Canadian Journal of Philosophy

Traditionally, incompatibilism has rested on two theses. First, the familiar Principle of Alternative Possibilities says that we cannot be morally responsible for what we do unless we could have done otherwise. Accepting this principle, incompatibilists have then argued that there is no room for such alternative possibilities in a deterministic world. Recently, however, a number of philosophers have argued that incompatibilism about moral responsibility can be defended independently of these traditional theses (Ginet 2005: 604-8; McKenna 2001; Stump 1999: 322-4, 2000 and 2002; van Inwagen 1983: 182-8; and Zagzebski 2000). Incompatibilists of this stripe are generally motivated by the concern that, if determinism is true, we are not genuine or ultimate sources of our actions and, hence, we are not responsible for what we do. Following Michael McKenna (2001), I shall call this view source incompatibilism. While the source incompatibilist's concern is rather vague as stated, it has given rise to a powerful argument against any attempt to reconcile moral responsibility and determinism. John Martin Fischer and Mark Ravizza (1998) have labeled this the Direct Argument, as it avoids the detour of alternative possibilities. I believe that this argument presents the greatest current challenge to compatibilism regarding determinism and moral responsibility. In fact, I shall argue that the Direct Argument cannot be defeated.

Instead, I conclude that it leads to a dialectical stalemate. While I am not alone on this point, the implications of this stalemate have not been [End Page 411] fully appreciated. In particular, I contend that the most plausible conclusion to draw from the current debate is that we should embrace agnosticism about moral responsibility: we simply do not know whether anyone is ever morally responsible for her actions.¹

My argument is cast against the backdrop of two relatively controversial assumptions. I do not plan on presenting a full defense of either here; however, I would like to give some indication of why I find each plausible enough as a starting point for our discussion. First, I assume that there is currently no compelling evidence that we possess a libertarian free will. Of course, it surely does seem to us that we are free and that, when we act, we could have done otherwise. And there are clearly a number of cases in which agents appear to demonstrate the ability to perform different actions in relevantly similar circumstances. But this cannot amount to compelling evidence for a libertarian account of free will. Otherwise, we would need to accept that the phenomenology of experience and anecdotal evidence concerning our abilities give us deep insight into the causal structure of the world. This I find wildly implausible. Neither direct introspection of the deliberative process nor casual observations of human behavior can tell us whether determinism is true.² Of course, it is possible that scientists may one day tell us that the world is indeterministic in just the right sort of way to allow us to exercise libertarian free will. Robert Kane (1996: 128-30), for one, is optimistic about such a prospect. Optimism, however, is not evidence, and we seem to be far away from understanding the inner workings of the mind well enough to make such a case.

While this assumption may appear innocuous enough, I believe that it presents a significant challenge to the commonplace belief that we are morally responsible agents. Indeed, it is hard to overstate the importance of this assumption for my argument. In particular, I take it to imply that any successful argument for incompatibilism about moral responsibility is also a successful argument for agnosticism about moral [End Page 412] responsibility. If moral responsibility requires a certain sort of psychological or neurological indeterminism — and I will argue that, for all we know, it may very well — and if, as I am assume, we do not yet know whether this requirement is met, then we do not yet know whether we are ever morally responsible for what we do. And so, unless every argument for incompatibilism can be decisively defeated, we should be prepared to admit our ignorance.³

Second, my argument for agnosticism relies on a particular version of the...