Abstract

The Direct Reference account of the semantics of singular terms is widely assumed to be inconsistent with the traditional Individualist account of psychological states. Because of this assumption, and because of the weight of the evidence for Direct Reference, Anti-Individualism has found supporters despite its counterintuitiveness. In this dissertation, it is argued that Direct Reference and Individualism are not genuinely inconsistent, but that the inconsistency emerges only with the additional assumption of Propositionalism--the orthodox, proposition-based framework for understanding thought and language. Consequently, it is not necessary to endorse Anti-Individualism in order to accept Direct Reference, so long as Propositionalism is accepted only in an attenuated form, if at all. ;It is argued that the best strategy for resolving the inconsistency is to retain all of Propositionalism apart from the Relational Analysis of belief ascriptions. An alternative analysis--the Quantified Relational Analysis --is developed and motivated. According to QRA, an ascription of the form X believes that S attributes the property of believing some member of the class of propositions containing the proposition that S and finer-grained versions of it. QRA is extended also to de re belief locutions. The conjunction of Direct Reference, Individualism, and the remnants of Propositionalism plus QRA is termed Russellian Non-Parallelism. ;Like all Direct Reference theories, Russellian Non-Parallelism has the consequence that co-referring names and other genuine terms are freely substitutable for each other, despite the datum that ordinary speakers intuitively regard such substitutions in belief contexts as suspect. These intuitions are accounted for both pragmatically and semantically. The pragmatic explanation specifies a mechanism whereby the audience sometimes infers from the choice of genuine terms the subject's way of thinking of the objects involved. The semantic explanation motivates an idiomatic reading of ascriptions, on which, under certain circumstances, the choice of co-referring terms in ascriptions literally indicates different belief contents. Russellian Non-Parallelism is then applied to enhanced substitution puzzles and to some other traditional problems, such as the contingent a priori. Finally there is a discussion of pertinent contemporary literature