The free-will defense of evil and suffering is a failure—so Hugh McCann contends. This verdict surprises, given the opinion of so many philosophers that Alvin Plantinga’s argument succeeds resoundingly. But it succeeds, maintains McCann, only because it abandons God’s providential control of human volition, thus calling into question his power to secure a good and truly redemptive outcome. This sounds like McCann must be a determinist of some sort, but that’s not the case either. He expressly identifies himself as a proponent of libertarian freedom. While he acknowledges that compatibilism would provide the kind of control over the world that he believes a perfect God should and must have, it sacrifices too much. Determinism violates the moral responsibility that properly belongs to creatures made in the image of God. If our actions were but the causal product of nature and nurture, “we would be reduced to the same status as our house pets: admirable in many ways, perhaps, but utterly bereft of the dignity of a moral being.”1
McCann takes up what many believe to be an impossible challenge: to formulate a robust understanding of divine sovereignty that respects, and indeed makes possible, the libertarian freedom of rational creatures. God must be accorded, he insists, “an active role as creator in the production of human action,” yet not in such a way as to compromise or nullify the liberty he bestows.2
But wait . . . we have yet to assess the merits of a position that appears to offer precisely what McCann is looking for. “If the standard free-will defense is to be made viable,” he explains, “we must find a way for God to know as creator what world he is creating. This requires that his knowledge of creaturely decisions and actions be prior to their actual occurrence.” This priority is logical, not temporal: that is to say, God’s knowledge of our deeds must not be “ontologically dependent upon or explained by our actually performing them.”3 Fortunately for us, a Jesuit theologian in the 16th century concocted a theory that seems to fit the bill. It goes by the name of Molinism.
Oh, is that William Lane Craig standing in the back frantically waving his hand? Yes, Dr Craig, do you have something to add?
Molinism teaches that God possess an exhaustive and complete knowledge of every possible world he might have created, including the free decisions and actions of the inhabitants of these worlds. He knows what every person will think, will, and do under every conceivable circumstance. Molinists call this middle knowledge. It’s as if God plays out in his mind every possible creation scenario. He comprehensively cognizes “in advance” the subjunctives of freedom. He knows all the what-ifs, maybes, and sure things. He knows all the causal permutations that lead to and flow from every event in each and every possible world.
Consider the Old Testament story of Jonah. Before God made the world, says the Molinist, he knew the subjunctive proposition “If Jonah were commanded to go to Nineveh and preach to the Assyrians, he would flee to Tarshish.” Not only this, but he also knew Jonah’s reactions, decisions, and actions in every possible set of circumstances. Now expand this middle knowledge to comprehend every prophet, every Israelite, every human being, both created and uncreated, actual and possible. McCann elaborates:
Molina referred to this kind of knowledge as scientia media or middle knowledge, because it falls between two other types of knowledge he attributed to God: natural knowledge and free knowledge. The first category contains all logical and conceptual truths, which Molina held, God knows simply by comprehending his own nature, in which they are embedded. Free knowledge is knowledge of propositions whose truth depends on how God exercises his own free will. Thus, given his decision to create tigers, the proposition that there are tigers belongs in this category. Middle knowledge is held to fall between the two in that, like natural knowledge, it consists of propositions whose truth or falsity is not up to God. God may decide the circumstances in which he will place Jonah, but it is Jonah who decides how he will act in them. And the truth is that in the circumstances we are considering, Jonah would decide to flee for Tarshish. This, for God, is a given from the Molinist perspective, just as it is a given that triangles, if there are to be any, shall have three sides. On the other hand, middle knowledge is like free knowledge in that it consists only of contingent propositions. The fact that there are tigers is contingent, and so is the fact that, presented with the divine commission to head for Assyria, Jonah would instead head for Tarshish.4
The advantages of middle knowledge are clear. It capacitates omniscience and sovereignty. The Creator freely actualizes the one specific world (one out of zillions and zillions) that best fulfills his providential and eschatological ends. Everything happens according to his will and permission, for it is this world and not another that he has chosen to bring into being:
Assuming it is a legitimate notion, middle knowledge restores God’s omniscience as creator, and does much to restore his sovereignty and providence in creating creatures with libertarian freedom. Armed with information about how any such creature would decide and act in the various circumstances in which he might be placed, God has the option of not creating the creature, or of creating him in whatever circumstances are called for by the subjunctives of freedom whose truth God wishes to be reflected in the actual world. Now of course the circumstances in which one creature is placed may depend in part on how others choose to exercise their freedom. But the willing of those others can in turn be providentially arranged, since they too fall under middle knowledge. In principle, then, nothing need occur in the actual world that does not have God’s prior recognition and consent, at least. He knows all of his options in advance, and once his creative decisions are made he knows by his free knowledge how things will go, not just how they could or would, if . . .5
This does not mean that everything happens according to God’s optimal desires. He no doubt would have preferred if Adam and Eve had not eaten of the apple; but an unfallen world, we may surmise, was unavailable for actualization. In all possible worlds free human beings always disobey their Creator and fall into alienation and wickedness; in every alternative cosmos there is sin and death. Let’s call this transworld fallenness. Similarly, God no doubt would have preferred if Jonah had gone directly to Nineveh, but in the specific world he chose to create that was not what was going to happen. “The best God could do was what, by way of tempest and leviathan, he eventually did: namely, alter Jonah’s circumstances so they would fit a true subjunctive of freedom that had the desired outcome.”6 And note: Jonah’s libertarian freedom is preserved. God does not cause him to flee to Tarshish nor compel him to change his mind (though the prospects of life in the stomach of the whale was no doubt a strong incentive). Jonah simply does what Jonah will always do in this specific scenario under these circumstances. God does not need to coerce anyone. People freely do what people do, and God plans accordingly—yet not in the sense of the grand master who always can find a way to win the game, no matter what move his opponent makes. The script is logically prior. God brings into being the specific history he wants. In this way the dual concerns articulated by McCann are fulfilled: divine creation of free human actions and the infallible accomplishment of his good purposes.7
But is the notion of divine middle knowledge plausible? McCann does not think so. He advances three objections. They boil down to the decisive difference between a possible world and a real world. If Jonah enjoys genuine libertarian freedom, does it make sense to speak of what he will do in a particular situation? God can run all the simulations he wants, but in the real world people remain free to do otherwise. Possible worlds are but abstractions. The real world is full of novelty, spontaneity, surprise:
We need to remember at this point that the world in which we actually exist (Wr, we’ll say)—that is, the concrete world of things like mountains, desks, and automobiles, and of real events like Booth’s assassinating Lincoln and Jonah’s making off for Tarshish—is not a possible world in the technical definition of that phrase. It is a real world—the only real world, we may suppose—which exists contingently, and whose existence theists hold to be owing to God’s creative will. The possible world that answers to the real world, namely, W0, is by contrast an abstract entity. It exists necessarily, and its existence (we may here presume) is utterly indifferent to that of Wr.8
Consider what it would mean for the Molinist Deity to actualize W0. He would need to know in advance that “he was going to create Jonah, that he would present Jonah with the commission to preach in Nineveh, and that Jonah would instead set sail for the ends of the earth—and similarly for every single contingent fact that pertains to the real world.”9 Creation becomes script. All is programmed. Where’s the fun in that?
Molinism neatly resolves the mystery of divine sovereignty and human freedom. It is easy to see why it might be attractive. But this is not a route that I can take. The image of God deliberating over scenarios of cosmic history and picking the one that best realizes his creative intentions strikes me as anthropomorphic and incompatible with the transcendence of the Christian God. That we, as temporal creatures, should think of God as knowing future events before they happen is no doubt inevitable. Even many of the Church Fathers appear to have thought of divine foreknowledge along these lines, yet the naïveté of this view becomes apparent as we ponder more deeply on divine eternity and the creatio ex nihilo.
The Molinist assumes that God can know a possible history constituted by free agents, even though it does not exist. This idea may seem initially plausible. Did not J. R. R. Tolkien imagine an (almost) fully consistent secondary world of Valar, Elves, Men, Ents, Balrogs, and Orcs? Yet even still, these creatures and characters only became real when Tolkien brought the story to page, at which point they assumed lives and identities of their own. In a letter to W. H. Auden, Tolkien describes the process of creation as one of surprise: “But I met a lot of things on the way that astonished me. Tom Bombadil I knew already; but I had never been to Bree. Strider sitting in the corner at the inn was a shock, and I had no more idea who he was than had Frodo.”
My analogy fails, of course—every analogy walks on three legs. The act of imagination is dependent upon the artist’s experience of a pre-existing world, and the story-teller may alter his narrative innumerable times. But the analogy points to the critical flaw in the Molinist proposal. The assertion of middle knowledge demands, as David Burrell incisively argues, that God possess a knowledge “on the one hand of what might have taken place but does not, and on the other of what has not yet come to pass.” It therefore presupposes “determinations to the object known which only existence could bestow. . . . For if it is only esse [the act of existing] which grants an individual to be the individual it is, then God cannot know such an individual apart from granting it to be.”10 God does not, and cannot, know Moses or the Virgin Mary or Al Kimel “until” he eternally brings these beings into reality. The creatio ex nihilo thus implies, avers Burrell, a “metaphysics of actuality”:
Everything turns, of course, on the ability to characterize the creator of all things as the cause of being, and to understand the capacity to act on the part of self-determining creatures to be a participation in existence freely granted by that primary cause. If, on the other hand, we think of the creator as surveying universes of fully determinate possible beings “before deciding” which one to create, we are at a loss to understand how such “creatures,” possessed of at best “intelligible existence [esse intelligibile],” could be understood to be free actors and hence “fully determinate.” Hence the picture we have of God as creator will imply a specific metaphysics, and vice-versa.11
A metaphysics of actuality, as opposed to a metaphysics of possibilism, requires that we reconceptualize the popular notion of divine foreknowledge. God does not know the future, for the future, by definition, does not exist. He knows, rather, that which he creates, or as Burrell puts it, “God knows—in God’s eternal now—what God does”; and what he does is bestow existence.12 We may not be able to comprehend how this is so—divine eternity and divine creation remain unfathomable to us—but we can see why it must be so.
God, who knows eternally and who knows by a practical knowing what God is doing, knows all and only what is, that is, what God brings into being. Yet by that knowledge, like an artist, God also knows what could be, although this knowing remains penumbral and general, since nonexistent “things” are explicitly not constituted as entities. By definition, an eternal God does not know contingent events before they happen; although God certainly knows all that may or might happen, God does not know what will happen. God knows all and only what is happening (and as a consequence, what has happened). That is, God does not already know what will happen, since what “will happen” has not yet happened and so does not yet exist. God knows what God is bringing about. Yet since our discourse is temporal, we must remind ourselves not to read such a statement as saying that God is now bringing about what will happen, even though what will have happened is the result of God’s action.
What these grammatical remarks underscore is the relentlessly actual character of God’s presence to the world. The eternity thesis has the primary advantage of releasing us from so-called “counterfactuals of freedom,” which amount to asking what “Sally” would do in response to a particular invitation, were she to be created. Besides being at a loss to know how to answer such a question, that her name is in scare quotes reminds us that the question itself is suspect. One cannot even refer to a possible individual as anything more than one of a kind, much less pretend to know how such a “one” would act freely, since such actions can only be predicated of existing persons. So to consider the statements themselves, asking whether the “counterfactuals of freedom” were true or false, would be to treat language without its mooring in reality, and thus represent an extreme “possibilism.”13
Given that we cannot picture God’s knowing of what-will-have-been, it is better to say that God does not now know what will-be and then correct the inevitable implications of that statement by insisting that God will not, however, be surprised by what in fact occurs, since what-will-have-been is part of God’s eternal intent, either directly or inversely, as in evil actions (ST 1.10.10). It remains that we will try, inescapably, to picture such an intent as God “peering into the future,” so we must then insist that the entire point of calling God’s knowledge of the “actual world” that of vision is to call to attention to the “contemporaneity” of an eternal creator with all that emanates from it, whereby what is, is so in its being present to its creator (ST 1.14.13). Yet the manner in which temporal realities are present to their eternal creative source is not available to us, so it is hardly surprising that we lack a model for God’s knowing eternally what-will-have-happened. For that reason I would prefer the arresting picture of God not knowing “the future” (which is also literally true) to the one which seems invariably to accompany the assertion that God does eternally know what will-have-happened—namely, that God can do what no one else can: peer into “the future.” God knows eternally what God brings about temporally; but this assertion does not entail that God now knows what God will bring about. The hiatus between eternity and time presents another manifestation of “the distinction” [between transcendent Creator and contingent cosmos], so it is only plausible in the perspective of creation and never comprehensible.14
At this point we may suspect that Burrell is presenting a version of open theism, but the suspicion is unwarranted. He is, rather, advancing a Thomistic understanding of the eternal Creator who transcends created space and time altogether. God knows all creatures in his ineffable, eternal act of making and in this act is present to all beings and to all history.
(21 July 2015 & 27 December 2016; rev.)
 Hugh J. McCann, Creation and the Sovereignty of God (2012), p. 94.
 Ibid., p. 92.
 Ibid., p. 84.
 Ibid., pp. 84-85.
 Ibid., p. 85.
 Yet we still find ourselves asking, Are human beings really free in the Molinist universe, given that history inexorably plays out just as God foreknows that it will. Herbert McCabe quips: the Molinist assertion of scientia media possesses “the unique merit of denying simultaneously both divine omnipotence and human freedom.” Faith Within Reason (2007), p. 72.
 McCann, p. 89.
 Ibid., p. 90.
 David B. Burrell, Knowing the Unknowable God (1992), p. 104.
 David B. Burrell, Freedom and Creation in Three Traditions (1993), p. 114.
 Ibid., p. 107.
 Ibid., pp. 105-106.
 Ibid., p. 110.