The Revised Reward Theory of Desire.

I propose and articulate a novel theory of desire, called the Revised Reward Theory. As the name suggests, the theory is based-and expands-on Arpaly and Schroeder's (2014) Reward Theory of Desire. The initial Reward Theory identifies desires with states of the reward learning system such that for an organism to desire some P is for its reward system to treat P as a reward upon receipt. The Revised Reward Theory identifies desires with a different state of the same system, such that for an organism to desire some P is for its reward system to expect or predict that P will be rewarding (roughly) prior to receipt. The difference amounts to equating desires with what we ultimately find rewarding or satisfying versus those that underlie our motivations to obtain that which we take ourselves to desire. I argue that the structure of the reward system is incompatible with the original Reward Theory but compatible with the Revised Reward Theory. I demonstrate that this difference has important philosophical implications. I focus on moral responsibility and demonstrate Arpaly and Schroeder's argument, that addiction can mitigate moral responsibility, turns on this precise difference. Arpaly, Schroeder, and I all ascribe to a meta-theory called 'natural kindism' which identifies mental kinds with neurocognitive kinds. This discussion, in addition to defending a theory of desire, is intended to act as a proof of concept for natural kindism as offering a powerful framework for relating empirical results to philosophical issues.