Talbott's Universalism Once More

Dr. William Lane Craig

William Lane Craig is Research Professor of Philosophy at Talbot School of Theology in La Mirada, California. He lives in Atlanta, Georgia, with his wife Jan and their two teenage children Charity and John. At the age of sixteen as a junior in high school, he first heard the message of the Christian gospel and yielded his life to Christ. Dr. Craig pursued his undergraduate studies at Wheaton College (B.A. 1971) and graduate studies at Trinity Evangelical Divinity School (M.A. 1974; M.A. 1975), the University of Birmingham (England) (Ph.D. 1977), and the University of Munich (Germany) (D.Theol. 1984). From 1980-86 he taught Philosophy of Religion at Trinity, during which time he and Jan started their family. In 1987 they moved to Brussels, Belgium, where Dr. Craig pursued research at the University of Louvain until 1994.

In the debate between universalism and particularism, three questions need to be addressed: (I) Has it been shown that it is inconsistent to affirm both that God is omniscient, omnipotent, and omnibenevolent and that some persons do not receive Christ and are damned? (II) Can these two affirmations be shown to be consistent? (III) Is it plausible that both affirmations are true? In this on-going debate with Thomas Talbott, I argue that Talbott has failed to show the above affirmations to be inconsistent, that while one cannot prove them to be consistent, it is plausible that they are, and that it is also plausible that both affirmations are in fact true.

"Talbott's Universalism Once More." Religious Studies 29 (1993): 297-308.

See also Talbott's Universalism.

NOTE: Dr. Craig uses logic symbols in this document. For users with Macintosh computer systems, for best results please set your browser to view this document using Western MacRoman encoding. In Netscape 3, use Options | Encoding and choose Western MacRoman. For Netscape 4, use View | Encoding and choose Western MacRoman.


In my No Other Name,{1} I asserted that detractors of Christian exclusivism are, in effect, posing a soteriological problem of evil, to wit, that the proposition

1. God is omniscient, omnipotent, and omnibenevolent

is inconsistent with the proposition

2. Some persons do not receive Christ and are damned.

Following the strategy of the Free Will Defense, I pointed out that if (1) and (2) are to be shown to be broadly logically inconsistent, then the anti-exclusivist must furnish some additional premiss(es) which meets the following conditions: (i) its conjunction with (1) and (2) formally entails a contradiction, (ii) it is either necessarily true, essential to theism, or a logical consequence of propositions that are, and (iii) its meeting conditions (i) and (ii) could not be rationally denied by a right-thinking person. These are very exigent conditions, and I confessed that I was "not aware of anyone who has tried to supply the missing premise which meets these conditions."{2}

Within a year, Thomas Talbott published his independently developed critique of Christian exclusivism in which he attempted to do just that.{3} The missing premiss is

(P3) Necessarily, God loves a person S (with a perfect form of love) at a time t only if God's intention at t and every moment subsequent to t is to do everything within his power to promote supremely worthwhile happiness in S, provided that the actions taken are consistent with his promoting the same kind of happiness in all others whom he also loves.

(P3) in conjunction with (1) entails

3. For any created person S and time t subsequent to the creation of S, God's intention at t is to do all he properly can to promote supremely worthwhile happiness in S,

and (3) is contradictory to

4. There is a person S and a time t subsequent to the creation of S such that it is not God's intention at t to do all that he properly can to promote supremely worthwhile happiness in S,

which is entailed by Talbott's

5. God will irrevocably reject some persons and subject those persons to everlasting punishment,

which Talbott would say is entailed by my (2). In accordance with condition (iii) above, Talbott must regard the truth of (P3) as rationally undeniable by a right-thinking person, and so also the entailment of (4) by (5) and (5) by (2).

Now in my response to Talbott,{4} I denied that Talbott had proven that the conjunction of (1), (P3), and (5) formally entails a contradiction, and I questioned the truth of (P3). More than that, I recalled arguments from my original piece intended to show that (1) and (5) are consistent and, moreover, even plausibly true.

This brief review is necessary because in his most recent contribution to the debate{5} Talbott subtly shifts the burden of proof from his own shoulders to those of the Christian exclusivist. There are two moves in the Free Will Defense: (I) to deny that the objector has shown (1) and the evil in question to be broadly logically inconsistent, and (II) to suggest a possibly true proposition that together with (1) entails the existence of the particular evil, thereby showing the co-existence of God and that evil to be broadly logically possible. The Theodicist will go on to argue that (III) the proposed proposition is not merely possible but plausible. By focusing in his most recent contribution almost exclusively on task (II), Talbott turns the spotlight away from (I) and from the extraordinarily difficult task with which he is confronted in meeting conditions (i), (ii), and (iii) above. Let us in the interest of clarity therefore address separately the three questions raised by the Free Will Defense and theodicy.

(I) Has Talbott Shown (1) and (2) to be Inconsistent?

Depending on how the terms in Talbott's premisses are defined, it is open to the Free Will Defender to point to lacunae at various places in Talbott's argument. For example, one might deny that Talbott has shown (P3) to be true, since God's omnibenevolence does not entail that He intends to do everything He properly can to promote a person's salvation (even in hell) if He has already offered that person sufficient grace for salvation and knows that further efforts would be futile. If Talbott amends (P3) such that omnibenevolence requires merely that God always wills the salvation of every person, then the Free Will Defender may deny that (5) entails a suitably amended (4), since God's judicial rejection of a person who has freely spurned His offer of salvation is consistent with God's willing that that person instead accept His offer. If Talbott should define "rejection" in such a way that it entails "not intending the rejected person's salvation," then the Free Will Defender will deny that (2) entails (5). All of these moves derive from the Free Will Defender's belief that what Talbott calls the Rejection Hypothesis is possibly true:

(RH) Some persons will, despite God's best efforts to save them, freely and irrevocably reject God and thus separate themselves from God forever.

If Talbott is to carry his argument, he must show that (RH) is broadly logically impossible.

With that in mind, consider now the opening section to Talbott's most recent contribution. Here Talbott distinguishes between broadly logical possibility and merely epistemic possibility, observing that a proposition's epistemic possibility does not suffice to establish its broadly logical possibility. He rightly points out that questions about the latter "are sometimes exceedingly difficult and, and as a practical matter, impossible to answer in a definitive way."{6} Talbott complains that I set out "a bewildering number of unsupported assertions to the effect that this or that is logically possible"{7} in an effort to deduce the possibility of what he calls the "Damnation Thesis":

(DT) There exists at least one sinner S such that S will never be reconciled to God and thus never be saved.

He charges that I confuse what it would take to demonstrate that, for all we know, (DT) is true and what is would take to demonstrate that (DT) is possible in the broadly logical sense. Talbott argues that I fail both to establish that (DT) is broadly logically possible and to undermine Talbott's argument that (DT) is broadly logically impossible.

The distinction to which Talbott draws our attention, namely, epistemic versus broadly logical possibility, is an important one, and one which I failed to delineate clearly in my critique of Talbott's universalism. Nonetheless, it does seem to me that Talbott is trying to foist upon me a project which I did not undertake in that critique, for I neither enunciated (DT) nor tried to provide a deductive proof of it, as Talbott represents. Rather my principal strategy was to attack Talbott's demonstration that (RH) is impossible. He argued that (RH) is not broadly logically possible because (i) the choice specified in (RH) is incoherent and (ii) even if such a choice were coherent, necessarily God would not permit it. I in turn tried to show that his arguments for (i) and (ii) fail.{8} My primary purpose was thus not to establish the possibility of (DT), but to deny that Talbott had shown (RH) to be impossible.

It is important to understand that at this stage of the argument the Free Will Defender is not obliged even to establish the broadly logical possibility of (RH); all he has to do is undercut the universalist's attempts to prove its impossibility. And the crucial point here is that in order for the Free Will Defender to accomplish that purely negative and defensive task, the epistemic possibility of (RH) is sufficient. Thus I am under no obligation to establish the broadly logical possibility of the Molinist hypotheses I set forth; all that is required to defeat Talbott's arguments against (RH) is that some of the "bewildering number" of suggestions I made be epistemically possible.

I.1. Is the Choice Specified in (RH) Incoherent or even Relevant?

Let us consider Talbott's first argument against (RH), that the choice involved is incoherent. Here two issues arise.

I.1.i. Has Talbott Proved that Transcircumstantial Damnation is Incoherent?

The debate here mirrors the seventeenth century Jesuit discussions concerning whether God has congruent grace for every possible person, that is, grace that is extrinsically efficacious but infallible in winning the free consent of the person to whom it is extended. According to the Congruist, God has congruent grace for every person He could possibly create. Now Talbott is a sort of modern day Congruist; nay, more than that, he is a hyper-Congruist. For he holds not merely that God has congruent grace for every possible person, but that, necessarily, God extends such grace to every person He creates, that it is logically impossible for any person to finally reject God's salvation. Congruism is a radical and traditionally controversial doctrine; its falsity certainly seems epistemically possible. But Talbott contends that hyper-Congruism can be shown to be true. His argument is that in every possible world God does His best to save every person and that there can be no motivation for any person to reject irrevocably God's best efforts to save him. I responded with two questions: (i) Is it not possible that the motivation for rejecting God is the will to self-autonomy? (ii) Is it not possible that rejection of God is due to a perversity which lacks any further motivation? So long as an affirmative answer to these questions is even epistemically possible, Talbott has failed to prove that (RH) is logically impossible and, hence, that (1) and (2) are inconsistent.

In response to my first suggestion, Talbott emphasizes that the decision to reject God must be "fully informed," which notion Talbott defines as a decision which "does not rest upon ignorance, or misinformation, or deception of any kind."{9} Given that God wills for me exactly what I at the most fundamental level will for myself--supreme happiness--, it follows that anyone in a position to make a fully informed decision would have the strongest conceivable motive not to reject God. Talbott's view is that in every possible world containing persons in need of salvation, God eventually places every person in circumstances in which he can make a fully informed decision about salvation and that that decision is always affirmative.

Persuasive as this argument may appear at first blush, a little reflection will show, I think, that it is question-begging. For suppose that the Free Will Defender responds that someone may be fully informed of God's intentions toward him, but out of a desire for self-autonomy refuses even supreme happiness because its price--bowing the knee to God--is too high. Why could not someone's hatred of God be so implacable that he chooses to reject God rather than be supremely happy? Talbott responds that such a case--Milton's Satan being a paradigm example--"hardly illustrates a fully informed decision to reject God."{10} Such a person still labors under so many illusions that his decision is less than fully informed.

This response raises the suspicion that in Talbott's view any decision to reject God is by definition not "fully informed," since a person who rejects God is by the nature of the case deceived and no "fully informed" decision rests upon "deception of any kind." We are then surely justified in questioning whether those who reject God must make "fully informed" decisions in this idiosyncratic sense. For such persons may be justifiably regarded as self-deceived, and it may not be within God's power to destroy such self-deceptions without destroying such persons' freedom. For self-deception, being rooted in the free will of the creature, may be as impenetrable to God's grace as the free will itself. Indeed, in the traditional Christian doctrine of sin the notion of self-deception lay at the heart of all man's sinful acts and especially his sinful rejection of God. God's removal of the deception in some persons could require abrogation of the freedom of the will itself, the freedom to deceive oneself. Talbott's argument is therefore either question-begging (because it rules out a fully informed rejection of God by definition) or unsound (because nothing requires that sinners' rejection of God be free of deception of any sort). Because Talbott's argument is not cogent, it remains epistemically possible that the will to self-autonomy may motivate rejection of God even under the best of circumstances.

What then of my second suggestion, the human perversity being what it is, perhaps no motivation is necessary for the decision to reject God? Talbott's response is that such an irrational decision cannot be characterized as free. Only a rational agent can act freely, so that someone who acts contrary to his own interest without any motive for doing so is not a rational agent and so is not capable of performing free actions. Again, however, this line of argument seems question-begging. We should not think of a decision sprung from perverseness as like a drug addiction or a quantum leap, but as something which issues out of one's own twisted moral character. Persons who just are evil may do wrong for its own sake and spurn God just to spurn Him. Since the decision arises from agent causation--albeit an evil, perverse agent--, I do not see why it cannot be described as free unless we simply define "free" to exclude such irrational acts. But if we adopt this idiosyncratic view of freedom, then why think that sinners must be free in that sense? There is a strong Christian tradition rooted in the New Testament that sinners are in fact not free, that they are slaves of sin and self. In this sense, those who reject God are not truly free. But because theirs is a self-bondage, rooted in their own will, which is capable of receiving God's grace, but, perversely, refuses to do so, they are culpable for that bondage. So, again, if Talbott's argument is not question-begging, it is still unsound because nothing has been shown to require that the decision to reject God be "free" in his peculiar sense. Hence, he has failed to show (RH) to be logically impossible.

Remarkably, Talbott in the end concedes, "at least for the sake of argument," that

(A) There exists at least one creaturely essence E such that, for any circumstances C in which the instantiation of E would be free in the matter of being reconciled to God, the instantiation of E would in fact freely refuse to be reconciled to God in C

is possibly true.{11} If (A) turns out to be true, then there are creaturely essences which have the property of transcircumstantial damnation, and not only hyper-Congruism but even ordinary Congruism is false. Talbott observes that (A), however, does not entail (DT), since God need not instantiate any of these creaturely essences. But here Talbott is shifting the burden of proof. For no one has claimed that (A) entails (DT). What is claimed is that no cogent argument has been given to show that (RH) is logically impossible. By conceding the possibility of (A), Talbott gives up one of his principal arguments against (RH), namely, that the choice specified in (RH) is incoherent.{12} That means that he shall have to rely on other arguments to prove the logical impossibility of (RH).

I.1.ii. Has Talbott Proved that (RH) Entails Transcircumstantial Damnation?

In my critique of Talbott's universalism, I made a sort of concession of my own, namely, I pointed out that even if Congruism were true, so that for any creaturely essence E there is a set of circumstances in which God can win the free, affirmative response of the instantiation of E to His salvific grace, (RH) is still epistemically possible because (RH) does not entail that (A) is true. I pointed out that even if for every created person S, there is a set of circumstances C in which S affirmatively responds to God's grace and is saved, it does not follow that there is a compossible set of circumstances in which all created persons freely embrace salvation. More than that, even if S1 would in C1 freely accept salvation and S2 would in C2 freely accept salvation and C1 and C2 are compossible, to claim that in C1 & C2, S1 would freely accept salvation or that in C1 & C2, S2 would freely accept salvation would be to commit the counterfactual fallacy of strengthening the antecedent. Thus the Free Will Defender could embrace Congruism, but maintain that it is (epistemically) possible that there are no feasible worlds in which all free persons accept salvation.

Talbott counters that without transcircumstantial damnation it is logically impossible that all feasible worlds should be such that in each of them some person rejects salvation. There is no feasible world in which every person freely accepts salvation only if every creaturely essence suffers from transcircumstantial damnation. So long as some essences lack this property, there is a world feasible to God in which all persons are freely saved. Talbott thus regards

B. There are feasible worlds in which some persons are freely reconciled to God; but for any feasible world W, if in W some persons are freely reconciled to God, then in W some persons irrevocably reject God

as "quite impossible".{13}

In opposition to (B) Talbott proclaims his "Glorious Feasibility Thesis"

(GFT) Necessarily, for any collection C of persons who do not suffer from transworld reprobation, there is a feasible world in which every member of C is freely reconciled to God.

In defense of (GFT), he argues, "For any two persons, S and S, if there is a feasible set of circumstances in which S is freely reconciled to God and there is a feasible set of circumstances in which S' is freely reconciled to God, then there is also a feasible set of circumstances in which S and S' are both freely reconciled to God."{14}

Notice, however, that even if sound, this argument does not establish (GFT); a "feasible set of circumstances" is not synonymous with a "feasible world," the latter differing from the former in maximality. The Free Will Defender could admit that there are circumstances in which S and S' are both freely reconciled to God without conceding that there is a feasible world in which all persons are freely reconciled to God.{15}

But is Talbott's argument in any case sound? In support of it, he explains that in C1, S would encounter a set of appearances such that, if God were to provide him with just those appearances, S would freely be reconciled to God; similarly in C2, S' would encounter a set of appearances such that, if God would provide him with just those appearances, S' would freely be reconciled to God. These two sets of appearances need not cohere with each other, and God can vary how things appear to S' without varying how things appear to S. So even if C1 and C2 are logically inconsistent, there would still be a feasible set of circumstances C3 consisting of God's providing S and S' respectively with the appearances conducive to their each being freely reconciled to God.

Talbott recognizes that this account might require God to engage in deception. Indeed, it might involve deception on a massive scale, with each person hermetically sealed in his own illusory world. Talbott contemplates "spiritual realms which have no ordinary physical connection with our universe"--assuming, indeed, that there exists such a common universe--and which are tailor-made, complete with "misleading appearances concerning the existence or the fate of other persons," for winning the free, affirmative response of its real, non-illusory denizen(s) to God's grace.{16} Talbott envisions for each person, if need be, "a billion lifetimes, a billion different realms and universes and sets of appearances, a billion ways (including a billion different forms of deception)" to secure a person's salvation.{17}

He admits that this defense of (GFT) might lead one to question whether God could engage in such deception and whether, therefore, C3 is really possible. In defense of the possibility of the envisioned scenario, Talbott offers the justification that God would have a morally sufficient reason for deceiving people (presumably, their salvation) and that such deception is merely temporary.

A number of things can be said in response to this argument. First, I think we must say at the very least that we have no firm moral intuitions for thinking a world involving states of affairs like C3 to be broadly logically possible. But without such a basis, we have no good grounds for thinking (RH) to be impossible. For my part, worlds based on systematic deception by God of human persons seem morally unconscionable and therefore impossible. As worthy an end as universal salvation is, God could not utilize immoral means for achieving such an end. And in deceiving human persons in the way Talbott envisions, God would be violating their personhood and so acting immorally. For God would be placing persons in situations in which, for example, their spouses, children, and friends, with whom they enter into supposedly meaningful and intimate relationships, turn out to be mere illusions. One's life struggles, the expressions of tenderness, the confessions of fears to a trusted companion, the sacrifices for persons one loved, one's apologies and extensions of forgiveness, the myriad emotions of sympathy, anger, shame, bitterness, compassion--all were an interaction with illusion, a world of maya. This constitutes a profound violation of human dignity. Talbott criticizes one of my proposals because God, in expunging from a person's memory the knowledge of lost family members, would not be providing that person with a supremely worthwhile happiness: "No loving father, for example--not even one whose daughter endures a brutal rape and murder and not even one whose son commits suicide--would want to remain blissfully ignorant about what happened . . . . and the idea that he might want to have all memory of a son or daughter obliterated from his mind--that he might prefer this over his anguish--is simply preposterous."{18} But consider: on Talbott's view, a loving father may find out that in fact he had no son or daughter at all, that he was in love with a chimaera; his family has been truly lost to him, for in fact he has no family--only the delusions caused by God. Such a one has been violated in the deepest core of his person by being thus deceived. Indeed, one is compelled to wonder whether such persons, when confronted by God with the truth about their illusory pasts, will continue to accept the salvation won by their delusion--which makes the difference between feasible circumstances and feasible worlds acute for Talbott's case. Talbott might say that so great a deception would not be required by God to bring all created persons to salvation--but then how do we know? Is it not epistemically possible that in order to achieve a significant number of saved without any lost, the deception would have to be that great, at least for one person? To even so deceive one person would implicate God in sin, and so any such world would be impossible. I find it a telling weakness of the responses to my Molinist perspective on Christian exclusivism that in the end they inevitably have recourse to systematic deception on God's part{19}--an expedient which is at the very least dubious and consequently no proof in any sense of the impossibility of (RH).

Second, even if there are feasible worlds involving states of affairs like C3, so that (GFT) is true, that fact does not imply the impossibility of (RH). For even if some such world is feasible for God, God may prefer not to create such a world based on deception, but to create instead a world in which people find themselves in real circumstances and are offered by God sufficient grace for salvation, people whose choices are respected by God. In order to provide grounds for thinking (RH) to be impossible, Talbott must show that God is morally obligated to choose C3-worlds rather than non-deceptive worlds. Again, we have no firm basis for thinking that God is so obligated. As I pointed out in my original article, C3-worlds could involve drawbacks that would influence God not to prefer them over non-deceptive worlds, even though in the latter some people freely reject His salvation. So long as it is (epistemically) possible that God prefer a straightforward world in which not all are saved over a deceptive world in which all are saved through delusion, then (RH) is also (epistemically) possible.

Third, it is not evident that Talbott has succeeded in averting logical invalidity in his argument. In order to avoid the fallacy of strengthening the antecedent, Talbott argues that if in C1 S would be freely saved and in C2 S' would be freely saved, then in C3 S and S' would both be freely saved. Let, then, Cn = "Circumstances n obtain," Sn = "Sn will freely accept salvation," and ACn = "The appearance of Cn obtains." In order to avoid the fallacy of strengthening the antecedent, Talbott infers from

6. C S1

7. C S2,


8. C S1 · S2.

He justifies this inference by supposing that

9. (AC1 · AC2) = C3.

We thus infer (8) from

10. (AC1 · AC2)  S1 · S2.

But how do we know that (10) is true? The answer is that to S1, C1 and AC1 are indistinguishable; similarly for S2 and C2 and AC2. So we may affirm

11. (C S1) É (AC S1),

12. (C S2) É (AC S2).

From (6), (11) and (7), (12), it follows that

13.(AC S1),

14.(AC S2).

But how do we move from (13), (14), to (10)? The answer seems to be: by strengthening the antecedent of (13) or (14), which is logically invalid. Hence, the argument for (8) is unsound.

It seems to me, then, that even if we were to concede Congruism, Talbott still has not shown (RH) to be logically impossible, since his refutation of the hypothesis that worlds in which all persons freely accept salvation are infeasible for God involves the dubious claim that God might deceive people in order to win their salvation, fails to show that God is morally obligated to prefer a deceptive world to a straightforward one in which not all are saved, and seems to be logically invalid.

It seems to me, therefore, that we can conclude that Talbott's first argument for the logical impossibility of (RH) fails, for he has not shown either that the choice allegedly envisioned in (RH) is logically incoherent nor that such a choice is required if all feasible worlds are to be such that in none of them are all free persons saved. I think we can safely conclude that Talbott has not met so far condition (iii) of a successful presentation of the soteriological problem of evil, not to mention conditions (i) and (ii).

I. 2. Could God Create Worlds in which Not All Persons are Saved?

It will be recalled that Talbott had a second major argument against (RH), namely, that even if there are no feasible worlds in which all persons are freely saved, necessarily God would not create a world in which some persons are damned.{20} Rather, if He created at all, He would create a world in which free will is sacrificed to achieve universal salvation. Talbott's argument for this contention was a curious one to the effect that God must choose a world in which all are saved because otherwise, the supreme happiness of the blessed would be undermined by their knowledge of the existence of the damned. It seemed to me that this problem could be averted by denying that the blessed have any such knowledge. It seems possible, for example, that God could expunge from their minds any knowledge of the damned or, better, that the overwhelming presence of Christ would drive from the minds of the blessed the consciousness of the damned. Again, the Free Will Defender need not prove that these hypotheses are any more than epistemically possible, but Talbott, if he is to defeat (RH), must prove them broadly logically impossible.

Talbott argues that my hypotheses are broadly logically impossible because in neither case would the blessed enjoy "supremely worthwhile happiness," which is essential to salvation. Supremely worthwhile happiness is, first, "the kind of happiness that could survive a complete disclosure of truth about the universe; and second, it is the kind that one possesses only when one is filled with love for others."{21} Before we look at Talbott's detailed argumentation, we should do well to reflect for a moment on this peculiar sort of happiness. I think it does make sense to speak of different degrees of worth of happiness. For example, the happiness of the sadist is not as worthwhile as the happiness of the care-giver. Here the worth of the happiness is related to the moral value of the action which gives rise to it. But it seems to me dubious and even false that supremely worthwhile happiness entails the ability to survive a full disclosure of the truth. For a happiness which would, ceteris paribus, be diminished by the disclosure of a tragic truth about a loved one seems more worthwhile than one which would survive undiminished. Indeed, on Talbott's own reasoning the happiness of a person who is filled with love for others must be diminished on learning the truth of a loved one's misfortune; otherwise, we could maintain that the blessed do love the damned, but that their happiness remains nonetheless undiminished by the knowledge of their terrible estate. Thus, supremely worthwhile happiness does not entail the ability to survive a full disclosure of the truth; if the truth is tragic, quite the opposite is the case.

The problem is that Talbott has conflated supreme happiness with supremely worthwhile happiness. A happiness which is supremely worthwhile need not, indeed, in some cases, cannot be supreme happiness, that is, happiness untinged by sadness. Aware of the fate of the lost, the blessed in heaven could have supremely worthwhile happiness without being supremely happy.

Talbott would perhaps contend that salvation entails supreme happiness which is supremely worthwhile. If the happiness of the blessed is to be both, then there must be no damned; otherwise, if the truth of their existence were disclosed, the supreme happiness of the blessed, precisely because it is supremely worthwhile, would be undermined. But such reasoning would be modally fallacious. The blessed's happiness being both supreme and supremely worthwhile does not entail being able to survive complete disclosure of the truth. If the blessed are unaware that the damned exist, their happiness can be supreme, but it can also be supremely worth-while because it would be diminished if this truth were known.

In order to make his argument stick, Talbott must maintain that happiness which is both supreme and supremely worthwhile entails one's being conscious of the full truth and yet one's happiness remaining undiminished by it. But I do not see any basis to think the happiness of a person who has a complete knowledge of the truth is any more worthwhile for that reason than that of a person who lacks such knowledge. If my son is listed MIA in Vietnam would my happiness be made more worthwhile by the discovery that he was executed by the North Vietnamese? No doubt, as Talbott says, I would want to know that painful truth and in a sense would even welcome it, despite the pain it would bring. But that does not entail that such knowledge makes my happiness more worthwhile than it would have been without it. The mere possession of more information seems irrelevant to the worth of one's happiness. What is relevant is how one's happiness would be affected by the disclosure of such painful knowledge. Happiness which, ceteris paribus, increases with news of a loved one's good fortune or decreases with the news of his misfortune is more worthwhile than a happiness which survives indifferent the disclosure of such information. But the mere lack or possession of information does not decrease or increase the worthwhileness of the happiness one experiences. Thus I do not see why supremely worthwhile happiness entails complete knowledge of the truth. Unaware of the existence of the lost, the happiness of the saved in heaven, could be both supreme (because it is untinged by sadness) and supremely worthwhile (because it would be diminished were they aware of the fate of the lost).

Talbott therefore needs to prove that it is logically impossible that the blessed should be unaware of the existence of the damned. Consider then my first hypothesis, that God removes or withholds from the blessed any knowledge of the existence of the damned. Talbott agrees that it is sometimes right to withhold painful information from someone, but he insists that this is always a concession to that person's poor health or psychological/spiritual immaturity. Since the blessed are presumably not characterized by such conditions, it would be immoral of God to withhold from them knowledge of the damned. Moreover, Talbott argues, the deception contemplated in the present hypothesis is immoral because it is an eternal, not merely temporary, deception.

To deal with Talbott's second point first, I think we are justified in resisting his characterization of the hypothesis as "deception." It would be deceptive of God to make the blessed believe that the lost were saved when in fact they are not. But that is not the proposal. God's merely erasing any memory of the lost does not involve God's deceiving the blessed about the state of the lost. The blessed entertain no false beliefs about the lost; they simply have no beliefs about them at all. The doctrine of progressive revelation teaches us that while God is morally bound to reveal to us nothing but the truth, He is not bound to reveal to us the whole truth or, for that matter, even the truth about some thing. Hence, I do not see that Talbott has proven that God's removing or withholding knowledge of the lost from the minds of the blessed in order to secure their supreme happiness is either deceptive or immoral.

As for Talbott's first point, it seems to me very mootable that appropriate withholding of needless, painful information is in every case a concession to poor health and immaturity. Is it due to poor health or immaturity that we hold closed coffin funerals, for example? Or why we do not convey to a friend a secret insult uttered by someone in the past? Or why we do not ask forgiveness of someone for having once despised him because we thought him ugly? It seems to me that there are plenty of occasions on which we withhold information from healthy, mature people simply because we love them and know that the disclosure of such information would do absolutely no good, but only hurt them. In any case, I fail to see why God is morally obligated to permit the saved to have knowledge of the damned.

What then of the second hypothesis, that the blessed know of the state of the lost, but are not conscious of it because of the overwhelming joy inspired by Christ's presence? Talbott rejoins that if the beatific vision drives from the mind of the blessed the consciousness of the lost, then those experiencing such a vision become less loving (presumably because they no longer love the lost). The beatific vision would "make the redeemed less loving and thus more calloused," which is incompatible with supremely worthwhile happiness.{22}

This objection seems a clear non sequitur. A person is less loving and more callous only if he fails to love all those persons of whom he is aware; but it would be fatuous to so describe someone for failing to love a person of whose existence he is completely unaware. Supremely worthwhile happiness only entails loving unconditionally every person of whom one is aware and is therefore compatible with the hypothesis in question.

Up to this point I have argued that Talbott has not shown that it is logically impossible for the damned to exist and the blessed to experience supreme happiness which is supremely worthwhile. But suppose I am wrong about this. It still does not follow that (RH) is false. For if, given certain counterfactuals of freedom, it is not feasible for God to actualize a world in which the blessed, despite the existence of the damned, experience supreme happiness which is supremely worthwhile, it does not follow that universalism is true. For happiness that is achieved at the expense of abrogating the free will of creatures is not supremely worthwhile either. Indeed, it seems to be plausible that this forced happiness is less worthwhile that the happiness achieved by the blessed's lack of awareness of the damned. So God may not be able to actualize in the blessed both supreme happiness and supremely worthwhile happiness. Perhaps Talbott would say that God would therefore be morally obligated to refrain from the creation of persons altogether. But I think such a response is obviously false: if God can achieve supreme happiness in the blessed, even if that happiness is not supremely worthwhile, His creating that enormously worthwhile supreme happiness is a lot better than creating no happiness at all! Hence, Talbott has not proved that it is logically impossible for God to bestow on the blessed a supreme happiness which involves their unawareness of the damned.

It seems to me, therefore, that Talbott has failed to show that it is logically necessary that God refrain from creating a world in which some persons freely reject His saving grace. It is epistemically possible that God erase from the minds of the redeemed any knowledge of the lost or that the presence of Christ and the happiness he brings should drive such knowledge from their consciousness. Therefore, Talbott's second argument for the logical impossibility of (RH) fails.

In summary, neither of Talbott's two arguments succeeds in showing (RH) to be broadly logically impossible. Therefore, no inconsistency has been shown to exist between (1) and (2).

(II) Can (1) and (2) Be Shown to be Consistent?

In any original piece, I asserted that the Free Will Defender can not only rebut attempts such as Talbott's to prove that (1) and (2) are broadly logically inconsistent, but that he can prove them to be consistent by specifying a proposition which is consistent with (1) and entails (2). Unfortunately, as Talbott emphasizes, the difficulty with this approach is that the universalist who denies the possibility of (2) will inevitably also deny the possibility of the new, proposed proposition. In the face of such modal scepticism, I must confess that I do not know how to construct a proof that (1) and (2) are consistent in the broadly logical sense.

Perhaps the best that the Free Will Defender can do at this point is to emphasize how modest his claim is, to hope that his proposed third proposition both is more perspicuously consistent with (1) than is (2) and is explanatorily superior to the bald assertion of (2), so that the consistency of (2) with (1) becomes more evident, and to refute any attempts to cast doubt on the possibility of the new, proposed proposition.

Consider, then, the proposition

C. Since there is no world (without over-riding deficiencies) feasible for God in which all persons are freely saved, God has chosen a world having an optimal balance between those who freely accept His grace for salvation and those who freely reject His grace for salvation.

The Free Will Defender is not claiming that (C) is true. In the actual world, there may be suitable feasible worlds in which all persons are freely saved; indeed, maybe the actual world is one of these and universalism is correct! Whatever true counterfactuals of creaturely freedom confront God in the actual world, in other possible worlds different sets of such counterfactuals (or world-types for God{23}) obtain. The claim of the Free Will Defender is that in at least one possible world W the world-type which confronts God logically prior to His creative decree in W is such that in no suitable feasible world is every free creature saved. The modesty of this claim is evident from the fact that the Free Will Defender can also freely grant the possibility of the claim

D. In every world of free creatures feasible for God, God wins the free, affirmative response to His saving grace on the part of every person.

That is to say, in at least one possible world W* the world-type confronting God in W* is such that in every world of free creatures feasible for Him all persons are freely saved. One can also admit the possibility of the opposite claim

E. In every world of free creatures feasible for God, no persons freely accept His saving grace.

That is to say, in at least one possible world W' the world-type confronting God in W' is such that in every world of free creatures feasible for Him He is unable to win a free response to His saving grace on the part of anyone (a hyper-Calvinist might hold the actual world to be such a world). The possibilities are endless. To assert that in some possible world or other God is confronted with a range of feasible worlds which does not include a suitable world in which everyone is freely saved is modest, indeed.

Moreover, (C) seems to be more perspicuously consistent with (1) than does (2). (C) seems consistent with God's omnipotence, since counterfactuals of creaturely freedom are outside of God's control and yet place no non-logical limit on His omnipotence.{24} It also seems consistent with God's omniscience, since God's decree to create a particular world is based on His middle knowledge, which includes knowledge of all counterfactuals of creaturely freedom. Finally, the claim seems consistent with His omnibenevolence, since God is said to choose the world with the optimal balance of saved and unsaved and to accord sufficient grace for salvation even to the unsaved, who He knew would reject Him. Furthermore, (1) makes the possibility of (2) more evident by giving some explanation of how a state of affairs like that described in (2) could obtain. We see that (2) could obtain when there is no suitable world available to God in which all persons are freely saved. What would damn people would be their own free rejection of God's saving grace. Despite this unhappy situation, God, possessing middle knowledge and omnibenevolence, could still be counted on to pick a world with an optimal balance of saved and unsaved. Thus, (C) helps to make (2) more evidently possible by making sense of how (2) is consistent with (1).

Consider now, by contrast, Talbott's position. Talbott is not asserting merely that in the actual world, all persons are saved, nor is he even asserting merely (D), that in every world feasible for God all persons are saved. He is not even claiming merely that in every possible world, every world feasible for God is one in which all persons are saved. Rather he is claiming that in no possible world whatsoever does a person freely reject God's offer of salvation. This is a radically immodest claim in contrast to (C) and is certainly in need of justification.

Of course, Talbott need not establish so radical a claim in order to neutralize this step in the Free Will Defense. He need do no more than give some reason why we should not think (C) to be logically possible. Here Talbott's arguments against the broadly logical possibility of (A) become relevant. He argues first, as we have seen, that there is no reason to think that it is possible that a person would make a fully informed and free decision to reject God, and, second, that there is no reason to think that it is possible that a fully-informed decision to accept God would not be free.

We have already looked at the first of these allegations, and what I suggested there certainly seems logically possible: that in some world or other the creaturely will to self-autonomy and the bent to self-deception is so strong that some persons irrevocably resist God; or again, that in some possible world people are so evil that they reject God for its own sake. Even if these are not the case in the actual world, are they not at least possible?

The only way in which God could save such people would be by a freedom-removing revelation of Himself. That brings us to Talbott's second point: he thinks the notion of a freedom-removing revelation is incoherent. He writes, "I fail to see how a knowledge of the truth, even where it renders certain actions psychologically impossible, in any way restricts one's freedom to perform such actions--as if those in possession of the beatific vision are no longer free agents."{25} Talbott's claim is problematic. The revelation of which we speak is less like the communication of knowledge than the presentation of an irresistible lure that overpowers the will. The notion of "psychological impossibility" needs clarification; in the case in which sinners are drawn irresistibly to God as iron filings to a magnet, it does seem to be freedom-removing. Indeed, I do not find it objectionable to affirm that the blessed in heaven in the presence of Christ no longer have the freedom to reject him, since the epistemic distance necessary for such freedom has been removed.

In the end Talbott admits that this is a "difficult and controversial" matter and is willing to concede that (A) is logically possible. But he is still not prepared to concede the logical possibility of (C). Talbott argues that we have no reason to suppose that it is possible that God values the free will of His creatures above their salvation. "If a knowledge of the truth, the ability to see things as they are, is incompatible with free agency, as some conceive of it, then so much the worse for the free agency so conceived," declares Talbott. ". . . we have no reason to believe it even possible that God would withhold a revelation of truth from some persons, or keep them in perpetual bondage to ignorance and illusion, merely to maintain forever the artificial kind of free agency that Craig imagines."{26}

This consideration has, I think, a strong emotional appeal. Those of us who have unbelieving family and friends no doubt often feel that if they will not freely give their lives to Christ, it would be worth it if God would simply overpower their wills and save them in spite of themselves; do doubt, too, we feel that if we had not freely yielded our lives to God, then we wish that God would have overpowered our wills and saved us anyway. But as strong as such feelings are, they do not change the fact that such an action on God's part amounts to salvation by divine rape. For God to subvert the will of someone who chooses to reject His grace would be to violate their personhood, and that God necessarily will not do. More than our own, God's heart also breaks for those of our family and friends who reject His love and salvation, but God will not force them to repent. Hence, I regard Talbott's conjecture as logically impossible; at the very least we must say that in some possible world in which God faces such a situation He chooses not to overpower sinners' wills, and that is all the Free Will Defense needs.

We come now to Talbott's attacks on the possibility of (B), which also have relevance to the possibility of (C). His opening move is to enunciate a proposition as equally plausible as (B) and yet incompatible with it, thereby removing any grounds for our thinking (B) to be logically possible. Suppose, says Talbott, we say that God's victory over sin is complete iff God manages to repair all of the damage that sinners do (both to themselves and others). In accord with this notion, Talbott proclaims the "Victorious God Thesis:"

(VGT) Necessarily, God would have created persons whom He knew would sin only if He also knew that He could achieve a complete victory over sin.

If all feasible worlds in which persons sin are worlds in which not all are freely saved, then (VGT) entails that either God desists from creating worlds in which He knew people would sin or else He removes their freedom with respect to salvation to the extent necessary to ensure universal salvation.

Once again, however, Talbott's argument seems question-begging, based on an idiosyncratic definition of terms. As Christian thinkers we certainly affirm that God shall have a complete victory over sin, and we will probably concur that God would not have created a sinful world unless He knew His victory over sin would be complete. But the notion of a complete victory over sin according to the New Testament is that every wrong ever committed is either efficaciously expiated by the blood of Christ or punished with its just dessert. In the end injustice will not prevail; the scales of God's justice will be balanced. Just as heaven is the triumph of God's grace and love, so hell is the triumph of His holiness and justice. God's complete victory over sin does not entail that He repair the damage sinners do to themselves; on the contrary it entails that unexpiated sin receive its just recompense in the sinner. Since Talbott's definition of "complete victory over sin" is not that of the New Testament, and Talbott gives no philosophical argument in support of (VGT), the Christian thinker ought to reject (VGT) as Talbott understands it. Thus, the mere enunciation of (VGT) does nothing to undermine the logical possibility of (B) or (C).

As for Talbott's attempt to oppose (GFT) to (B) or (C), I have already commented on what seems to me the logical impossibility of so profound and massive a divine deception.

Finally, as we have seen, Talbott tries to undermine (C) by trying to show that it is incompatible with the supremely worthwhile happiness of the blessed, which attempt, I have argued, is a singular failure.

So in answer to the question of this section, we must admit that we have no proof of the consistency of (C) with (1) and therefore of the consistency of (2) with (1). But (C) does seem prima facie to be consistent with (1) and can help us to see how (2) can also be consistent with (1). Moreover, no good reason exists to reject the logical possibility of (C). Modal scepticism is in any case a two-edged sword: to the extent that it undermines our confidence of the logical possibility of (C) it also makes Talbott's attempt to prove that (1) and (2) are not true in any possible world a virtual impossibility.

III. Are (1) and (2) Plausibly True?

We come at last to the Theodicist's question concerning, not what is true in some possible world, but what is true in the actual world. I suggested that not only are (1) and (C) possibly true, but they may very well be in fact true and therefore (2) true as well. Notice that like the case of the Free Will Defender, the Theodicist's case involves a negative and a positive aspect: he asserts first, that his interlocutor has not shown the co-existence of God and the evil in question to be implausible and, second, that their co-existence is plausible. In order to undercut his opponent's argument, he need only succeed in the first, defensive move; success in the second move would be the coup de grâce. Notice, too, that the Theodicist also makes a modest claim: he does not assert that (C), for example, is true or can be proved to be true, but merely that it may well be true. The Theodicist does not claim to know the actual reason why the evil in question exists, but he offers suggestions which are not mere possibilities, but plausible explanations.

In confronting the soteriological problem of evil, the Theodicist can make the same moves that Plantinga makes concerning the probabilistic problem of evil, such as noting the fogginess of the notion of probability, the relativity of probabilities to background information, and so forth.{27} But I should like to confront the problem more directly; namely, I do not think that Talbott has shown (C) to be implausible at all--indeed, I think this proposition could very well be true.

What is implausible about (C)? Consider its assertion that in no world (without overriding deficiencies) feasible for God do all people freely accept His salvation. Even if we conceded that Congruism were true, that would not undercut the plausibility of (C), since it may well be the case that the con-creation of all the optimally conducive circumstances for salvation for all persons is not feasible for God. I think we must say that Talbott's objection to this thesis, which appeals to God's deceiving people into salvation, is, if not invalid, very implausible. Given that God must work with feasible worlds which are less than ideal, it seems not at all implausible that some people in every world which is significantly populated would reject Him, despite their being accorded sufficient grace for salvation.

On behalf of the plausibility of this aspect of (C), certain background doctrines revealed in Scripture can be adduced. On the Christian world view, man is regarded as significantly free vis à vis God.{28} He is able to rebel against God and is held accountable for it. Man in his fallen state is, as Luther put it, bent in upon himself, and this will to self-autonomy is so strong that man would rather plunge into self-destruction than bend the knee to God. We learn from Scripture that sin has a hardening effect upon man's heart: the longer and more determinedly he resists God's grace and the drawing of the Holy Spirit, the more likely he is to continue in such a state of rebellion against God (Heb. 3.7-13; cf. Ex. 7-9, Rev. 16.9, 11, 21; 22.11). In my critique of Talbott's views, I charged that he lacked a serious doctrine of sin, and I reiterate that charge here. Contrary to Talbott, the New Testament picture of sin is that those who choose it in opposition to God go from bad to worse. A biblical doctrine of sin therefore lends credibility to the possibility that some persons can freely reject God forever. So do the New Testament warnings against apostasy, which indicate that a person who has received salvific grace can reject it and that if he does so, he will be irretrievably lost (Heb. 6.4-8; 10.26-31). If a person who has known God's salvation can irrevocably reject it, a fortiori so can a person who has never known it irrevocably reject God's gracious initiatives for salvation. I have already rejected as impossible the idea that God might deceive or "jump start" sinners repeatedly in order to win their acceptance of salvation, since so to act would not treat their personhood with the respect it deserves. Moreover, it seems quite plausible, in view of the mind-boggling complexity of providentially ordering a world of free creatures, that if there are feasible worlds in which all persons freely accept God's saving grace then such worlds are deficient in other respects (for example, by having only a few persons in them). When one contemplates the incalculable difficulty of getting a significant number of free creatures to give themselves freely to God and without deception, then it seems quite plausible that in any world feasible for God involving a large number of free persons, some persons would choose to reject Him.

Consider, then, (C)'s second affirmation, that God has chosen a world with an optimal balance between those who freely accept and those who freely reject salvation. Given His omnibenevolence, it seems that God, in the moment logically prior to His creative decree, would choose a world from among those feasible worlds having such a balance.{29} Therefore, the actual world is such a world. Talbott, of course, argues that there are no feasible worlds in which persons are allowed to reject God. But his argument on this score, based on the supremely worthwhile happiness of the saved, is not very compelling. Even if the first option I suggested strikes one as implausible, the second, based on the overwhelming joy inspired by Christ's presence, seems to me very plausible. The Scriptural descriptions of theophanies and epiphanies (Ex. 33.17-23; Rev. 1.12-17; 21.3-4), as well as the mystical experiences of Christians down through church history, make it entirely believable that the presence of God should drive from consciousness an awareness of the state of the damned. Nor has Talbott given us any good reason to question the worthwhileness of this divinely inspired happiness or to think that it is less worthwhile than a happiness achieved by the abrogation of creaturely freedom.

In sum, once Talbott's arguments against the logical possibility of (C) are seen to be unsound, not much remains of them to render (C) implausible though possible. On the contrary, background considerations from a biblical world view make it altogether believable that although God is desirous of saving all human persons, God was limited in His choice of worlds to those in which universal salvation did not obtain and that He accordingly created a world in which the balance between saved and lost is as favorable as He could achieve.


Once the differing projects of defense and theodicy are properly delineated, it can be seen that Talbott's endeavor to prove the logical impossibility of Christian particularism does not succeed. While he is correct that a positive demonstration of the consistency of (1) and (2) by the Free Will Defender has not been forthcoming, the mere epistemic possibility of (RH) undercuts Talbott's own claim that they are inconsistent. Neither of Talbott's major thrusts to demonstrate the broadly logical and, hence, epistemic impossibility of (RH) can be deemed successful. Moreover, the Molinist Free Will Defender's (C) appears prima facie to be consistent with (1) and entails (2), and Talbott's attempts to show its broadly logical impossibility are no more successful than his objections to (RH). Finally, the Theodicist seems justified in regarding (C) as a plausible explanation of how it is that universalism does not obtain in the actual world.


{1}William Lane Craig, "'No Other Name': a Middle Knowledge Perspective on the Exclusivity of Salvation through Christ," Faith and Philosophy 6 (1989): 172-188.

{2}Ibid., p. 180.

{3}Thomas Talbott, "The Doctrine of Everlasting Punishment," Faith and Philosophy 7 (1990): 19-42; idem, "Providence, Freedom, and Human Destiny," Religious Studies 26 (1990): 227-245.

{4}William Lane Craig, "Talbott's Universalism," Religious Studies 27 (1991): 297-308.

{5}Thomas Talbott, "Craig on the Possibility of Eternal Damnation," Religious Studies 28 (1992): 495-510.

{6}Ibid., p. 497.


{8}With respect to (i) see my summation on p. 304 of Craig, "Talbott's Universalism"; with respect to (ii) see my prospective statement on p. 306.

{9}Talbott, "Eternal Damnation," p. 500.


{11}Talbott's admission of the possibility of (A) is problematic in an interesting way. He cannot consistently affirm that (A) is possible in the broadly logical sense because he also holds that it is broadly logically impossible that God permit any person to irrevocably refuse to be reconciled with God, so that (A) involves what Thomas Flint calls a "collapsing counterfactual," i.e., a counterfactual which entails that its antecedent is false (Thomas P. Flint, "Middle Knowledge and the Doctrine of Infallibility," in Philosophical Perspectives, vol. 5: Philosophy of Religion, ed. James Tomberlin [Atascadero, Calif.: Ridgeway Publishing, 199], pp. 373-393). For if in no world does God permit persons to refuse reconciliation with Him, then necessarily if it is true that the instantiation of E were to refuse reconciliation under any freedom permitting circumstances he were in, then God does not instantiate E in such circumstances. A collapsing counterfactual cannot be true, and so Talbott cannot consistently admit the logical possibility of (A). If he does admit it, then he must reject his claim that necessarily God does not permit persons to refuse reconciliation with God. If Talbott holds (A) to be merely epistemically possible, then he is admitting that for all we know (A) is in fact true. Such an admission nullifies his first argument against (RH).

Talbott cannot adopt Flint's solution of asserting that the counterfactual only counterfactually implies the falsity of its antecedent, but does not entail it, for then there are possible worlds in which God permits people to refuse reconciliation with God. Nor can Talbott adopt Flint's second solution of replacing (A) with something like

(A') There exists at least one creaturely essence E such that, for any circumstances C in which the instantiation of E would be free in the matter of being reconciled to God, the instantiation of E would not in fact freely consent to be reconciled to God in C.

because (A') still involves a collapsing counterfactual. Rather he must affirm something like

(A") There exists at least one creaturely essence E such that, for any circumstances C, if the instantiation of E were to be created in C, the instantiation of E would not freely consent to be reconciled to God.

(A") does not imply that E's instantiation would be damned, but that, on Talbott's view, he would be saved without his free consent in some C. Such a position still gives away Talbott's first argument against the coherence of a choice to reject God.

{12}Perhaps Talbott would say that he concedes the possibility of (A) only under an artificial and false conception of "freedom," according to which a stunning revelation might be freedom-removing. But under a correct understanding of freedom, (A), or better, (A") is false. According to this interpretation the instantiation of E would not freely consent to reconciliation in the artificial sense, but he would in the correct sense. Such an interpretation of Talbott's position, however, still undercuts his argument for the broadly logical impossibility of (RH), since he admits that his understanding of "freedom" is "a difficult and controversial-controversial matter" which he does not try to settle (Talbott, "Eternal Damnation," p. 502). Therefore, it is epistemically possible that his understanding of "freedom" is incorrect, thereby undermining his argument that (RH) is not logically possible.

{13}Talbott, "Eternal Damnation," p.504.

{14}Ibid., p. 505.

{15}The difference between the envisioned circumstances C3 and a possible world is evident not only from Talbott's belief that even God could not create rational agents ex nihilo with perfect understanding and clarity of vision, but more importantly from the fact that God's deception, to be moral, must be merely temporary, so that those who freely accepted salvation in C3 are freed from their illusions. This latter circumstance is very important, since it is not at all evident on Talbott's view that persons who freely accept salvation in C3 could not, upon being confronted with the fact of their deception, including the illusory existence of those they loved, rebel against God and reject His salvation.

{16}Talbott, "Eternal Damnation," p. 507.


{18}Ibid., p. 509.

{19}See also David P. Hunt, "Middle Knowledge and the Soteriological Problem of Evil," Religious Studies 27 (1991): 3-26, who envisions the elect interacting with a world populated by soulless automata.

{20}I have tried to state this in such a way that it does not involve a collapsing counterfactual, as does Talbott's statement, "If every creaturely essence suffered from transworld reprobation and God . . . knew this, they either He would refuse to instantiate any essences at all or, if he did instantiate some of them, he would have a morally sufficient reason not to leave created persons entirely free with respect to salvation . . ." (Talbott, "Eternal Damnation," p. 504).

{21}Talbott, "Eternal Damnation," p. 507.

{22}Ibid., p. 510.

{23}On the notion of a world-type, see Thomas P. Flint, "The Problem of Divine Freedom," American Philosophical Quarterly 20 (1983): 255-264.

{24}See Thomas P. Flint and Alfred J. Freddoso, "Maximal Power," in The Existence and Nature of God, ed. Alfred J. Freddoso (Notre Dame, Ind.: University of Notre Dame Press, 1983), pp. 81-113.

{25}Talbott, "Eternal Damnation," p. 502.

{26}Ibid., pp. 502-503.

{27}Alvin Plantinga, "The Probabilistic Argument from Evil," Philosophical Studies 35 (1979): 1-53.

{28}See helpful discussion in D. A. Carson, Divine Sovereignty and Human Responsibility, New Foundations Theological Library (Atlanta: John Knox, 1981), pp. 18-35.

{29}I assume that an optimal balance exists. If there does not, then God is no more obliged to create such a balance than He is to create the "best of all possible worlds."

Copyright (C) William Lane Craig. All Rights Reserved.