Semantics & pragmatics: Kripke’s reply to Donnellan

Jeff Speaks

March 29, 2005

1 Kripke’s approach to Donnellan’s distinction
 1.1 Kripke’s aims
 1.2 The question of whether Donnellan disagrees with Russell
 1.3 Kripke on Grice’s distinction between speaker’s meaning and expression meaning
2 Kripke’s defense of Russell
 2.1 Against the strong interpretation: ‘the’ is not ambiguous
 2.2 Against the weak interpretation: ‘Russell languages’ + Gricean pragmatics
3 Some further questions
 3.1 ‘The Messiah’
 3.2 A further problem for Russell’s theory: abnormal uses

1 Kripke’s approach to Donnellan’s distinction

1.1 Kripke’s aims

Kripke is skeptical of Donnellan’s claim to have refuted Russell’s theory of definite descriptions. His aim, however, is not to defend Russell’s theory of descriptions; rather, he takes on the more limited aim of arguing the negative claim that Donnellan’s examples and arguments do not, by themselves, refute Russell’s theory of descriptions.

The interest of his defense of this claim goes beyond the particular issue of Russell’s theory of descriptions. Kripke is interested in certain methodological points that arises in arguments in the philosophy of language; and, in getting clear about these methodological points, he presents, using the work of Grice, a compelling framework for thinking about questions of meaning and reference.

1.2 The question of whether Donnellan disagrees with Russell

Kripke begins part 3 of his paper by asking whether Donnellan’s examples, or even Donnellan’s views, really contradict those of Russell. As Kripke says, this may seem like a bit of a stupid question. After all, Donnellan spends most of his paper criticizing Russell, and saying why he thinks Russell is wrong. But the question is not whether Donnellan thinks that Russell’s view is false; that’s obvious enough. The question is whether there is anything in Donnellan’s discussion of definite descriptions which really conflicts with Russell’s views about definite descriptions.

This question turns out to be a very tricky one, because, on inspection, it is less clear than it might have seemed what Donnellan’s view about definite descriptions really amounts to. Here, for example, is one view about definite descriptions which really would conflict with Russell’s view of definite descriptions:

All definite descriptions of the form the F are ambiguous. Sometimes (when used attributively) they have the meaning and reference that Russell’s theory of descriptions would ascribe to them, while at other times (when used referentially) they are more like Russell’s logically proper names, and refer to whatever object the speaker intends to pick out by a use of the name. Which meaning definite descriptions have on an occasion is determined by the intentions of the speaker.

Such a view would not be crazy; it would just be claiming that definite descriptions are a bit like ‘bank’, which really is such that it has two different meanings and is such that the intentions of speakers determine which of the two meanings a certain use of it has. And this view, that definite descriptions are ambiguous, would conflict with Russell’s view, since Russell did not think that ‘the’ was ambiguous.

But is this Donnellan’s view? It seems not, for he writes

“It does not appear plausible to account for this [the two different uses of definite descriptions] ...as an ambiguity in the sentence. The grammatical structure of the sentence seems to me to be the same whether the description is used referentially or attributively: that is, it is not sytactically ambiguous. Nor does it seem at all attractive to suppose an ambiguity in the meaning of the words; it does not appear to be semantically ambiguous. (Perhaps we could say that the sentence is pragmatically ambiguous ...)” (58-59)

So it seems that, although the ambiguity theory would conflict with Russell’s view, we cannot ascribe it to Donnellan. So how should we understand Donnellan’s view?

Kripke is on to a genuine unclarity in the view that Donnellan puts forward. On the one hand, Donnellan claims that there are two different analyses of definite descriptions. This suggests a strong interpretation of Donnellan’s view, which is basically the ambiguity view sketched above. The advantage of this interpretation of Donnellan is that it clearly does lead to a conflict with Russell. The disadvantage of this interpretation is, of course, that he explicitly disclaims it. Nonetheless, it does seem true that, as Kripke says, Donnellan’s disclaiming the ambiguity claim runs against the grain of most of Donnellan’s paper; so perhaps the disclaimer was just a mistake on Donnellan’s part. The main question facing the strong interpretation is: do Donnellan’s arguments support the idea that definite descriptions are ambiguous between attributive and referential interpretations?

Suppose, though, that we take seriously Donnellan’s claim that he does not mean to posit an ambiguity in definite descriptions, and thus consider a weaker interpretation. On this interpretation, Donnellan is not claiming anything about the meaning of definite descriptions, and is not claiming anything about what definite descriptions refer to on particular occasions. Rather, he is claiming something about what speakers mean and refer to when they are using definite descriptions. The key question for the weaker interpretation, then, is: does it really pose a challenge to Russell? Is there any conflict between the weak interpretation of Donnellan’s position, and Russell’s view of definite descriptions?

To sum up, then, we have two basic questions we should be asking here about Donnellan’s arguments: Do they establish that definite descriptions are ambiguous? And, if they do not, do they establish anything which is in conflict with Russell’s theory of the meaning of definite descriptions?

1.3 Kripke on Grice’s distinction between speaker’s meaning and expression meaning

We need some framework for discussing these questions, and in particular for understanding this distinction Kripke invokes between the strong interpretation, which claims something about the meanings and references of definite descriptions, and the weak interpretation, which only makes claims about what speakers mean and refer to when they use definite descriptions.

A first step, as Kripke notes, is to distinguish, following Grice, between what a speaker’s words mean, on a given occasion, and what the speaker meant, by using those words, on a given occasion. Kripke’s example (3b) of how these two notions can come apart: one burglar says to another: “The cops are around the corner.” What do the words uttered by the burglar mean? Pretty obvious; they mean that the police are around the corner. But the speaker may well have meant: “We can’t wait around here any more; let’s split!”

Another important example is that of the magician, which he borrows from Grice: ‘It looked red.’ Also: ‘He has excellent penmanship, and always arrives at class on time.’

Kripke, again following Grice, thinks of this as being different than simple cases of ambiguity. He considers the example: “The cops are inside the bank.” ‘Bank’ may mean different things on different occasions of use of this sentence; and this is a matter of a difference in word-meaning, not just in speaker’s meaning.

From these notions of speaker’s meaning and ordinary, expression meaning we can extrapolate to notions of speaker’s reference and semantic reference, which will be the reference of expressions used by speakers. We already have a handle on what ordinary reference -- the reference of expressions -- is. Kripke defines speaker’s reference thus:

“...we may tentatively define the speaker’s referent of a designator to be that object which the speaker wishes to talk about, on a given occasion, and believes fulfills the conditions for being the semantic referent of the designator. ...the speaker’s referent [is] the correct answer to the question, ‘To whom were you referring?”’ (238)

The simple case and the complex case defined in terms of speaker’s reference and semantic reference. The question is then: are referential uses simple cases, or complex cases?

Given this framework, we can recast our questions about the two interpretations of Donnellan as follows:

Strong interpretation. Does the fact that speakers can use a description ‘the F’ to refer to something which is not F show that definite descriptions are ambiguous?

Weak interpretation. Does the fact that speakers can use a description ‘the F’ to refer to something which is not F show that Russell’s theory of descriptions is false?

2 Kripke’s defense of Russell

2.1 Against the strong interpretation: ‘the’ is not ambiguous

We can explain the Donnellan phenomena either by positing an ambiguity, or by appealing to the distinction between speaker’s reference and semantic reference. Kripke thinks that we should prefer the latter on three grounds:

(i) It is always better to account for linguistic phenomena without positing an ambiguity (especially, he might add, in cases in which ordinary speakers cannot easily be made to notice the ambiguity). ‘It is very much the lazy man’s approach to philosophy to posit ambiguities when in trouble.’ (243)

(ii) The pragmatic account is needed anyway to explain a wide variety of facts, as Gricean examples show.

(iii) In order to account for all referential/attributive phenomena, Donnellan would have to posit ambiguities in cases where it is not at all plausible, and where ‘the’ is not a part of the sentence. (E.g. ‘Smith, get out of the way of that car!’ when Jones is the one in the way of the car. See also bottom p. 237.)

Kripke also offers an independent argument against the thesis that definite descriptions are ambiguous. He claims that the ambiguity theory cannot make sense of the use of ‘he’ in dialogue ‘Her husband is kind to her’/‘No, he isn’t; the man you’re referring to isn’t her husband.’ (p. 247)

2.2 Against the weak interpretation: ‘Russell languages’ + Gricean pragmatics

Suppose then that we reject the idea that the phenomena to which Donnellan drew our attention show that definite descriptions are ambiguous. We can still ask: does the fact that definite descriptions can be used in these two ways count against Russell’s theory?

Kripke thinks not. He argues the point by suggesting that, if Russell’s theory of definite descriptions were correct for some language L, we would still expect observation of speakers of L to indicate that some of their uses are referential (in the sense that the speaker’s reference of a description ‘the F’ is sometimes something which is not F).

He describes three different ‘Russell languages’:

The weak Russell language. Sounds just like English, but sentences involving definite descriptions are stipulated to have the truth conditions assigned to them by Russell’s theory.

The intermediate Russell language. Like the weak Russell language, with the additional assumption that sentences containing definite descriptions are stipulated to have the same meanings as their Russellian expansions.

The strong Russell language. The definite article is banned from the language, rather than uttering sentences involving ‘the’, speakers must utter sentences like ‘There is a unique man in the corner drinking a martini, and every man in the corner drinking a martini is interesting.’

Would we get referential uses in these three languages? If so, then the fact that we get referential uses in English does not seem to cut against Russell’s analysis of the meaning of definite descriptions.

The example of the ambiguous D-language, which contains different words ‘the’ and ‘ze’ for the supposed two meanings of ‘the’ in English. A problem for Donnellan: in such a language, it seems that ‘the’ would still admit of referential uses.

Kripke’s conclusion:

“Under these circumstances, surely general methodological principles favor the existing account. The apparatus of speaker’s reference and semantic reference ...is needed anyway to explain the Smith-Jones case; it is applicable to all languages. Why posit a semantic ambiguity when it is both insufficient in general and superfluous for the special case it seeks to explain? ...

It is very much the lazy man’s approach to philosophy to posit ambiguities when in trouble. If we face a putative counterexample to our favorite philosophical thesis, it is always open to us to protest that some key term is being used in a special sense, different from its use in the thesis. We may be right, but the ease of the move should counsel a policy of caution: Do not posit an ambiguity unless you are really forced to, unless there are really compelling theoretical or intuitive grounds to suppose that an ambiguity really is present.” (243)

3 Some further questions

3.1 ‘The Messiah’

Return to an interesting example from the beginning of the paper: ‘the Messiah.’ This seems to be just the kind of case Donnellan would have wanted all along, for here we have intuitions that the word itself really exhibits a referential/attributive distinction (or at least fails to work as Russell’s theory predicts). What should we say about this? Maybe just: it’s operating as a disguised name. But then what happens to the thesis that names are just disguised Russellian descriptions? This is a question to which we will return in our discussion of Naming and Necessity.

3.2 A further problem for Russell’s theory: abnormal uses

The example of an utterance of ‘The table is covered with books’ in a room with many tables, some covered with books and some not, and no contextual indication that one of the tables is more relevant to the conversation than any of the others. The problem of explaining why this use is unintelligible; how this could be explained on a theory in accord with the strong interpretation of Donnellan. (The example is from Ramachandran, “A Strawsonian Objection to Russell’s Theory of Descriptions”, Analysis 53.)