Abstract

The language of thought hypothesis holds that propositional attitudes are causes of behavior and identical to relations between subjects and tokens in a language-like system of mental representation. Its main support is Jerry Fodor's argument that our only plausible psychological models of cognitive processes presuppose it. This dissertation examines criticisms of Fodor's argument by three philosophers: Christopher Peacocke, Stephen Stich, and Daniel Dennett. Peacocke argues that cognitive processes do not require a language of thought since the states realizing them need have no syntactic structure. I argue that he mistakenly assumes LTH requires a correlation between features of content and syntax, and that his description of how psychological processes could be realized without an LT is inadequate. Stich argues that cognitive psychology commits us only to holding that there are syntactically structured mental representations that need have no content. I show that Stich is mistaken about the commitments of cognitive psychology, that the availability of the purely syntactic theory he prefers would not impugn current cognitive theory's models, and that there are sound reasons for ascribing contents to mental representations. Dennett raises the problem of what the difference is between those organisms having mental representations and those not. I consider Fodor's theory on this, amend it, and offer such a difference. Dennett also objects that the contents of mental representations are inappropriate for propositional attitudes, but I show he only establishes that some are. I consider two other of Dennett's arguments purporting to show that the intentional strategy has no realist implications, argue that neither motivates rejecting Fodor's argument, but contend that one does support a weaker claim: that we need not draw realist implications if we are willing to accept instrumentalism about Fodor's models. Finally, I argue that future psychology might not support Fodor's argument, hence that our commitment to LTH on the basis of current psychology must be to that extent provisional