The compatibilist says that I "could have done otherwise" just in case, if I had chosen to do otherwise, I would have done otherwise. But is this right?
Imagine someone who has a psychological defect that forces him to make certain choices. We can suppose that when such a person is presented with the option of doing X, he is unable to choose to do anything other than X. He's a compulsive Xer. For example, perhaps you've gradually introduced curry into his diet, so that now he has an addictive desire for curry which he cannot control. So he always chooses to eat curry. Notice: his obsession is so strong that it does not merely cause him to X, it also prevents him from choosing not to X.
We can note two features of this person:
So here we have a person who is not able to act otherwise. Yet the counterfactual "If he had chosen to act otherwise, he would have acted otherwise" is true. If this is correct, then compatibilist's account of what it means to be able to act otherwise must be wrong. To be able to act otherwise, it is not enough for the counterfactual "If you had chosen to act otherwise, you would have acted otherwise" to be true. In addition, you have to be able to choose to act otherwise.
Another case of this sort would be a case where my choices are caused by an evil scientist's neural manipulations. My actions in that case may correspond perfectly to my choices. But intuitively I am not acting freely, and I could not have acted otherwise than I do in fact act, because I could not have chosen otherwise than I did. Nonetheless, it can still be true that if I had chosen to do otherwise, e.g., because the evil scientist made me choose something else, then I would have done that other thing instead. So here again the compatibilist seems to be wrong. It can be the case that if I had chosen to do otherwise, I would have done otherwise; but in fact I could not have chosen to do otherwise; hence in such a case it's wrong to say that I could have done otherwise.
(See Kane pp. 29-30, and pp. 45-7 of Taylor's article, for presentations of this objection.) Taylor describes someone whose choices are manipulated by a scientist and says:
This is the description of a man who is acting in accordance with his inner volitions, a man whose body is unimpeded and unconstrained in its motions, these motions being the effects of those inner states. [So it's a case the compatibilist would count as acting freely. But] It is hardly the description of a free and responsible agent. It is the perfect description of a puppet. To render someone your puppet, it is not necessary forcibly to constrain the motions of his limbs, after the fashion that real puppets are moved. A subtler but no less effective means of making a person your puppet would be to gain complete control of his inner states, and ensuring, as the theory of soft determinism does ensure, that his body will move in accordance with them. (p. 46-7)
Now, determinism threatens to show that this is always our situation. We are always causally determined to choose as we do. Hence, even if it's the case that if we had chosen otherwise we would have done otherwise, that doesn't help us very much. We're always causally determined to choose as we do, so it's never the case that we could have chosen otherwise.
A more sophisticated form of compatibilism accepts that criticism. It proposes the following:
(C3) You do X freely iff:
- you do X, and
- doing X is what you want and choose to do (as before), and
- if you had chosen to do something other than X, you would have succeeded in doing that other thing (as before), and
- if there were a good reason for you to act otherwise, you would have chosen to act otherwise
This account does rule out people who are unable to choose otherwise. Such people don't count as acting freely, because they do not satisfy condition (d) of the account. They are not sensitive to reasons in the right way. It doesn't matter what reasons the obsessive-compulsive, or the addict, or the scientist's victim have for not doing X. They would still choose to do X anyway. X is the only choice they are able to make.
We've already seen that being causally determined to do X does not prevent one from satisfying conditions (b) and (c) of this account. Is being causally determined to do X also compatible with satisfying condition (d)? It seems that it is. For it may be that one is causally determined to do X because one is causally determined to have good reasons to do X. If one had been causally determined to have good reasons to do something else, instead, then one may have gone ahead and done that other thing. All that's important, according to this compatibilist, is that your choices are sensitive to and track your reasons. The mechanisms that produce those choices have to be reasons-responsive mechanisms. They have to be such that, if you had had different reasons, they would have produced different choices. If your choices are produced by reasons-responsive mechanisms, and you also satisfy conditions (b) and (c), then you do act freely, on this account. It is not necessary that your choices or actions be causally undetermined.
This theory sounds pretty good. It's very sophisticated, and it avoids most of the problems that we've discussed so far. However, there are difficulties for it, too.
One difficulty is this. Not just any reasons-responsive mechanism will do. Suppose my choices are being caused by the neural manipulation of a benevolent scientist. This scientist always causes me to choose and act in the way that accords with the reasons I have. If there is a good reason for me to X, the scientist causes me to choose to do X. If there is a good reason for me to Y, the scientist causes me to choose to do Y. And so on. In this case, my choices are produced by a reasons-responsive mechanism, but I do not seem to be choosing or acting freely. Perhaps we can get around this problem by requiring that the reasons-responsive mechanism be located entirely inside the agent.
A second difficulty is this. On the current view, it sounds like I can act freely only if I always choose to do what I have good reason to do. I have to always choose to do "the right thing." But if I'm free, can't I also choose to do the wrong thing? Can't I choose to do something which I recognize I don't have good reasons to do? Sure, maybe that would be foolish or evil. But it does seem like it ought to be in my power. The current view says that such a choice would not be free, because it would not have been produced by a mechanism that responds to my reasons in the right way. But it's hard to see why choosing to do the wrong thing has to be less free than choosing to do the right thing.