Abstract

The thesis discusses certain questions concerning the semantics and pragmatics of definite descriptions. It begins with a brief look at the famous Russell-Strawson debate. I argue that, at least on one interpretation of Strawson's position and for a large class of cases, Strawson's account of definite descriptions is less satisfactory than Russell's account. ;The implications for Russell's theory of the more recent criticisms raised by Keith Donnellan in the paper "Reference and Definite Descriptions" are explored. Donnellan distinguishes two uses of descriptions, which he calls the referential use and the attributive use. I attempt to determine exactly how this distinction should be understood and argue that there are two ways in which a speaker can use a description to refer. In an attributive use the speaker intends to communicate a particular mode of identification to his audience, whereas in a referential use the speaker simply intends his audience to identify the object in some way or another. ;The distinction is explained in terms of what the speaker intends to communicate, not in terms of what he says, and this raises the question whether the distinction bears in any way on the semantics of definite descriptions. I address this question in the remaining chapters and look at Donnellan's criticisms of Russell and Strawson and at his account of the semantics of descriptions. Several alternatives to Donnellan's position are described. ;I critically examine Donnellan's grounds for the position arguing that he fails to show Russell's theory of descriptions is mistaken. I look at his arguments for this claim in "Reference and Definite Descriptions", and examine his more recent attempt to defend his position. ;Although I argue that Donnellan fails to show that the theory of descriptions is mistaken, I defend a conclusion similar to Donnellan's. My argument, which is similar to an argument of Howard Wettstein's, concerns the use of non-uniquely denoting definite descriptions. Although a Russellian view can go a long way toward accommodating the kind of phenomena Donnellan describes, in the final analysis I believe that Donnellan's position is more or less correct: Russell's theory does not apply to all uses of definite descriptions