Abstract

In "Reference and Definite Descriptions", Keith Donnellan introduces the notion of referential and attributive uses of definite descriptions--phrases such as "the author of Waverly". Here and in subsequent papers he argues that Russell's Theory of Descriptions is inadequate as a semantic analysis for all uses of definite descriptions. According to Russell's theory, definite descriptions are quantifier phrases, not singular terms. Donnellan's arguments suggest that definite descriptions are ambiguous. When they are used attributively, definite descriptions should be read as quantifier phrases. When they are used referentially, they should be read as genuine singular terms. ;I argue that it is possible to draw a sharp distinction between referential and attributive uses of definite descriptions. Following Donnellan, I argue that whether a description is used referentially or attributively depends on the intentions of the speaker concerning the truth-conditions of the proposition he wishes to express. This provides us with a precise way to determine whether a definite description is used referentially or attributively. ;I turn next to the question: Does Donnellan's referential/attributive distinction show that definite descriptions are semantically ambiguous? I distinguish two positions which have emerged in the literature. What I call The Strong Referential Thesis maintains that definite descriptions are singular terms, when they are used referentially. On this account, definite descriptions are semantically ambiguous. What I call The Weak Referential Thesis maintains that definite descriptions, when used referentially, are properly analyzed according to Russell's Theory of Descriptions. This thesis maintains that in referential cases, a speaker's utterance semantically expresses one proposition, but he may succeed in communicating some other proposition. This account attempts to defend Russell's Theory of Descriptions against the apparent counterexamples offered by Donnellan and others. ;I present two arguments in support of The Strong Referential Thesis. First, I argue that The Weak Referential Thesis results in the ascription of contradictory beliefs to speakers. I call this The Argument from the Principle of Charity. Second, I argue that The Weak Referential Thesis cannot, in all cases, explain how speakers may succeed in communicating a determinate proposition other than that which their utterances semantically express