In this chapter I shall consider, and criticise, Fodor's argument from realism about the status of propositional attitude psychology to the conclusion that some version of causal co-variance semantics must be correct, and so to the conclusion that the language of thinking is Mentalese. I shall also consider how plausible it would be to develop a semantics for natural language directly, not via a prior account of the semantics of thought. It is important that some form of language-based semantics should be defensible, I shall argue, if it is to be possible to defend any robust version of the thesis that public language is implicated in conscious thinking.

From mental realism to Mentalese

In the last chapter I outlined a number of criticisms of Fodor's form of causal co-variance semantics. It is hard to be confident, however, that Fodor can have no adequate reply to those objections. It is even harder to be confident that there can be no other version of causal co-variance theory which might avoid the objections. So there is no sense in which Fodor's approach to the semantics of thought can be regarded as having been definitively refuted. Moreover, Fodor has available to him, in addition, a general argument for claiming that some version of causal co-variance semantics must be right, since it is the only kind of semantic theory which stands any chance of being genuinely atomistic (see his 1994, ch. 1).