Mind-Reading and Computational Tractability

This paper examines the problem of the specification of the intentional content in the attribution of psychological states. It will be argued that such a specification cannot be complete, both in the case of interpreting the minds of human beings, and in that of other animals. The incompleteness of the specification of the intentional content belongs to the physiology of mindreading, not to its pathology. In favour of this thesis are put forward both epistemological and computational arguments. From the epistemological point of view, it will be suggested that the complete specification of the intentional content makes it impossible the psychological attribution. From the computational point of view, it will be argued that the completeness would make the problem computationally intractable.