Show Summary Details

Page of

PRINTED FROM OXFORD HANDBOOKS ONLINE ( © Oxford University Press, 2018. All Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a title in Oxford Handbooks Online for personal use (for details see Privacy Policy and Legal Notice).

date: 13 December 2018


Abstract and Keywords

This essay describes the Molinist picture of providence, a picture that results from attempting to combine a strong, traditional notion of divine providence with a libertarian account of freedom. After showing how the Molinist picture seemingly allows one coherently to resolve the tensions that this combination engenders, the essay recalls the two general objections that have typically been offered against the Molinist view and briefly sketches the two alternative views of providence (open theism and Thomism) that typically flow from these objections. Two other general anti-Molinist arguments—less common but perhaps of greater potential interest—are then examined. Finally, directions in which the discussion of Molinism might most profitably proceed are suggested.

Keywords: providence, Molinist, Molinism, middle knowledge, Thomism, open theism, freedom, Christian, libertarian


Molinism is a much-debated philosophical picture of how a provident God can exercise sovereign control over his world while honoring the genuine freedom he has bestowed upon his creatures. Originally developed by the sixteenth-century Jesuit theologian Luis de Molina, Molinism was extremely controversial at its inception, and narrowly (or providentially?) escaped condemnation by the Catholic authorities in Rome. The controversy surrounding the picture of divine providence offered by Molina has remained ever since. Among analytic philosophers of religion, the debate concerning the coherence, the plausibility, and the value of the Molinist picture has remained hot (and at times scalding) for more than a generation.

In this essay, the Molinist account of providence is elaborated, the current state of the debate surrounding it evaluated, and the prospects for future developments suggested. We first consider the general notion of divine providence, the problems that notion seems to engender, and the manner in which Molinists suggest those problems can best be resolved.

Divine Providence: Control, Freedom, and the Appearance of Tension Between Them

Christians agree on at least the general lineaments of the notion of divine providence.1 Like their Jewish and Muslim brethren, Christians traditionally think of God not as a distant deistic demiurge but as a creator whose omniscience, omnipotence, and perfect love are manifested in his world, a world that develops according to his plan. As the product of an omnipotent creator, his world is truly and fully subject to his sovereignty; as the product of an all-knowing creator, it manifests his wisdom; and as the product of a perfectly good creator, it exhibits the order, beauty, and loving-kindness one would expect of a craftsman who cares for all that he has made.

The heart of this traditional Christian notion of providence is nicely stated at the start of Chapter 5 of the Westminster Confession:

God, the great Creator of all things, doth uphold, direct, dispose, and govern all creatures, actions, and things, from the greatest even to the least, by his most wise and holy providence, according to his infallible foreknowledge, and the free and immutable counsel of his own will, to the praise of the glory of his wisdom, power, justice, goodness, and mercy.

According to this tradition, then, God is not merely the one who brought all things into being. All aspects of his creation depend upon his sustaining presence, and all are subject to his control. As the last line of the passage just quoted makes clear, God’s plan for the world is not merely all-encompassing but fully in accord with the divine perfections of justice, goodness, and mercy. And that plan, the quotation clearly implies, is not one that God makes up as he goes along. Rather, he foreknows all that will occur, and thus is neither surprised by what happens nor forced to alter his intentions as events move in directions unanticipated by him.

This picture of providence has been dominant throughout the history of Christianity, though (as we shall see) it has been called into question in recent years. On reflection, it is not difficult to see why such a picture (rooted both in biblical and classical Greek sources) should seem at least prima facie attractive to thoughtful Christians. After all, the heart of Western monotheism is a conviction that God is the all-knowing, all-powerful, all-good creator of all that is. How, one might wonder, could such a deity who chooses to create not know precisely what he is doing and what he could have done? How could he have created a world that was not crafted to mirror his infinite power, wisdom, and love? Must we not assume that, in Cardinal Newman’s memorable phrase, “He knows what He is about”2, and can we not rest confident both in his overall plan and in our place therein?

As we have seen, part of this picture is that God governs all in accord with the free and immutable counsel of his will. This, of course, assumes that God has free will. And this assumption naturally leads those of a philosophical bent to reflect upon just what freedom might amount to, in God’s case or in ours. Many Christians have insisted that the notion of freedom now commonly referred to as libertarianism is the one that is most consonant with a Christian outlook.

Though a complete and precise characterization of libertarianism is notoriously difficult to provide, the basic idea is that external determination of a person’s action (especially causal determination by some factor not subject to the person’s causal control) is incompatible with that action’s being free. Libertarians insist that some of our actions are free; hence, they deny that those actions are determined, mediately or immediately, by events not under our causal control. My actions (or at least my free actions) are the ones that I initiate and control. Actions caused by something or someone external would be up to them, not up to me, and simply could not be actions for which I am responsible—for which I could properly be praised or blamed. And what goes for us goes for God as well; his freedom too must be understood in this libertarian way.

Again, this is a fairly coarse-grained depiction of the libertarian picture. Libertarians disagree about many things: the precise nature and degree of causal relations present within a free act; the precise connection between the beliefs and desires we have and the actions we perform; the frequency (or infrequency) of free acts; whether some genuine actions are unfree; and so on. But enough, I hope, has been said for readers to sense the type of position on freedom that libertarians champion.

I hope as well that enough has been said to see why there at least appears to be a tension between providence (understood in the manner delineated here) and freedom (interpreted in a libertarian manner). On the one hand, providence postulates complete divine foreknowledge of and control over all that occurs, including human actions. Libertarianism, on the other hand, seemingly insists that external determination of an action, even by a deity, is incompatible with that action’s being free. To endorse providence, then, it seems we need to deny that there are any free actions, at least as understood by libertarians; to endorse libertarian freedom, it seems we must deny (or at least limit) God’s providence. For Christians antecedently disposed to accept both a strong traditional notion of providence and the libertarian account of freedom, the prospect of rectifying matters by surrendering (or at least significantly modifying) one or the other of the two must surely be unsettling. Is there, they might naturally wonder, a way to hold onto both?

The Molinist Account of Providence

Molina and his followers contend that the two can indeed be upheld. Once we understand the full panoply of knowledge that God would possess, we can see that he can indeed exert full providential control of his world, and do so in large part via free creatures who enjoy full libertarian freedom.

Consider, Molinists say, God’s knowledge of his world. Some of the truths God knows—for example, mathematical truths such as two plus three equals five—are necessary truths, ones that could not have been other than they are and that are in no sense the result of any free decision on God’s part. Knowing such truths can be seen as part of God’s very nature; hence, Molina labeled this God’s natural knowledge.

On the other hand, many truths of which God is aware—for example, Libby will freely watch a baseball game on TV tomorrow—are neither necessary nor beyond God’s power to control. There are many possible worlds in which Libby does not freely watch a baseball game tomorrow, and there are many ways in which God could have prevented her doing so (most radically, by deciding never to create Libby at all). Contingent truths of this sort, then, are true only because God freely allowed them to be true. Such propositions, said Molina, are elements of God’s free knowledge. And this, obviously, is where foreknowledge of the contingent events that occur in our world belongs.

Molina insisted, though, that natural knowledge and free knowledge do not exhaust the knowledge that God possesses. Natural knowledge tells God what must be; free knowledge tells him what will (but need not) be. But are there not also truths about what would be? If God is truly all-knowing, then he must know not only what his free creatures will do in the situations in which they in fact will find themselves, but also what they would do in any situation in which they could be placed. Consider Libby and her decision to watch the baseball game. If (contrary to fact) Libby were to turn on the TV only to discover that the game in question was cancelled due to rain, is there not a fact of the matter as to what she would freely do (e.g., watch a movie) instead? And would not God, being omniscient, know this fact?

Molinists contend that there are such truths and that God would indeed know them. And, according to Molina, the existence of truths of this sort—counterfactuals of the form “If agent S were in situation C, S would freely perform action A”—force us to posit a third category of divine knowledge. For these conditionals (which are now frequently referred to as counterfactuals of creaturely freedom) are, like elements of free knowledge, contingent truths; but, like elements of natural knowledge, they are not in any way under God’s control, given the fact that the creatures they are about have libertarian freedom. As Molinists are wont to say, such truths are prevolitional, meaning simply that they are truly independent of any exercise of God’s free will. So the contingent but prevolitional truths really belong in a middle category between the necessary, prevolitional truths that make up natural knowledge and the contingent, postvolitional truths (i.e., those that are subject to God’s will) that constitute free knowledge. Not surprisingly, Molina gave the name middle knowledge to this third category.

The notion of middle knowledge is the cornerstone of the Molinist position. For it allows us to see how God can both foreknow and exercise control over all that occurs. Since both natural knowledge and free knowledge are prevolitional, they can be thought of as present to God when he is deciding what creative act to perform.3 Given his middle knowledge, God knows exactly how his free creatures would react in any situation in which he might place them. He knows, for example, that if he were to create Libby and put her in such-and-such circumstances, she would freely watch the game; he knows that if the game were cancelled, she would opt for the movie instead; and so on. And God knows the true counterfactuals of creaturely freedom about every possible creature and every situation in which he or she might be placed.4 So, once he decides which creatures to create and in which situations to place them, he foreknows exactly what those creatures will do. For example, by middle knowledge, he knows what Libby would do if she had the option of watching the game; once God decides to put her in the situation that gives her that option, it follows that he knows what she will do. So middle knowledge, when combined with God’s full creative decision (concerning which beings to create in which circumstances), simply entails foreknowledge. Given middle knowledge, then, foreknowledge no longer appears to be a mystery.

Nor does it seem to be a threat to our freedom. For foreknowledge is not simply a function of God’s free choice; rather, it is a function of his choice concerning what to create (over which he does have control) and truths about what we would freely do (truths over which he has no control). If one thought of the counterfactuals about Libby as utterly beyond her control, one might have doubts about whether or not middle knowledge safeguards her freedom. But no Molinist worth his or her salt thinks of the counterfactuals that refer to a creature as manacles she is powerless to escape. For the counterfactuals of creaturely freedom say only what the creature would in fact freely do, not what she would have to freely do. It may be true that Libby would freely watch the game, but it is also true that, having turned on the TV, she has full libertarian freedom to (say) watch the movie instead, and were she to do so, the counterfactual that is in fact true about her would never have been true. In other words, God is powerless regarding which counterfactuals of creaturely freedom are true, but we are not.5 Hence, middle knowledge is fully consistent with our freedom.

Furthermore, say Molinists, it is also consistent with a strong notion of divine providential control. True, given the reality of libertarian freedom, God cannot causally determine everything that takes place. But the decision to create beings with libertarian freedom was God’s in the first place. And assuming that decision, say Molinists, middle knowledge affords God a type of control fully consistent with that demanded by the traditional notion of providence. For the Molinist God, thanks to middle knowledge, is not a risk taker. He never has to deal with acting in situations where he knows only what probably would result from his actions. Rather, he knows precisely how his free creatures would react to any action on his part. If he were to put Libby in situation A, she would freely do X; if he were to put her in situation B, she would do Y; and so on. God can then decide which situation to put her in (say, A or B) depending on which result (X or Y) can more readily be woven into a world that satisfies his creative intentions. God need not worry about things heading off in an unintended direction—about his being surprised by unexpected actions on the part of his creatures. With middle knowledge, he truly does know what he is about. By exercising his providential control through the free actions of his creatures, he (and we) are assured that the world that results will fully manifest his wisdom and love.

Objections and Alternatives: A Brief Introduction

The Molinist picture of providence has been controversial ever since its introduction. What are the causes of this controversy? And what alternatives to it have the critics proposed?

These are questions that cannot be examined in the depth they deserve in this essay.6 Let me, though, briefly indicate both the general problems with Molinism that objectors have alleged and the two major types of alternatives that have been offered.

As we have seen, the crucial Molinist move is the suggestion that God has middle knowledge—knowledge of contingent truths (and especially of the counterfactuals of creaturely freedom) over which he has no control. Critics have suggested that there is reason to doubt either that there are such truths or that they are beyond God’s control. Let us consider each point, and the alternative picture of providence that each naturally engenders, in turn.

First: Why doubt that there are true counterfactuals of creaturely freedom? The arguments here can get quite complicated, but the fundamental objection has been that there are insufficient grounds for any such truths—that there is nothing that makes or brings about their truth. Counterfactuals of creaturely freedom, recall, are supposed to be providentially useful to God in the creative process. Prior to the decision to create Libby and to put her in such-and-such circumstances, God knows what she would do in such-and-such circumstances. But, say the critics, how can this be? If libertarianism is correct, and circumstances do not determine free actions, how can there be any fact of the matter as to what Libby would do in those circumstances prior to her actually being in those circumstances and making up her mind? This family of objections frequently goes under the name of the grounding objection, though the singular “the” is obviously misleading, since such terms as ground, bring about, cause, make true, and so on are not clearly synonymous and since each individually can be interpreted in a number of different ways. Trying to spell out all the available variations that have or could be fashioned on the basic form of the grounding objection would be a herculean task, and may not be worth all the effort. After all, hardly anyone claims that his or her grounding argument constitutes a refutation of middle knowledge, and even the most ardent proponents of the grounding objection acknowledge that there are responses available to render Molinism coherent.7 Still, for many, one or another version of the grounding objection is powerful enough, in their minds, to render the claim that God has knowledge of counterfactuals of creaturely freedom unattractive.

Those who reject Molinism for this reason typically embrace in its stead what has come to be called open theism. Open theists (or openists) believe that, since libertarianism is correct but God has no middle knowledge, we cannot tell the Molinist story of how God knows and controls the future. Indeed, say openists, we need to acknowledge that God simply cannot know and control the future (at least not in the way that Molinists and other traditional advocates of the concept of providence affirm) if that future contains free actions. Libertarianism requires that the future be genuinely open: Libby truly could watch the game and truly could watch the movie instead. No one, not even God, could know what she either will or would freely do. Hence, the traditional notion of divine providence needs to be abandoned if we wish to make room for human freedom. In its place, open theists suggest, we need a picture of a God who, lacking both (complete) foreknowledge and (any) middle knowledge, interacts with creatures whose freedom he truly respects, creatures whom he guides (rather than forces or manipulates) toward good ends. Such a God inevitably takes risks if he creates free beings, for his knowledge of how they will react to his initiatives is limited. Perhaps God knows what Libby would probably do, given her formed character, if he were to put her in such-and-such circumstances, but there are no guarantees here; people do act out of character, or in ways that antecedently would appear improbable even to a noetically unsurpassable creator. God’s knowledge and power give him considerable control over his world, but not the degree of control that most in the Christian tradition have asserted.

The other common objection to Molinism is to deny not the truth of counterfactuals of creaturely freedom or even God’s knowledge of them but rather their prevolitional status. Yes, such critics say, there are (contra the Open Theists) truths as to what Libby would do if placed in various circumstances. But to think that such truths are beyond God’s control is all but sacrilegious. And the sacrilege flows primarily from the libertarian picture of freedom that Molinists assume. Libertarianism causes problems because it conflicts with one of the most central and indispensable principles of traditional Christian belief: that God is genuinely in control of everything that goes on in his universe. Molinists, in deed if not in word, deny this control when they pronounce counterfactuals of creaturely freedom to be truths that are fully independent of God’s will. But the fundamental problem is with libertarianism, not with the counterfactuals. Libertarianism suggests that we free creatures and God are competitors in a zero-sum game of actualizing the world.8 God performs certain free actions, and there is nothing we can do about them. We perform other free actions, and there is nothing he can do about them (short, of course, of simply putting us out of the free-action-performing business entirely). Increase divine free activity and the sphere for free human activity shrinks; magnify human freedom and divine sovereignty and control wanes. No true Christian, the objector says, can comfortably accept so radical a diminishment of the Creator.9

Such an objection to Molinism is most commonly given by advocates of a very different picture of divine providence, one that is commonly (albeit controversially) called the Thomist account.10 For the Thomist, then, God’s sovereignty requires that God have greater control over our actions than libertarians have typically allowed. Some Thomists suggest that God’s arranging of natural causes that determine our actions is fully in accord with their freedom. But this, to my understanding, has never been more than a minority view among Thomists. Libertarians are right, they say, in resisting the common contemporary compatibilist view of freedom—the view that actions that are functions of the laws of nature and prior states of the natural world can still be free provided that the crucial determining events are of the right sort (such as the agent’s own, fully embraced beliefs and desires) and bring about the action in the normal way.11 What libertarians have failed fully to appreciate, though, is that God is not just another natural cause. God’s relation to his universe is utterly unique, and his determining of his creatures’ actions no more robs them of their freedom than does (say) Euripides’s authorial determination of Medea’s actions mean that she lacks freedom within the world of the play. As Aquinas famously put it,

it does not of necessity belong to liberty that what is free should be the first cause of itself, as neither for one thing to be cause of another need it be the first cause. God, therefore, is the first cause, Who moves causes both natural and voluntary. And just as by moving natural causes He does not prevent their actions from being natural, so by moving voluntary causes He does not deprive their actions of being voluntary; but rather is He the cause of this very thing in them, for He operates in each thing according to its own nature.12

Many Thomists feel comfortable (as Aquinas seems to here) speaking of God as the source of the very being both of ourselves and of our actions, and hence as the one whose will supernaturally causes the free actions we perform. Others feel that such language is at best misleading; God determines our decisions, they believe, without there being any causal connection between his act of will and our decisions. But whatever the precise explanation embraced here, Thomists concur in maintaining that the type of absolute metaphysical freedom endorsed by most libertarians, a view that insists that I, not God, am the final ontological arbiter of my actions, cannot be sustained by the circumspect Christian.13 And so (pace the open theists) it is fine to think that God has complete control over and foreknowledge of his world, and (though some Thomists minimize this point) it is even fine to think that God knows what every free creature would do in any situation, but (pace the Molinists) God’s knowledge of such conditionals cannot be thought of as middle knowledge. God knows what Libby would freely do because (and only because) he knows what he himself would determine that she do.

Needless to say, Molinists have not found any of the arguments against the theory of middle knowledge convincing and have vigorously defended both the strong traditional notion of providence and the libertarian account of freedom, a combination that neither their Thomist nor their openist critics can embrace. In addition, as is discussed in greater detail later, Molinists have endeavored to show the theological fecundity of middle knowledge by applying the Molinist view to a number of specific theological topics (such as the problem of evil and the Incarnation). My own view is that the Molinist defense has been more than adequate and that much of the recent discussion of, for example, the latest variations on variations of the grounding objection has done little to push the debate forward.14 Still, a couple of novel objections to Molinism—one offered by Dean Zimmerman and one that has thus far not been delineated and dissected carefully—are worthy of further commentary.

Zimmerman’s Anti-Molinist Argument15

Zimmerman’s “Yet Another Anti-Molinist Argument” begins from the typical Molinist assumption that counterfactuals of creaturely freedom are best thought of as having complete antecedents—that is, antecedents that state the complete causal history of the world at the time the agent in question is free to act. Such an antecedent (call it C) might easily differ from another antecedent (call it C*) in only some insignificant manner regarding some minor event in the history of the world prior to T, the time at which the agent is free. Indeed, C* could differ from C only with respect to some such insignificant event long before the existence of the relevant agent (call her S). Now Molinists typically think that counterfactuals of creaturely freedom are logically independent of one another. The fact that our agent S performs some action (call it A1) in one set of circumstances does not entail that she would perform the same action in some distinct set of circumstances. So it could be that, though (CA1) is true, (C*A1) is false; if S were placed in C*, she would perform some other action (say, A2) instead.16

Indeed, it is conceivable that, for every action S might perform at the time in question, there is a variation on C that, like C*, is such that, were S placed in that variation, she would perform that action. That is, it could be that (CA1), (C*A2), (C**A3), and so on. Indeed, it is conceivable that the counterfactuals be of this sort not only for the actions available to our agent at T but for all actions she might perform at any time. If that were the case, our agent S would be (as Zimmerman puts it) transworld manipulable; God would in effect have (as he also puts it) a remote control over all the actions our agent performs.

Now, if it is possible that our agent S be transworld manipulable, it is also possible that every possible agent be transworld manipulable. That is, it is possible that there be a world in which the counterfactuals of creaturely freedom are so arranged that God’s remote control gives him total control not only over S but over every agent he might create. Zimmerman refers to such worlds as voodoo worlds and contends that, since he has not used any assumption in his argument that Molinists should doubt, they also should not doubt the possibility of voodoo worlds. Yet Zimmerman contends that even the possibility of voodoo worlds should at the very least pose a severe challenge to Molinism. For, how could a creature in a voodoo world be genuinely free? Zimmerman thinks that one of the necessary conditions of a creature’s being free is that it not be under the control of another person. Yet God has absolute control over what every creature does in a voodoo world. So no creature in a voodoo world is free—and, paradoxically, this is a consequence of the counterfactuals of creaturely freedom in that world. In granting that the counterfactuals of freedom could be such as to eliminate the possibility of God’s creating anyone with freedom, is not the Molinist putting himself in an all-but-untenable position?17

This is, to be sure, an intriguing argument. How should the Molinist respond? The first point to make, I think, is to remind ourselves of just how bizarre a voodoo world would be. Indeed, the idea that even one possible creature was transworld manipulable is hard to swallow. How credible is it to think that, even though S performed A1 in C, she would have performed A2 in C*? To return to our earlier example: how plausible is it to think that, even though Libby freely watched the baseball game, given the actual history of the world, she in fact would have watched the movie instead if there had been (to take one of Zimmerman’s favorite examples) some miniscule change in the swirl of dust in some pre-big-bang space-time? Not credible at all, I would think. Agents act for reasons, and Libby’s reasons are identical regardless of the pattern of preprimordial dust. Surely the world in which she acts differently though the circumstances are different only with respect to distant dust is a rather distant world! And what holds for Libby surely holds for our agent S as well; given that (CA1) is true, so in all likelihood is (C*A1). The odds that S is fully manipulable (in the manner Zimmerman envisions) even with respect to a single action is tiny. Even smaller are the odds that she is transworld manipulable. And voodoo worlds—worlds where every possible creature is transworld manipulable? Well, the Jules-Verne-o-Scope that would afford one more than the faintest glimpse of so distant a world has not yet been invented.18

There is one superficially powerful argument for thinking that the Molinist cannot consistently take this line—cannot hold that, if (CA1) is true, then (C*A1) is also (in all probability) true, and hence that voodoo worlds of the sort Zimmerman describes must be exceedingly distant from the actual world. Suppose our situation C were one in which it would not be bizarre or irrational or in any sense strange for S either to do A1 or to do A2—presumably, a case where the reasons for and against doing A1 are more or less evenly balanced. Suppose S in fact does A1 in W, the actual world. Then there is a presumably a nearby world Y in which S does A2 in C. (Y must be near to W because, as noted, there would have been nothing at all unexpected in S’s doing A2 rather than A1.) But now consider what the Molinist is committed to saying about world Y. In Y, it must (according to the Molinist) be true that, since S has done A2 in C, then in all likelihood S would have done A2 in C* as well. (Recall that C* differs from C only with respect to that ancient dancing dust.) So there is a world (call it Z) close to Y in which S is in C* and S does A2. So there is a world (Y) close to the actual world in which (C*A2) is true. And this disproves the Molinist claim that, if (CA1), then (C*A1) is also in all probability true, while (C*A2) is true only in rather distant worlds.

Since I called this a “superficially powerful” argument, I must think that it’s not really a powerful one. Why not? Because the argument shows only that “rather distant” needs to be interpreted carefully. What the Molinist should say might be expressed as follows:

If (CA1), then in all likelihood (C*A2) is true only in worlds more distant from the actual world than those in which (C*A1) is true.

Even though (C*A2) is true in a world close to the actual world, since (C* and A2) is true in world Z, there is every reason to think that there is another world (call it Z*) closer to W than is Z in which (C* and A1) is true. What would be bizarre is not for Z to be relatively close to W, but for Z to be closer to W than is Z*.

Note also that, since Y is not feasible (i.e., is not a world God could indeed create given the true counterfactuals of creaturely freedom) if W is, Z’s closeness to Y gives us no reason to think that Z is feasible either.19 But Zimmerman’s case requires that worlds such as W and Z both be feasible. So the relative closeness of Z to W gives us no reason to think that both would be feasible and so no reason to expect that God has the degree of control Zimmerman’s story suggests.

All of this is interesting (or not), one might say, but strictly speaking it is beside the point. At most, all that we have shown is that voodoo worlds are exceedingly unlikely to be actual, not that they could not be actual. But the mere possibility of such a world is all that Zimmerman’s argument requires. And what have we said to justify the Molinist’s questioning this possibility?

Well, nothing. And it seems to me that the mere possibility of a voodoo world is hard for a Molinist to deny. Some have at least come close to denying it, though. William Lane Craig, for example, spends a considerable portion of his response to Zimmerman’s argument by calling into question what he calls the Recombination Principle—the thesis that “every (strictly) consistent combination of counterfactuals of creaturely freedom constitutes a possible world-type for God”—and thus calling into question the genuine possibility of voodoo worlds:

Well, maybe there are worlds like that [i.e., worlds in which two counterfactuals of creaturely freedom with minor causally irrelevant differences in their antecedents differ in truth value], but it seems to me that here we begin merely to speculate. When we go even further to speculate whether relative to some world or other the counterfactuals of freedom combine to constitute a world-type for God which affords him perfect control of every possible action of every possible free creature, then we are completely out of our depth. It is epistemically possible that no such world-type confronts God in any world.20

Such a position, though, seems strange for a Molinist to take. Suppose (CA1) is true; in all likelihood, then, (C*A1) is also true. But surely the two are not logically linked; surely (CA1) does not entail (C*A1). So there is a world (however distant it might be) in which both (CA1) and (C*A2). And what is true of this counterfactual of creaturely freedom is surely true of all others. So surely there is a possible world in which the true counterfactuals of creaturely freedom are all oddly paired in this way. And that means that there is a world in which “the counterfactuals of freedom combine to constitute a world-type for God which affords him perfect control of every possible action of every possible free creature.” So far as I can see, nothing in such reasoning seems unacceptably speculative or epistemically daring from the Molinist perspective.

The best Molinist response to Zimmerman, then, is not to question (as Craig does) the possibility of voodoo worlds, but rather to question (as Craig also does) their relevance to the viability of Molinism. Voodoo worlds are ones where every possible creature is transworldly manipulable, and this might indeed pose an embarrassment to Molinism if the transworld manipulability of a creature were inconsistent with that creature’s freedom. But Zimmerman offers no argument for the claim that the two are inconsistent—and he himself admits (in effect) that he has no argument for the claim.21 Rather, as he puts it, he tells “stories with which … I tried to badger the reader into agreeing with me,” stories that, to Zimmerman’s mind, support the claim that “being free analytically entails not being under the complete control of another.”22 But to my mind, none of these stories (all, in effect, regarding God’s or some other agent’s using his remote control to dial up the actions he wants) should move a Molinist, or anyone who is well informed but genuinely undecided in the debate over providence. Knowledge of counterfactuals of creaturely freedom should give God considerable control over his creatures’ free actions; any “odd couple” of such counterfactuals should give him complete control over the specific free action in question; and the bizarre combination of counterfactuals we see in voodoo worlds should give him complete control over all free actions. Maybe we would feel less free if we knew we were in a voodoo world, but it is hard to see why that feeling should be thought to have any substantive force, especially when we remind ourselves that a voodoo world’s being actual is not a fact we are powerless to control. This is simply a consequence of the fact that the counterfactuals of freedom true about me are not one and all truths that I am powerless to make false.23

Much more could be said about Zimmerman’s intriguing contribution to the debate. It seems clear to me, though, that, for all its ingenuity, it constitutes yet another anti-Molinist argument that does little damage to the Molinist position.

The Power Over the Past Argument

There is one other general anti-Molinist argument, what I call the Power Over the Past Argument (POPA), which might well be worthy of more attention than it has thus far received. This argument is based on the fact that there are many prior events (or true propositions, or actual states of affairs) that Molinists typically contend are still under the control of certain agents. Consider the case of Jesus’s prophecy that Peter would deny him. At the time of Peter’s denial (which I assume was a free act on his part), it was already the case that

  1. (1) Jesus prophesied that Peter would deny him.

Yet this fact about the past, Molinists (like, in fact, many non-Molinist libertarians) contend, did not constrain Peter’s action. Being free, he had the power not to deny Jesus; and (assuming that Jesus could not have issued a false prophecy) had Peter exercised this power, Jesus never would have issued the prophecy. So at the time of Peter’s action, (1) is true, and Peter has the power so to act that (1) would have been false. Schematically, letting Z stand for some truth about the past, Px stand for the claim that a certain agent has the power to perform an action x, and Ax stand for the proposition that the agent in question in fact performs that action, we have in this case an instance of the following three-part conjunction:

(CPP) Z & Px &(AxZ).

But this is hardly the only type of instance of (CPP) that Molinists (and perhaps some other libertarian theists as well) will countenance. Consider:

  1. (2) A colony of ants moved into Paul’s yard last Saturday.

  2. (3) Abraham was born.

  3. (4) Christ’s human nature was assumed by the Son.

  4. (5) God knew prior to creation that (C → A) is a true counterfactual of creaturely freedom.

In each of these cases, it is easy to imagine (at least from a Molinist perspective) that the proposition in question is true but that some agent has the power to perform an action such that the proposition in question would not have been true had that action in fact been performed. The truth of the relevant past-tense proposition, then, in fact places no restrictions on the freedom of the relevant agent. For example, Plantinga famously told the story of Paul, whose refraining from mowing his yard preserves the colony of ants that entered his property last Saturday; were he to mow (which he can indeed do), God would have seen to it that the ants never entered Paul’s lawn last Saturday. So while it is a fact that they entered, Paul has access to worlds in which they did not.24

The last of our examples is the one that leads to the POPA. Molinists typically say that God can have—and use—knowledge of what a free creature S would do (namely, action A) in situation C, but such middle knowledge (and even God’s use of such knowledge in making his creative decisions) is fully compatible with S’s freedom to do other than A in C; and had S exercised such freedom, God’s middle knowledge would have been other than it in fact is. So though God’s middle knowledge is providentially useful to him, it leaves S free. Now, our objector asks, if all this is so, what is to prevent God’s not merely knowing but in fact deciding which counterfactual of creaturely freedom is to be true prior to God’s creative decision? That is, why not add the following as yet another instance of (CPP):

  1. (6) God willed prior to creation that (CA) be a true counterfactual of creaturely freedom.

That is, why not say that God chose to bring it about that (CA) is true, but that S does indeed have the power, if placed in C, to refrain from doing A; and if S had so refrained, (CA) would have been false? God would thus be the source of the truth of the counterfactuals of creaturely freedom, but without limiting the freedom of his creatures. For if (1) through (5) are facts about the past that cohere with the freedom of the agents involved, why not think that the same holds for (6)?

Clearly, to view (6) as another instance for which (CPP) holds is to abandon Molinism, because central to Molinism is the claim that the true counterfactuals of creaturely freedom are prevolitional—that is, not under God’s control. If God’s free will determines the true creaturely world-type, then the counterfactuals of creaturely freedom are true only postvolitionally. To affirm (6), then, is to abandon Molinism. The challenge here, of course, is for the Molinist to explain why such an abandonment is not called for, given the standard Molinist position on (1) through (5). What principled reasons—other than “it means the end of Molinism”—can Molinists offer for not saying of (6) what they say of (1) through (5)?

Before addressing this question directly, we should first not that the POPA would probably be offered in slightly different ways by advocates of the alternative positions on divine providence discussed previously. Open theists would feel no attraction to (6) but would say that Molinists, given their approval of (1) through (5), should approve of (6) as well; the absurdity of (6) thus gives further evidence of the absurdity of Molinism. Thomists, on the other hand, would presumably think of (6) as very plausible, at least so long as the antecedents of the counterfactuals of creaturely freedom do not include God’s causal activity, and would be suggesting that consistency requires the Molinist to embrace (6)—and thereby abandon Molinism for Thomism.

There are a number of directions Molinists might follow in response to the POPA, but I suspect that most would go in one of two directions. Some might say that embracing (6), and seeing it as akin to (1) through (5) as an instance of (CCP), would primarily be deficient morally, while others might claim its defects are more fundamental and metaphysical. Let us briefly consider each response in turn.

Suppose, as our Thomist proponent of the POPA does, that God’s free will is the source of the true counterfactuals of creaturely freedom. Then it was truly up to God, for example, whether or not Adam would eat the apple (thereby disobeying God) if placed in the Garden of Eden set of circumstances. We know what happened (assuming, of course, the literal truth of the Genesis story); Adam did indeed eat, and all hell (so to speak) broke loose. So God must have known, prior to creation, that

  1. (7) If Adam were placed in the Garden situation, he would sin (and all hell would break loose).

But if God not only knows but also determines which counterfactuals of creaturely freedom are true, then he freely decided to make it the case that (7) was true. Why, one might ask, would a good God do that? Why bring it about that a free creature would sin—that he would disobey his creator’s command—when one could just as easily have brought it about that he would obey? Is not a God who would act in this way a far cry from the morally perfect God of the Anselmian tradition?

One could imagine a Thomist advocate of our POPA responding in a couple of ways, one incoherent and one merely hard to swallow. The incoherent reply would be: God in fact did will that Adam obey if placed in the Garden—that is, prior to creation, God in his goodness chose to make it true that

  1. (8) If Adam were placed in the Garden situation, he would obey.

But his making (8) true put no limitations upon Adam’s freedom, and, alas, Adam employed his freedom to sin. Adam’s sinning, of course, means that God knew from eternity that he would sin if placed in the Garden, and this in turn means that, because of Adam’s sinning, God willed prior to creation that (7) be true. But his so willing is a logical consequence of Adam’s choice and hence casts no shadow on God’s goodness.

The incoherence of this reply is, I hope, obvious. For it holds that God makes (8) true prior to creation and then (?), consequent to Adam’s action, makes (7) true prior to creation instead. So which—(7) or (8)—did God know when making his creative decisions? The answer cannot really be “both”; yet that seems to be the answer this reply is offering. So this is not a coherent means of answering the moral objection.

The only other way for the Thomist advocate of the POPA to reply, it seems to me, is to say that God brought about only (7), and even though he could have brought about (8), he is in no way morally deficient for having done so. God brought about all the true counterfactuals of creaturely freedom, and as with everything that he willed, he brought about their truth out of his own wisdom, goodness, and love. We may not understand why it was better for him to make it the case that Adam would sin if placed in the Garden, rather than make it the case that he would obey, but we need to have trust that it was better. The counterfactuals of creaturely freedom that are in fact true are the best that could have been true, or, at the very least, God’s making these counterfactuals true, as opposed to his making any other set he might have made true, was as good and loving a creative action as he could have performed.

To my mind, this is a coherent reply. But it is also a highly implausible one. At least I suspect most Molinists (and many non-Molinists) will find it so. When we consider a counterfactual such as (7), and assume that it is true, our natural response is not to thank our lucky stars that (7) rather than (8) is true; our strong feeling is in the opposite direction. Molinists typically suggest that God is, so to speak, stuck with the counterfactuals of creaturely freedom he finds true, and the “stuck with” locution strongly suggests that all of us see that the set God finds true is not that than which none greater can be conceived. Proof here is (as usual) hard to come by. Still, most Molinists (and many non-Molinists) would probably find the argument against their Thomist opponents here convincing.

So much, then, for the moral response to the POPA. Even if effective, this response would have little purchase on the openist advocate of the POPA, since open theists could well agree that (6) would have implications hard to reconcile with divine moral perfection. The problem, the open theist might say, is that the Molinist still has not indicated why (6) should be treated differently from (1) through (5). And unless such an indication can be offered, the Molinist still seems to be committed to the admittedly implausible claim that God can decide which counterfactuals of creaturely freedom are true.

The best Molinist response to this challenge, I think, is to claim that coherent stories explaining how (1) through (5) can be instances of (CPP) all seem to presuppose the standard Molinist picture of a God who knows but does not determine counterfactuals of creaturely freedom. For example, the elaborate justification of (1) offered in my Divine Providence assumes that the relevant counterfactuals are prevolitional and that their truth limits the prophecies God can make.25 No such limitations would be plausible if the truth of the counterfactuals were itself something that was up to God. Here, then, the Molinist seems to have grounds to distinguish (6) from (1) through (5), for one clearly cannot, on the assumption that counterfactuals of creaturely freedom are not under God’s control, proceed to assert, as (6) asserts, that they are under his control.

Much more could be said here, but I believe this is the direction the conversation on POPA would go. To me, it seems evident that it fails as a refutation of Molinism. Still, examining it further might push the discussion in a new and interesting direction.

Promising Directions for Future Discussion

So some further discussion of voodoo worlds and the POPA is probably warranted. I am not convinced, though, that the most progress is likely to be found by pursuing such paths. If these are not the most promising roads, then what approaches should we take in the discussion of Molinism? I have four interrelated suggestions.26

First, I think our conversations should take place with the understanding that there are three serious competitors for our support, not just (as some imply) two. Open theism has been the most-discussed alternative to Molinism in recent years, at least if one limits one’s ear to conversations within Anglo-American Protestant academic circles. But the Thomist (or Calvinist, or Augustinian, or …) alternative, one where it is the status, not the existence, of true counterfactuals of creaturely freedom that is up for debate, has been far more prominent historically, and this alternative picture of divine providence still needs to be seen as a genuine competitor today.27

Second, we need to be open to the possibility that the best way forward in evaluating Molinism may involve looking back—not back to the comparatively recent battles on which some would have us fixate, but back on the long controversy concerning the theory of middle knowledge, especially (though not exclusively) among Catholic philosophers. Most of us contemporary combatants are not primarily historians and have only a halting familiarity with the works engendered (even into the early twentieth century) by the controversy. Perhaps the pro- and anti-Molinist works that appeared in the three and a half centuries prior to Plantinga’s “rediscovery” of Molinism in the 1970s have little to add to the current discussion. Are we, though, in a position at this point to make such a judgment? I suspect not. For all we know, some of these now-all-but-forgotten discussions contain insights that would push the current conversation in (for us) novel and illuminating directions. Greater efforts (by those equipped to do the job) to connect the history of the dispute with its current manifestations should surely be encouraged.28

Third, advocates of each of the three general accounts of providence should devote more of their energy to questions of internal philosophical development, questions that are conducted primarily with allies, not with opponents, as one’s principal interlocutors. Molinists might do well to consider more carefully precisely what is and what is not included in the antecedents of counterfactuals of creaturely freedom.29 Open theists could profitably focus more of their attention on the issue of whether God’s providence operates via elaborate prearranged contingency plans or through improvisational divine reactions to unforeseen circumstances.30 Thomists have surely not reached consensus concerning the precise explanation as to how divine determination of human actions takes place. And these are but three of what are undoubtedly many unsettled issues within each of the camps. Unfortunately, such intramural matters have all too often been given short shrift, while too much attention has been devoted to foreign relations—to launching arguments against one’s “enemies” or to parrying attacks from outsiders. Advocates of each of the three major positions need to spend a bit less on (so to speak) their nuclear arsenals and a bit more on basic infrastructure.

Finally, the greatest hope for progress might well be found in directing most of our energy toward examining the ramifications of the different general views of providence for specific elements of Christian (or, more generally, theistic) faith. Several years ago, I put the point this way:

We need to look carefully at how [open theism] and its rivals fare when applied to particular Christian doctrines, beliefs, and practices. How appealing is the openist picture of biblical authorship or of prophecy? Can Molinists offer us an attractive account of God’s responsiveness to prayer, or of divine love? What can Thomists say about the problem of evil, or about hell? And we need to examine these questions, not with the expectation that we can, even in the most kindly Christian way, grind our opponents into the dust, but rather with the hope of gradually building a nondecisive but persuasive cumulative case for or against one of the views. We need to seek, not to win by a knockout, but to win on points.31

I continue to believe that this is the approach that is most likely to prove most fruitful. Fortunately, much work of this sort has been going on for a while now, especially within the Molinist camp. Plantinga (as usual) led the way here, applying (albeit unknowingly, at least at first) Molinist resources to address the problem of evil.32 William Lane Craig has offered Molinist-motivated explications of biblical inspiration and salvation outside the church.33 Michael Rea has suggested that Molinism might have implications for our understanding of original sin.34 Robert Hartman has maintained that Molinism is helpful in solving the theological problem of moral luck.35 And I have tried to apply the general Molinist picture to such Christian doctrines and practices as papal infallibility, prophecy, prayer, and the Incarnation.36 Similarly, efforts to employ the open theist perspective to illuminate prayer, salvation, revelation, and related issues have been made, most notably by John Sanders and William Hasker.37 And Thomists have hardly been averse to doing applied work as well, as Hugh McCann’s recent discussions of God’s relationship to human sin attest.38

The way forward, it seems to me, is to foster the continuation of these explorations in applied Molinism, Thomism, and open theism. Clearly, more needs to be done. Each of the three camps contains a variety of viewpoints, and endeavors at application have sometimes led to strong resistance from within.39 On the other hand, “outsiders” have sometimes challenged adherents to one or another of the three general views to relate their providential perspective to a specific Christian doctrine, and some such challenges remain unmet.40 And surely there are unplowed fields here—other elements of belief or practice where attempts at application might prove rewarding.

These, then, are the directions in which I hope the debate will be pursued. If such paths are taken, the debate concerning Molinism should continue to be lively and productive well into the future.41


Adams, Robert. (1987). “Middle Knowledge and the Problem of Evil.” In Robert Adams, ed., The Virtue of Faith and Other Essays in Philosophical Theology. New York: Oxford University Press, 77–93.Find this resource:

Adams, Robert. (1991). “An Anti-Molinist Argument.” In James E. Tomberlin, ed., Philosophical Perspectives: Vol. 5. Atascadero, CA: Ridgeway, 343–353.Find this resource:

Craig, William Lane. (2011). “Yet Another Failed Anti-Molinist Argument.” In Ken Perszyk, ed., Molinism: The Contemporary Debate. Oxford: Oxford University Press, 144–162.Find this resource:

Flint, Thomas P. (1998). Divine Providence: The Molinist Account. Ithaca, NY: Cornell University Press.Find this resource:

Flint, Thomas P. (2004). “Risky Business: Open Theism and the Incarnation.” Philosophia Christi 6: 213–233.Find this resource:

Flint, Thomas P. (2009). “Divine Providence.” In Thomas P. Flint and Michael C. Rea, eds., The Oxford Handbook of Philosophical Theology. Oxford: Oxford University Press, 262–285.Find this resource:

Flint, Thomas P. (2011). “Molinism and Incarnation.” In Ken Perszyk, ed., Molinism: The Contemporary Debate. Oxford: Oxford University Press, 187–207.Find this resource:

Flint, Thomas P. (2011). “Whence and Whither the Molinist Debate: A Reply to Hasker.” In Ken Perszyk, ed., Molinism: The Contemporary Debate. Oxford: Oxford University Press, 37–49.Find this resource:

Freddoso, Alfred J. (1988). “Introduction.” In Luis de Molina, ed., On Divine Foreknowledge: Part IV of the Concordia. Ithaca, NY: Cornell University Press, 1–81.Find this resource:

Hasker, William. (1989). God, Time, and Knowledge. Ithaca, NY: Cornell University Press.Find this resource:

Hasker, William. (2004). Providence, Evil and the Openness of God. London: Routledge.Find this resource:

Hasker, William. (2011). “The (Non-)Existence of Molinist Counterfactuals.” In Ken Perszyk, ed., Molinism: The Contemporary Debate. Oxford: Oxford University Press, 25–36.Find this resource:

McCann, Hugh. (1995). “Divine Sovereignty and the Freedom of the Will.” Faith and Philosophy 12: 582–598.Find this resource:

Molina, Luis de. (1988). On Divine Foreknowledge: Part IV of the Concordia. Trans. and ed. by Alfred J. Freddoso. Ithaca, NY: Cornell University Press.Find this resource:

Perszyk, Ken, ed. (2011). Molinism: The Contemporary Debate. Oxford: Oxford University Press.Find this resource:

Plantinga, Alvin. (1985). “Self-Profile” and “Replies to My Colleagues.” In James E. Tomberlin and Peter van Inwagen, eds., Alvin Plantinga. Dordrecht: D. Reidel, 3–97, 313–396.Find this resource:

Plantinga, Alvin. (1986). “On Ockham’s Way Out.” Faith and Philosophy 3: 235–269.Find this resource:

Sanders, John. (1998). The God Who Risks: A Theology of Providence. Downers Grove, IL: InterVarsity Press.Find this resource:

Zimmerman, Dean. (2009). “Yet Another Anti-Molinist Argument.” In Samuel Newlands and Larry Jorgensen, eds., Metaphysics and the Good: Themes from the Philosophy of Robert Merrihew Adams. Oxford: Oxford University Press, 33–94.Find this resource:

Zimmerman, Dean. (2011). “An Anti-Molinist Replies.” In Ken Perszyk, ed., Molinism: The Contemporary Debate. Oxford: Oxford University Press, 163–186.Find this resource:


(1) The order of presentation (and some of the wording) in this and the next section is based on my “Divine Providence,” in Thomas P. Flint and Michael C. Rea, eds., The Oxford Handbook of Philosophical Theology (Oxford: Oxford University Press, 2009), pp. 262–285.

(2) John Henry Newman, Prayers, Verses and Meditations (Ft. Collins, CO: Ignatius Press, 2002), p. 339.

(3) The language here suggests that God is in time—that there is a time when he is deciding what to do, and a later time at which he does it. Many Molinists, though, think of God as atemporal, and view the use of temporal metaphors (based, obviously, on the fact that our decision-making is typically a temporally extended process) as a useful but potentially misleading way of describing what is essentially a dependence relation, not essentially a temporal one. For more on this point, see my Divine Providence: The Molinist Account (Ithaca: Cornell University Press, 1998), pp. 37 and 174–176. My presentation and discussion of the Molinist view is immeasurably indebted to the introduction offered by Alfred Freddoso in Luis de Molina, On Divine Foreknowledge: Part IV of the Concordia, tr. and ed. by Alfred J. Freddoso (Ithaca, NY: Cornell University Press, 1988), pp. 1–81.

(4) I am speaking a bit loosely here in referring to possible creatures. Molinists are not committed to the odd claim that, in addition to all the actual creatures in existence there is also a horde of merely possible creatures hanging around! For a stricter statement of the Molinist view, see my Divine Providence, pp. 46–47.

(5) At least, we are not powerless with respect to counterfactuals with true antecedents. Molinists need not think that Libby can do something about counterfactuals where the antecedent describes a situation she’s in fact never in. All they need insist is that, if she were in such a situation, she would have power over both the counterfactual and her own action.

(6) For more extended discussions, see both my article “Divine Providence” and Divine Providence, Chapters 3 through 7. For two of the classic anti-Molinist arguments discussed there, see Robert Adams, “An Anti-Molinist Argument”, in James E. Tomberlin, ed., Philosophical Perspectives, 5 (Atascadero, CA: Ridgeway Publishing, 1991), 343–353, and William Hasker, God, Time, and Knowledge (Ithaca, NY: Cornell University Press, 1989). For Hasker’s more recent thoughts on the argument against Molinism, see his “The (Non-)Existence of Molinist Counterfactuals”, in Ken Perszyk, ed., Molinism: The Contemporary Debate (Oxford: Oxford University Press, 2011), 25–36.

(7) See, for example, Robert Adams’ remark in note 22 of the version of “Middle Knowledge and the Problem of Evil” printed in his The Virtue of Faith and Other Essays in Philosophical Theology (New York: Oxford University Press, 1987), pp. 77–93. See also Hasker’s Providence, Evil and the Openness of God (London: Routledge, 2004), p. 197.

(8) The language here is borrowed from David Burrell. See, for example, his Freedom and Creation in Three Traditions (Notre Dame, IN: University of Notre Dame Press, 1993), p. 112.

(9) Hugh McCann, one of the foremost contemporary advocates of the Thomist view, nicely summarizes and expands upon this point:

to the extent God does not exert active control over my decisions, whether through other events or direct involvement, He does not control them at all. He can therefore achieve His ends only by reacting to what I do, and to that extent His plans are subordinated to mine. In addition to weakening His sovereignty, this situation also threatens God’s omniscience. It suggests He can know how I will act in the circumstances in which I am placed only by observing my actions. As creator, He is in the dark. He can know what the possibilities are, but if my freedom makes for more than one, then even His knowledge of the world He is creating appears to depend on my action—an unsatisfactory situation to say the least. These problems can be avoided if God is able to exercise creative control in my actual choice.

See Hugh McCann, “Divine Sovereignty and the Freedom of the Will,” Faith and Philosophy 12 (1995), p. 586.

(10) The next two paragraphs are taken in part from my “Divine Providence.”

(11) For a clear expression of the Thomist rejection of contemporary compatibilism, see, for example, Theodore J. Kondoleon, “The Free Will Defense: New and Old,” The Thomist 46 (1983), p. 19.

(12) Thomas Aquinas, Summa Theologica, Part I, Question 83, Article 1, ad 3. Translation by the Fathers of the English Dominican Province.

(13) The words quoted are from McCann, p. 593.

(14) Some of the more interesting discussions can be found in Ken Perszyk’s excellent work Molinism: The Contemporary Debate (Oxford: Oxford University Press, 2011). See especially the papers by Hasker, Merricks, Zimmerman, and myself in Chapters 1 through 6. While I think each of these papers is of high quality and makes some points that specialists in the debates of the last thirty years might see as novel, I think even the authors would probably confess that they do little to change the rhetorical landscape in any fundamental way.

(15) Dean Zimmerman, “Yet Another Anti-Molinist Argument,” in Samuel Newlands and Larry Jorgensen, eds., Metaphysics and the Good: Themes from the Philosophy of Robert Merrihew Adams (Oxford: Oxford University Press, 2009), pp. 33–94.

(16) I use the single-line arrow to represent counterfactual conditionals. So “(CA1)” is shorthand for “If it were the case that S was in C, S would perform action A1.”

(17) For a very helpful condensed outline of the argument, see Dean Zimmerman, “A Précis of ‘Yet Another Anti-Molinist Argument,’” in Ken Perszyk, ed., Molinism: The Contemporary Debate (Oxford: Oxford University Press, 2011), pp. 140–143.

(18) Jules Verne-o-scopes were patented by David Kaplan; see his “Transworld Heir Lines,” in Michael J. Loux, ed., The Possible and the Actual: Readings in the Metaphysics of Modality (Ithaca, NY: Cornell University Press, 1979), pp. 88–109.

(19) The notion and terminology of feasibility was first introduced in my doctoral dissertation, Divine Freedom (University of Notre Dame, 1980). See also my “The Problem of Divine Freedom,” American Philosophical Quarterly 20 (1983), pp. 255–264.

(20) William Lane Craig, “Yet Another Failed Anti-Molinist Argument,” in Perszyk, pp. 144–162; the quotation is from p. 156. A world-type for God is the set of counterfactuals of creaturely freedom God knows to be true. World-type terminology was introduced in my Divine Freedom and “The Problem of Divine Freedom”; see note 19.

(21) See p. 83 of “Yet Another Anti-Molinist Argument”; see also pp. 178 and 184 of Zimmerman’s “An Anti-Molinist Replies,” in Perszyk, pp. 163–186.

(22) Zimmerman, “An Anti-Molinist Replies,” p. 178.

(23) See note 5.

(24) See Alvin Plantinga, “On Ockham’s Way Out,” Faith and Philosophy 3 (1986), pp. 235–269.

(25) See Chapter 9 of my Divine Providence.

(26) What follows is a somewhat abbreviated version of the suggestions made in my “Whence and Whither the Molinist Debate,” in Perszyk, Molinism: The Contemporary Debate, pp. 37–49.

(27) The variety of names at the beginning of this sentence should remind us (if reminders are necessary) that none of the three positions on providence is monolithic. Molinists can and do differ among themselves over lots of significant matters: for example, the prevalence of free action, the causal relationship (if any) between agents and their actions, the manner in which reasons influence free actions, and so on. Similar differences are found among open theists and among Thomists.

(28) Some such efforts, of course, have taken place. For one attempt to connect a relatively obscure historical figure to the current debate, see Donald Wayne Viney, “Jules Lequyer and the Openness of God,” Faith and Philosophy 14 (1997), pp. 212–235. Needless to say, following the course I advise here will not be easy. Few philosophers are both sufficiently conversant with the intricacies of the contemporary debate over Molinism and sufficiently equipped (linguistically and historically) to examine the works I have in mind and determine whether they can add significantly to the conversation. (I can say, with confidence and regret, that I am most assuredly not among this few.)

(29) I discuss this point in “Divine Providence,” pp. 276–277.

(30) For some of the issues relevant to this point, see my review of Hasker’s Providence, Evil and the Openness of God in Philosophia Christi 8 (2006), pp. 493–496.

(31) See p. 214 of my “Risky Business: Open Theism and the Incarnation,” Philosophia Christi 6 (2004), pp. 213–233.

(32) See Alvin Plantinga, The Nature of Necessity (Oxford: Clarendon Press, 1974), Chapter 9. For some of his later thoughts on issues related to Molinism, see his “Self-Profile” and “Replies to My Colleagues”, in James E. Tomberlin and Peter van Inwagen, eds., Alvin Plantinga (Dordrecht: D. Reidel, 1985), pp. 3–97 and 313–396.

(33) On inspiration, see William Lane Craig, “‘Men Moved By The Holy Spirit Spoke From God’ (2 Peter 1:21): A Middle Knowledge Perspective on Biblical Inspiration,” Philosophia Christi 1 (1999), pp. 45–82. On salvation outside the church, see his “‘No Other Name’: A Middle Knowledge Perspective on the Exclusivity of Salvation Through Christ”, Faith and Philosophy 6 (1989), pp. 172–188.

(34) See Michael Rea, “The Metaphysics of Original Sin,” in Peter van Inwagen and Dean Zimmerman, eds., Persons: Human and Divine (Oxford: Clarendon Press, 2007), pp. 319–356.

(35) Robert Hartman, “Applying Molinism to the Theological Problem of Moral Luck,” Faith and Philosophy 31 (2014), pp. 68–90.

(36) On the first three issues, see Chapters 8 through 11 of my Divine Providence. On the Incarnation, see “‘A Death He Freely Accepted’: Molinist Reflections on the Incarnation,” Faith and Philosophy 18 (2001), pp. 3–20; “The Possibilities of Incarnation: Some Radical Molinist Suggestions,” Religious Studies 37 (2001), pp. 125–139; “Should Concretists Part with Mereological Models of the Incarnation?,” in Anna Marmodoro, ed., The Metaphysics of the Incarnation (Oxford: Oxford University Press, 2011), pp. 67–87; and “Molinism and Incarnation,” in Perszyk, Molinism: The Contemporary Debate, pp. 187–207.

(37) For Sanders, see especially Chapter 8 in his The God Who Risks: A Theology of Providence (Downers Grove, IL: InterVarsity Press, 1998). For Hasker, see especially “Appendix: Replies to My Critics” in his Providence, Evil and the Openness of God; see also The Triumph of God over Evil (Downers Grove, IL: InterVarsity Press, 2008).

(38) See his “The Author of Sin?” Faith and Philosophy 22 (2005), pp. 144–159. For criticism of this view, see Katherin Rogers, “God Is Not the Author of Sin: An Anselmian Response to Hugh McCann,” Faith and Philosophy 24 (2007), pp. 300–309. For McCann’s reply, see his “God, Sin and Rogers on Anselm: A Reply,” Faith and Philosophy 26 (2009), pp. 420–431.

(39) For example, my Molinist-inspired views on the Incarnation have not (alas) elicited universal acclaim even among Molinists. See, for example, William Lane Craig, “Flint’s Radical Molinist Christology Not Radical Enough,” Faith and Philosophy 23 (2006), pp. 55–64. Similarly, Craig’s discussion of the exclusivity of salvation through Christ has proven controversial, with noteworthy Molinist opposition to his view. For one such dissent, see Raymond J. VanArragon, “Transworld Damnation and Craig’s Contentious Suggestion,” Faith and Philosophy 18 (2001), pp. 241–260.

(40) In my “Risky Business: Open Theism and the Incarnation,” for example, I argue that, though both Molinists and Thomists can (assuming the truth of their respective views) make sense of the traditional idea that Jesus Christ was both divine and significantly free, it is quite difficult to see how open theists can plausibly do so. To the best of my knowledge, open theists have not yet attempted to deal with this issue.

(41) I am grateful to Trenton Merricks for very helpful comments on an earlier version of this essay.