Foreknowledge and Freedom

In The Dilemma of Freedom and Foreknowledge, Linda Trinkaus Zagzebski outlines a variety of solutions to the dilemma of the title – how God’s omniscience, which includes omniscience regarding the future, is compatible with human freedom. Is it possible to account for God’s foreknowledge in such a way that His foreknowledge doesn’t have a causal, determining relation to my actions? She believes that philosophers have shown that freedom is not logically incompatible with divine foreknowledge, but she also recognizes that there’s a difference between demonstrating logical consistency and making a claim plausible. She offers various proposals that move from logical consistence to plausibility.

One proposed solution depends on a Thomist account of divine knowledge, inflected with elements of an Ockhamist account of foreknowledge. Thomas claims that God’s knowledge is simple, a single unified act of knowing, and that, given the reality of divine simplicity, His direct object of knowledge is His own essence. He doesn’t know or believe discrete bits of information. He knows Himself, and in knowing Himself knows all that He has and can do. In knowing Himself, He knows not only everything that is the case, but also everything that would be the case in all possible worlds.

As she summarizes Thomas’s position, “Primarily and essentially God knows only himself. To know anything else primarily would be to focus the divine gaze on the imperfect. But in his simple and direct intuition of his own essence, God knows secondarily everything else. This is because God’s essence contains exemplars of the infinitely many ways his essence can be represented in finite reality: ‘God knows Himself as primarily and essentially known, whereas He knows other things as seen in His essence.’” More succinctly, “God’s knowing all contingent truths is like seeing them through the ‘features’ of his own essence.”

This breaks the causal link between God’s foreknowledge and my actions: “God’s belief at t1 [time 1] does not strictly imply my act, since the numerically same belief occurs in worlds in which I do not perform the act in question. What is essential to God’s belief at t1 is its primary object, namely, God’s own essence. It is only an accidental property of that belief that it is a belief that I do S rather than that I do not do S. The numerically same belief occurs in worlds in which I do not do S.” Thus, God’s foreknowledge no more determines my act than His foreknowledge determines my not-acting in an alternate world.

Zagzebski’s argument takes some odd turns. Start with that last claim: God’s knowledge of His own essence includes knowledge of things that are not in fact the case, of events and actions that do not occur because they are events and actions of a possible world rather than the actual world. We can’t say that God “knows” falsehoods; falsehoods don’t count as knowledge. So, in knowing both the world that is and the worlds that are not, God must somehow be able to distinguish between knowledge of a possible and knowledge of the actual world. We don’t want to say that God knows all actual facts and all possible facts, but doesn’t know the difference between actual and possible. Once we make that distinction, though, we are in some tension with the claim that God knows His essence in a single simple epistemic act; the knowing seems differentiated, at least between the actual and the possible. And then, it seems, we must ask whether or not God’s foreknowledge is identically related to states of affairs that are actual and states of affairs that are possible.

Zagzebski makes an argument for God’s independence of time, thus challenging the Ockhamist claim that God is temporal. She doesn’t think this is of the essence of the issue; she thinks that the key point in Thomas’s account of foreknowledge is not timelessness but the nature of God’s knowing, the fact that God knows everything else through knowing Himself. But the argument for timelessness slides from speaking of God’s “independence” from time to His “non-relation” to time. She writes, for instance, “Since God’s independence of everything outside himself can never be lost at any time in any world, his aseity requires the truth of at least the following. For every x outside of God, God is essentially independent of x. For every time t and world w, it is possible at t-in-w that God has no relation to x.” I don’t think so: God is indeed essentially independence of every x, but it is not possible for any x to exist at all in any t or w, without being in some relation to God – the relation of a dependent creature on the Creator. I suspect the “no relation” means “no relation of dependence,” because the import of the actual language is the opposite of her intention: If x can exist independently of God, the God is not the Creator and preserver of everything outside Himself.”

The account of Thomist divine epistemology also takes a strange twist. She draws an analogy between God’s knowing of things through knowing His essence and the way we know secondary objects through knowing a primary objection: “a visual act can have a secondary object as well, and a secondary object is not essential to it or, at least, need not be. Let us suppose that the secondary object of V is the features of the young man’s mother. You see the mother’s features through the features of the son. The seeing of the secondary object is not a distinct act of seeing, given the way it is related to the primary object, yet in this case it is plausible that the secondary object is only accidentally related to V That is, if you had not seen the mother in the son it would still have been the same visual act. V could have occurred without the secondary object; it would have been a complete visual act, in fact, that complete visual act, without it. So now we have a case in which an act of seeing is essentially directed toward one object and accidentally directed toward another object.”

This leads Zagzebski to use the category of “accident” to describe secondary objects of God’s knowledge. Essentially He knows His own essence; other “objects are accidental properties of the divine epistemic state.” Again, “if it is not reasonable to say that certain beliefs God has at t are distinct from certain other beliefs God has at t, it is not reasonable to propose principles for distinguishing the accidentally necessary ones from the nonaccidentally necessary ones.” This takes us back to the earlier questions about the distinction of possible and actual worlds. But more obviously, it raises some questions about the overall thrust of the argument: Something seems amiss in an argument that begins from divine simplicity and aseity and ends with talk of “accidents” of God’s epistemic state.