William Hasker has presented influential arguments against divine foreknowledge and middle knowledge. I argue that his objections are fallacious. With respect to divine foreknowledge, three central issues arise: temporal necessity, power entailment principles, and the nature of free will. In each case Hasker's analysis is defective. With respect to divine middle knowledge, Hasker presents four objections concerning the truth of counterfactuals of freedom. Against Hasker I argue that such propositions are grounded in states of affairs belonging to the actual world logically prior to its full instantiation and are contingently true or false.
Source: "Hasker on Divine Knowledge." Philosophical Studies 67 (1992): 57-78.
Introduction
The most positive feature of William Hasker's recent God, Time, and Knowledge, it seems to me, is its focusing the wide-ranging discussion of the diverse issues raised by the problem of divine foreknowledge and future contingents down to a few, central issues. In this paper, I shall argue that the adherent of God's foreknowledge of future contingents has successfully defended his position on these issues against Hasker's attacks.
I. Theological Fatalism
In the dispute over the argument for theological fatalism --Necessarily, God has always believed that p; Necessarily, if God has always believed that p, then p; Therefore, necessarily p--Hasker isolates three central issues separating the disputants: temporal necessity, power entailment principles, and the nature of free will. Let us examine each of these.
Issue 1: Temporal Necessity
Key to the argument for theological fatalism is the claim that God's past belief is characterized by a sort of temporal necessity that renders it counterfactually inviolable to future events. Since the Ockhamist and Molinist claim that this is not the case--that it lies within my power to act in such a way, that were I so to act, God's belief would have been different from what it in fact was--, it is incumbent upon the theological fatalist to show why such a position is not possible, that is to say, why God's past belief must be categorized as a "hard" rather than a "soft" fact.
Hasker's procedure is to explicate the distinction between hard and soft facts by a series of steps. First, he explicates the notion of an immediate or future-indifferent proposition:
H1: An elementary proposition is future-indifferent iff it is conceptually consistent with there being no times after the present, and also with there being times after the present.
Next he explicates the notion of a "hard fact," asserting,
H5: Any future-indifferent proposition that is true is a hard fact.
"No argument will be given for (H5)," he says, "since it is a common assumption among those who discuss hard and soft facts that true propositions that are 'really about the past' are such that it cannot be in anyone's power to render them false."{1} But Hasker's procedure is all too quick here. For Plantinga has shown that on Ockhamist principles a fact's being future-indifferent is no guarantee of its being hard,{2} and I have elsewhere provided examples of future-indifferent facts which in various contexts are plausibly regarded as soft.{3} Hence, (H5) is simply unacceptable, and since Hasker gives no argument for it, we are free to reject it out of hand.{4}
But even if we accepted (H5), God's past belief is prima facie not future indifferent and therefore not a hard fact. Hasker tries to elude this consequence by substituting the name "Yahweh" for "God" in "God has always believed p." As a non-connotative proper name, "Yahweh" carries with it none of the connotations of infallibility that "God" does, so that "Yahweh has always believed p" is future-indifferent and hence expresses a hard fact. But such an argument seems clearly inconclusive. For Hasker assumes that proper names are non-connotative. He does not even mention, much less refute, views that take proper names to have a Fregean sense. On such a view, one may be unaware that the sentence "Yahweh has always believed p" expresses the proposition "God has always believed p" due to a conceptual inadequacy on one's part.{5} Until Hasker offers some refutation of such alternative theories of proper names, his argument is ineffectual. But secondly, even if "Yahweh" were a non-connotative proper name, how is this relevant to the issue of whether "Yahweh has always believed p" expresses a hard or soft fact? For on Direct Reference theories "Yahweh" refers to the same individual as "God" and the two sentences express the same singular proposition; if the referent believes p, then p is true. What is of importance here is the referent, not the means of fixing the reference.
Therefore, it seems to me that Hasker's analysis of temporal necessity is not at all compelling and that no incoherence in the Ockhamist/Molinist position has been shown.
Issue 2: Power Entailment Principles
Hasker distinguishes the notion of "bringing it about that" from any sort of causal relation or mere counterfactual dependence. "The core idea in the notion of 'bringing about' is the notion of something's being the case in consequence of what an agent does. . . ."{6} The power to "bring it about that" lies somewhere in between counterfactual and causal power.
Now Hasker wants to argue on the basis of his power entailment principle
PEP5: If it is in S's power to bring it about that P, and "P" entails "Q" and "Q" is false, then it is in S's power to bring it about that Q
that the Ockhamist solution to theological fatalism, namely,
1. S has it within his power to act in such a way that, were he to act in that way, God would not have believed that p,
entails the assertion of S's ability to bring about the past. More specifically, S's ability to do something other than what God foreknows he will do entails S's ability to bring about God's past beliefs.
The question here is whether (PEP5) is true. I doubt that it is. Consider Thomas Flint's objection that S's power over P could be causal, but his power over Q merely counterfactual:
If two propositions are logically equivalent and I have power over the truth of one of them (i.e., its truth is up to me), then it does seem clear that the truth of the other one is within my power as well; what does not seem clear is that I need to have power in the same sense of 'power' over the second as over the first. Suppose I have causal power over the truth of one of two logically equivalent propositions; is it not sufficient that I have counterfactual power over the other? Is that not enough for me to say that each of them is such that its truth is up to me?{7}
If Flint is correct, then it is "up to me" what God believes concerning some free action of mine, but I do not have the power to bring about God's past belief concerning that action. Hasker replies, "On the one hand, power to bring about need not be causal power; on the other hand, the counterfactual dependency relation (and therefore also 'counterfactual power') is not 'enough for me to say that each of them is such that its truth is up to me'."{8} This reply misses Flint's point. We may agree that "bringing about" does not imply causal power and that my counterfactual power over something does not imply that thing is up to me. Flint's point is that the composite state of affairs of S's being able to bring it about (even non-causally) that P and its being the case that P ® Q implies that Q is up to S, even though S cannot bring about Q. Accordingly, Flint would accept no more than
PEP5*: If it is in S's power to bring it about that P, and "P" entails "Q" and "Q" is false, then it is up to S whether it be the case that Q.
Hasker thus fails to show why power to "bring it about that" is closed under entailment.
Second, counterexamples to (PEP5) can be offered, although space does not permit a discussion of them here.{9} But even if we concede (PEP5), whether it has deleterious consequences for divine foreknowledge depends on how we adjudicate issue three.
Issue 3: The Nature of Free Will
Hasker argues that in order to avoid theological fatalism, the Ockhamist must claim that one has the power to bring about the past, worse than that, the "power to bring about past events that have not occurred."{10} Hasker is willing to concede for the sake of argument the power to bring about actual past events, but "What needs to be explained, but has not been explained, is how it is possible that God has always believed a certain thing, and yet it is in someone's power to bring it about that God has not always believed that thing."{11} Hasker thinks the Ockhamist must hold that S has the power to bring it about that whereas it was true at t1 that God had always believed p, it was no longer true at t2 that God had always believed p. Thus, S must have the power to eliminate the past fact of God's believing p, which is the power to alter the past, an evident absurdity.
Hasker recognizes that Ockhamists protest that they assert no such power, and this fact, which bewilders him, leads Hasker to infer that Ockhamists have a different concept of power and freedom than the standard libertarian analysis. When Hasker speaks of power,
The power in question is the power to perform a particular act under given circumstances, and not a generalized power to perform acts of a certain kind. . . . In general, if it is in N's power at T to perform A, then there is nothing in the circumstances18 that obtain at T which prevents or precludes N's performing A at T.
----------
18It will be recalled that the circumstances that obtain at T comprise all and only the hard facts with respect to T.{12}
In this sense of power, one does not have it within his power to act differently than God foreknows one will. In a different sense of power, in the sense of general abilities, "I may perfectly well have a power . . . to do something even though it is either logically or causally impossible that I exercise the power under the circumstances that obtain at a particular time."{13} But the problem with this sense of power, he argues, is that it is insufficient for libertarian free will. In this sense of power,
. . . Peter can have the power to refrain from sinning even though it is logically impossible that he should exercise that power under the existing circumstances. But if one has the 'power to do otherwise' only in that sense--the sense in which having the power does not guarantee that it is possible for the power to be used--then the central idea of libertarianism . . . has been lost. Once again, we see that the compatibilist on foreknowledge cannot consistently affirm libertarian free will.{14}
It is remarkable how clearly the echoes of Richard Taylor's fatalism resound through these passages.{15} Hasker's analysis of the notion of "within one's power"--which Taylor complained his critics never understood--is virtually the same as Taylor's and is thus infected with the same deficiencies.
The best way to get at this problem is by drawing some helpful distinctions which were well-known to medieval discussants of these issues. Foremost is the distinction between the sensus compositus and the sensus divisus of a proposition. Hasker's failure to differentiate these senses leads him into muddles. For example, consider the problem of the unchangeability of the past and future. Hasker tries to explain that the unchangeability of the past is not a mere tautology and the changeability of the future not a self-contradiction because the past is a concrete totality which is, while the future is a realm of mere possibilities.{16} This affirmation of an A-theory of time does not, however, bring any clarity to the logical issues raised. Utilizing the medieval distinction between the senses, however, consider the proposition
2. A future event can fail to occur.
In sensu diviso, (2) means
3. Possibly, an event, which is future, will fail to occur
and is true if the event is contingent. But taken in sensu composito, (2) means
4. Possibly, an event which is future will fail to occur,
which is necessarily false. Thus, what is at issue with regard to the misleading notion of "altering the future" is whether one has the power to prevent a future event in sensu diviso. One can prevent the event, but were one to do so, then the event would not be future. To say that one cannot prevent a future event in sensu composito is merely to assert that one cannot bring it about that the event both will and will not occur--hardly a restriction on human freedom! Now consider
5. A past event can have failed to occur.
In sensu composito, (5) means
6. Possibly, an event which is past has failed to occur,
which is a self-contradiction. In sensu diviso, (5) means
7. Possibly, an event, which is past, has failed to occur.
It is clearly this latter sense that is at issue when Hasker raises the question concerning the "power to bring about past events that have not occurred"--otherwise, this phrase would be as self-contradictory as "square circles." The so-called unalterability of the past in sensu composito amounts to nothing more than the logical impossibility of bringing it about that an event has both occurred and not occurred. This trivial sense is irrelevant to considerations of power and freedom. The really interesting question is whether we have it within our power to postvent a past event in sensu diviso. In such a case one can bring it about that an event, which is past, did not occur, but were one to do so, then it would not have been a past event.
In so far as such postvention of the past relies upon retro-causation, we may certainly agree with Hasker that considerations of time and objective becoming rule out causal postvention of the past. But Hasker seems to have forgotten that the "bringing about" relation is non-causal. In this weak sense of "bring about," we do according to (PEP5), have power over the past, for as Freddoso has shown, we bring about the past truth of future-tense propositions by bringing about the truth of present-tense propositions which entail them.{17} It was Taylor's failure to discern this power over the past in sensu diviso that proved fatal to his fatalism.
But is there not a similar fatal fallacy in theological fatalism? Consider
8. An event foreknown by God can fail to occur.
In sensu composito, this means
9. Possibly, an event which is foreknown by God will fail to occur,
which is self-contradictory. But in sensu diviso, (8) means
10. Possibly, an event, which is foreknown by God, will fail to occur,
which may be true. Thus, my ability to prevent the event is not the ability the bring about the self-contradictory state of affairs that God foreknew the event and the event does not occur. It is the power to prevent the event, which is foreknown by God, and were I to do so, it would not have been foreknown by Him.
On the assumption of (PEP5), the above implies that one has it within one's power to bring it about that the past should be different than it is, in that one can bring it about that God should have different beliefs than He has. This is not the power to alter or eliminate past events in sensu composito, which is absurd, but the power to bring it about that the past would have been different. For by acting differently now, one brings about the truth of different present-tense propositions and indirectly the past truth of different future-tense propositions. Since God is essentially omniscient, one thereby indirectly brings it about that He believed different propositions than He does. What is objectionable about that?
Hasker would reply that it is not within my power under the circumstances to act differently now. But the fallacy in in this reply may be seen by means of a second distinction, closely related to the first, which the medievals discerned, that between necessitas consequentiae and necessitas consequentis or the necessity of a hypothetical inference versus the necessity of the consequent of the hypothetical. Thus the proposition
11. If God foreknew Peter would sin, then Peter cannot refrain from sinning,
properly understood, means
12. Necessarily, if God foreknew Peter would sin, then Peter does not refrain from sinning.
Hasker is misled by (11) into asserting a necessitas consequentis which he interprets as a abridgement of Peter's personal power. But what is impossible is not Peter's refraining from sin, but the composite state of affairs of God's foreknowledge of Peter's sin and Peter's refraining. That is to say, the proposition
13. Peter can refrain from the sin which God foreknew he would commit
is false in sensu composito, but true in sensu diviso.
Of course, (13)'s truth in sensu diviso implies that a backtracking counterfactual is in order here, in that since the composite state of affairs is impossible, Peter's power to refrain implies that were he to refrain, the circumstances (God's foreknowledge) would have been different. Such a counterfactual is justified since there are no possible worlds in which God errs. Of course, Hasker will insist, as the footnote in the above citation reminds us, that the circumstances he is talking about involve exclusively hard facts so that while the Ockhamist solution works for logical fatalism, it fails for theological fatalism. But such a reply only throws us back to the question of whether God's past belief is a hard fact, and we have seen that Hasker's inadequate analysis of that notion failed to provide any convincing argument against the Ockhamist position.
In short, the Ockhamist does not at all operate with a non-libertarian understanding of power or freedom. Once the proper distinctions are drawn, we see that Hasker has in no wise shown that one does not have the power to bring it about that God should have believed differently than He did.
II. Middle Knowledge
The doctrine of middle knowledge plays a foundational role in discussions of divine prescience, providence, and predestination. But Hasker lodges four objections against the doctrine of middle knowledge:{18} (i) What, if anything, is the ground of the truth of counterfactuals of freedom? (ii) Crucial counterfactuals of freedom, if true at all, are necessarily true, which is incoherent. (iii) Counterfactuals of freedom cannot guide God's creation of the world because it is only by deciding which world to create that God settles which world is actual and therefore which counterfactuals are true. (iv) Either the truth of counterfactuals of freedom is brought about by the relevant agent or not. But it cannot be brought about by the agent; and if it cannot be brought about by the agent, then the agent's freedom is obviated. Therefore, there are no true counterfactuals of freedom. Let us consider then each of these objections.
Objection (i)
Hasker wants to know what makes counterfactuals of freedom true. So stated, this is not much of an objection; it is just a question which ought to prompt further philosophical inquiry. Ignorance of an answer to the question demonstrates no incoherence in the position. In any case, it seems to me that the answer is that counterfactuals of freedom are true in virtue of what makes any non-truth-functional proposition true, namely, correspondence. Tarski's T-schema for truth, Tp º p, applies to counterfactuals just as it does to any atomic proposition. The proposition, "If I were rich, I should buy a Mercedes," if true, is true in virtue of the fact that if I were rich I should buy a Mercedes. True counterfactuals correspond to reality and are therefore true; false counterfactuals fail to correspond and are therefore false.
Of course, if might be said that this answer only pushes the question back a notch: now we must ask, what makes certain counterfactual states of affairs obtain? Hasker says,
In order for a (contingent) conditional state of affairs to obtain, its obtaining must be grounded in some categorical state of affairs. More colloquially, truths about 'what would be the case . . . if' must be grounded in truths about what is in fact the case.{19}
For example, ". . . the truth of causal conditionals, and of their associated counterfactuals, are [sic] grounded in the natures, causal powers, inherent tendencies, and the like, of the natural entities described in them."{20}
Hasker's principle, as stated, is clearly false because we can entertain counterfactuals about what the world would be like were different laws of nature or boundary conditions to obtain. For example, consider
14. If a meter stick were set in motion relative to the aether, then it would undergo a FitzGerald-Lorentz contraction.
This counterfactual is true, but not virtue of what is in fact the case, since the classical aether does not exist. It might be said that the categorical state of affairs which in part grounds it is the state of affairs
15. The aether has the property of immobility.
But the problem is that (15) is in fact false, since there is no aether and merely possible objects neither exist nor have properties. What is true is rather
15'. If the aether existed, it would have the property of immobility.
But (15') is itself a counterfactual state of affairs, so that one counterfactual state of affairs is grounded by another.
Perhaps Hasker would merely recast his principle, however, such that a counterfactual state of affairs must be ultimately based on the individual essences of the things referred to in the counterfactual proposition. Because the essence "aethericity" includes the property of immobility, (15') is true and because (15') is true, (14) is true. But again, one can think of counterfactuals from the natural world for which this does not seem to be the case. Consider, for example, the Einstein-Podolsky-Rosen thought experiment with twin photons traveling in opposite directions. If we measure the momentum of photon 1, then photon 2 must possess the same momentum, even though no measurement is carried out on it. But we could just as easily have measured instead the position of photon 1, and then photon 2 would have had a precise position. So photon 2 must possess simultaneously both position and momentum. Notice that counterfactual reasoning plays a key role in this argument. Since quantum physics prohibits our measuring both the momentum and position of photon 1 simultaneously, all the physics allows us to assert is
16. Since the momentum of photon 1 is measured to be a certain value, photon 2 has a similar value
or
17. Since the position of photon 1 is measured to be a certain value, photon 2 has a similar value.
But what the thought experiment requires us to say is that if, say, (16) is true, it is also true that
18. If we had chosen to measure instead the position of photon 1, then photon 2 would have possessed a certain value for its position.
To most thinkers, (18) seems intuitively obvious, but one will search in vain for anything in the natures of quantum entities to ground it. Now maybe Einstein, Podolsky, and Rosen were wrong to assume (18); maybe (18) is false. But it is certainly not obviously false, and the three scientists could hardly be called irrational or their position incoherent because they accepted it. In the same way, one who accepts the truth of counterfactuals of freedom can hardly be said to be embracing an incoherency.
And how do we know that counterfactuals of freedom do not satisfy Hasker's principle? Plantinga has defended the possibility of transworld depravity--that every creaturely essence is such that, if exemplified, its exemplification would have committed moral evil.{21} More recently, Kvanvig has argued that creaturely essences contain all the relevant counterfactuals of freedom concerning what their exemplifications would do in any circumstances.{22} On such views counterfactuals of creaturely freedom are grounded in the relevant individual essences of the agents referred to in the propositions. Against Kvanvig, Hasker objects, "But this is fatal to the theory. No individual chooses, or is responsible for, what is contained in that individual's essence."{23} But this objection does not tell against a view like Plantinga's, according to which creaturely essences have properties involving counterfactuals contingently, and Kvanvig could avoid the objection by making the counterfactual properties world-indexed. If creaturely essences possess counterfactual properties, then it could be maintained that counterfactuals of creaturely freedom are grounded in individual creaturely essences.
Of course, it might still be asked why individual creaturely essences have the counterfactual properties they do. But why think that volitional counterfactual properties or states of affairs must be grounded in their relevant categorical counterparts at all? Perhaps this at best characterizes only causal counterfactual states of affairs. The demand for a ground for volitional counterfactual states of affairs seems misguided. It implicitly presupposes that libertarianism and agent causation are false doctrines. To see the point, consider the libertarian claim "Jones freely chose x." If a compatibilist were to demand what makes this proposition to be true, the libertarian might well respond that nothing makes it to be true, that it simply is true in virtue of the fact that Jones freely chose x. But suppose the compatibilist presses him further, demanding why that state of affairs obtains. If Jones's choice was undetermined, then why did not some other state of affairs obtain, say, Jones's freely choosing y? The libertarian will respond that the compatibilist has missed the whole point. Jones himself is the cause of his choice and there is nothing further that makes it the case that Jones freely chose x; to ask for that is implicitly to deny the very liberty the libertarian presupposes. But in the same way, the proposition, "If Jones were in C, he would freely choose x" is true in virtue of the fact that the counterfactual state of affairs it describes obtains. To demand "But what makes it the case that if Jones were in C, he would choose x?" implicitly denies Jones's liberty. There is no further ground of why Jones would freely choose x if he were in C. To think that there must be such is to deny the hypothesis of Jones's free causal agency. Hence, Hasker's query is simply misconceived.
Objection (ii)
Hasker notes that counterfactuals are true or false relative to a world. According to the possible worlds semantics for counterfactual discourse, one is to consider the sphere of possible worlds most similar to the actual world in which the antecedent of the counterfactual is true. Better, all one has to consider are the initial world-segments of such worlds up to the time specified in the counterfactual, since what happens after that time can scarcely be relevant to the truth of the counterfactual. If the consequent is also true in all such antecedent-permitting world-segments, then the counterfactual is true. But Hasker argues that if the antecedent is maximally specified, then the restriction "most similar to the actual world" becomes superfluous. For there is only one sphere of possible world-segments which permits such maximally specified antecedents. An antecedent-permitting world-segment could not have some feature which made it more or less similar to the actual world because all such features are already taken account of in the maximally-specified antecedent. But then no matter what possible world one chooses as one's reference point, it will be that same sphere of worlds which will be closest to that world. Hence, if a counterfactual is true, it is true in all antecedent-permitting world-segments regardless of which possible world is one's reference point. There is thus no possible world in which the counterfactual is false. It is therefore necessarily true, which contradicts the hypothesis that there are true counterfactuals of freedom.
What this objection overlooks is that shared counterfactuals are themselves a measure of the similarity between worlds.{24} Thus, if some counterfactual is true in the actual world, there still are antecedent-permitting worlds which are farther from the actual world than the sphere of antecedent-permitting worlds in which the consequent is universally true, namely, those worlds in which the consequent is false. But those worlds may be closer to some other possible world; hence, in that world the counterfactual which is true in the actual world is false.
Hasker retorts that this answer "violates the reason for introducing the comparative-similarity notion in the first place--that reason being . . . to secure that counterfactuals are evaluated in worlds sufficiently similar to the actual world in noncounterfactual respects."{25} But if that was the motivation behind the similarity relation, it only follows that the motives of those who drafted possible world semantics for counterfactuals were thwarted. But as Plantinga explains, it follows neither that such semantics fails to correctly specify the truth conditions of counterfactuals nor is viciously circular.{26}
Objection (iii)
Which counterfactuals are true depends on which antecedent-permitting, initial world-segments are most similar to the actual world. But which world is actual, Hasker continues, depends in part on God's decision about what to create. Therefore, God could not have been guided by the truth of counterfactuals of creaturely freedom in deciding what world to actualize, since such propositions are true only as a consequence of which world is actual.
What this objection fails to appreciate is that parallel to the logical sequence in God's knowledge--natural knowledge, middle knowledge, free knowledge--there is a logical sequence in the instantiation of the actual world as well. In the first logical moment of God's natural knowledge, all broadly logically necessary states of affairs already obtain. In the second logical moment of God's middle knowledge the actual world is even more fully instantiated than at the first moment. For now all those states of affairs corresponding to true counterfactuals of creaturely freedom obtain. For example, the state of affairs If Peter were in C, he would deny Jesus three times obtains. Then comes logically the divine decree to create, and God freely actualizes all remaining states of affairs of the actual world. In the third logical moment, God possesses free knowledge of the actual world, which is exemplified in all its fullness (tenselessly speaking). Only at this point can the actual world as such be said to obtain.
It is therefore misleading to say that prior to the divine decree the actual world does not obtain simpliciter, for certain aspects do and other aspects do not. And those states of affairs that do obtain are sufficient for the truth of counterfactuals of creaturely freedom, since the latter correspond with reality as it thus far exists and since possible worlds can be ranked in their similarity to the actual world as thus far instantiated in terms of degree of shared counterfactuals, thus supplying the truth conditions for a possible worlds analysis of the truth of counterfactuals of freedom. Once it is appreciated that there is a logical sequence in the instantiation of the actual world just as there is in God's knowledge, then objections to middle knowledge based on counterfactuals' being true "too late" to facilitate such knowledge vanish.
Hasker complains that such an answer leaves us unable to explain the fact that those counterfactuals of creaturely freedom are true which are true. But this merely reiterates his first objection concerning the ground of the truth of counterfactuals of freedom and so fails to advance the discussion.
At any rate, objections to middle knowledge based on its alleged incompatibility with the possible worlds account of the truth conditions of counterfactuals strike me as very unimpressive. That account was drafted without any consideration of the peculiar situations engendered by theism (compare the way in which the existence of an Anselmian God upsets our intuitions about broadly logical modality!{27}) or middle knowledge. The account may simply be inadequate for the concerns of the philosopher of religion. In fact, I think it is evident that the possible worlds semantics for counterfactual conditionals is defective, for that account cannot adequately handle counterfactuals with impossible antecedents.{28} If the detractor of middle knowledge is to refute that doctrine, then, he will have to come up with a lot stronger arguments than its alleged incompatibility with current semantical theories.
Objection (iv)
Hasker's fourth argument involves a tortuously formulated illustration{29} which has fortunately been reduced with Hasker's approbation to three crucial premisses by Thomas Flint:{30}
I. If E brings it about that "Q" is true, then E is a token of an event-type T such that [(some token of T occurs) € ® Q] and [ ~ (some token of T occurs) € ® Q], and E is the first token of T which occurs.
II. Counterfactuals of freedom are more fundamental features of the world than are particular facts. (Hence, worlds which differ from the actual world with regard to factual content are closer than those which differ from it with regard to counterfactuals of freedom.)
III. If it is in S's power to bring it about that P, and "P entails "Q" and "Q" is false, then it is in S's power to bring it about that Q.
On the basis of these premisses Hasker argues as follows: Let A € ® B be a true counterfactual of freedom about me and let A be true. Let us assume that I can bring about the truth of this counterfactual by performing the action specified in B. If premiss (I) is correct, then I can bring about the truth of A € ® B (i.e., Q) only if it is the case that, if I had not performed the action specified in B (i.e., E) then A € ® B would have been false (i.e., A € ® B). But if (II) is correct, this necessary condition will never be satisfied because the closest worlds to the actual world will always be worlds in which it is the case that A rather than worlds in which it is the case that A € ® B. So I cannot bring about the truth of any counterfactual of freedom. Moreover, the Molinist holds that A & € ® B entails A € ® B. So according to (III), if it lies within my power to bring it about that A & € ® B (i.e., P), it is also within my power to bring it about that A € ® B (i.e., Q). But since I cannot bring about the truth of A € ® B, it follows that I cannot bring it about that A & € ® B. This is not due to my inability to bring about A, since A is already the case; so it must not be within my power to bring it about that € ® B. Since this abrogates my freedom, we must deny the original assumption, that there are any true counterfactuals of freedom.
Now I think it is very apparent that the inference drawn in (II) is a non sequitur. In one sense, counterfactual states of affairs about creaturely freedom are more fundamental than states of affairs about particular facts, namely, the former already obtain logically prior to God's decree while the latter are logically subsequent to it. Thus, prior to God's decree, it is the case that if Peter were to be in C, he would deny Christ three times, but it is not the case that Peter is a Galilean fisherman. (The same could be said, as well, about certain counterfactual states of affairs about natural kinds. Freddoso, for example, would say that logically prior to God's decree it is the case that if water were cooled to 0o C., it would freeze, but it is not the case that most of the Earth's water is saline. {31} ) But even though counterfactual states of affairs about creaturely freedom are thus logically prior to states of affairs about particular facts, they are no less contingent, for creatures could choose to act differently and then other counterfactuals of creaturely freedom would be true. Thus, "fundamental" in the sense of logical priority in the instantiation of the actual world has nothing at all to do with the resolution of vagueness between worlds to determine which are most similar to the actual world.
Why, then, should the defender of middle knowledge be committed to preservation of counterfactuals of freedom at the expense of the laws of nature in determining which worlds are most similar to the actual world? Suppose, for example, that it is true that if my phone were to ring, I would pick it up, and that it does ring. Which world is more similar to the actual world: one in which I do not pick it up when it rings or one in which a flying pink elephant crashes through my office window destroying my telephone--which I would have picked up had it rung? Hasker's (II) would require us to say that the second of these two worlds is more similar to the actual world. But there is nothing in the doctrine of middle knowledge that commits its defender to so silly a supposition.
In his reply to Flint, Hasker argues that the Molinist is committed to such a thesis because counterfactuals based on the laws of nature are only "would-probably" conditionals, whereas counterfactuals of freedom are necessitation conditionals. "Would-probably" counterfactuals are, selon Robert Adams, whose analysis Hasker approves, conditionals in which the consequent would be probable if the condition specified in the antecedent were to obtain. In such conditionals there is no guarantee that if the antecedent obtains, the consequent also obtains. By contrast, in counterfactuals of freedom, if the antecedent condition obtains, then the consequent condition definitely obtains as well. Now Hasker apparently thinks that modern physics has proved that all the fundamental laws are probabilistic rather than deterministic in character. Therefore, counterfactuals of freedom surely have to be weighted more heavily than counterfactuals backed by the laws of nature in determining the relative closeness of possible worlds.
But this response (wholly apart from the false assertion that all natural laws are probabilistic rather than deterministic{32}) is multiply confused: (1) The Molinist is under no obligation to accept Robert Adams's analysis of probabilistic counterfactuals. In fact, I should say that Plantinga has convincingly refuted Adams on this score, that it is the whole conditional that is probable—Probably ( f 
y )--rather than the consequent alone, and that probability does not specify a certain value in a broad range of truth values, but registers our degree of epistemic certainty about which of two truth values the proposition possesses.{33} But the superiority of Plantinga's analysis aside, the point is that if Hasker is to refute middle knowledge, he has to do so either by showing some incoherence on Molinism's own assumptions or else refute those assumptions. But he cannot simply import without argument analyses of counterfactuals which Molinists would reject and then show that on that analysis, middle knowledge fails. (2) Counterfactuals backed by laws of nature are no more "would-probably" conditionals than are counterfactuals of freedom. The determinateness of the counterfactual's truth value is not affected by the determinacy of the causal relations involved. Alethic bivalence is just a different category from causal determinacy. This is evident in that some Molinists, like Freddoso, would say that even a counterfactual about causally indeterminate events such as
19. If a photon were fired through the aperture at t, it would strike the screen at coordinates <x, y>
is bivalent and may, for all we know, be true.{34} I think the source of Hasker's confusion may be his conflation of a proposition's certainty and its definiteness.{35} Definiteness refers to its possession of one of two truth values; certainty does not characterize the proposition itself but is our degree of conviction as to which truth value it has. Thus, (19) may be utterly uncertain to us, but nonetheless definitely true. In the sense that (19) is definitely true, the consequent is guaranteed on the antecedent, regardless of causal indeterminacy. Really there are no such things as "would-probably" counterfactuals in Hasker's sense; there are bivalent counterfactuals which we know to be true or false to different degrees of probability. (3) Even if Hasker were correct, I still fail to see what relevance this has to the resolution of vagueness between possible worlds. How, on his analysis, does it follow that a world in which I do not pick up my phone when it rings is less similar to the actual world than a world in which the quantum motions of the subatomic particles in my telephone all happen to coincide so that my phone "tunnels" through my office wall instead of ringing--though it remains true that if it rang, I would pick it up?
In his reply to Flint, Hasker suggests that counterfactuals of freedom need not be, after all, more fundamental than counterfactuals based on the laws of nature, just so long as they are far more fundamental than particular facts.{36} But I do not see that the defender of middle knowledge need be committed even to this. Is it obvious, on the same hypothesis about the actual world above, that a world in which I do not pick up the phone when it rings is less similar to the actual world than one in which, say, a short-circuit prevents my phone from ringing? I fail to see why the Molinist need make such a judgment.
What is especially curious about Hasker's argument based on (II) is that it seems to commit him to the use of backtracking counterfactuals in this case. For his argument based on (II) implies that if I were not to do the action specified in B, then it would not have been the case that A, that is ~B ~A. In our example, were I not to pick up the phone, it would not have rung, even though it did ring. But the use of backtracking counterfactuals requires some justification for a special resolution of vagueness, such that worlds involving adjustments of the past are more similar to the actual world than worlds without such adjustments. It is precisely to such backtracking counterfactuals that the defender of divine foreknowledge of future contingents appeals in rebutting theological fatalism, and he is careful to offer justification for their appropriateness there.{37} But Hasker has failed to justify why a special resolution must be always employed if it is to be within one's power to negate the consequent of any counterfactual of freedom.
But now an even deeper problem surfaces. For if ~B ~A is true, then do I not have it within my power to bring it about that that counterfactual of freedom is true, which contradicts Hasker's hypothesis? To avert that conclusion, Hasker must show that it is not within my power to perform ~ B. Hasker argues that I cannot bring it about that ~ B because to do so is to bring about A & ~B, which entails A ~ B, which it is not within my power to bring about. But this line of argument seems patently inconsistent. For we have already seen that Hasker is committed to the backtracking counterfactual ~ B ~ A, so that to bring it about that ~B is not to bring it about that A & ~ B, but to bring it about that ~ A & ~ B. By (III), then, what lies within my power is to bring it about that ~A ~ B, which does not contradict either A B or B ~ A.
The source of Hasker's error appears to be his belief that if A is already given, then my ability to perform ~ B implies the ability to bring it about that A & ~ B. He infers from my inability to bring about the composite state of affairs (A & ~ B) and the givenness of A that it is not within my power to bring it about that ~ B. The reader will recognize that this is just the same, old argument for theological fatalism dressed up in a new guise, only in Hasker's hands it becomes a mish-mash of inconsistent elements from both Molinism and theological fatalism. I conclude that Hasker has provided no good reason for thinking that the doctrine of middle knowledge is incoherent and therefore not a possible solution to the problems of divine prescience, providence, and predestination.
Endnotes
{1}William Hasker, God, Time and Knowledge, Cornell Studies in the Philosophy of Religion (Ithaca, N.Y.: Cornell University Press, 1989), p. 89.
{2}Alvin Plantinga, "On Ockham's Way Out," Faith and Philosophy 3 (1986): 235-69.
{3}William Lane Craig, "'Nice Soft Facts': Fischer on Foreknowledge," Religious Studies 25 (1989): 235-46.
{4}Hasker incorrectly asserts that Freddoso now thinks that God's past belief is a hard fact (Hasker, God, Time and Knowledge, p. 95; cf. idem, review of On Divine Foreknowledge, Faith and Philosophy 7 [1990]: 358-59). He fails to notice that the Molinist definition of a hard fact is different than the Ockhamist's and amounts to the causal closedness of the past. But the past is still counterfactually open, and thus in the Ockhamist sense God's belief remains a soft fact. See Alfred J. Freddoso, "Introduction" to On Divine Foreknowledge, by Luis de Molina, trans. with Notes by A.J. Freddoso (Ithaca, N.Y.: Cornell University Press, 1988), pp. 59-60.
{5}See Alvin Plantinga, The Nature of Necessity, Clarendon Library of Logic and Philosophy (Oxford: Clarendon Press, 1974), p. 86; idem, "The Boethian Compromise," American Philosophical Quarterly 15 (1978): 129-38.
{6}Ibid., p. 107.
{7}Thomas Flint, "In Defense of Theological Compatibilism," Faith and Philosophy 8 (1991): 240.
{8}Hasker, God, Time, and Knowledge, p. 109.
{9}See my discussion in Divine Foreknowledge and Human Freedom, Studies in Intellectual History 19 (Leiden: E. J. Brill, 1990), pp. 89-90. I should be willing to accept
PEP5': If it is in S's power to bring it about that P, and "P" entails "Q" and "Q" is false, and Q is a consequence of P, then it is in S's power to bring it about that Q.
I should be inclined to say that it is within in S's power to bring about God's past beliefs about S's free actions in the same sense that Socrates had it within his power to make Xantippe a widow by drinking the hemlock cup.
{10}Hasker, God, Time, and Knowledge, p. 129.
{11}Ibid., p. 130.
{12}Ibid., p. 134.
{13}Ibid., p. 135.
{14}Ibid., p. 141.
{15}Richard Taylor, "Fatalism," Philosophical Review 71 (1962): 56-66; idem, "Fatalism and Ability I," Analysis 23 (1962-1963): 25-27; idem, "A Note on Fatalism," Philosophical Review 72 (1963): 497-99.
{16}Hasker, God, Time, and Knowledge, p. 126.
{17}Alfred J. Freddoso, "Accidental Necessity and Power Over the Past," Pacific Philosophical Quarterly 63 (1982): 54-68.
{18}Hasker, God, Time and Knowledge, pp. 29-52.
{19}Ibid., p. 30.
{20}Ibid.
{21}Plantinga, Nature of Necessity, p. 188.
{22}Jonathan L. Kvanvig, The Possibility of an All-Knowing God (New York: St. Martin's Press, 1986), pp. 124-25.
{23}Hasker, God, Time, and Knowledge, p. 32.
{24}See Plantinga, Nature of Necessity, pp. 177-78.
{25}Hasker, God, Time, and Knowledge, pp. 35-36.
{26}Plantinga, Nature of Necessity, p. 179; Alvin Plantinga, "Reply to Robert Adams," in Alvin Plantinga, ed. James Tomberlin and Peter Van Inwagen, Profiles 5 (Dordrecht: D. Reidel, 1985), p. 378.
{27}See Thomas V. Morris, The Logic of God Incarnate (Ithaca, N.Y.: Cornell University Press, 1986), pp. 111-19.
{28}See my "'Lest Anyone Should Fall': A Middle Knowledge Perspective on Perseverance and Apostolic Warnings," International Journal for Philosophy of Religion 29 (1991): 65-74.
{29}The counterfactuals employed by Hasker, for example, are not in the canonical form of counterfactuals of creaturely freedom--if S were in C, S would freely decide to do A--and so could be dismissed by the Molinist as neither true nor false. But Flint has done Hasker the service of freeing the essential argument from its illustrative infelicities.
{30}Thomas P. Flint, "Hasker's God, Time and Knowledge," Philosophical Studies 60 (1990): 104-05; William Hasker, "Response to Thomas Flint," Philosophical Studies 60 (1990): 118.
{31}Alfred J. Freddoso, "The Necessity of Nature," in Midwest Studies in Philosophy, vol. 11, ed. P. French, T.E. Uehling, Jr., and H. Wettstein (Minneapolis: University of Minnesota Press, 1987).
{32}Modern physics rests on the twin pillars of Quantum Theory and Relativity Theory, but these two bodies of law are irreconcilable with each other. Relativistic laws are not probabilistic.
{33}Plantinga, "Reply to Robert Adams," pp. 380-381.
{34}Freddoso, "Introduction," pp. 28-29.
{35}Consider his statement ". . . how can those psychological facts provide good grounds for the assertion that the agent definitely would (as opposed, say, to very probably would) respond in that way?" (Hasker, God, Time, and Knowledge, p. 24) Cf. p. 31 where he seems to confuse taking psychological facts as evidence that a proposition is true and taking such facts as making a proposition true.
{36}Hasker, "Response," pp. 118-19.
{37}If God's beliefs are merely inerrant in the actual world, then that inerrancy warrants a special resolution of vagueness; if God is essentially omniscient, then no special resolution is required to justify backtracking counterfactuals, since no worlds exist in which God errs, so that the standard resolution suffices.
