Borderline consciousness, phenomenal consciousness, and artificial consciousness: a unified approach

Abstract

Borderline conscious creatures are neither definitely conscious nor definitely not conscious. In this thesis, I explain what borderline consciousness is and why it poses a significant epistemological challenge to scientists who investigate phenomenal consciousness as a natural kind. When these scientists discover more than one overlapping kind in their samples of conscious creatures, how can they identify the kind to which all and only conscious creatures belong? After assessing three pessimistic responses, I argue that different groups of scientists can legitimately use the concept of phenomenal consciousness to refer to different kinds, in accord with their empirical interests. They can thereby resolve three related impasses on the status of borderline conscious creatures, the neural structure of phenomenal consciousness, and the possibility of artificial consciousness. The thesis has three parts: First, I analyse the concept of borderline consciousness. My analysis counters several arguments which conclude that borderline consciousness is inconceivable. Then I explain how borderline consciousness produces the multiple kinds problem in consciousness science. Second, I assess three recent philosophical responses to this problem. One response urges scientists to eliminate the concept of consciousness, while another judges them to be irremediably ignorant of the nature of consciousness. The final response concludes that scientific progress is limited by the concept's referential indeterminacy. I argue that these responses are too pessimistic, though they point to a more promising approach. Third, I propose that empirically constrained stipulation can solve the multiple kinds problem. Biologists face the same problem because of their longstanding controversy over what counts as a species. Building on new arguments for stipulating the reference of species concepts, I demonstrate that this use of stipulation in biology is neither epistemologically complacent nor metaphysically capricious; it also need not sow semantic confusion. Then I defend its use in consciousness science. My approach is shown to be consistent with our understanding of natural kinds, borderline cases, and phenomenal consciousness.