Bad Attitudes: Rethinking the Role of Propositional Attitudes in Cognitive Science

Mental representation is a central concept in cognitive science. It is the causal story involving “symbolic structures” that form the explanatory schema of classical cognitive science. Objects that populate the particular ontology of a scientific theory must have an instrumental role in explaining the processes in question. As such, it is requisite that the explanatory vehicles within a theory be in some way informative with regards to a particular inquiry. Due to this methodological presupposition, the entities invoked in a theory can be evaluated in terms of their explanatory value or their status as causal-explanatory entities altogether. In pursuing a more explanatory cognitive science, I will be advancing an eliminative program which addresses the need for representational language but questions the status of propositional attitudes as causal objects in computational explanations. Propositional attitudes are sentences consisting of an intensional verb followed by a content clause; for example, “x believes that y.” It is not my aim simply to reiterate Paul Churchland’s claim that folk psychology is an empirical theory (and thus falsifiable). Rather, I will show that (1) propositional attitudes as causal entities are inadequate and largely uninformative for explaining representation computationally; and (2) understanding the dynamics of networks of neurons (though I will rely on artificial models) with relation to representation will require description in non-propositional terms using a dynamic neurobiological model. I will argue that the functional story, which will inform us about how the brain represents, will be