ALFRED J. FREDDOSO, University of Notre Dame

This outstanding book, which incorporates but goes beyond Hasker's extensive previous work on the subject, is a genuinely pivotal contribution to the lively current debate over divine foreknowledge and human freedom. If you plan to plunge into this debate at any time in the foreseeable future, you will have to take account of God, Time, and Knowledge.

Hasker's book has three commendable features worthy of immediate note. First, it contains a carefully crafted overview of the recent literature on foreknowledge and freedom and so can serve as an excellent introduction to that literature. Second, it is tightly reasoned and brimming with brisk arguments, many of them highly original. Third, it correctly situates the philosophical dispute over foreknowledge and freedom within its proper theological context and in so doing highlights the intimate connection between the doctrines of divine omniscience and divine providence. This is especially significant because much of the recent literature has been oblivious to the complications that arise once we take seriously the traditional theistic tenet that God, far from being a passive observer of the universe, is in fact its sovereign and provident governor. In short, a comprehensive account of freedom and foreknowledge must not only cohere with but also illuminate the notion of divine providence.

Yet it is precisely here that God, Time, and Knowledge is controversial and to my mind profoundly disturbing. For in opposition to every important classical Christian philosopher, Hasker denies in the end that God has infallible knowledge of exactly how the contingent future will turn out. True, he is not the only distinguished contemporary Christian philosopher to have travelled this road; Richard Swinburne, John Lucas, and Peter Geach have preceded him. What's more, Hasker provides the most extensive argument I have seen for the claim that the denial of divine foreknowledge is fully consistent with Christian orthodoxy and, more specifically, with "an affirmative, constructive and (I believe) religiously satisfying conception of God's knowledge of the world and of his providential governance of the world" (p. 186). Only Geach's Providence and Evil is in the same class here. But Geach does not attempt an exhaustive refutation of the alternatives, as /100/ Hasker does; nor does he argue in such detail for the 'risk-taking' account of divine providence.

Still, I cannot hide my dismay. The likes of Justin Martyr, Tertullian, Origen, Augustine, Anselm, Bonaventure, Aquinas, Scotus, Ockham, Luther, Calvin, Molina, Banez, Suarez, Arminius, Leibniz, and Edwards surely realized that they could spare themselves a lot of philosophical grief if only they would repudiate divine foreknowledge and with it the traditional understanding of divine providence, according to which every event that transpires in the universe, including every free action, is either knowingly intended or knowingly permitted by God prior to creation. Yet not one of these Christian intellectual heroes so much as entertained such a drastic expedient; to the contrary, the very thought of it would have appalled them. Were they less enlightened than we are about the Christian Faith as it pertains to providence and foreknowledge? Were they, as Hasker intimates (p. 191), the unwitting victims of an over-hellenized theology? (Even Luther and Calvin?!) It verily takes one's breath away to suppose so. Yet Hasker and his co-travellers apparently do suppose so.

I will return to this matter below. First, however, I want to trace the line of reasoning by which Hasker reaches his novel conclusion.

After cursorily surveying the historically important attempts to reconcile God's foreknowledge with human freedom (Chapter 1), Hasker devotes the bulk of the book (Chapters 2-9) to an extended argument for the thesis that no philosopher has ever succeeded in showing that God's putatively exact and infallible knowledge of future contingents is compatible with human freedom conceived of in a strongly libertarian way.

He first (Chapter 2) attacks Molina's theory of middle knowledge, according to which God's free knowledge of the actual (or absolute) contingent future results from His free act of will as guided by His antecedent natural knowledge of necessary truths and His antecedent middle knowledge of conditional future contingents, i.e., propositions specifying, for any possible created agent A and complete set of circumstances C in which A is able to act indeterministically, how A would in fact act if placed in C. Hasker begins with Molinism because it is the only theory that promises to harmonize God's foreknowledge with strongly libertarian freedom in a way that preserves the traditional understanding of providence.

I will not here rehearse or impugn Hasker's many objections to the theory of middle knowledge, in part because I have already had occasion to identify what I take to be problematic about his main line of criticism, and in part because this section of God, Time, and Knowledge has generated incisive rebuttals by Thomas Flint and Rod Bertolet.1

Having polished off Molinism, Hasker next (Chapter 3) argues that comprehensive simple foreknowledge (free knowledge in Molina's sense) is by /101/ itself useless, since it comes too late in the order of explanation to guide God's choice of an all-embracing providential plan. Through His simple foreknowledge God knows what will in fact be effected by all the causal activity in the universe--including His own, since according to traditional Christian theology, every effect produced in the universe has God Himself among its immediate causes.2 Hence, His simple foreknowledge, unlike His natural knowledge and middle knowledge, is subsequent to His free act of will and thus cannot guide that act.3 So even before confronting the various 'compatibilist' replies to the 'incompatibilist' thesis that there is a conflict between foreknowledge and libertarian freedom, Hasker takes himself to have undermined the main theoretical reason for affirming comprehensive simple foreknowledge in the first place, viz., the crucial role it supposedly plays in God's governance of the world.

As I see it, this chapter constitutes an important advance which, along with recent defenses of middle knowledge, should create some intellectual discomfort for those many strong libertarians who affirm God's foreknowledge but reject middle knowledge. Hasker and the Molinists may occupy opposite ends of the foreknowledge/providence spectrum, but they are united in their disdain for what they take to be the feeble positions in the middle.

After carefully laying out the incompatibilist argument (Chapter 4), Hasker scrutinizes four compatibilist replies which attribute to God's infallible beliefs about absolute future contingents some sort of asymmetric dependence on what free agents will do, and which in this way try to subvert the incompatibilist assertion that because God's beliefs are themselves part of the fixed and 'accidentally necessary' past, they keep human actions from being free.

This section begins (Chapter 5) with a critique of the Ockhamist contention that God's beliefs about future contingents, while part of the past, are not part of the accidentally necessary past--so that human agents even now have the power to determine just which propositions about their future free actions God has always and infallibly believed, and just which such propositions He has never believed. Ockhamists thus divide the past into (i) an accidentally necessary part composed of 'hard' facts, propositions which are not dependent for their truth on what will happen in the future and which thus can no longer be made false, and (ii) a contingent part made up of 'soft' facts, propositions which are future-dependent and which even now can be (though they will not be) made false by free agents. As the Ockhamists see it, God's past beliefs about future contingents are soft facts; and so even if God has always and infallibly believed that, say, Peter will refrain from watching the fifth game of the NBA Finals tonight, Peter still has the power to watch the game and with it the power to bring it about that God has never believed that he will refrain from watching it. /102/

Hasker painstakingly examines the much-discussed though elusive distinction between hard and soft facts and concludes that on any plausible rendering of that distinction God's beliefs about the contingent future will, pace the Ockhamists, turn out to be hard facts and thus elements of the accidentally necessary past. I find Hasker's argument here unpersuasive, since it turns on a rather doubtful use of the notion of conceptual necessity. Still, I accept his conclusion, having come to believe on other grounds that the hard fact/soft fact distinction, though perhaps useful in specifying the causal prerequisites of free action, provides no warrant for denying that every truth about the past is accidentally necessary.

Unlike Ockhamism, the other three compatibilist replies do not depend on partitioning the past into hard and soft facts. The second reply, propounded by George Mavrodes, is that, for all we know, none of the past is accidentally necessary, and so human agents may well have the power to prevent events that have already occurred from ever having occurred--including events that any version of the hard fact/soft fact distinction would count as paradigmatic hard facts. Hasker (Chapter 7) eschews the obvious (and, I believe, correct) rejoinder that this reply, even if coherent, is so wildly outlandish that only sheer desperation could prompt a compatibilist to adopt it. He responds instead--more insidiously, as it turns out--that Mavrodes's proposal can be true only if we read the term 'power' in a weaker sense than that presupposed by the incompatibilist argument. To return to our example, assume that Peter retains his general causal ability to get up, walk over to his television set, tune it to the channel featuring the game, etc. Some might say that this is all the power he needs in order to be free to watch the game, even despite God's having believed from eternity that he will refrain from watching it. Not so, counters Hasker. God's past belief logically (though not causally) precludes Peter's exercising the relevant causal abilities in his concrete circumstances, and so Peter lacks the 'full-blooded' power to watch the game. Hence, when he refrains from watching it, he does not do so freely. (I will return to this in a moment.)

The third compatibilist reply concedes that God's infallible beliefs about future contingents are part of the accidentally necessary past, but denies that they thereby deprive free agents of the power (in the sense required for freedom) to do otherwise. This reply, while not the exclusive property of Molinists, finds a natural home within the theory of middle knowledge. So assume for now that Molinism is true and that God believed from eternity, on the basis of His middle knowledge and free act of will, that Peter will refrain from watching the game tonight. And assume further, as we did above, that there are no causal barriers to Peter's watching the game, so that when he in fact refrains from watching it, he does not do so by natural necessity. Then, says the Molinist, despite God's past belief, Peter has the power (in /103/ the sense required for freedom) to watch the game. However, pace Mavrodes and the Ockhamists, this is not to say that Peter has the power (Power A) to bring it about that God never had the belief in question; after all, God's having had that belief is now accidentally necessary. Rather, what Peter has is the power (Power B) to do something, viz., watch the game, such that if he had been going to do it in the relevant circumstances, then God would have known from eternity, via His middle knowledge, that Peter would freely watch the game if situated in those circumstances. Thus, God's middle knowledge and free knowledge, though not directly brought about by what Peter will freely do, are nonetheless 'counterfactually sensitive' to what he will freely do. So even though God's belief about Peter is now part of the fixed past, still, as long as the causal prerequisites for free action are satisfied, Peter has the power to watch the game; therefore, when he in fact refrains from watching it, he does so freely. And the theory of middle knowledge provides a metaphysical model on which all these claims come out true.

Hasker replies in Chapter 6 by invoking an "unassailable" (p. 114) power entailment principle according to which Peter's having Power B entails his also having the admittedly objectionable Power A. Thomas Flint has shown that this power entailment principle, far from being indisputable, cannot even be reasonably assessed until we know more precisely what the slippery locution 'bring about' means as it occurs in the principle.4 However, here I want to take a different tack by displaying how intimately Hasker's use of the principle in Chapter 6 is connected with his later appeal to logical preclusion in Chapter 7.

The 'logical preclusion' argument goes roughly as follows: "Because God's infallible belief is part of the accidentally necessary past, it is an uneliminable circumstance of Peter's action. What's more, that belief is incompatible with, and thus logically precludes, Peter's watching the game. But if an uneliminable circumstance logically precludes an agent's performing a given action, then the agent lacks the power (in the sense required for freedom) to perform that action; for having the power to perform the action would entail having the power to eliminate an uneliminable circumstance. Hence, Peter lacks the power to watch the game; and so when he refrains from watching it, he does not do so freely."

Hasker's power entailment principle merely serves to codify this argument. But how are we to evaluate the argument? Consider another example adduced by Hasker: Thomas wants to marry Edwina but is logically precluded from doing so by the fact that she is already married and that, given the laws governing the institution of marriage, it is impossible for anyone to be married to more than one person at the same time; hence, Thomas lacks the power (in the sense required for freedom) to marry Edwina.

Now ask yourself just which of his basic causal powers Thomas is precluded /104/ from exercising solely by virtue of the fact that Edwina is already married. The correct answer, I think, is: "None at all." Thomas can (if he is willing to lie) secure a marriage license, appear with Edwina before a duly designated official, profess the marriage vows, and, in general, carry on exactly as if there were no legal impediment to the marriage. To be sure, the law renders him unable to marry Edwina. However, this 'inability' is due not to a lack of causal power on Thomas's part, but to the absence of a legal circumstance, viz., Edwina's being unmarried, that is required in order for his exercise of the relevant causal powers to count as his marrying Edwina. The question now becomes: Is there an analogous 'inability' in the case of Peter? And the answer, it seems to me, is that there is not. For, as in the case of Thomas, Hasker is willing to admit that Peter retains all the basic causal capacities required for his watching the game, and that all the external causal prerequisites (a properly functioning television set, a sufficient supply of electricity, etc.) are satisfied. Yet in Peter's case there is no obvious analogue of the legal constraints that were operative in the case of Thomas and Edwina. Given that Peter is fully capable of moving his body in the ways appropriate for watching the game and that none of the external causal prerequisites is absent, there seems to be nothing that would keep his exercise of the relevant causal powers from counting as his watching the game. In brief, it is extremely difficult to see how the 'circumstance' of God's past belief could play in this case a role corresponding to that played in the other case by the circumstance of Edwina's already being married.

In the case of Thomas and Edwina, then, we have a clear picture of how logical preclusion figures in Thomas's inability to marry Edwina, as well as a clear understanding of the nature of that inability; Thomas can exercise his basic causal powers all right, but his doing so will still not count as his marrying Edwina. By contrast, in Peter's case we have no such clarity at all. Hasker seems to suggest that even though Peter, like Thomas, retains the relevant causal powers, he, unlike Thomas, is not able to exercise them--and this because of God's past foreknowledge. But how is this supposed to work, exactly? Does God's foreknowledge somehow insinuate itself into the causal order to literally prevent Peter from exercising his power to watch the game? No, since as Hasker emphasizes, his own incompatibilist argument "in no way depends on the assumption that God's belief causes human actions" (p. 141). Well, then, just how does it work?

These considerations explain why I am unimpressed by Hasker's charge that "the central idea of the Molinist position, as explicated by Freddoso, seems to be that any determinism that results from divine foreknowledge is not a 'serious' determinism, because it is not causal" (p.141). True enough, but only because I do not have even a faint idea of how logical preclusion by divine foreknowledge--as opposed to, say, logical preclusion by legal /105/ statute--is supposed to result in an agent's inability to do something. Nor am I swayed by Hasker's accusation that "a compatibilist such as Molina or Freddoso is claiming that Peter can have the power to [watch the game], even though it is logically impossible that he should exercise that power under the existing circumstances" (p. 141). I admit, of course, that there is no possible world in which God believed from eternity that Peter will not watch the game and in which Peter watches the game nonetheless. No surprise there. But I deny that there is any more interesting sense in which it is impossible for Peter to exercise his power to watch the game. More precisely, Peter's exercising the basic powers required to watch the game is fully compossible with all the other causal activity, including God's, that has ever occurred in the past or is anywhere occurring in the present. This is as much as any strong libertarian could possibly hope for. So even while conceding that the notion of a basic causal power needs further analysis,5 I find Hasker's complaints entirely unmoving.

The fourth compatibilist reply is the so-called 'eternalist solution', according to which God's knowledge of future contingents is not part of the past at all, but is instead located in the presentness of God's eternity, which admits of no past or future. Thus, God's beliefs about future contingents have no necessity of a sort that rules out human freedom. In reply, Hasker argues that even though the doctrine of God's eternality is intelligible (Chapter 8) and perhaps even sufficient to defuse the incompatibilist argument as it stands (Chapter 9), it does not by itself provide a sufficient foundation for the traditional understanding of divine providence. Interestingly, this is the same conclusion Molina had reached in Disputation 49 of Part IV of his Concordia. And I must confess that it seems right to me, too.

This brings us at last to Hasker's defense of the 'risk-taking' (or, as I prefer to call it, the 'damage control') account of divine providence (Chapter 10). Hasker realizes that he is on shaky ground here. A sure sign of this is that he becomes exceedingly, though rather delightfully, tendentious. So allow me to reply in kind.

On the risk-taking account of providence, God lacks exact and infallible knowledge of the contingent future. Yet given His thorough familiarity with present causal tendencies and His clear grasp of His own providential designs, He is almost sure about how the future will turn out. In fact, He is even pretty sure about whether or not we human beings (including, presumably, Jesus Christ) will do freely what He intends us to do; but He is strong enough to make us do it anyway, if it suits Him. As Hasker puts it, "God is perfectly capable of making someone an 'offer he can't refuse'" (p. 196). So even if the world begins to go really badly, God, though disappointed, is fully capable of controlling the damage. What's more, His prophecies about future free actions--including sinful ones, such as Peter's denial of Christ--are almost /106/ sure to be fulfilled; and, once again, even if the improbable happens and God turns out to have been mistaken in so prophesying, He is powerful enough to put things back on track. Likewise, even allowing that some of His ends (e.g., the triumph of grace over sin) depend crucially on specific free actions being performed by specific human beings (e.g., Abraham, Moses, Jeremiah, Mary the Mother of God, Peter, Paul, etc.), He can be almost sure, given His unusually high degree of knowledge and power, that those ends will be realized. And if all this strikes you as excessively anthropomorphic and as coming dangerously close to turning God the Father into the Godfather, then according to Hasker you have not been sufficiently dehellenized.

Not content simply to promote his own watered-down account of providence, Hasker heaps scorn upon the traditional account, according to which "our most ennobling achievements are just the expected printouts from the divine programming" (p. 199). In his zeal, he even resorts to the "Hitler" defense. (One can imagine a medieval Hasker conjuring up the "Genghis Khan" defense.) After running roughshod over hundreds of pages of the best scholastic theology by declaring ex cathedra that those who adhere to the traditional account cannot distinguish what God intends from what He merely permits, Hasker concludes that they "cannot avoid saying ... that God specifically chose Hitler to become leader of the Third Reich and instigator of the Holocaust" (pp. 199-200). Really now. More to the point, ask yourself whether Hasker's risk-taking account fares any better with regard to Hitler. Once Hitler accedes to power and gets the Holocaust rolling on its grisly way, even the risk-taking God, who is after all pretty knowledgeable, should have a crystal-clear idea of the further specific evils that are almost certain to occur. So if He does not intervene early on to stop Hitler, this can only be because of some worthy (though very hidden) purposes He has in mind. In that case, would it not be just as true on the risk-taking account as on the traditional account that "God has deliberately and with full knowledge chosen that these good purposes shall be fulfilled through a plan that entails the actual occurrence (not just the possibility) of specific evils" (p. 200)? Let's face it. Hitler is a problem for everyone.

Finally, after implicitly saddling the traditionalist with Eleonore Stump's incredibly strong suggestion that the sufferings of each human person are outweighed by a greater good which those very sufferings produce for that same person, Hasker endorses Michael Peterson's more congenial 'risk-taking' theodicy, according to which a world created by God might be literally teeming with genuinely gratuitous evils. I would have thought--with, say, Aquinas--that the most plausible theodicies lie somewhere between Stump's and Peterson's.

I am not convinced, then, by Hasker's closing proclamation that "[Peterson's] theodicy, and the understanding of divine providence which it involves, are clearly acceptable as judged by the canons of orthodox, mainstream /107/ Christian theology" (p. 205). In a word, even if, contrary to what I have urged, Hasker's arguments for incompatibilism were indeed compelling, it would be still be incumbent upon Christian philosophers to preserve the classical understanding of divine providence, which is a linchpin of the Christian Faith and of the traditions of intellectual inquiry it has inspired. As I see it, this is the main theological lesson to be learned from Hasker's remarkably provocative book.

1. See Luis de Molina, On Divine Foreknowledge (Part IV of the "Concordia"), translated, with an introduction and notes, by Alfred J. Freddoso (Ithaca, NY: Cornell University Press, 1988), pp. 75-78; Thomas P. Flint, "Hasker's God, Time, and Knowledge," Philosophical Studies 60 (1990): 103-115; and Rod Bertolet, "Hasker on Middle Knowledge," Faith and Philosophy 10 (1993), Number 1. In fairness to Hasker, I should also mention his "Response to Thomas Flint," Philosophical Studies 60 (1990): 117-126, as well as the amicus brief filed by Robert Merrihew Adams in "An Anti-Molinist Argument," Philosophical Perspectives 5 (1991): 343-353.

2. For more on this, see my "God's General Concurrence with Secondary Causes: Why Conservation is not Enough," Philosophical Perspectives 5 (1991): 553-585.

3. In "Divine Providence and Simple Foreknowledge," Faith and Philosophy 9 (1992): Number 3, David Hunt argues that if we reject the claim that God has complete providential sovereignty over all events, including free actions--a claim he stigmatizes as `hyper-Calvinist' but might just as well have called `Augustinian' or `Thomistic'--we may coherently maintain that simple foreknowledge can contribute to "the strongest providential control compatible with there being free agents other than God." Perhaps. But Hunt is mistaken in his undocumented assertion that he is defending the "traditional" position.

4. See Flint, "In Defence of Theological Compatibilism," Faith and Philosophy 8 (1991): 237-243; and "Hasker's God, Time, and Knowledge," esp. pp. 112-114. For the record, the relevant power entailment principle is this: If p is true and entails q, then if it cannot be in anyone's power to bring it about that p is false, it cannot be in anyone's power to bring it about that q is false. So given that (i) God's past belief entails that Peter will refrain from watching the game and that (ii) Peter cannot make it the case that God never held that belief, Peter cannot make it false that he will refrain from watching the game.

5. For more on this point, see Nelson Pike, "A Latter-Day Look at the Foreknowledge Problem," International Journal for Philosophy of Religion, forthcoming.