Christopher Hill

I will be concerned in these pages with the views that Gilbert Harman puts forward in his immensely stimulating paper Self-Reflexive Thoughts.¹ Harman maintains that self referential thoughts are possible, and also that they are useful. I applaud both of these claims. An example of a self referential thought is the thought that every thought, including this present one, has a logical structure. I feel sure that this thought exists, for I have entertained it on a number of occasions. Moreover, I feel that it is extremely useful. Without deploying it, how could we tell the whole truth about the nature of thoughts?

This is a book about sensory states and their apparent characteristics. It confronts a whole series of metaphysical and epistemological questions and presents an argument for type materialism: the view that sensory states are identical with the neural states with which they are correlated. According to type materialism, sensations are only possessed by human beings and members of related biological species; silicon-based androids cannot have sensations. The author rebuts several other rival theories, and explores a number of important issues: the forms and limits of introspective awareness of sensations, the semantic properties of sensory concepts, knowledge of other minds, and unity of consciousness. The book is a significant contribution to the philosophy of mind, and has much to say to psychologists and cognitive scientists.

In this paper I celebrate the virtues of Hurlburt and Schwitzgebel's path-breaking book on introspection, but I also exp-ress dissatisfaction with a few of its recurring themes. The main body of the paper consists of seven theses about the way in which the study of introspection should be conducted. Thus, to a large extent, the paper is a methodological proposal, though it also makes a number of concrete claims about the nature of introspection, and about the epistemological status of its deliverances. The methodology I endorse is quite different than the one that Hurlburt advocates, but even so, it is compatible with assigning a large role to Descriptive Experience Sampling. Equally, while I am no fan of Schwitzgebel's radical scepti-cism about introspection, he and I are of like mind on a number of spe-cific epistemological issues, and we share the sense that it would be useful to draw on other areas of cognitive science in extending Descriptive Experience Sampling and refining it