Homunculi and self-representing representations

Representational models of the mind have trouble accounting for the thing that has representations. Their picture of the mind is usually Humean, in which nothing but representations and transformations of representations have a place. It is difficult to build an account of the subject, the thing to whom representations represent, out of these resources – difficult, but perhaps not impossible. In the second edition, Kant at least opened the way to a possible solution to this problem that stays within the ambit of representations. Traditionally, philosophy and cognitive science have concerned themselves with how representations represent, how they are representations of something. In this chapter, we are going to focus on their representing to someone.

The representational model of the mind in one form or another has been so dominant for the past four centuries, at least in European philosophy and its offshoots, that even those who do not accept it always end up discussing it. Representational states are, minimally, states or processes that contain or convey information about some other states or processes or objects, real or imagined. In the representational model, mental states are treated as representations (information-bearing states), propositional attitudes are viewed as consisting in computational relations among representations, and mental activity is a matter of performing transformations on representations. There is nothing in any of this that Kant would have rejected.