UC San Diego

Functional localization is a primary goal of neuroscience, but there is a long history of debate about whether it is possible. That is, can one decompose the brain into parts, whose concerted functioning explains psychological states and behavior? These debates continue to the present day. In my dissertation, I criticize "absolutism," a meta- theoretic assumption about functional explanation that has shaped dialectic throughout the history of debates about localization. Absolutism is the view that each neural area performs a univocal function, and performs that function regardless of the contexts in which it operates. Functional accounts in neuroscience have traditionally been absolutist. This was true of phrenology's attempts to isolate personality traits in parts of the brain, and remains true in current discussions about whether, for instance, a particular area of the brain is the "face- recognition" area. More deeply, absolutism has been seen as virtually definitive of localization--those denying absolutist function ascriptions have tended to deny localization altogether. I argue, based on a detailed case study of perceptual area MT, that absolutism should be rejected. MT is traditionally construed as the "motion detector" of the visual system, but recent evidence strongly suggests that MT performs different functions-- including, most compellingly representing depth--depending on the perceptual information available. I suggest that this evidence is best read as showing that MT is context- sensitive in its functioning. That is, it performs different functions depending on the context. Context- sensitivity is often considered to be incompatible with giving powerful theories, and several theorists, motivated by this concern, have attempted to save absolutism by modifying the kinds of functions posited. I argue that these attempts fail, and that, appropriately understood, contextualist theories can be seen as providing functional explanations that meet the epistemic goals of functional decomposition