This is part seven of eleven in a series responding to the essays in Ben Arbour’s edited volume, Philosophical Essays against Open Theism (Routledge, 2019).
In this post I tackle chapter 7 by Robert B. Stewart, “On Open Theism Either God Has False Beliefs, or I Can Know Something That God Cannot” (pp. 110–118). This is a relatively short essay, but it poses a significant challenge against open theism. It’s a challenge that I tackle directly on pp. 47–48 of my recently published Open Theism book (Cambridge, 2024).
In what follows, I state Stewart’s argument against open theism and then explain precisely why the objection fails. It relies on an unstated assumption about the way in which truth relates to time. This assumption is not only one that most open theists would reject because it conflicts with their view of the future as open-ended, but it’s also an assumption that can be challenged on independent grounds. In short, Stewart’s objection fails because it begs the question. That is, it depends on an assumption that should not be conceded without argument, not even by someone who is hypothetically neutral with respect to open theism.
1. Stewart’s Objection
As Stewart puts it, “[t]he thesis of this paper is that on open theism either God has some false beliefs or I can have propositional knowledge of something that God does not know” (p. 110). Either way, God is not omniscient in the sense of believing all and only truths. On the first horn of the dilemma, God fails to be omniscient because He does not believe only truths—He believes some falsehoods. On the second horn of the dilemma, God fails to be omniscient because He does not believe all truths—there are truths we can know that God doesn’t.
The issue that drives Stewart’s objection has to do with God’s propositional attitudes toward future possibilities on open theism. If the future is genuinely open-ended, and known by God as such, then we can imagine God currently waiting on some future contingency to resolve, say, whether there is a sea battle tomorrow. God, open theists say, knows the probabilities of such an event occurring. Suppose the probability of a sea battle tomorrow is quite high, 90%. By virtue of knowing the probability, is it correct to say that God believes that a sea battle occurs tomorrow?
It would seem so: “Very few would accept that 100% certainty concerning an event is required for one to believe a proposition related to said event” (p. 112). Indeed, to suppose that belief requires 100% certainty “raises the bar too high,” for if this were true then “we could never have any historical or scientific knowledge” (p. 112), since knowledge entails belief.
Moreover, if we consider how we normally ascertain whether someone has a belief or not, one standard indicator is that the person “would be surprised if it failed to be true” (p. 113). Another is that “[i]f circumstances were to arise in which it made a practical difference whether p was true or false, he would act as if it were true” (p. 113). But then, after referencing quotes from open theist scholars Greg Boyd and John Sanders, Stewart points out that open theists themselves concede that “God ‘expects’ some things to come to pass and is ‘mildly surprised’ when they do not” and “acts on his expectations” (p. 113) It seems reasonable to conclude, then, that God on open theism has beliefs about highly probable future contingent events that may very well turn out otherwise than God expects. By Stewart’s reckoning, this implies that on open theism, God has (or at least could have) false beliefs (p. 113).
Stewart considers a response from Greg Boyd to the effect that God only believes that which is 100% certain and thus doesn’t have beliefs about future contingent events but only about the probabilities of future contingent events. Stewart regards this as linguistically implausible, however: Given that belief pairs with expectation, if that E occurs tomorrow is 90% probable, then “the only responsible way to interpret the notion of expectation is to assume that God expects [E]” and thus that God believes that E occurs tomorrow (p. 114).
But suppose we run with Boyd’s suggestion despite its prima facie implausibility. In that case, Stewart argues, “I can know something that God does not know.” For example, “I know that my wife [of 30 years] will love me tomorrow.” This knowledge is obviously justified for Stewart and we have little reason for thinking that the proposition in question is not true. Moreover, since, as open theists insist, “love must be a free act,” this example (allegedly) shows that Stewart can have “the exact sort of knowledge of the future that open theists typically insist God cannot have—knowledge of a future free human choice” (p. 114).
Stewart considers the rebuttal that the proposition <Stewart’s wife will love him tomorrow> is not true because it, like other propositions concerning future free human choices, it is not adequately “grounded” in anything real since future events do not exist and the events in question are not causally determined (p. 115). After considering several different theories of truth, Stewart backs off from his claim that he can know something that God cannot know and argues instead that he can know something logically or chronologically before God knows it. How so? Well, Stewart has a highly justified belief that his wife will love him tomorrow. Even if that belief isn’t yet true, strictly speaking, as soon as tomorrow arrives it becomes true and Stewart will then have known that proposition. God will also know that proposition tomorrow, but in God’s case He won’t even believe it until tomorrow. Stewart reasons that because his belief is logically and chronologically anterior to the truth of that proposition whereas God’s belief is logically posterior to its truth, his (Stewart’s) knowledge of that proposition is at least logically prior to God’s knowledge of it (p. 116). This seems wrong, however, because God is supposed to be the greatest possible knower, and so His knowledge should not (it seems) be posterior in any sense to creaturely knowledge.
Stewart concludes his essay by summarizing his argument in the form of a dilemma:
We are faced with a dilemma. If God has beliefs about future free human choices, then on open theism, God has [or at least could have] some false beliefs. If God does not presently have beliefs about future free human choices but will have at some point in the future, then it seems that I can presently know something that he cannot know at this time, or I can know something logically prior to God’s knowing it. (p. 117)
2. Stewart’s Dilemma Refuted
I’m now going to respond to Stewart and refute his dilemma. Before explaining how it goes wrong, let me state my appreciation of Stewart’s argument. His objection against open theism deserves to be taken seriously. It’s not a cheap shot. If Stewart’s argument were sound, it would pose a significant problem for open theism. Furthermore, Stewart does a good job of trying to consider how open theists might respond to his dilemma and adjusts his argument accordingly. Good for him. Unfortunately, as I stated above, his dilemma (specifically, the first half of it) only gets off the ground if one makes a highly questionable assumption. It’s an assumption, however, that can seem highly plausible, so open theists shoulder the burden of showing why it is questionable.
The assumption is what Patrick Todd calls retro-closure (see this post for discussion):
retro-closure (RC): p → HFp (i.e., if p is the case then it has always been the case that p will be the case)
What RC says, in essence, is that the future has always been alethically settled. There has always been a complete, true story of the future because, however the story plays out, it has always been true that it was going to play out that way. Here’s how RC informs Stewart’s dilemma: Stewart reasons that if God had an expectation/belief about a future contingent event and that event did not turn out as God expected it to, then God had a false belief. That is, because the event turned out a certain way it was (per RC) previously true that it was going to turn out that way. Hence, if God expected that it was going to turn out otherwise, then God had a false belief.
Now, as Stewart points out, one way open theists can try to resist this conclusion is by arguing that God’s expectations about future contingents are not, strictly speaking, beliefs. Perhaps God only believes things He is 100% certain about and so doesn’t have any beliefs about how future contingents turn out but only beliefs about their probability of turning out one way or another. This is the main response that Stewart focuses on.
I don’t think this is a good tack for open theists to take. For one thing, it doesn’t challenge the underlying assumption of retro-closure, an assumption that, if accepted, would force open theists into what’s been called limited foreknowledge open theism (LFOT) as opposed to the much more plausible open future open theism (OFOT). (On pp. 11–12 of my book, I argue that OFOT is decisively better than LFOT.) So, open theists ought to challenge RC and not let that assumption stand. Moreover, as Stewart himself notes, it is counterintuitive to deny that belief pairs with expectation: If God expects event E to occur, then it is natural to say that God believes E will occur, or at least that E will probably occur. In short, taking the “God doesn’t have any beliefs about how future contingents turn out” line is biting a bullet. If open theists don’t have to bite that bullet, then, all other things equal, they shouldn’t. By arguing against RC, open theists not only show that the first half of Stewart’s dilemma is false, but they also render the second half moot. There simply is no problem with God having beliefs about how future contingents turn out and things occasionally turning out otherwise than God believes they will.
To argue against RC it suffices to point that it’s a conceptual non sequitur. That p is the case does not entail that it has always been the case that p will be the case, for it may well be that p was not always previously true but later became true. There is no inconsistency in that supposition. So, RC is a non sequitur. If understood as an a priori necessary claim, it is false.
To put it another way, suppose that open futurism is true. Open futurism is the view that the future is objectively open-ended and that, consequently, there is no such thing as a unique actual future. What we call “the future” is a branching array of possible futures. Now, if open futurism is true and the truth of p is a future contingent, then p is not now true even if p subsequently becomes true. The reason why p is not now true is because, since its truth is a future contingent, there remain objectively possible futures in which p is never true. If we suppose that p is true now, before the contingency regarding p‘s truth is resolved, then we countenance a logical contradiction, one in which p both is true and yet never is true. To avoid this contradiction, we either have to reject RC or reject open futurism. To simply reject open futurism at this point would beg the question against open theism. So, we would need a reason for affirming RC, and it can’t be that RC is a conceptual truth, because it’s not.
Are there reasons that can be given for affirming RC? Yes, but they are metaphysical and/or pragmatic reasons, not conceptual ones.
Metaphysically, RC follows if either causal determinism or a linear block ontology of time are true. Open theists (especially those of the OFOT variety), of course, reject both of those metaphysical hypotheses, and they have lots of good company in doing so. Other metaphysical reasons that might be offered for RC include the claims that (a) possible worlds necessarily contain a complete, determinate history and there is a unique possible world which is the actual world, and (b) that tenseless truth does not vary over time just like it doesn’t vary over space. As for (a), this is simply question-begging against OFOT. Open theists can coherently reject the idea that possible worlds necessarily contain a complete, determinate history. They can argue, instead, that possible worlds are intrinsically tensed and therefore change from moment-to-moment. In other words, there is always a unique actual world, but which world is actual at any given time is non-constant. As for (b), this has some intuitive appeal, for we generally think that truth doesn’t vary over space—if p is true here, then it’s also true there. But why should we think that truth relates to time just like it does to space? If anything, time is more metaphysically fundamental than space. So, we have no compelling reason to think that the alethic properties of space transfer over to time.
On the pragmatic front, there is really only one argument that’s ever given on behalf of RC. It’s an argument from how people often talk about successful bets. Sam places a bet that the horse Eclipse will win the Derby, Eclipse goes on to win, and then Sam brags to his friends, “See, I was right!” Supposedly, in saying “I was right” Sam is endorsing the idea that the proposition <Eclipse will win the Derby> was true when the bet was made (and at all times prior to that). But this is an incredibly weak argument. In the first place, deriving a general metaphysical conclusion (RC) from the mere fact that people happen to talk a certain way is very problematic. If this were a good way of arguing, then we ought to be able to prove geocentrism from the mere fact that people typically talk about “sunrises” and “sunsets” as if the sun were revolving around the Earth. In the second place, if Sam were endorsing the idea that <Eclipse will win the Derby> was true when he placed the bet, then there would be no way to resolve the bet, for—as we have already seen—RC is a conceptual non sequitur. From the mere fact that Eclipse wins, we cannot conclude—unless we fall back on one of the already discussed metaphysical arguments for RC—that it was true that Eclipse was going to win. For all we can tell empirically, that proposition may have only become true just before Eclipse actually won, in which case Sam would have actually lost had he been betting that the proposition was true when he placed the bet. Finally, we can give a perfectly adequate explanation of our betting practices without supposing that anyone ever bets on the current (or prior) truth of a proposition about future betting outcomes. A bet, after all, is nothing more than and outcome-based payoff scheme. If the outcome I bet on results, I win. If it doesn’t, I lose. That’s all. Thus, whether <Eclipse will win the Derby> was true previously or not is strictly irrelevant. All that matters for purposes of adjudicating the bet is whether Eclipse does in fact win or not. When Sam boasts, “I was right” he’s most likely not saying that a certain proposition was previously true, rather, he’s merely saying something like “I was right to have made that bet.”
Summing up, Stewart’s dilemma only gets off the ground if we assume RC. Not only, however, is RC provably false as a conceptual claim, but there is no good non-question-begging metaphysical or pragmatic argument for RC. Open theists are, thus, well with their epistemic rights in rejecting RC. Consequently, there is nothing problematic in an open theist’s affirming (1) that God has expectations/beliefs about whether certain future contingencies occur or not, (2) that things sometimes fail to pan out as God expects them too (cf. Isaiah 5:4 and Jeremiah 3:7), and (3) that God never had any false beliefs. From an open futurist perspective, if E is a future contingent event with probability k of occurring, then what’s true beforehand is neither <E will occur> nor <E will not occur>—for it remains an open question whether E occurs—but rather propositions like <E might-and-might-not occur> and <E will-with-probability-k occur>. E’s eventual nonoccurrence only shows that when the contingency resolves, the latter two propositions become false (after having been true) and the will not proposition becomes true (after having been either false or neither-true-nor-false). Since God’s knowledge perfectly tracks truth, as contingencies resolve the content of God’s beliefs changes accordingly. At no time, then, does God ever have false beliefs.