RANDOLPH CLARKE ON THE POSSIBILITY OF RATIONAL FREE ACTION (Received in revised form 26 March 1996) Is it possible for a human action to be both free and fully rational? Galen Strawson (1986, 1994) has argued that it is not. Rational free action, he claims, would be possible only for an agent who was causa sui. Since evidently nothing can be the cause of itself – at least no human agent can – it is impossible that we perform actions that are both free and rational. The impossibility holds, Strawson maintains, whether determinism is true or not. I argue here that Strawson has failed to make a case for the impossibility of rational free action. I share with Strawson the con- viction that freedom is incompatible with determinism; my concern will be the possibility of rational free action given indeterminism. Section One summarizes Strawson’s impossibility argument. In Sec- tion Two I discuss the explanation of nondeterministically caused events, including rational explanation of nondeterministically caused actions. Section Three sketches an account of free will that appears to provide for both fully rational action and greater agent-control than what could be available given determinism. In Section Four, I examine the extent to which such a view can meet Strawson’s challenge. I. STRAWSON’S CHALLENGE Strawson notes that, when it comes to free action, we are particularly interested in rational free actions, or actions performed for reasons (even if the reasons are not the agent’s best reasons). He proposes a certain view (one with which I have no quarrel) of the relation of reasons to action in the case of rational action. When an agent a performs an action for certain reasons, Strawson says, a’s desires and beliefs (or a’s being or coming to be in a certain desire-and-belief state, or events in a characterizable in terms of desiring and believing) must, in Philosophical Studies 88: 37–57, 1997. c 1997 Kluwer Academic Publishers. Printed in the Netherlands.