Abstract

This dissertation works out a new approach to understanding what makes a representation pictorial and what makes a representation imagistic. Over the last thirty years, the most common approach to these problems has been to claim that what makes a representation pictorial is that normal perceivers can perceive it in certain ways. By contrast, my approach singles out structural features of representational systems as that which distinguishes pictures from other kinds of representations. Pictorial systems are those that are transparent, relatively syntactically sensitive, relatively replete, and semantically rich. All pictures are images, but the converse relation does not hold because images need not be transparent, though they satisfy the weaker condition of being mimetic. This approach is similar to, but enjoys advantages over, Nelson Goodman's work on the topic. ;The account of what makes a representation pictorial or imagistic is then brought to bear on the imagery debate, which concerns whether the brain implements imagistic systems of representation, specifically in the early stages of perception. A review of what we know about the workings of the visual and auditory systems reveals that no representations in the brain are pictorial, though some are imagistic in quite limited respects. Though they fail to be imagistic in many interesting respects, the early stages of the visual and auditory systems are highly isomorphic to what they represent. ;The role of isomorphism in perceptual states has been much discussed in both philosophy and cognitive neuroscience. The last two chapters of the dissertation argue that perceptual representations of traditionally secondary qualities like colors and sounds are isomorphic to those properties to a lesser degree than representations of traditionally primary qualities like shapes. This distinction in how such properties are represented can affect out intuitions concerning the nature of those properties and the nature of perceptual experience. Furthermore, it is argued that isomorphism is an important feature of perceptual representations because of the role it plays in making information cognitively available. This role for isomorphism is maintained even if one is an information theorist, and therefore thinks that lawful causal generalizations fix the content of representations