Neo-Molinism and the Infinite Intelligence of God

Classical Molinism holds that, since God is omniscient and knows all truths, he knows not only what every agent will do in the future, but also what every agent would have done in every other “possible world.” In this essay I argue that classical Molinism overlooked a whole category of truths that an omniscient God would know: namely, the category of truths about what free agents might and might not do. The open view of the future affirms the truth of classical Molinism but also incorporates this previously neglected domain of truths an omniscient God must know. The open view can thus be described as a form of Neo-Molinism. Neo-Molinism, I argue, offers God the same providential advantage as classical Molinism, but it does so while avoiding some of the logical pitfalls of classical Molinism.

One of the most frequent and most emotionally charged objections to the open view of the future is that, if God doesn’t know with certainty all that shall come to pass, he cannot ensure that his purposes for a individual’s life or for world history will be accomplished. As Bruce Ware puts it, the open view of God posits a “limited, passive, hand-wringing God” who can do little more than hope for the best. (1) In his view, what is “lost in open theism is the Christian’s confidence in God.” And he continues:

When we are told that God…can only guess what much of the future will bring…[and] constantly sees his beliefs about the future proved wrong by what in fact transpires….Can a believer know that God will triumph in the future just as he has promised he will? (2)

Both on a scholarly and popular level, this concern seems to fuel much of the passion with which some evangelicals have recently come against the open view. It permeates Bruce Ware’s book, God’s Lesser Glory and is found throughout other recent polemics against the open view.

In this paper I shall demonstrate that this concern is unfounded. Indeed, I shall show that the concern itself is rooted in a thoroughly anthropomorphic view of God, for it is premised on the assumption that God’s intelligence is limited, just as human intelligence is. My claim is that if we consistently think through the implications of infinite intelligence, we will see that a being who possesses this attribute loses no providential control because he knows a partly open, partly settled future rather than one that is exhaustively settled.

Of course, a being who exhaustively controlled the future would possess more providential control of what comes to pass than a being who did not. Any view that admits that agents possess libertarian freedom has to sacrifice exhaustive divine control. But in this case we are debating predestination, not foreknowledge. Whatever could be argued against open theists on this account could also be argued against classical Arminians. The issue I’m concerned with in this essay is foreknowledge. And the question I’m asking is, Does God lose any providential advantage by knowing some of the future as composed of possibilities rather than knowing it as exhaustively certain? My argument shall be that, if we grant that God is infinitely intelligent, the only conclusion we can come to is that he does not.

To argue my case, I shall first compare the open view with that form of Arminianism that has ascribed to God the most providential control: namely, Molinism (§1). Whereas advocates of simple foreknowledge affirm that God knows all that shall come to pass, including what free agents will do, the Molinist argues that, to a significant degree, God chooses what shall come to pass but without sacrificing libertarian freedom. This view thus ascribes to God far more providential control than simple foreknowledge. I shall argue that the open view can be thought through in such a way that it ascribes to God roughly the same level of providential control as Molinism. Indeed, I shall argue that the open view is so close to Molinism that it could (and perhaps should) accurately be labeled “neo-Molinism.” For the purposes of this paper, I shall use “open theism” and “neo-Molinist” interchangeably.

The neo-Molinist perspective differs from classical Molinism in only one respect—though it is a very significant respect. Neo-Molinism modifies the standard Molinist understanding of God’s middle knowledge by maintaining that it includes “might-counterfactuals” (viz. knowledge of what agents might or might not do in given situations) as well as “would-counterfactuals” (viz. knowledge of what agents would do given situations). This modification, I shall argue, allows open theism to avoid four major objections that have been raised against classical Molinism.

Second, I shall apply the neo-Molinist account of God’s knowledge to the issue of God’s infinite intelligence (§2). I shall argue that for a being who knows all “might-counterfactuals” as well as all “would-counterfactuals,” there is no loss of providential control. God can anticipate from all eternity “might-counterfactuals” as perfectly as he does “would-counterfactuals”—indeed, as perfectly as he does future certainties (what shall occur).

Finally, I shall apply this argument explicitly to the different ways Molinists and neo-Molinists conceive of God working in the world (§3). I shall here argue that all but one of the reasons one could give for arguing that Molinism offers God a providential advantage over the open view of God is misguided. I shall thus argue that the neo-Molinist construal of the open view of God presents a model of God’s relationship to the world that has most of the providential advantages of classical Molinism, but without the attendant difficulties. At the very least, the providence control ascribed to God by open theists is far greater than that ascribed by simple foreknowledge Arminians. And thus, if successful, I will have shown that the neo-Molinist view presents an understanding of God that is as far from a “limited, passive, hand-wringing God” as any Arminian view could be.

The Open View as Neo-Molinism – God’s “Middle Knowledge”
In the sixteenth century Jacob Molina proposed that between God’s knowledge of all logical possibilities, on the one hand, and God’s factual knowledge of what shall come to pass, on the other, we should posit another category of knowledge: God’s knowledge of what would come to pass in any conceivable set of circumstances. This he called God’s “middle knowledge,” for it is “acquired” in the logical moment between God’s knowledge of logical possibilities and God’s knowledge of facts (what shall come to pass).

There are three primary arguments for God’s middle knowledge. First, Scripture depicts agents as free and morally responsible on the one hand, while depicting God as sovereignly in control of the world on the other. The best way to render these two facts compatible, argue middle knowledge theorists, is to assume that God possesses counterfactual knowledge of creaturely free acts. In their view, God knows what every agent would do in every possible world and then creates that world which best achieves his creational objectives.

Second, Scripture gives a number of examples of God claiming to know what various people would do in different situations (e.g. 1 Sam. 23:6–10; Jer. 38:17–18). This further suggests that God possesses counterfactual knowledge.

And third, since God is omniscient, he must by definition know from all eternity the truth-value of all propositions. Hence he must eternally know the truth-value of all counterfactual propositions, including counterfactuals of creaturely freedom.

This position, it is argued, has an advantage over Calvinism, for in affirming libertarian freedom it avoids all the paradoxes that attend to compatibilism. But it also has an advantage over the simple foreknowledge position, for God doesn’t simply know what is going to take place: he chooses what is going to take place on the basis of his middle knowledge. To be sure, as noted above, God does not have exhaustive control over what transpires, as in Calvinism, but he has much more control over what transpires than if he simply “found out,” as it were, how creatures were going to choose to freely act in the future. The view doesn’t allow God to always get his way—no view that affirms libertarian freedom can. But it does allow God to choose that “possible world” in which his creational objectives are best achieved, given that he chose to create agents who possess libertarian freedom.

Difficulties with Classical Molinism
Despite its advantages, there are a number of problems with the classical Molinist position, four of which may be mentioned presently.

First, many have argued that it is difficult to render the Molinist account of middle knowledge philosophically plausible. In this view, every possible decision any possible free agent might ever make in any possible world is an eternal fact. From all eternity, the facticity not only of the future world, but of all possible worlds, exists. It is an unalterable fact that agent x shall do y in situation z, but would do a, b or c in situations u, v and w.

Rendering this eternal facticity intelligible is no easy matter, however. It is a contingent facticity, but it is not clear what brings it about. It cannot be brought about by God’s will, for that would constitute determinism, something Molinists want to avoid. But neither does it seem that this eternal facticity can be brought about by created agents, for created agents are not eternal, and Molinists (and most others) generally deny retroactive causation. What is more, unactualized would-counterfactuals cannot be said to be brought about by created agents for the simple reason that they are facts that agents don’t choose.

We are left then with the unappealing alternative of denying that anything grounds the eternal facticity of counterfactuals of creaturely freedom. From all eternity the fact of what every conceivable free agent would do in every possible situation was simply there—without anything bringing it about. Though it is supposed to be contingent, it exists as an eternal metaphysical surd. This position, many argue, is at best counter-intuitive. It could also be charged with being dualistic, for the facticity of the world, and of every possible world, is construed as an uncreated definite reality that eternally co-exists alongside of God. For all intents and purposes, it occupies the status of an essential attribute of God. Yet it is supposed to be contingent.

Along the same lines, many have argued that it’s not clear that the middle knowledge position is consistent in its view of libertarian freedom, and this is the second potential problem with this position. On most Arminian accounts, an agent can be said to possess morally responsible libertarian freedom if it lies within their power to do otherwise, given the exact same set of antecedent conditions. But how then can we meaningfully say that an agent could have done otherwise if all they shall ever do, and all they would have ever done in any possible world, is a groundless unalterable fact an eternity before they even exist?

Stated otherwise, how can an agent be said to be self-determining when they don’t ground the eternal facticity of what they shall do, and ever would have done in different circumstances? How can we meaningfully say that it is possible for an agent to act otherwise when it has from all eternity been impossible for them to act otherwise in a given possible world? The facticity of all their acts eternally precedes them, brought about by nothing, as we have seen. If agents possess self-determining or libertarian freedom, it seems they must be the ones who resolve possibilities into facts. And to accomplish this, it seems they must exist.

Third, in at least one important respect, Molinism seems to have a more difficult task in confronting the problem of evil than either the simple foreknowledge view or the open view of God. In Molinism, it seems God could refrain from creating people he is certain will use their freedom to damn themselves to hell. If God always strives for maximal goodness, then we must conclude either that it is better for damned individuals to exist in hell than not to exist at all, or that the cosmos as a whole is better for including damned individuals. For a variety of reasons, neither is an attractive alternative.

Yet a fourth possible criticism may be raised, though it has not to date been part of the standard critique of Molinism. There is certainly some scriptural warrant for holding that God knows would-counterfactuals of creaturely freedom, as Molinists have pointed out. But there is, I would argue, an even greater wealth of Scripture which warrants the conclusion that God knows might-counterfactuals of creaturely freedom as well.

For example, Scripture frequently depicts God as speaking and thinking of the future in terms of what might or might not happen (e.g. Exod. 4:1–9; Exod. 13:17; Ezek. 12:3; Jer. 26:3; Matt. 26:39 ). (3) Since God is the only one in a position to know with certainty the nature of the future, this should serve as a reliable indicator that some of the future consists of possible events that might or might not take place. Similarly, God frequently changes his mind, even after he’s explicitly declared his intentions. This too suggests that some of his particular intentions are “maybes” until they are acted upon (e.g. Exod. 32:10–14; Deut. 9:13–29; 1 Chron. 21:15; Jer. 18:7–10; 26:2–3, 19; Jonah 3:10, cf. 4:2; Joel 2:12–13).

Along similar lines, God often “tests” people in order to find out what they will decide to do, suggesting that their future actions are “maybes” until he tests them (Gen. 22:12; Exod. 16:4; Deut. 8:2; 13:1–3; Judges 2:22; 2 Chron. 32:31; ). Scripture also frequently depicts God as experiencing regret (Gen. 6:6; 1 Sam. 15:10, 35), disappointment, frustration, and unexpected outcomes (Exod. 4:10–15; Jer. 3: 6–7, 19–20; Ezek. 22:29–31; Isa. 5:1–5), again suggesting that the future is to this extent composed of possibilities rather than certainties. Its more difficult to conceive of God experiencing such things if the future is exhaustively settled in his mind than if it is in part composed of possibilities. (4)

Would-Counterfactuals and Might-Counterfactuals (5)
While some open theists follow Aristotle and others in maintaining that future tensed contingent statements are neither true or false, I personally suspect that this is a mistake. I am in agreement with the Molinist view that propositions expressing counterfactuals of creaturely freedom have an eternal truth-value, and thus that the omniscient God eternally knows this truth-value. However, I see no reason to embrace the classical Molinist assumption that counterfactuals are exclusively about what agents would or would not do. For these do not exhaust the logical possibilities of counterfactual propositions, and thus do not exhaust the content of the counterfactual propositions an omniscient being would know.

On a counterfactual square of oppositions, the logical antithesis of the statement, “agent x would do y in situation z” is not the statement, “agent x would not do y in situation z.” This is a contrary proposition, not a contradictory proposition. The logical antithesis of “agent x would do y in situation z” is rather the statement, “agent x might not do y in situation z.” This latter statement also has an eternal truth-value and hence must be known by an omniscient being.

The point is that would-counterfactuals do not exhaust the category of counterfactuals: there are also might-counterfactuals. Propositions about both categories of counterfactuals have an eternal truth-value that must be known by God. Hence I see no reason to restrict God’s middle knowledge to knowledge of would-counterfactuals, or, what comes to the same thing, to conclude that all might-counterfactuals are false.

If we include might-counterfactuals in God’s middle knowledge, we arrive at the following neo-Molinist position. Between God’s pre-creational knowledge of all logical possibilities and God’s pre-creational factual knowledge of what will come to pass is God’s “middle knowledge” of what free agents might or might not do in certain situations as well as of what free agents would do in other situations. Each might and would counterfactual is indexically referenced to a possible world. If it is true that agent x might or might not do y in situation z, it is false that agent x would do y in situation z, and vice versa. On the basis of this knowledge, God chose to have actualized the possible world that best suited his creational objectives. If the world God creates is a world in which some might-counterfactuals are true (as I argue is in fact the case) then this world must by definition contain open possibilities. It is, in short, a world in which there are some things free agents might and might not do.

To speak more precisely, this world would actually be a delimited set of possible worlds, any one of which might be actualized, depending on the choices free agents make. In such a world, God’s knowledge of what will be and what would be would not exhaust what God knows: God would also know what might or might not be. In short, the future, in such a world, would be partly open.

To the best of my knowledge, Molinism never developed this possibility. It was assumed that would-counterfactuals exhausted the category of God’s counterfactual knowledge. By logical necessity, all propositions expressing might-counterfactuals were assumed to be false. With the bulk of the classical traditional, Molinists assumed that omniscience logically entailed exhaustively definite foreknowledge. To this classical assumption they simply added the claim that God has exhaustively definite knowledge of what would come to pass in all other possible worlds.

Our discussion above reveals the arbitrariness of this assumption. Unless might-counterfactuals are self-contradictory, they cannot be false by definition, and thus God’s omniscience cannot by definition consists only of facts and would-counterfactuals. As paradoxical as it sounds, the standard charge that the open view of the future diminishes God’s omniscient is tantamount to saying that God’s omniscience is somehow diminished by virtue of the fact that in this view he knows some might-counterfactuals to be true. But unless it can be shown that might-counterfactuals are by definition false, it cannot be argued that God’s knowledge by definition rules them out. Hence, so far as I can see, we have no reason to conclude that the sovereign God could not create a world in which some might-counterfactuals were true if he wanted to.

The Open View as Neo-Molinism
The contention of open theists is that we have good reasons to believe that God decided to create a world in which some might-counterfactuals were true. If accepted, we arrive at a neo-Molinist perspective that has all the explanatory power of classical Molinism, ascribes to God most of the providential advantages of classical Molinism, but which avoids the four common objections that have plagued classical Molinism.

Before considering how the neo-Molinist view is able to ascribe most of the providential advantages of classical Molinism, let us briefly consider how it avoids the objections to classical Molinism.

First, if we accept that some might-counterfactuals are eternally true, we no longer have the problem of an ungrounded eternal facticity to possible worlds that include libertarian freedom, and there is no longer any problem accounting for libertarian freedom itself. Determinate aspects of any possible world are grounded in God’s will. For any possible world God might create, there may be things that he decides would certainly come to pass if he were to create them, and thus things that shall come to pass in the world God decides to create. Insofar as any possible world includes might-counterfactuals, there simply is no eternal facticity that needs to be accounted for. There are only eternal possibilities of what these agents might or might not do in various situations.

Second, because it allows that propositions expressing might-counterfactuals have a truth-value, the neo-Molinist account has no problem accounting for libertarian freedom. Agents themselves transition possibilities into facts. Hence, insofar as they are free, their actions are brought about by their own free agency. We thus do not have to wrestle with an ungrounded eternal facticity that precedes an agent’s own free decisions.

Does this entail that would-counterfactuals cannot also be true of creatures who possess libertarian freedom? Not at all. It simply means that would-counterfactuals cannot be the only truth that applies to creatures who possess libertarian freedom. Insofar as would-counterfactuals apply to future free agents, they do so because the actions of these agents flow either from the determinate character God has given them (habitus infusus), in which case they are not morally responsible for them, or from the determinate character they will freely acquire (habitus acquisitas) if they choose to pursue a certain possible course of action, in which case they are morally responsible for them.

In either case the would-counterfactuals are not ungrounded, as in classical Molinism. Rather, they are indexically referenced to a possible world in which God and/or the agent herself brings them about. From all eternity God knows that, if he chooses to create free agent x, she will have the basic characteristics of a, b and c (habitus infusus). And from all eternity God knows that if agent x freely follows a certain possible life-trajectory, she will become the kind of person who would do x in situation z (habitus acquisitas). In other worlds, the would-counterfactuals for which agent x is morally responsible are contingent upon the might-counterfactuals for which she is morally responsible. Where the would-counterfactuals are true of an agent, the might-counterfactuals are false, and vice versa.

Third, if we accept that the eternal destiny of people is a “might-counterfactual,” we do not have to account for why God creates people he knows will end up in hell. God creates beings he knows might put themselves in hell, but their fate is not settled at the moment of their creation. From the neo-Molinist perspective, it is only this “might” that allows for the possibility of a loving, morally responsible choice to accept God’s grace. This also explains why throughout Scripture God is depicted as genuinely hoping for and striving for the salvation of all people, and why he is deeply grieved when certain people refuse to yield to his saving influence.

Finally, if we accept that some might-counterfactuals are true, we can now easily account for those passages of Scripture that either specifically attribute to God or imply that he faces a future partly composed of possibilities, not certainties. In this view, God thinks and speaks of the future in terms of what may or may not happen simply because the future is partly composed of events that may or may not happen. So too, God truly changes his intentions because his intentions were conditioned upon what might or might not come to pass. He tests people “to know” their heart because their character and actions are genuinely unsettled relative to the issue he is testing them on. And he genuinely regrets things, experiences disappointment, frustration or unexpected outcomes because the future was not exhaustively settled an eternity beforehand.

Neo-Molinism and the Infinite Intelligence of God
The neo-Molinist model of God’s knowledge and relationship to the world clearly gives God a much greater providential advantage than does the simple foreknowledge model. Instead of simply knowing what shall come to pass, God to a large extent chooses what will come to pass on the basis of his middle knowledge. But does it give God as much of an advantage as the classical Molinist account? If God chooses to create a class of possible worlds, any one of which might come to pass, is he not significantly less in control of the world than if he chooses to create one possible world in which everything that shall come to pass is certain to him ahead of time?

Divine Control and Libertarian Freedom
It is important to first get clear on what kind of control we are talking about in asking this question. Both classical Molinism and neo-Molinism ascribe libertarian freedom to agents. Hence we are not talking about control as “power to exhaustively control free agents.” Rather, the kind of providential control we are talking about is God’s “power to integrate the decisions of agents into his sovereign plan for the world.” We are talking about God’s ability to perfectly anticipate and plan an ideal response for every decision every free agent might ever make. We are, in short, talking about God’s ability to ensure that his objectives for world history will be maximally realized, given that agents possess libertarian freedom.

Classical Molinism clearly ascribes to God this sort of control. Though God knows that there is no feasible (as opposed to merely logical) possible world in which free agents choose exactly as he would wish, he knows and chooses to create that one possible world in which he knows he can most effectively respond to the decisions of agents to achieve his creational objectives. By knowing exactly what agents would choose in every possible circumstance, he decides what agents will choose in the circumstance in which he places them, and thus he can perfectly anticipate how he will respond to their choices as a means of maximally achieving his will.

Can the neo-Molinist model ascribe to God this level of providential control? Because she affirms that God creates a world that includes might-counterfactuals, the neo-Molinist clearly cannot affirm that God knows in every instance what an agent will choose in every possible circumstance. But does this entail that God cannot perfectly anticipate how he will use each agent’s free decisions as a means of maximally achieving his will, as in the classical Molinist account? Does it entail that God “can only guess what much of the future will bring” or that he is a “limited, passive, hand-wringing God” who can do little more than hope for the best? The answer, I submit, is that it does not.

The Infinite Intelligence of God
We must consider the implications of ascribing to God infinite intelligence, something all orthodox Christians assume to be true. Simply put, all orthodox Christians believe there is no limit to God’s intelligence.

Now, we humans are more confident when we only have to prepare for one forthcoming event than we are when we have to prepare for a multitude of possible forthcoming events because we possess finite intelligence. Our intelligence gets spread thin the greater the number of possibilities we must confront. Hence our ability and confidence to respond ideally to a possible future event is lessened in proportion to the number of alternative possibilities we have to consider alongside this event.

If God were limited in this fashion, the neo-Molinist model would certainly provide God with much less providential advantage than the classical Molinist model. Indeed, given the innumerable might-counterfactuals that characterize our world, we might be inclined to conclude that such a God could indeed “only guess what much of the future will bring” and that he must be a worrisome, “limited…hand-wringing God” who can only hope for the best. But if God is not limited in this fashion, this inference is altogether unfounded.

If God is not limited in intelligence, he doesn’t have to divide up his intelligence between possibilities he is facing the way humans do. If his intelligence is infinite, it doesn’t get “spent” as he considers various possibilities. Indeed, he can attend to each of any number of possibilities as though each one was the only possibility he had to consider. All of his infinite intelligence is fixed on each and every possibility equally. Hence, God can perfectly anticipate and be as prepared for any possible future event as he is for any certain future event. In other words, for a God of infinite intelligence, there is no distinction that can be made between possibilities and certainties in terms of providential advantage. Any view that denies this thereby concedes that it doesn’t presuppose that God is infinitely intelligent.

In the neo-Molinist model, then, we may say that from all eternity God planned his perfectly wise response to what he knows shall occur as well as to what he knows may occur. And because of his infinite intelligence, his response to the latter is not one iota less perfect than his response to the former. Whatever comes to pass, the neo-Molinist may say with confidence that “from all eternity, God was preparing for just this event,” whether the event was certain to occur or was simply a possibility. Indeed, with classical Molinists, neo-Molinist may conceive of God sovereignly choosing the world he created on the basis of his middle knowledge of creaturely free acts as well as his knowledge of how he could responsively weave these acts into his sovereign plan. The only difference is that the neo-Molinist does not exclude might-counterfactuals from God’s middle knowledge. But he is as perfectly prepared for the might-counterfactuals as he is for the would-counterfactuals as well as the future certainties. For an infinitely intelligent being, there is no difference in terms of preparedness in any of these categories of events.

Clearly, the criticism that God must worry and can only guess at what the future will bring if the future includes might-counterfactuals is predicated on a limited, anthropomorphic view of God.

Classical Molinism, Neo-Molinism, and Maximal Providential Control
It should now be clear why the neo-Molinist model can claim to offer God much greater providential control than does the simple foreknowledge view. Not only does it allow God a greater capacity to choose what comes to pass on the basis of his middle knowledge, it empowers God to perfectly affect what comes to pass as it is unfolding on the basis of his infinite intelligence. But does the neo-Molinist account grant God anything like the level of providential control as the classical Molinist account?

Prima facia, it may seem that it can not. Since in the neo-Molinist model the world God chooses to create is actually a set of possible worlds, God can’t ensure that the world that comes to pass will best suit his providential purposes. Agents may choose courses of action from within the options God gives them in each situation which are non-ideal, not just relative to God’s ideal, but relative to other choices they could have made in the exact same situation. Yes, in neo-Molinism God can respond perfectly to whatever comes to pass. But there will be always be possibilities that agents could have actualized, rather than the ones they did in fact actualize, which would have better suited God’s providential purposes.

It thus seems that a God who knew ahead of time what agents would do in every conceivable situation (any “possible world”) would have a distinct advantage over a God who knew some of the future as a “maybe.” It seems the God of classical Molinism can select which combination of all conceivable situations to actualize on the basis of how they fit into his providential plan, something that cannot be said for the God of neo-Molinism.

Though the argument may initially sound compelling, I shall offer several considerations that suggest that it is largely misguided.

Would-Counterfactuals and Might-Counterfactuals
It is of course true that in neo-Molinism there will be possibilities that agents could have actualized, rather than the ones they in fact actualized, that would have better suited God’s providential purpose. But this much is true of classical Molinism as well. Both models affirm libertarian freedom. Hence both models affirm that it is the agent and the agent alone who reduces all possible actions down to one in a given conceivable situation with their decision. Both models therefore contend that God must respond to non-ideal decisions in working to bring about the maximal achievement of his will.

Both models therefore further affirm that the world God creates could have, and should have, been better than it is, for agents could have and should have made different decisions. But both models also contend that the world that comes to pass is the best possible, given the non-ideal decisions agents make. God has as much providential control as is logically possible in a world populated with free agents, for he eternally knows as certain, or as though certain, not only all that shall come to pass, but all that would or might have come to pass in different circumstances. And he created this world from among all possible worlds with this knowledge in mind.

The points on which the two models differ do not affect their views on the scope of God’s providential control. The classical Molinist maintains that God’s middle knowledge knows only “would-counterfactuals,” while the neo-Molinist argues that his middle knowledge must also include “might-counterfactuals.” For the neo-Molinist, God’s knowledge of “would-counterfactuals” that were once “might-counterfactuals” is acquired as agents themselves transition “maybes” into facts. But note, because of God’s infinite intelligence, God’s eternal knowledge of “would- counterfactuals” does not increase his providential control over “might-counterfactuals.”

In the neo-Molinist view, God eternally knows and perfectly prepares for all counterfactuals, whether “might-counterfactuals” or “would-counterfactuals,” as though they were certainly going to come to pass. And because God is infinitely intelligent, there is no difference between these two classes of counterfactuals in terms of his ability to perfectly prepare for them. Hence, this difference in the two models does not grant the classical Molinist model greater providential advantage.

Unilaterally Actualizing a Possible World
Still, the classical Molinist maintains that God is able to unilaterally reduce all feasible possible worlds down to one actual world, while the neo-Molinist can only maintain that, insofar as agents possess libertarian freedom, God can only reduce all possible worlds down to a set of possible worlds. Does this not give God greater assurance of maximally achieving his creational objectives in the world? I believe that several further considerations suggest that the answer is no.

First, we must again remember that in both models it is ultimately the agent alone who is responsible for transitioning a set of possible actions into one actual action in a given situation (assuming the agent possesses libertarian freedom in the situation). In both models the actual decision the agent makes may end up being non-ideal compared to other decisions the agent could have, and should have, made. And in both models, God is nevertheless in control, for he created this particular world with an eternal view to how he would ideally respond to this decision and incorporate it into his sovereign plan.

The fact that the neo-Molinist holds that it was genuinely possible for the agent to choose differently at the time he made the decision to act as he did, while the Molinist holds that the agent ruled out these possibilities in eternity, before the agent existed, is irrelevant. For both models grant that in this particular situation the agent didn’t choose differently. And this actual decision reveals the fact that in this situation the agent wouldn’t choose differently. The fact that the neo-Molinist maintains that this “wouldn’t” was (perhaps) preceded by a “might” or “might not” doesn’t put God at any providential disadvantage, for as we have seen, there is no distinction for an infinitely intelligent being between “maybes” and “certainties” in terms of his providential control. The neo-Molinist can therefore affirm, just as strongly as the Molinist, that God created this world, as opposed to any other possible world, with an eternal view toward how he would respond to the way (it turns out) an agent would act in a particular situation.

Open Sets and Possible World Sets
But, someone might yet argue, in the Molinist account God chooses to actualize any particular situation with the certainty that an agent would choose as it turns out they do, whereas in neo-Molinism God’s knowledge is only of what the agent might or might not choose in a particular situation. This seems to empower God to select all the best situations, together with the decisions agents make in those situations, which God knows he can most effectively integrate into his providential plan. Because he knows some of the future as composed of might-counterfactuals, however, the God of neo-Molinism isn’t empowered in this fashion. Though he restricts possible worlds down to a set of possible worlds, he’s nevertheless stuck with this set. He can’t choose which from among this set to actualize. And this again seems to give classical Molinism a providential advantage.

The problem with this argument is that, in point of fact, Molinism does not empower God to select all the best situations in creating the actual world. The root of this mistake is confusing logically possible worlds with feasible worlds. Logically, God can piece together any logically possible combination of behaviors he chooses. But the class of possible worlds God must consider as he prepares to create the world is the class which contain libertarian free agents who in fact act in particular ways in particular situations. These alone are the feasible possible worlds out of which God can create the actual world.

Hence, God can’t simply piece together a world he wants to actualize by selecting all the best situations and choices agents make in various situations. In classical Molinism, God must rather consider all the counterfactuals that actually characterize a possible world as a whole. The world which includes agent x making a great choice in situation z is also the world in which agent x makes a very poor choice in situation q. Though there is no logical necessity that z and q go together with agent x, in the feasible possible worlds in which agent x exercises libertarian freedom they may. Hence God can’t select z without also selecting q. God must consider the total combination of choices agents make in each feasible possible world as he contemplates which world to actualize. It is this move which gives Molinism a significant advantage over Calvinism in coming to terms with the problem of evil.

It should be clear from the above that we cannot compare Molinism to neo-Molinism on a situation-by-situation basis, as it were. We have to take a total world perspective in comparing the two, and on this score it seems neither view offers a providential advantage over the other. In both models God is “stuck with” working with sets of decisions, some of which are non-ideal relative to other decisions the agent could have and should have made in each particular situation.

The difference between the two models is that classical Molinism holds that God interacted with each feasible set of decisions in eternity, as it were, whereas neo-Molinism holds that God also interacts with a sub-set of possible decisions agents face in each situation as the situation unfolds. The classical Molinist conceives of the feasible world sets being resolved in eternity, and then actualized, while the neo-Molinist goes further and conceives of a sub-class of these feasible sets being resolved in time, as they are actualized. But in both models God is “stuck with” non-ideal sets and, in any given situation, with actualized decisions that are part of these sets that are non-ideal. Yet, since God is as perfectly able to prepare from eternity an ideal response to possible decisions made in time just as perfectly as he is a decision purportedly made in eternity, there is no providential advantage to be had by thinking he does one rather than the other.

Is There a Better Feasible World?
In the end, the disagreement between the two models amounts to a speculative disagreement about whether there existed a better feasible world that could have been actualized had God possessed knowledge of how agents would act in every conceivable situation. The Molinist might argue that, on the open theists’ account, there could be feasible worlds in which the total set of decisions free agents make is better than the ones they end up making in the actual world. At the very least, it seems that if God has to prepare for a variety of possible responses to a set of possible choices, the resources available to him as he prepares this response will be less than if he knew with certainty what exactly free agents will choose in the future. Yes, God’s intelligence is unlimited in the open view, and so he can eternally prepare a response to whatever comes to pass. But he has to do this for a large number of possible responses, and this seems to compromise the effectiveness of his response to any one of them. To be sure, God doesn’t have to spread his intelligence thin, but he does have to spread his resources thin in preparing a possible response. Hence, it seems the set of choices that God must work with in classical Molinism is less detrimental to God’s providential control than the set God must work with in the neo-Molinist account.

At present, I’m inclined to agree with this. This may be the one area where Molinism offers a providential advantage over neo-Molinism. If in fact it were logically possible for God to know what classical Molinism claims God knows, it may be that he would have this providential advantage. This concession would still warrant the conclusion that neo-Molinism ascribes to God far greater providential control than simple foreknowledge. But it would admit that neo-Molinism offers less providential control than the classical Molinist view.

Still, the issue isn’t quite as clear cut as it may first appear, which is why I concede only to being inclined to accepting this conclusion. Two closely related points are relevant.

First, its exceedingly difficult to argue about what non-actualized worlds were and were not feasible, for we have no clear way of distinguishing between merely logically possible worlds and feasible worlds. That is to say, we have no way of discerning whether a possible total set of counterfactuals of creaturely free actions were genuinely feasible, or merely logically possible. Hence, even in the classical Molinist framework, it’s not self-evident that there existed a feasible world in which God could more effectively utilize resources better than he does in the world on the neo-Molinist account. It is possible that the set of decisions God must work with in this world prevents more optimal lines of preparation. Of course a better set is conceivable. But how can we decide that it’s feasible?

Second, the ambiguity involved in arguing about non-actualized worlds is clearly revealed in the fact that this argument has been turned against classical Molinism. As noted earlier, if any non-actualized world seems feasible and seems better than the one that is instantiated, it is one that rules out people using their free will to damn themselves. This forces the question of why God didn’t refrain from creating people he knew on the basis of his middle knowledge would damn themselves to hell?

The only answer a Molinist can give is that God couldn’t do this without losing other important values which belong to the total world set in which these individuals are damned. This is a difficult, though not impossible, supposition to accept. But it at least points to the difficulty of arguing against someone’s position on the grounds that their view precludes a better feasible world.

The neo-Molinist has a parallel answer to why God didn’t create a world in which he only had to prepare a response for what was going to come to pass rather than a world that included a set of things that could come to pass. The answer is simply that he couldn’t do so without losing other important values which belong to this world. Unlike the Molinist response to this challenge, however, the neo-Molinist is able to specify what value would be lost if God were to rule out might-counterfactuals. It is the value of libertarian freedom. And it is precisely this answer which allows the neo-Molinist to avoid the difficulty of explaining why God creates individuals he knows will go to hell.

The classical Molinist of course insists that a feasible world need not include might-counterfactuals in order to include libertarian freedom, but this is just another way of saying that we disagree about what worlds are and are not feasible. To the neo-Molinist, the feasible world they hold up as an example of the greater providential control their view offers God is simply not feasible.

Hence, while I concede the possibility that a God who possessed would-counterfactual knowledge of all future agent decisions might have a providential advantage over a God who knew their future decisions as might-counterfactuals, there is no way a Molinist can demonstrate this is so, since they already concede that any feasible world confronts God with non-ideal sets of decisions. But, even more fundamentally, the neo-Molinist sees no reason to accept that such a world is genuinely feasible.

We have seen that Molinism and neo-Molinism both share the conviction that God possesses middle knowledge and that this grants God maximal providential control over a world inhabited by agents possessing libertarian freedom. In both models, God has as much control as is logically possible, given his decision to create a world in which agents possess libertarian freedom.

The two views differ in that the neo-Molinist supplements the classical Molinist view that God knows the truth-value of all propositions expressing would-counterfactuals with the supposition that God also knows the truth-value of all might-counterfactuals. Because God is infinitely intelligent, however, his ability to anticipate might-counterfactuals is no less perfect than his ability to anticipate would-counterfactuals. Though this may entail that God cannot as effectively utilize the world’s resources in preparing for a future response, it may nevertheless be concluded that the neo-Molinist can in principle ascribe to God the same level of providential control as the classical Molinist.

Even if the neo-Molinist concedes that Molinism affords God this providential advantage—purchased, according to the neo-Molinist, by regarding an impossible world as feasible—it nevertheless stands that the neo-Molinist account affords God far more providential advantage than the simple foreknowledge view. Hence, it should be clear that the caricature of the God of open theism as a “limited, passive, hand wringing God” who can only “guess…what the future may bring” can no more be applied to neo-Molinism than it can to classical Molinism. In both models God is the sovereign creator and governor of this world who, because of his middle knowledge and infinite intelligence, is able to perfectly anticipate all that shall and all that may comes to pass.

From my vantage point, the distinct advantage of neo-Molinism over classical Molinism is that the neo-Molinist can affirm all this while avoiding the philosophical and biblical difficulties that attend to classical Molinism.

End Notes

(1) B. Ware, God’s Lesser Glory: The Diminished God of Open Theism (Wheaten, Ill.: Crossways Books, 2000), 216.

(2) Ibid., 20–21.

(3) For a more thorough presentation of the openness motif, see G. Boyd, God of the Possible (Grand Rapids, Mich.: Baker, 2000).

(4) While defenders of classical theism explain these depictions as anthropomorphic or phenomenological, the passages themselves read like straightforward narrative. Moreover, it is not at all clear what passages that depict (say) God changing his mind are anthropomorphic depictions of—if in fact God doesn’t change his mind. If there is a paradigm that allows us to accept the straightforward meaning of the text, then, all other things being equal, I submit it should be embraced, especially if it can affirm maximal divine control with undermining libertarian freedom. I shall argue that the neo-Molinist model of God and the world allows us to do just this.

(5) This essay was written in 2002. Now (in 2008) I would not refer to “would” and “might” counter-factuals, for I now see that this way of speaking presupposes there’s a settled future that is factual which these statements are “counter” to. I would thus rather speak of “would-factuals” and “might-possibilities.”

Tags: ,

Related Reading

Podcast: Is God Capable of Total Control?

Greg talks the consequences of creating autonomous persons.

How do you respond to Acts 2:23 and 4:28?

Question: Acts 2:23 and 4:28 tell us that wicked people crucified Jesus just as God predestined them to do. If this wicked act could be predestined, why couldn’t every other wicked act be predestined? Doesn’t this refute your theory that human acts can’t be free if they are either predestined or foreknown? Answer: In Acts…


Revelation 13:8 refers to “everyone whose names have not been written before the foundation of the world in the book of life.” How does that square with open theism?

Three possibilities exist in terms of reconciling Revelation 13:8 with open theism. 1) First, the “from the foundation of the world” clause can attach to either “everyone whose names have not been written” or to “the lamb that was slain.” For example, the TNIV translates this passage “All inhabitants of the earth will worship the…

What is the significance of Exodus 16:4?

The Lord commands the Israelites to gather only enough bread for one day while in the wilderness. “In that way,” the Lord says, “I will test them, whether they will follow my instruction or not.” Testing people to find out how they will resolve their character only makes sense if God is not certain of…


Lighten Up: I’m Not Worried Frank

Well, my dear friend Frank Viola has been spouting off again about how my “logic will be shredded, excoriated, and turned into confetti before a watching world” when we host our debate on Open Theism this fall. I’m not too worried though, since Frank studied the art of debate in the clinic featured in the…

Podcast: If Open Theism is True, Does it Make Sense to Pray For Intercession?

Greg talks about prayer and freedom. Specifically, if free will is so important, why would God override it in answer to prayer?