Foreknowledge and Libertarian Freedom: The Self-Reference Problem

I have held a certain argument against the compatibility of libertarian free will and exhaustive divine foreknowledge which I thought I saw a way out of recently, but in writing this post discovered that the “way out” was susceptible to a modified version of the same argument.

The argument is as follows: if God has complete divine foreknowledge, then for God to do otherwise than he does would involve God causing his beliefs about the future to have been false (and thus, not knowledge). Thus, God could not choose to do otherwise than he does while retaining his complete foreknowledge. To put it another way: God’s complete foreknowledge includes foreknowledge of things which God does. Given that God knows what he will do, he must do that which he knows he will do, and so does not have libertarian free will.

The above argument works as is, however, only if God acts temporally. That is, if God’s actions follow each other in time, and are done in response to other temporal events, as is the case with our actions. If, on the other hand, God acts once, or all at one moment from his perspective, then it is possible for him to have libertarian free will with respect to his actions.

If God’s actions are temporal, then when God acts he already has the knowledge of what he is about to choose to do. He cannot, therefore, act otherwise. If God’s actions are done all at once, along with his knowing the future, then his actions may be libertarianly free. This requires more than just God’s acting in a moment, however. If God acts all in a moment, but foreknows in a prior moment, then the problem remains. If God acts without foreknowledge, then he effectively acts blindly. It is necessary, therefore, if we are to retain both God’s libertarian freedom and his foreknowledge, that God act in light of his foreknowledge of what will happen under certain circumstances. He then acts all in a moment choosing what he will do. Having done so, he may as well have complete divine foreknowledge, but it cannot affect what he does, since he has already acted. He may, in his actions, know what set of actions he is choosing and thus what the future will be, and may therefore perform each action in light of all the other actions which he is performing, and this, it seems to me, is as close to complete divine foreknowledge as we can get while retaining God’s libertarian free will—and is close enough for me (not that I endorse the position, given that I don’t actually believe that God has libertarian free will).

This counter, which I believe is a variety of Molinism, is supposed to allow for God having libertarian free will, but it is unclear as of yet whether it falls to the same argument as I started with if we modify the argument a little.

The problem is that Molinism requires God to have knowledge of what people will freely choose under certain conditions. If this includes God, then he has knowledge of what he will, in fact, do, and so is no longer libertarianly free. If it does not include God, then the question is why not? It cannot be due to the fact that his circumstances do not exist, or are unknown, since God, at least, could rigorously specify his circumstances which we have vaguely specified as choosing between various actions. If there is some fact as to what God will do in these circumstances, then, having complete foreknowledge of what would happen in various circumstances, i.e., knowledge about all facts about what will happen or be done under any possible circumstance, God will know what God will do under the circumstance at hand. Thus, even if God’s foreknowledge is limited to what will happen under various circumstances, God is still caught in his foreknowledge such that he cannot have libertarian freedom.

The point of arguing against God’s libertarian freedom is that if God does not have it, we do not need it in order to be morally responsible, or for any other purpose. If God is good and not libertarianly free, then moral agency does not require libertarian free will, else God would need it in order to be good. At this point, I believe I have shown that the options for belief regarding libertarian freedom are:

1. Open theist: deny divine foreknowledge, affirm libertarian freedom. 
2. Determinist: affirm divine foreknowledge, deny libertarian freedom.

The question is which we should choose. The choice is unproblematic for me, since I do not see libertarian freedom as logically possible. Likewise, do not think that the A-theory of time, which open theism relies on, is logically coherent either (more on that in a later post).