Mental representation and causal explanation.

Mental causation has been a concern in the philosophy of mind since Descartes. Intuitively, thoughts are causes of behavior, and they are causes of behavior in virtue of their mental properties. The computational theory of mind views thoughts as symbol tokenings, and thus as causes. However, if the computational theory of mind is correct, the causal efficacy of mental properties is problematic. A representation tokening causes further representation tokenings or behaviors in virtue of local computational properties of the representation. Mental properties could explain mental causation as well, if they could be identified with, or they supervened upon, causally relevant computational properties of representations. But on plausible construals of the nature of mental properties, they do not. If mental properties are assigned relevance in our mental lives, the result is a picture in which the effects of mental events are overdetermined by their mental and physical properties. Since such overdetermination is implausible, the causal efficacy of mental properties should be denied. A number of philosophers have proposed sufficient conditions for causal relevance and argued that mental properties meet those conditions. The role of mental properties in laws or counterfactuals is taken to be pivotal. But there are serious problems with each of the proposed accounts. A property can play an explanatory role, even if it does not play a causal-explanatory role. The point of assigning mental properties to representations is to account for a system's information processing capacities. Mental properties can play this explanatory role without accounting for cause-effect relationships. The causal efficacy of mental properties can be denied, while an explanatory role for mental properties is maintained