Teleology and Degrees of Freedom1

There is a debate in philosophy of mind about the nature of reason explanations of action, and this volume is testament to a resurgence of interest in non-causal accounts. In Teleological Realism: Mind, Agency, and Explanation,2 I have proposed a non-causal account according to which common-sense reason explanations of action are irreducibly teleological in form. I claim that we explain behavior by citing the state of affairs towards which the agent was directing her behavior, i. e., by citing the purpose or goal of the behavior. I will not be defending that account of action explanation here, but will be assuming it and applying it to the free will debate. I will argue that the teleological account of action explanation leads to a view of free will with some interesting and attractive features. Philosophical accounts of free will typically propose some sort of criterion for determining which behaviors count as free, e. g., that the agent could have done otherwise, or that the behavior was in accord with the agent’s second-order volitions. The account is then usually used to answer the question of whether we have free will, and especially whether we can have free will even if determinism is true. To test these accounts, philosophers make arguments of various sorts, but the primary method is with cases. Here’s the recipe for an objection to an account of free will: come up with a case where the theory says that the agent is free, but our intuition says that the agent is unfree, or vice versa. I won’t question the method of cases in principle; it’s what we have and it makes some sense. On the picture of analytic philosophy as conceptual analysis, it makes perfect sense. Philosophers put forward a precisely spelled out version of our concept of free will; but if such an account is at odds with firm intuitions in particular cases, then it can’t be what we mean. We know that the situation is not really so simple as that. Our ordinary concept of free will might be messy and not perfectly consistent; accordingly, any particular theory might get a few cases ›wrong‹ as judged by our intuitions. Instead of supposing that we are doing straightforward conceptual analysis, it makes more sense to assume that we are aiming for something like Rawlsian reflective equilibrium, in which we consider