a language. But according to Chomsky, the initial state is already very rich and complex.

3. Arguments for the innateness of syntax (more precisely, of Universal Grammar).

(1) the "poverty of the stimulus" argument: the child is not exposed to enough linguistic data to learn a specific language from scratch as rapidly as they actually do

(2) empirical evidence 1: finding a shared structure in all natural languages (according to the Principles and Parameters view, UG consists of universal principles that all languages share, with a few parameters that experience shows us how to set; the differences between natural languages are just different parameter-settings)

(3) empirical evidence 2: one generation is all it takes to transform a pidgin language into a creole (and, similarly, to enable a child to transform an awkward made-up or ill-understood sign language into a full-blown language).

(4) empirical evidence 3: some people, as a result either or brain damage or of congenital problems, lack grammatical capabilities while still having normal intelligence and learning ability in other respects. Other people have limited or damaged intelligence, but flawless grammatical capabilities (and will chatter on grammatically about nonsense). This suggests that learning and understanding a language rests on a specialized "modular" ability, not just on general intelligence.


Last update: April 25, 2007. 
Curtis Brown  |  Philosophy of Language   |  Philosophy Department  |   Trinity University