How to separate conceptual issues from empirical ones in the study of consciousness

Modern consciousness studies are in a healthy state, with many progressive empirical programmes in cognitive science, neuroscience, and related sciences, using relatively conventional third-person research methods. However not all the problems of consciousness can be resolved in this way. These problems may be grouped into problems that require empirical advance, those that require theoretical advance, and those that require a re-examination of some of our pre-theoretical assumptions. I give examples of these, and focus on two problems--what consciousness is, and what consciousness does--that requires all three. In this, careful attention to conscious phenomenology and finding an appropriate way to relate first-person evidence to third-person evidence appears to be central to progress. But we may also need to re-examine what we take to be "natural facts" about the world, and how we can know them. The same appears to be true for a trans-cultural understanding of consciousness that combines classical Indian phenomenological methods with the third-person methods of Western science.