Abstract

This dissertation begins with an analysis of the recent history of the cognitive neuroscience of vision. This analysis is intended to test the basic assumptions of computational functionalism. Functionalists have argued that since psychological properties can be realized by a variety of structures, those structures are irrelevant to a proper characterization of their properties. By considering applications of functionalist methodology in cognitive science, I argue that functionalist models of mental life can be faulted for being either too minimal, or for making empirical and methodological claims that cannot be supported. In opposition to the functionalist autonomy thesis, Chapters Three and Four argue that taking neuroscientific evidence seriously can justify a revision of psychological taxonomies. ;Though the approach taken here with respect to the relationship between neuroscience and philosophy echoes some arguments presented in support of theory reductionism, it does not presuppose taking a reductionist or eliminativist stance on either ontological or methodological questions. In opposition to traditional eliminativist or reductionist suggestions, it is argued that naturalizing the philosophy of mind does not imply the complete subordination of philosophical inquiry to neuroscientific evidence. As evidence for this general claim and for the relevance of philosophical analysis to scientific inquiry, Chapters Four and Five present an analysis of uses of the notion of information in the neuroscience of vision. This analysis draws on past work in the logic of perception, relying in particular on Jaakko Hintikka's analysis of identification. Chapter Five considers the so-called two-pathways theory of visual identification in cognitive neuroscience, concluding that developments in semantical analysis can make an important difference at the level of concrete neuroscientific investigation. Specifically, the scientifically misleading claim that two kinds of visual information are necessary for the implementation of two modes of identification is corrected.