Abstract

The problem for this dissertation is whether animals have mental states. Intentional attribution is useful in predicting and explaining animal behavior, but this is not enough to establish that animals actually do have mental states. Intentional attribution is also useful for predicting and explaining the behavior of subjects which we are quite sure do not have mental states, as well as for subjects whose mental status is unclear. So we cannot conclude, simply from the fact that animals act as if they have mental states, that they actually have them. We need a test for determining when intentional attribution is merely instrumentally useful and when it should be interpreted realistically. In this dissertation, I develop such a test. I argue that, for cases in which intentional ascription works in predicting and explaining the behavior of some subject, we need an explanation of this fact. The best explanation for the success of intentional explanation for animals is that animals have intentional states, causally responsible for their behavior. The argument is, therefore, a version of inference to the best explanation. In the course of the argument I argue that having mental states is a matter of being something it is like something to be. I also take up the question of whether it makes sense to say that mental states are causally relevant to behavior, given that we have, in principle, a complete causal explanation for all behavior in purely physical terms. I argue that we can make sense of mental causal explanations because we can have multiple causal explanations of the same explanandum. I also develop a counterfactual test for causal relevance, and argue that mental states pass it. Finally, I discuss the objection that language is necessary for mental states, so that non-linguistic animals cannot have mental states. I argue that although language is indeed important, it is not necessary. Language is necessary in order to have concepts, but one can have mental states even without concepts