Imagine a person so stubborn that no matter what God does, no matter what world God places them in, they freely refuse Him. Every appeal is wasted. Every kindness is ignored. In every possible arrangement of history, this person turns away. If even one such person could exist, then the argument of the previous two chapters has a hole in it. God cannot save everyone freely, because at least one person would reject Him no matter what.

That is not a caricature. It is the sharp edge of a famous idea from one of the greatest philosophers of the twentieth century: Alvin Plantinga. In his landmark work God, Freedom, and Evil (1974), Plantinga introduced a concept he called "transworld depravity." He built it to solve a different problem (the problem of evil, not the problem of universalism), but it reaches into our question with real force.¹ If transworld depravity is possible, then maybe God simply cannot guarantee that every free person will say yes to Him. The non-coercion argument of Chapter 15 would be in trouble.

This is the most serious technical objection in this book, and I want to take it seriously. We will look at the hole, and then we will close it.

Why It Matters

The case I have been building runs through a chain. God wants all people saved (Chapter 4). God knows what each person would freely do in any situation (Chapter 14). So God can arrange, for each person, the appeal most likely to win a free yes (Chapter 15). The chain leads to a strong conclusion: a God with this kind of knowledge and this kind of love can save everyone freely.

But what if the chain breaks at a key link? What if there is some person for whom no sequence of appeals would work? Not because God is weak, and not because the appeals are poorly chosen, but because the person, in every possible arrangement, would freely refuse? Then God's hands are tied. He cannot save that person without forcing them. And if He forces them, He violates their freedom, and the whole case collapses.²

This is why the transworld-depravity objection matters so much. It attacks the universalist hope at the exact point where freedom and love meet. If even one creature is "transworld damned" (if they would freely reject God in every world God could make), then universal salvation is not just unlikely. It is impossible without overriding freedom.³

The stakes reach further still. If transworld damnation is real, then the Molinist framework itself, which has been doing so much work in this book, turns into a weapon against us. Kenneth Keathley describes the Molinist picture cleanly. From the infinite set of possible worlds, God knows which ones are feasible. These are the worlds that would actually unfold as He intends, given the free choices of real people. God then freely chooses among those feasible worlds.³⁵ The universalist adds one more step: God chooses a feasible world in which all are freely saved. But if transworld damnation is real for even one person, then no such feasible world exists. Molinism, the very tool that powered the non-coercion argument, would rule out the universalist's conclusion.

R. Zachary Manis has seen this tension sharply. In his careful analysis of Molinism and the problem of hell, Manis argues that if Molinism is true, God has the power to bring about any feasible world. That includes one where all are saved. Molinism and universalism, he writes, "would seem to be a match made in heaven." But he also sees the anti-universalist's escape route: the claim that no feasible world exists where all are saved and heaven is filled. That escape route, Manis shows, requires a claim that is "simply incredible." The Molinist anti-universalist must hold that there is no feasible universalist world with a large enough population. He must also hold that the lost in one feasible world are lost in every feasible world. And he must hold that this is not merely possible but actually true.³⁹

That is why this objection cannot be dodged. It has to be answered on its own terms. And the answer, as we will see, has to do with time. Plantinga's worry assumes a fixed set of chances. But what happens when the chances are not fixed? What happens when God has eternity?

The Argument

Move 1. The Conceptual Move: What Transworld Depravity Means

Plantinga's concept runs like this. Think of all the possible worlds God could create. In each one, a given person faces different situations, different choices, different temptations. Plantinga suggested that a person might suffer from "transworld depravity" if, in every one of those worlds, they would freely go wrong at least once. No matter where God places them, no matter what circumstances surround them, they make at least one bad choice.⁴

Plantinga used this idea to solve the problem of evil. His point was simple: if free creatures can suffer from transworld depravity, then God could not create a world with free creatures and no moral evil. Even an all-powerful God cannot force a free creature to always choose rightly. That is not a limit on God's power. It is just what freedom means.⁵

But the concept has a darker cousin: transworld damnation. A person has transworld damnation if, in every feasible world God could create, they would freely reject God forever. They would never say yes. Not because they lacked the chance, not because they were tricked, but because in every situation, they would freely choose against God for all eternity.⁶

Notice the difference. Transworld depravity is about making some wrong choice. Transworld damnation is about making the final wrong choice: the permanent rejection of God. It is transworld damnation that threatens our hope, because if even one actual person suffers from it, God cannot save everyone freely.

Move 2. The Logical Move: The Objection in Full

This is a clean argument, and both premises are hard to deny outright. Premise (1) looks almost trivially true. If freedom is real, then surely it is possible that someone would always refuse. We cannot rule it out by logic alone. Premise (2) follows from what freedom means: a forced "yes" is not a free "yes."⁷

Kirk MacGregor, a leading Molinist philosopher, states the worry plainly. If the human soul has libertarian freedom, then "it is possible for certain souls to separate themselves from God during their earthly sojourn." And they might "forever continue their rejection of any divine overtures to save them." If God's middle knowledge tells Him that this is what they would do in every feasible world, then God faces a stark choice: create a world with some people who are never saved, or create no world at all.⁸

Move 3. The Metaphysical Move: The Infinite-Opportunity Answer

The transworld-damnation worry looks powerful at first glance. But it rests on a hidden assumption that, once you see it, changes everything. The assumption is this: the person faces a finite set of chances to say yes.

Think about what the objection really asks. It asks whether there is a person who, across every feasible world, would reject God at every moment forever. In other words, it asks whether a finite creature could hold out against an infinite series of perfectly fitted appeals from a God who knows them completely.

This is where time becomes the key. If God has only a finite window (say, the person's earthly life), then transworld damnation looks more plausible. Maybe there are people who, in any arrangement of a seventy-year life, would never turn to God. That is a hard claim to prove, but it is at least imaginable.

But the universalist is not limited to a seventy-year window. The case I have been building says that death is not the deadline on grace (Chapter 19), and that God has all the time eternity allows (Chapter 21). So the real question changes. It is not "could this person refuse in a single life?" It is "could this person refuse across an endless series of fresh, perfectly suited appeals?"⁹

Eric Reitan and John Kronen, in their philosophical defense of universalism, God's Final Victory, press exactly this point. They present what they call the "Argument from Infinite Opportunity." The argument is simple in structure. God need only do three things: strip away barriers to free choice that He can remove, keep each person in existence until they freely choose God, and leave the choice of God open at every moment. If God does these three things, then across an infinite timeline, every person will freely choose salvation. The probability of any person holding out forever drops toward zero.¹⁰

Timothy Stratton makes a connected point from the Molinist side. If God is practically rational, then He not only intends ends but also knows and supplies the means to reach those ends. Middle knowledge gives God exactly that: knowledge of what each person would freely do in every set of circumstances. A God who has this knowledge and who has eternity to work with is not guessing. He is composing, for each soul, the most effective possible series of appeals, each one crafted with full knowledge of what that person needs.³⁶ The question is whether there could be a person for whom no such series would ever work, even across infinite time. That is a very strong claim.

Here is an analogy. Imagine flipping a coin. On any single flip, heads is perfectly possible. But ask this: is it possible to flip a fair coin an infinite number of times and never get tails? The answer is no. The probability of never-once-tails shrinks toward zero as the flips increase. It does not reach zero in any finite number, but across an infinite series, an outcome that has any chance at all will eventually come.¹¹

Now, a free choice is not a coin flip. I want to be careful here. A person is not a random event. Libertarian freedom means the person is the source of their choice, not a mechanism governed by probability.¹² But the comparison still helps at a key point: the question is whether a nonzero chance, repeated endlessly, yields a near certainty. And it does. Even for a nonrandom agent, the claim that they would never once freely yield, across an infinite series of ever-better appeals, is a very strong claim. It amounts to saying that this person's refusal is so deep that no amount of light, love, or truth could ever crack it. Not even once. Not even after ten thousand years. Not even after ten trillion.

Move 4. The Probabilistic Move: Possible Is Not Probable

Here is where the careful thinker draws a line that makes all the difference. There is a gap between what is logically possible and what is probable. The objection needs only the first. But the universalist's hope needs only to deny the second.

Is transworld damnation logically possible? I am willing to grant that it might be. If freedom is real, I cannot prove by logic alone that every person must eventually say yes. That is why I hold conviction, not proof (Chapter 2).¹³

But is transworld damnation probable? That is a very different question. And the answer, I believe, is no. It is vanishingly improbable.

Consider what would have to be true for transworld damnation to hold. A finite person, created for relationship with God, would have to sustain a free refusal through an endless series of encounters with a God who knows them completely, who loves them perfectly, and who crafts each appeal to address exactly what holds them back. Each appeal removes another layer of ignorance, heals another wound, shines another beam of light. And this person says no. Every single time. Forever.

Hunter Coates, drawing on the work of Reitan and Kronen, makes this vivid with an analogy. Imagine a penny in a box, with glue on one side. Every time you shake the box, there is a chance the penny lands glue-side down and sticks. Shake the box enough times, and the penny will eventually stick. Now suppose you shake the box forever. The chance that the penny never sticks approaches zero. It does not matter how small the chance is on any single shake. Over infinite shakes, even a tiny chance becomes a near certainty.¹⁴

Coates extends the analogy. Replace the penny with a person in hell who has libertarian freedom. Replace the box-shake with a fresh encounter with God's love. At each encounter, there is some chance (however small) that the person freely says yes. Over infinite encounters, the chance that they never once say yes becomes, in Coates's words, "infinitesimal." That is, the probability gets closer to zero than any real number, though it never quite reaches it.¹⁵

Now, Coates writes from a position of stronger confidence than I do. He says "mathematically guaranteed." I would say "practically certain" or "overwhelmingly probable." The distinction matters, because I keep libertarian freedom, and a truly free choice is not a coin flip. MacGregor makes this correction carefully: libertarian freedom presupposes agent causation, and a nonrandom event does not follow the same probability rules as a random one.¹⁶ I agree. But even with that correction, the weight of the reasoning holds. A never-ending refusal, sustained against ever-better appeals, is not something we should expect from a finite creature made for God.

Andrew Hronich sharpens the point further. Hronich introduces a concept he calls "transworld sanctity," the mirror image of transworld damnation. A person has transworld sanctity if every world in which they always choose wrongly or finally reject God is a world God cannot bring about. On Hronich's view, given rational freedom (freedom aimed at the good), it is plausible that no creature suffers from transworld damnation. He endorses the philosopher Joshua Rasmussen's result: if the number of possible persons is infinite, then the probability that no evil-free feasible world exists is itself zero.¹⁷

That is the mirror of Craig's transworld damnation. Where Craig says some people would reject God in every feasible world, Hronich says it is plausible that every person would accept God in at least one. And if God has infinite feasible circumstances to offer each person, the force of the point multiplies. Hronich begins his analysis by laying out the options: either all creatures suffer from transworld damnation, some do, or none do. If all do, then no one can be saved, which is absurd. If some do, God can choose not to create those particular people. And if none do, universalism follows naturally.⁴³

I want to note an honest distinction here. Hronich's argument, like Coates's, runs toward necessary universalism, the claim that all must be saved. I do not go that far. I keep libertarian freedom, and with it I keep the bare logical possibility that someone could hold out. That is the one open question (Chapter 17). But the probability argument still does real work. It shows that transworld damnation, while perhaps logically possible, is so unlikely that it should not trouble the person who holds the universalist hope as a reasoned conviction.¹⁸

Move 5. The Constructive Move: What Survives

The transworld-damnation objection says: maybe God cannot save everyone freely, because at least one person would refuse in every world. The answer is: that claim rests on the assumption that the person faces a bounded set of chances. Remove the bound (give God eternity and endless, perfectly fitted appeals), and the probability of a never-ending refusal drops toward zero.

This answer does not require us to deny freedom. It does not require us to claim that the will is determined. It does not even require us to claim that universal salvation is logically necessary. All it requires is this: that the combination of perfect knowledge, perfect love, and unlimited time makes eternal refusal so unlikely that a reasonable person may hold the hope with confidence.

What does this mean for the non-coercion argument? It means the argument stands, not as a logical proof, but as a powerfully grounded conviction. God may not be able to guarantee, by a single act at a single moment, a free yes from every person. But a patient God, armed with perfect knowledge and endless love, with all the time eternity allows, drives the holdouts toward zero.¹⁹

Think of what we have built across these chapters. Chapter 14 gave us the mechanism: God knows what each person would freely do in any set of circumstances. Chapter 15 gave us the method: God uses that knowledge to offer each person the appeal most fitted to win them, without overriding their freedom. This chapter closes the gap: even if, for some person, no single appeal is sure to work, an infinite series of appeals, each perfectly fitted, makes the holdout vanishingly unlikely.³⁷

Keathley is not a universalist. He describes the Molinist framework as a way for God to predestine events through free choices. But notice what the framework makes possible: if God can "set the table" for each person's free choice, and if He has unlimited time to keep setting the table, then the question of transworld damnation becomes the question of whether any person could resist every table-setting, forever. The machinery that Keathley describes, combined with the patience that eternity allows, points toward the larger hope.

Freedom is intact. The person's "yes" is their own. God does not force it. He waits for it, works toward it, offers the light again and again. And the light is not getting dimmer. Under the divine-presence model (Chapter 12), the fire is God's love, and the appeal grows more compelling over time, not less, because each encounter peels back another layer of blindness.²⁰

Hope is intact. The shepherd goes after the one lost sheep "until he finds it" (Matthew 18:12). There is no moment when the shepherd gives up. There is no world where the search ends in failure. The shepherd seeks until the sheep is found.

Jesus does not say the shepherd searches for a while and then stops. He says the shepherd searches until he finds. That is the picture. Patient love, across whatever time it takes, aimed at every lost one.²¹

The Philosophers at the Table

This question sits at a crowded intersection of modern philosophy of religion, and the voices deserve a careful hearing.

Alvin Plantinga, the originator of the concept, did not build transworld depravity for the sake of hell. He built it to defend God against the logical problem of evil, and he did so brilliantly. His free-will defense remains one of the most celebrated arguments in analytic philosophy. But Plantinga himself never pressed transworld depravity into a doctrine of eternal damnation; his interest was in the possibility of moral evil, not in the finality of anyone's rejection of God.²²

William Lane Craig, the foremost contemporary defender of Molinism, takes the concept further. Craig has argued that transworld damnation is a real possibility within the Molinist framework. On his view, some people may be such that, in every feasible world, they would freely reject God, and God, in His providence, ensures that those who would be saved in any feasible world are saved in the actual world.²³ Craig uses this to defend the justice of God in a world where not all are saved. I respect the argument. But I think the infinite-opportunity answer shows that transworld damnation, while a logical possibility Craig is right to name, is not a probability we should rest on.

MacGregor engages the question at length, replying to R. Zachary Manis's critique of what Manis calls the "no-feasible-universalist-world hypothesis." MacGregor argues that Manis's objections to a world where some are lost can be met. He contends that on a Molinist model with transworld damnation, God still treats each person as an end, not a means, because He gives each one optimal grace and respects their free choices.²⁴ MacGregor is not a universalist. But his analysis is helpful for our purposes, because it shows how the Molinist machinery handles the concept. The universalist simply takes the next step: if God has eternity to work with, then the number of feasible worlds expands, and the chance that no universalist world is feasible drops toward zero.

Manis himself presses the other side of this argument with real force. He notes that the Molinist anti-universalist faces a steep burden. The defender of hell must claim not only that there is no feasible world where all are saved and heaven is well populated, but also that those who are lost in one feasible world are lost in every feasible world. And this combined claim must be not merely possible but actually true.

Manis is not a universalist, and I want to say that clearly. He is a careful philosopher writing about the problem of hell, and his conclusion is more modest than mine. But his argument is a gift to the universalist case, because it shows that even a non-universalist philosopher finds the anti-universalist's Molinist defense to be, in his word, "incredible."⁴⁰

MacGregor states the worry honestly. But notice his framing: "during their earthly sojourn" and "to forever continue their rejection." The first phrase limits the initial encounter to this life. The second assumes the rejection is forever sustained. What the universalist adds is this: why assume the appeal stops at death? And if it does not stop, why assume the rejection can last forever against an ever-growing flood of light?²⁵

Jerry Walls presses this line of critique from a different angle. Walls, who is not a universalist, calls Craig's claim that all those who never hear the gospel suffer from transworld damnation "extremely implausible." As Walls notes, it strains belief to suppose that all the people who lived and died in regions the gospel had not yet reached would have rejected it if they had heard it. He points to places like Korea and sub-Saharan Africa, where Christianity flourishes today, and asks whether it is really credible that none of the ancestors of these believers would have responded to the gospel had they heard it.⁴¹

There is also a biblical reason for doubting transworld damnation. In Matthew 11:21–24, Jesus tells the cities of Chorazin and Bethsaida that if the miracles done in them had been done in Tyre and Sidon, those cities "would have repented long ago in sackcloth and ashes." Taken at face value, this means the people of Tyre and Sidon were only contingently lost. They refused God under one set of circumstances but would have accepted Him under another. That is the opposite of transworld damnation.⁴²

MacGregor also engages Reitan's claim that libertarian freedom, given infinite time, guarantees universal salvation. MacGregor pushes back, noting that libertarian choices are not random events and that the agglomeration step (moving from "each individual will probably be saved" to "all will be saved") is not automatic. He appeals to Talbott's concept of psychological impossibility, noting that a genuinely hateful person might, on libertarian freedom, never choose to stop hating God.²⁶ I grant the force of this reply. It is exactly the kind of pushback that keeps me at conviction rather than proof. But I note that MacGregor's scenario requires a person whose hatred is self-sustained against an infinite series of healing encounters, a claim that grows harder to believe the more seriously we take God's love and knowledge.

Thomas Talbott, the philosopher whose inconsistent triad anchors Chapter 5, comes at the question from the other direction. Talbott argues that a truly free will, freed from ignorance, deception, and disordered desire, has no remaining motive to reject God. On his view, the only things that drive us away from God are exactly the things that God's presence heals.²⁷ That is a powerful thought. The book will develop it fully in Chapter 17.

Counter-Objections

First Objection: Probability Is Not Certainty

"Even infinitely many tries cannot guarantee success if each try could fail. You say the chance of eternal refusal approaches zero, but it never reaches zero. So you have not solved the problem."

I grant this. And I have been careful to say so throughout. The probability of eternal refusal is vanishingly small, but it is not strictly zero. That is why I hold conviction, not logical proof. The case this book builds is cumulative and probabilistic, not deductive (Chapter 2).²⁸

But consider what the objector is left with. If the probability of eternal refusal approaches zero, then the person who bets on it is betting on an outcome that is less likely than almost anything we can name. It is the theological equivalent of betting that a fair coin will come up heads every time, forever. You cannot prove it will not happen. But it would be strange to build your doctrine of God on the expectation that it will.

The honest question is: which position requires more faith? The universalist says a loving God, working with perfect knowledge over infinite time, will win every heart. The traditionalist says at least one person will hold out against that love forever. I know which claim seems more extraordinary to me.²⁹

Second Objection: Maybe Refusal Is Self-Reinforcing

"Your argument assumes that each fresh appeal is independent of the last. But maybe refusal gets harder to break over time, not easier. Maybe each 'no' makes the next 'no' more likely. A heart can harden, and hardening feeds on itself."

This is the strongest residual pushback, and it deserves a careful answer. Michael McClymond presses a version of this worry with real force. He charges that Talbott's picture of postmortem salvation looks like "the postmortem equivalent of waterboarding or electroshock, at ever-increasing levels of intensity, until the poor sufferer finally 'cries uncle' and surrenders to God's demand to be recognized." Though the universalist speaks of love, McClymond says, "this seems more like torture."³⁰ James De Young presses a related charge: universalism, he argues, "is more deterministic and coercive than Calvinism is."⁴⁴ These are serious objections. If the persistent appeals of God amount to breaking down the will by force, then the universalist hope is no better than the necessity it claims to refuse.

There is something right about this worry. We know from experience that habits of refusal can deepen. A person who turns away from God once may find it easier to turn away again. The biblical language of "hardening" (as in Pharaoh's heart, Exodus 7:13) takes this seriously.

But the hardening worry assumes a specific picture: that the person is left alone with their refusal, sinking deeper into it as time passes. That is not the picture the universalist draws. Reitan and Kronen address this directly. Even if a critic argues that "it is a fact of human psychology that our choices produce habits which dispose us to make similar future choices," they reply that under the model of infinite opportunity "it would be madness to think" that damnation is forever sustained. The person who insists on eternal refusal is making a claim about an infinite timeline, and across that timeline the probability that the refusal never breaks remains vanishingly small.⁴⁵ Under optimal grace (Chapter 15), God does not simply repeat the same appeal. He adapts. He knows what caused the hardening, and He addresses it. Under the divine-presence model (Chapter 12), the person is not hiding from God in a dark corner. They are standing in the full light of God's love. That light does not dim over time. It grows clearer.³¹

The hardening of a heart in this life happens in darkness, in ignorance, in the absence of God's full presence. But what happens when the ignorance is removed? When the deception is stripped away? When the person sees God as He truly is? The conditions that produce hardening are exactly the conditions that God's presence dissolves. That is the argument of Chapter 17, and it is the deepest answer to the hardening worry.³²

God is not in a hurry. He is patient. And His patience is not passive waiting. It is active, searching love, fitted to each soul, with all the time eternity allows.

Conclusion: The Hole Closed, the Hope Held Humbly

The transworld-depravity objection is the most technical challenge in this book, and it deserves the respect I have tried to give it. Plantinga's concept is real. The logical possibility it names cannot be disproven by force. And if freedom matters (and it does), then we cannot simply declare that everyone must say yes.

But "possible" is not "probable." The chance that a finite person, made for God, would freely refuse Him at every moment across an infinite series of perfectly fitted appeals, each one crafted by a God who knows them better than they know themselves, is vanishingly small. It is so small that resting the doctrine of eternal hell on it looks less like caution and more like a bet against love.

The constructive conclusion: God may not be able to guarantee a free yes from a given person at a given moment. But a God who knows every heart, who crafts every appeal, who has all of eternity, and who never gives up, will, across that eternity, see every refusal melt. Not because the creature had no choice. But because the light was too bright, the love too patient, and the truth too clear for any darkness to hold forever.

I think the person who objects to universalism on the ground of transworld damnation is making a specific bet. They are betting that somewhere in all of creation there is a soul so set against God that no amount of love, truth, beauty, or healing could ever reach it. Not after a thousand years. Not after a trillion. Not ever. That is a remarkable bet. It is a bet against the character of the God who said, "I have loved you with an everlasting love" (Jeremiah 31:3).³⁸

Jesus told a parable about a persistent widow who kept asking a judge for justice "and he will not" (Luke 18:4). In the end, the unjust judge gave in. How much more will a just and loving God, who never tires, whose appeals never weaken, whose knowledge never fails, press the case for every soul?³³

The cumulative case is growing. Freedom is real. God's knowledge is perfect. God's grace is fitted to each soul. And time is on love's side. The next chapter asks the deepest question of all: could anyone really refuse God forever?³⁴

Notes

↑ 1. Alvin Plantinga, God, Freedom, and Evil (New York: Harper & Row, 1974); see also Plantinga, The Nature of Necessity (Oxford: Clarendon Press, 1974), ch. 9. Plantinga introduced transworld depravity as part of his free-will defense against the logical problem of evil. The concept was not originally soteriological, but its implications reach directly into the question of universal salvation.

↑ 2. The non-coercion argument is developed in Chapter 15. Its central claim is that God, via middle knowledge, can arrange for each person the appeal most likely to win a free response, without overriding freedom. Transworld damnation, if actual, would break this chain for at least one person.

↑ 3. The distinction between transworld depravity (some moral wrong in every world) and transworld damnation (final rejection in every world) is important. The first is Plantinga's original concept; the second is the stronger version that bears on universalism. See William Lane Craig, "No Other Name: A Middle Knowledge Perspective on the Exclusivity of Salvation through Christ," Faith and Philosophy 6, no. 2 (1989): 181–84, for an early application to salvation. Reitan and Kronen describe this as "the pivotal" proposition in Craig's defense of eternal damnation: "some persons might be characterized by 'trans-world damnation'; that is, that they would finally and irrevocably reject God if God actualized any feasible possible world that they are 'inhabitants' of" (God's Final Victory, 194 n. 18). As Manis notes, "Craig cannot take the analogous line with respect to transworld damnation without running into theological absurdity," because if all creaturely essences suffered from transworld damnation, no one could be saved (Manis, Sinners in the Presence of a Loving God, 133 n. 40). Hronich confirms that "Craig does not believe that all essences suffer from transworld damnation, whereas Plantinga believes it is possible that all essences suffer from transworld depravity" (Hronich, Once Loved Always Loved, ch. 2, n. 35).

↑ 4. Plantinga, God, Freedom, and Evil, 47–49. Plantinga defined transworld depravity in terms of a creature who would freely perform at least one morally wrong action in every feasible world. The technical definition is: a person P suffers from transworld depravity if, for every world W that God could weakly actualize and in which P is significantly free, P would go wrong with respect to at least one action.

↑ 5. This is the heart of Plantinga's free-will defense: it is logically possible that God could not create a world with free creatures and no moral evil. That possibility is sufficient to defeat the logical problem of evil. See Plantinga, God, Freedom, and Evil, 44–53.

↑ 6. "Transworld damnation" is the term used by Craig and others to denote the soteriological extension of Plantinga's concept. See Kirk R. MacGregor, Molinist Philosophical and Theological Ventures (Eugene, OR: Wipf & Stock, 2022), ch. 15, where MacGregor treats transworld damnation as part of the "no-feasible-universalist-world hypothesis."

↑ 7. The argument has the form of a modus tollens against the universalist's claim. If "God can save all freely" entails "there is a feasible world in which all are saved," and if transworld damnation shows no such world is feasible, then the universalist's claim is false. The question is whether the minor premise (transworld damnation is actual for some person) is well supported.

↑ 8. MacGregor, Molinist Philosophical and Theological Ventures, ch. 15. MacGregor's precise framing: "If the human soul possesses libertarian freedom, as Molinism holds, then it is possible for certain souls to separate themselves from God during their earthly sojourn and to forever continue their rejection of any divine overtures to save them." He explores two versions of this hypothesis, the second incorporating transworld damnation.

↑ 9. This is the crucial shift. On the traditionalist picture, the "test" of a human life is bounded by death. On the universalist picture developed in this book, God's offer extends beyond death and across all the time eternity allows. See Chapter 19 and Chapter 21.

↑ 10. Eric Reitan and John Kronen, God's Final Victory: A Comparative Philosophical Case for Universalism (London: Continuum, 2011), 160–67, esp. §8.3 "Libertarian Freedom and the Infinite Opportunity Argument." Reitan and Kronen present their formal "Argument from Infinite Opportunity" (AIO) on pp. 160–61: God need only (a) strip away all salvation inhibitors He can remove prior to salvation, (b) sustain all persons in temporal existence until they choose communion with God, and (c) leave the choice of communion an "open choice" at every moment. Given (a)–(c), "the choice whether to embrace communion with God is not made once, but is confronted every moment of an existence of potentially infinite duration. At each moment of this indefinite existence, there is an even chance that this person will choose communion with God — at which point the person is saved. Under these conditions, the person's salvation becomes a mathematical certainty" (161). AIO's support for universalism is, by the authors' own admission, probabilistic rather than deductive (129), which is why this book holds conviction rather than necessity.

↑ 11. In probability theory, if an event has probability p > 0 at each trial, and the trials are independent, then the probability of the event never occurring in n trials is (1 − p)ⁿ, which approaches 0 as n approaches infinity. The coin-flip analogy is widely used in the philosophical literature to make this point accessible. Reitan and Kronen give it a formal rendering using a "possible world tree" in which "the number of possible worlds in which Fred remains unsaved becomes a progressively smaller percentage of the possible world segments" and "as the timeline approaches infinity, within this tree the percentage of world segments in which Fred remains unsaved approaches zero" (God's Final Victory, 162). See also Hunter Coates, Grace Abounds: A Holistic Case for Universal Salvation (2024), ch. 11, who develops the analogy at length.

↑ 12. This is MacGregor's important correction. "Libertarian freedom presupposes agent causation," and agent causation means the agent is the ultimate source of the choice, not a random mechanism. Mathematically, a nonrandom event with a nonzero intrinsic probability may fail to happen even given infinite time. See MacGregor, Molinist Philosophical and Theological Ventures, ch. 15. The point is well taken: the probability argument is not a strict proof. But it still generates a rational conviction.

↑ 13. The book's shape as conviction rather than proof is stated in Chapter 2 and runs through every chapter. I part from Hart on precisely this point: Hart denies that a truly free will can finally reject God; I allow the bare logical possibility but argue it is overwhelmingly improbable. See Chapter 13 for the parting from Hart on freedom.

↑ 14. Coates, Grace Abounds, ch. 11. Coates attributes the penny-and-box analogy to Reitan and Kronen and develops it at length. The glue on the penny represents the natural draw toward God; the box-shake represents a fresh encounter with God's love. Over infinite shakes, the penny sticks. Over infinite encounters, the person freely yields.

↑ 15. Coates, Grace Abounds, ch. 11. Coates writes: "The percentage of 'possible worlds' in which the penny does not stick heads-side down approaches zero as the timeline approaches infinity." He further argues that the number of persons makes no difference: "the same analysis can be applied to every penny. Hence, the same outcome will be obtained whether there are one hundred billion pennies in this box or two pennies in this box."

↑ 16. MacGregor, Molinist Philosophical and Theological Ventures, ch. 15. MacGregor notes that the "detachment rule for probabilities" (inferring p from high probability of p) does not apply to nonrandom events in the same way, and that the "agglomeration step" (from high probability of individual salvation to high probability of universal salvation) is also not automatic. He cites Jonathan Kvanvig, Destiny and Deliberation: Essays in Philosophical Theology (Oxford: Oxford University Press, 2011), 23–24.

↑ 17. Andrew Hronich, Once Loved Always Loved: The Logic of Apokatastasis (2023), ch. 2 ("Determined to Believe?"). Hronich proposes "transworld sanctity" as the opposite of transworld damnation: a person has transworld sanctity if every world in which they always choose wrongly or finally reject God is a world God cannot weakly actualize. He endorses Joshua Rasmussen's result that the probability of there being no evil-free feasible worlds approaches zero as the number of possible persons approaches infinity. Hronich's argument runs toward necessary universalism; this book takes the force of his reply while keeping conviction rather than necessity.

↑ 18. The distinction between conviction and proof is load-bearing for this entire book. See Chapter 2. The probability argument does not prove that transworld damnation is impossible; it shows that it is so improbable that a reasonable person may hold the universalist hope with confidence.

↑ 19. This is the cumulative weight of Chapters 14–16 taken together: middle knowledge provides the mechanism (Chapter 14), optimal grace provides the appeal (Chapter 15), and the infinite-opportunity argument closes the transworld-depravity hole (this chapter). The full weighing of the cumulative case is the work of Chapter 29.

↑ 20. The divine-presence model, developed in Chapter 12, holds that the "fire" of hell is God's own unveiled love, experienced as torment by those who resist it. Over time, as resistance softens, the same love that burns also heals. See R. Zachary Manis, Sinners in the Presence of a Loving God: An Essay on the Problem of Hell (Oxford: Oxford University Press, 2022).

↑ 21. The parable of the lost sheep (Matthew 18:12–14; cf. Luke 15:3–7) is one of three "lost" parables in Luke 15. In each, the seeker does not give up. The shepherd leaves the ninety-nine to find the one; the woman sweeps the house until she finds the coin; the father watches the road until his son comes home. The pattern is relentless, personal pursuit.

↑ 22. Plantinga's free-will defense is widely regarded as having solved the logical problem of evil (the claim that God and evil are logically incompatible). See J. L. Mackie, "Evil and Omnipotence," Mind 64 (1955): 200–212, for the original statement of the problem, and Plantinga, God, Freedom, and Evil, for the defense. Mackie himself conceded in The Miracle of Theism (1982) that Plantinga's defense was successful against the logical form of the argument.

↑ 23. Craig, "No Other Name," 181–84. Craig argues that God, via middle knowledge, has arranged the actual world so that every person who would freely accept Christ in any feasible world is placed where they will in fact hear and respond. Those who are not saved, on this view, would not have been saved in any feasible world. The universalist's response: this calculation changes if the "set of circumstances" includes postmortem encounters across eternity.

↑ 24. MacGregor, Molinist Philosophical and Theological Ventures, ch. 15. MacGregor appeals to the Doctrine of Double Effect: God does not intend anyone's damnation, even if He foresees it. He also argues that being in God's "repletive presence" (even in hell) is an intrinsic good, better than nonexistence. This is a subtle and honest engagement.

↑ 25. MacGregor's framing assumes that the relevant appeal is bounded by the "earthly sojourn." The universalist's case removes that bound. See Chapter 19.

↑ 26. MacGregor, Molinist Philosophical and Theological Ventures, ch. 15. MacGregor constructs a formal version of Reitan's argument and identifies the controversial premises: (1) libertarian freedom contains a random element, and (3) if this random element does not cover salvific choices, then all persons will eventually be saved. He argues that agent causation rules out (1), and that without randomness, the probability argument does not yield necessity. I agree that the argument does not yield necessity; I claim it yields conviction.

↑ 27. Thomas Talbott, The Inescapable Love of God, 2nd ed. (Eugene, OR: Cascade, 2014). Talbott's central contention is that the motives for rejecting God (ignorance, deception, bondage to desire) are exactly the conditions that the full encounter with God removes. His sharpest formulation of this claim is quoted at length in Walls, Hell: The Logic of Damnation, 116: "As long as any ignorance, or deception, or bondage to desire remains, it is open to God to transform a sinner without interfering with human freedom; but once all ignorance and deception and bondage to desire is removed, so that a person is truly 'free' to choose, there can no longer be any motive for choosing eternal misery for oneself." Walls concedes that "this issue cannot be settled on empirical grounds" and engages Talbott's argument at length; see Walls, Hell, 115–36. Manis observes that "Talbott remains convinced that self-deception is at most a temporary barrier to reconciliation to God" and that "God need only allow one to experience the natural consequences of rejecting Him" (Manis, Sinners in the Presence of a Loving God, 204 n. 34). The full development of Talbott's claim is the work of Chapter 17.

↑ 28. See Chapter 2 for the full statement of the book's epistemic posture: the case is cumulative and probabilistic, yielding overwhelming conviction rather than deductive proof. This posture is the parting from Hart's necessitarianism and what keeps freedom real.

↑ 29. This is a comparative-plausibility argument. Both positions (universalism and traditionalism) rest on claims about what free persons will do across eternity. The question is which claim is more extraordinary: that perfect love, with perfect knowledge and infinite time, will win every heart, or that at least one heart will hold out forever against that love. The universalist bets on love; the traditionalist bets against it.

↑ 30. Michael J. McClymond, The Devil's Redemption: A New History and Interpretation of Christian Universalism, 2 vols. (Grand Rapids: Baker Academic, 2018), V 2, p. 956. McClymond is responding to Talbott's account of postmortem conversion in The Inescapable Love of God. McClymond writes: "With the best of motives, God employs the postmortem equivalent of waterboarding or electroshock, at ever-increasing levels of intensity, until the poor sufferer finally 'cries uncle' and surrenders to God's demand to be recognized. Though Talbott speaks of love, this seems more like torture." He further charges that "the effort to extend grace to all undermines the principle of grace" (V 2, p. 956). The charge is powerful against a model of ever-increasing pain. But it does not land on the model this book defends: optimal grace (Chapter 15) works through persuasion fitted to the person, not through escalating torment. The "hardening" motif in Scripture is also complex. In the Exodus narrative, the text alternates between "Pharaoh hardened his heart" (Exodus 8:15, 32; 9:34) and "the LORD hardened Pharaoh's heart" (Exodus 9:12; 10:20, 27; 11:10). Both formulations appear, suggesting a dynamic interplay between human choice and divine action, not a simple one-way mechanism.

↑ 31. Under optimal grace (Chapter 15), God does not offer the same appeal twice. He adapts, because He knows (via middle knowledge) what each person would freely do in each new situation. So the "hardening" objection assumes a static appeal; optimal grace is a dynamic one, fitted to the person's changing condition.

↑ 32. Talbott's argument, developed fully in Chapter 17, is that the causes of hardening (ignorance, deception, disordered desire) are precisely what God's full presence removes. A heart cannot harden against a light it fully sees; hardening happens in shadow, and God's presence is the end of shadow.

↑ 33. The parable of the persistent widow (Luke 18:1–8). Jesus tells it "to the effect that they ought always to pray and not lose heart" (v. 1). The a fortiori logic is explicit: if an unjust judge will relent under persistent asking, how much more will a just and loving God respond? The universalist reads the persistence as a picture of God's own relentless pursuit of each soul.

↑ 34. Chapter 17 takes up the question this chapter has prepared: whether anyone could hold out against God's love forever. That chapter owns the "one open question" at the object level. The cumulative weighing of the whole case is reserved for Chapter 29.

↑ 35. Kenneth Keathley, Salvation and Sovereignty: A Molinist Approach (Nashville: B&H Academic, 2010), ch. 1. Keathley summarizes the Molinist framework in terms of three "moments" of God's knowledge: natural knowledge (everything that could happen), middle knowledge (everything that would happen given free choices), and free knowledge (everything that will happen in the world God chooses to create). The universalist's concern is that, if transworld damnation obtains, no feasible world features universal salvation, so God's creative choice is constrained.

↑ 36. Timothy A. Stratton, Human Freedom, Divine Knowledge, and Mere Molinism (Eugene, OR: Wipf & Stock, 2020), and his collaboration with Jacobus Erasmus on "best feasible freedom-permitting worlds." Stratton argues that if God is practically rational, He "not only intends ends, but also knows and supplies the necessary means to achieve those ends," and that middle knowledge is the means by which God accomplishes His creative purposes through free agents. Stratton and Erasmus posit that "the actual world is the best feasible freedom-permitting world," leaving room for what they call "hopeful universalism."

↑ 37. The three chapters together form a chain: mechanism (middle knowledge, Chapter 14), method (optimal grace, Chapter 15), and the closing of the transworld-depravity gap (this chapter). The chain yields a conviction-grade assurance that God can save all freely. The full weighing of the cumulative case is reserved for Chapter 29.

↑ 38. Jeremiah 31:3 (ESV). The Hebrew ‘ahavat ‘olam ("everlasting love") carries the sense of a love that has no end and no conditions that could extinguish it. If God's love for a person is truly everlasting, then the question is not whether God will stop pursuing them but whether they can hold out against that pursuit forever. The universalist says the odds of holding out are vanishingly small.

↑ 39. R. Zachary Manis, Sinners in the Presence of a Loving God: An Essay on the Problem of Hell (Oxford: Oxford University Press, 2022), 122–34. Manis observes that "Molinism and universalism would seem to be a match made in heaven" (123), since middle knowledge gives God the power to actualize any feasible world, including one where all are saved. But the anti-universalist must then show that no such feasible world exists with adequate population and that those who are lost in one feasible world are lost in every feasible world. Manis calls this combined claim "simply incredible" (133–34). He concludes that "if this is the price Molinism charges to shore up anti-universalism, the anti-universalist is best advised to keep shopping" (134). Manis is not a universalist; he is engaged here as an interlocutor whose critique of the anti-universalist position serves our argument.

↑ 40. Manis, Sinners in the Presence of a Loving God, 133–34. Manis further points out a crucial difference between Craig's argument and Plantinga's original: "Craig cannot take the analogous line with respect to transworld damnation without running into theological absurdity," because if all creaturely essences suffered from transworld damnation, no one could be saved (133 n. 40). Plantinga could claim universal transworld depravity (that everyone does some wrong in every world) without absurdity, but Craig cannot claim universal transworld damnation (that everyone finally rejects God in every world). The asymmetry matters. Gordon Knight, cited by both Manis and Hronich, sharpens the moral point: "God, in creating persons whom he knows will never accept Christ, shares a responsibility for the resultant suffering in hell." See Knight, "Molinism and Hell," in The Problem of Hell: A Philosophical Anthology, ed. Joel Buenting (Farnham: Ashgate, 2010), 112. Stephen Kershnar adds that if a person is truly damned in every world, "this would be true only if their rejection of God's grace followed from their essential nature and this would so greatly restrict their act-options as to undermine their responsibility for rejecting God's grace." See Kershnar, "Injustice of Hell," 118, cited in Andrew Hronich, Once Loved Always Loved: The Logic of Apokatastasis (2023), ch. 2.

↑ 41. Jerry L. Walls, Hell: The Logic of Damnation (Notre Dame: University of Notre Dame Press, 1992), 96–97. Walls finds it "extremely implausible" that Craig's transworld damnation thesis could serve as a theodicy rather than a mere defense. He writes that "it is exceedingly hard to entertain seriously the notion that all the persons who lived and died in countries the gospel did not reach for centuries would have rejected it if they had heard it." Walls is not a universalist; he defends the possibility of postmortem salvation, and he reads Craig's view more charitably by suggesting that the "unsaved" in the hypothesis could refer to "those persons who have decisively rejected, or will decisively reject God's grace, offered in optimal conditions, either in this life or after death" (97). See also Hronich, Once Loved Always Loved, ch. 2, citing Walls's response. Hronich adds a further practical observation: if those who have never heard the gospel suffer from transworld damnation, then sending missionaries to them is pointless, since they would never have accepted the gospel anyway.

↑ 42. Matthew 11:21–24 (ESV). This passage is used against transworld damnation by both Manis (Sinners in the Presence of a Loving God, 134 n. 42) and Hronich (Once Loved Always Loved, ch. 2). The counterfactual Jesus states ("if the mighty works done in you had been done in Tyre and Sidon, they would have repented") implies that the people of Tyre and Sidon were only contingently lost. Craig dismisses this as "probably religious hyperbole" (Craig, Only Wise God, 137), but both Hronich and the plain reading of the text resist this move. As Hronich observes, while Jesus certainly intended to shock His hearers, that does not require His counterfactual to be false.

↑ 43. Hronich, Once Loved Always Loved, ch. 2. Hronich also proposes a stronger version (T₂) at the level of creaturely essences: "A creaturely essence E possesses transworld sanctity if, and only if, every possible world in which the instantiation of E faces one or more moral choices and either (a) always freely chooses wrongly or (b) eventually and freely rejects God and spurns His grace is a world that God cannot weakly actualize." He credits the formal structure of T₂ to Talbott's framework in "Providence, Freedom and Human Destiny," 235. See also Craig, "No Other Name," 182, for the anti-universalist's attempt to block this move: Craig argues that even if each individual could be saved in some feasible world, "it does not follow that there are worlds which are feasible for God in which all individuals are saved." The universalist's reply, developed in this chapter, is that infinite time and optimal grace expand the feasible set beyond what Craig's bounded-life assumption allows.

↑ 44. James B. De Young, Exposing Universalism: A Comprehensive Guide to the Faulty Appeals of Universal Reconciliation (Eugene, OR: Wipf & Stock, 2020), pp. 66, 217–19, 238–39. De Young charges: "By insisting that all those in hell must repent, UR is more deterministic and coercive than Calvinism is." The charge rests on the assumption that guaranteed outcomes require coerced choices. But the model this book defends separates the two. God does not force the choice; He fits the appeal, and eternity gives the appeal time to work. A guarantee that rests on the power of love combined with unlimited time is not determinism. It is patience. See Chapter 26 for the full engagement with De Young's philosophical case.

↑ 45. Reitan and Kronen, God's Final Victory, 162–63 (§8.4 "Objections to AIO(3)"). A critic might argue that "our choices produce habits which dispose us to make similar future choices." Reitan and Kronen reply that defenders of eternal damnation "are not merely asserting that Fred is alienated from God up to some time T, but that Fred is damned at every time T. Insofar as this claim is about an infinite timeline, the mathematics of infinity applies." They press the analogy: "Given an indefinite (if not infinite) timeline, the question becomes whether we will eventually reach some time T at which we choose communion with God. This is akin to asking whether, given the ability and willingness to flip a fair coin indefinitely (even if that should take a trillion tries), we will eventually toss a coin that lands heads-up. The answer to the question above is such a resounding yes that we have what amounts to a guarantee. Certainly, under these conditions it is not merely reasonable to believe that the coin will eventually land heads-up — it would be madness to think otherwise" (163). Though the argument is couched in probability rather than proof, its force is considerable.

The Transworld-Depravity Objection

The Question on the Table