Abstract

Arguments for the claim that we think in a distinct language of thought are common in the philosophical literature from Plato to the present. In this dissertation, I examine the philosophical foundations of this currently popular "Language of Thought" Hypothesis , evaluating both the empirical results and the a priori grounds that have been presented in support of it. After presenting an historical survey of philosophical motivations for LOT, I discuss a number of psychological experiments in the area of bilingualism and that of the integration of pictorially and linguistically encoded information. I argue that the empirical results in these areas do not establish, as many have claimed, that there exists a special language of thought, distinct from any natural language or system of pictorial representation. Rather, I show, the data can be taken to support LOT only if they are interpreted in the light of some questionable a priori assumptions about the nature of language, pictures, and concepts. I examine and undermine these assumptions as well as a number of a priori defenses of LOT, and argue that my alternative account of mental representation, the Multiple Codes Model, is not only consistent with the empirical data, but also preferable on a priori grounds