Jerry Walls’ Argument for Libertarian Free Will

In Jerry Walls’ “Why No Classical Theist, Let Alone Orthodox Christian, Should Ever be a Compatibilist” (Philosopia Christi, Vol. 13, No. 1, 2011), he offers an interesting argument to the conclusion that some actually existing humans have libertarian free will. Here is his argument, numbering in the original:

(6) If God is necessarily perfectly good, He eliminates all evil He can properly eliminate in all possible worlds.
(7) In all possible worlds in which persons are not free or are only free in the compatibilist sense, God could properly eliminate all moral evil except that evil necessary for creatures properly to appreciate good (or similar purposes).
(8) Therefore there are no possible worlds in which persons are free only in the compatibilist sense , and in which there is moral evil beyond what would be necessary for creatures properly to appreciate good (or similar purposes).
(9) Our world contains much appalling moral evil that could not plausibly be thought necessary for creatures properly to appreciate good (or similar purposes).
(10) Therefore, in our world persons must be free in the libertarian sense.

There are a lot of issues I have with this argument that I will not address in this post. My main concern is that I don’t think Walls establishes his conclusion. I will not challenge the truth of the conclusion (though I do not believe it is true—indeed, I believe it is necessarily false!); rather, I don’t think it follows from its premises. I think the reason why it doesn’t follow is instructive, hence, this post!

I begin with my upshot, and then I’ll explain below how I get there.

Upshot: Walls’ argument trades on a confusion between indeterminism and incompatibilism. Jerry’s conclusion demands the truth of the latter but, at best, he’s only demonstrated the former. Put another way, Walls seems to think that showing that a world is not determined but that (some) people in that world are free is sufficient to prove that their freedom is libertarian. On the contrary, showing such a thing is only a necessary condition for concluding that their freedom is libertarian.

First, let’s get three concepts under our belt: (1) compatibilism, (2) libertarianism, and (3) incompatibilism.

Compatibilism: The thesis that free will (and/or moral responsibility) is compatible with the truth of determinism. We might put it this way: a consistent model in which (a) determinism is true in world w and (b) someone is free in w can be constructed.¹

Libertarianism: The thesis that (a) incompatibilism is true and (b) someone is free/morally responsible. The conditions that need to be met for (b) need not concern us at the moment. What will be important for my argument is unpacking (a).

Incompatibilism: This is the negation of compatibilism. Whereas compatibilism says that it is possible that free will is true and determinism is true in some model, incompatibilism says no such possible model exists. Therefore, incompatibilism is, if true, necessarily true.²

Now let’s walk through Walls’ argument.

So let’s grant (6).

Premise (7) is troublesome. It buries whether a subtle but important state of affairs is in fact true, or whether it need not be actually true. Sometimes, someone might be presupposing or conversationally implying that we live in a deterministic world by uttering the phrase, “We have compatibilist freedom.” In these cases, if “We have compatibilist freedom” is true, then “our world is deterministic” is also true. Call this utterance Context-α

Other times, this is not the case. You will note that, according to the definition of ‘compatibilism,’ the actual truth of determinism is not required for the truth of compatibilism. All it says is that a model can be constructed on which ‘world w is determined’ is true and ‘someone is free in w‘ is true as well. How can such a model be constructed? The details are beyond the scope of this entry. But assume for simplicity’s sake that it turns out that the kind of freedom that is needed for moral responsibility is something like this: “The ability to choose what we want to do for the right reason.” Let us now suppose that indeterminism does not undermine the proper exercise of the control that we need to be free and/or morally responsible. Then, this freedom could be had in both indeterminist and determinist worlds. Therefore, this freedom is compatibilist freedom, though having it does not entail that the actual world is deterministic. Thus, someone might utter the phrase, “We have compatibilist freedom” and not mean or presuppose that the actual world is deterministic. They could be agnostic about whether it is. Call this utterance Context-β

Now look at (7) again. Let’s ignore the “not free” disjunct. What is the context of Jerry’s use of “people are free in the compatibilist sense, ” Context-α or Context-β? It seems obvious to me that he means it in the Context-α sense. If Jerry meant it in the Context-β sense, it’s not clear that the conditional is true. For we could have compatibilist freedom (in our simplistic sense defined above) yet the world be indeterministic. We could choose A (because we want to and for the right reason) in our world and yet in another world with the same laws and past leading up to our choice, some alternative possibility obtains, and we do not choose A. But then, it’s not clear that God could stop all the evil. Indeed, suppose God could not have (or voluntarily gives up) exhaustive foreknowledge of future contingents, then he couldn’t know our future, contingent choices even though our freedom is compatibilist (i.e., if we had it in a deterministic world, we would be free and/or morally responsible). For all we know, there may have been very good reasons for God to create the world indeterministic in this way.

So, it seems to me that (7) presupposes that the world in question is deterministic. We should read it thus

(7*) In all possible worlds, w, in which persons are . . . only free in the compatibilist sense, [and determinism is true in w,] God could properly eliminate all moral evil except that evil necessary for creatures properly to appreciate good (or similar purposes).

But from (7*), (8) does not follow. Let w be a world that is indeterministic and the freedom had is compatibilist (i.e., in Context-β, if the world were deterministic people would be free, but the world might not actually be deterministic). In w, our future free choices are contingent. That is, the future in w is (metaphysically) open. In w, God does not know our future free choices (though he could if he decided to determine them). God, then, cannot stop all the freely committed evils while also leaving the world indeterministic. But, as stipulated, the freedom had in w is compatibilist. Hence, (8) is false. (A similar argument could be run assuming God has exhaustive foreknowledge, if that is even compatible with the problem of knowing future contingents.)

Now, (9) claims that our world is a world with evils that “can’t plausibly be thought necessary for creatures properly to appreciate good.” I think (9) is false or, at least, that it’s a bad “noseeum” inference. But leave this aside. What we’ve seen is that, if Jerry is right, then our world can’t be deterministic.

Now for the conclusion, (10). Jerry concludes that since our world can’t be deterministic, and (suppressed premise) people are free in our world, then their freedom must be libertarian. But this is too quick. It is not the case that if S is free in world w, and world w is indeterministic, that freedom is incompatible with determinism. For indeterminism is a contingent truth, if true at all. Many if not all compatibilist philosophers admit that our world might be indeterministic. Indeed, some think it is. However, they think that if determinism turned out to be true, this wouldn’t affect us very much in terms of our ordinary practices of holding responsible, blaming, praising, etc. The truth of indeterminism is insufficient to demonstrate incompatibilism.

To conclude that libertarianism is true, Jerry needed to prove incompatibilism. That is, that no worlds are worlds where determinism is true and someone is free in that world. But for all Jerry has said, there can be such worlds. Consider again:

(6) If God is necessarily perfectly good, He eliminates all evil He can properly eliminate in all possible worlds.
(7) In all possible worlds in which persons are not free or are only free in the compatibilist sense, God could properly eliminate all moral evil except that evil necessary for creatures properly to appreciate good (or similar purposes).
(8) Therefore there are no possible worlds in which persons are free only in the compatibilist sense , and in which there is moral evil beyond what would be necessary for creatures properly to appreciate good (or similar purposes).

Now, it is consistent with (6)–(8) to say this: “Yes, I agree. But, I am a compatibilist. So:

(9*) There is a world, w, in which persons are free in the compatibilist sense and w is a world without evil.”

(I assume Jerry thinks that (9*) is possible on the assumption that w is determined by God and some humans in w have compatibilist freedom.)

Jerry’s (10) requires that (9*) is necessarily false, but nothing he says in (6)–(8) rules out (9*)-type worlds. He can’t move from the contingent premise, “There is unjustifiable evil in @,” to the modal premise inherent in his conclusion, “So deterministic worlds with free persons are impossible.”

However, things are not that bad for Walls. He only needs to adjust his conclusion from (10) to:

(10*) Therefore, the actual world is indeterministic.

(10*) is, if true, bad enough for Calvinism. I don’t think (10*) is true, but arguing against it is beyond the scope of this entry.

_______________

¹ Technically, the compatibilist’s job is even easier. It need not be that w is determined, w could be indeterministic with pockets of determinism. Compatibilism would only need to show that there is some possible deterministic pocket in which does a free (and/or morally responsible) action. Things get complicated, however, because libertarianism could in principle agree to the above, so long as the free person in the deterministic pocket can trace her freedom back to a prior libertarian free action that in some sense grounds her freedom in the pocket. So, the compatibilist would have to make clear that derivative (libertarian) freedom doesn’t ground the free/responsible act in the pocket.

² Kevin Timpe presents the incompatibilist thesis thus: “The incompatibilist . . . maintains that [a fully determined world containing a free person] is impossible and that free will is possible only if determinism is false”. Similarly, Kadri Vihvelin states, “Incompatibilism is usually understood as the claim that the truth of determinism entails the non-existence of free will: that there is no possible world where determinism is true and someone has free will.” In modal logic terms, if compatibilism is a proposition scoped by a ‘<>’, then incompatibilism is the negation of this, i.e., ‘~<>’, which is equivalent to ‘[]~’. So proving incompatibilism amounts to proving a claim scoped by the necessity operation, ‘[]’.