H. P. Grice, "Definite descriptions in Peano and in the vernacular"

theory of descriptions, an analysis, initially developed by Peano and Russell, of sentences containing descriptions. Descriptions include indefinite descriptions such as ‘an elephant’ and definite descriptions such as ‘the positive square root of four’. On Russell’s analysis, descriptions are “incomplete symbols” that are meaningful only in the context of other symbols, i.e., only in the context of the sentences containing them. Although the words ‘the first president of the United States’ appear to constitute a singular term that picks out a particular individual, much as the name ‘George Washington’ does, Russell held that descriptions are not referring expressions, and that they are “analyzed out” in a proper specification of the logical form of the sentences in which they occur. The grammatical form of ‘The first president of the United States is tall’ is simply misleading as to its logical form. According to Russell’s analysis of indefinite descriptions, the sentence ‘I saw a man’ asserts that there is at least one thing that is a man, and I saw that thing – symbolically, (Ex) (Mx & Sx). The role of the apparent singular term ‘a man’ is taken over by the existential quantifier ‘(Ex)’ and the variables it binds, and the apparent singular term disappears on analysis. A sentence containing a definite description, such as ‘The present king of France is bald’, is taken to make three claims: that at least one thing is a present king of France, that at most one thing is a present king of France, and that that thing is bald – symbolically, (Ex) {[Fx & (y) (Fy / y % x)] & Bx}. Again, the apparent referring expression ‘the present king of France’ is analyzed away, with its role carried out by the quantifiers and variables in the symbolic representation of the logical form of the sentence in which it occurs. No element in that representation is a singular referring expression. Russell held that this analysis solves at least three difficult puzzles posed by descriptions. The first is how it could be true that George IV wished to know whether Scott was the author of Waverly, but false that George IV wished to know whether Scott was Scott. Since Scott is the author of Waverly, we should apparently be able to substitute ‘Scott’ for ‘the author of Waverly’ and infer the second sentence from the first, but we cannot. On Russell’s analysis, ‘George IV wished to know whether Scott was the author of Waverly’ does not, when properly understood, contain an expression ‘the author of Waverly’ for which the name ‘Scott’ can be substituted. The second puzzle concerns the law of excluded middle, which rules that either ‘The present king of France is bald’ or ‘The present king of France is not bald’ must be true; the problem is that neither the list of bald men nor that of non-bald men contains an entry for the present king of France. Russell’s solution is that ‘The present king of France is not bald’ is indeed true if it is understood as ‘It is not the case that there is exactly one thing that is now King of France and is bald’, i.e., as -(Ex) {Fx & (y) {[Fy / y % x)] & Bx}. The final puzzle is how ‘There is no present king of France’ or ‘The present king of France does not exist’ can be true – if ‘the present king of France’ is a referring expression that picks out something, how can we truly deny that that thing exists? Since descriptions are not referring expressions on Russell’s theory, it is easy for him to show that the negation of the claim that there is at least and at most (i.e., exactly) one present king of France, -(Ex) [Fx & (y) (Fy / y % x)], is true. Strawson offered the first real challenge to Russell’s theory, arguing that ‘The present king of France is bald’ does not entail but instead presupposes ‘There is a present king of France’, so that the former is not falsified by the falsity of the latter, but is instead deprived of a truth-value. Strawson argued for the natural view that definite descriptions are indeed referring expressions, used to single something out for predication. More recently, Keith Donnellan argued that both Russell and Strawson ignored the fact that definite descriptions have two uses. Used attributively, a definite description is intended to say something about whatever it is true of, and when a sentence is so used it conforms to Russell’s analysis. Used referentially, a definite description is intended to single something out, but may not correctly describe it. For example, seeing an inebriated man in a policeman’s uniform, one might say, “The cop on the corner is drunk!” Donnellan would say that even if the person were a drunken actor dressed as a policeman, the speaker would have referred to him and truly said of him that he was drunk. If it is for some reason crucial that the description be correct, as it might be if one said, “The cop on the corner has the authority to issue speeding tickets,” the use is attributive; and because ‘the cop on the corner’ does not describe anyone correctly, no one has been said to have the authority to issue speeding tickets. Donnellan criticized Russell for overlooking referential uses of theory of descriptions theory of descriptions 914   914 descriptions, and Strawson for both failing to acknowledge attributive uses and maintaining that with referential uses one can refer to something with a definite description only if the description is true of it. Discussion of Strawson’s and Donnellan’s criticisms is ongoing, and has provoked very useful work in both semantics and speech act theory, and on the distinctions between semantics and pragmatics and between semantic reference and speaker’s reference, among others.