Reverse Inference and Mind-Brain Identity

Reverse inference is a widespread procedure of reasoning from patterns of brain activation to the engagement of specific mental processes. One of the main attractions of reverse inference is the apparent possibility it opens for exceeding the limits of behavior based procedures in psychology. Underlying this motivation is an implicit assumption of mind-brain identity according to which behavior does not play a constitutive role in the classification of mental kinds. A widely accepted consideration, however, against mind-brain identity is that
while identity is a one-one relation, there appears to be mounting evidence that the mind-brain relation is one-many. In this paper we examine three recent strategies for responding to this consideration, positing that two of them fail, while the third may be successful but at the cost of giving the behavioral criteria of mental kinds a constitutive role to play in the classification of brain kinds. For this reason, the third strategy does not yield an account of mind-brain identity capable of grounding a positive answer to the question of whether reverse inference can exceed the limits of behavior based procedures in psychology. It follows that philosophical defenses of mind-brain identity may be useless for the purposes of science. Nevertheless, they are not irrelevant to science, since their failure to ground a specific scientific research strategy such as reverse inference may constitute an important negative lesson concerning this strategy.