Failed solutions

The problem of free guarantees is actually quite easy to solve if Molinism is true. God could use middle knowledge to ensure that he never creates any human who does not eventually accept his offer of salvation. If God necessarily adopts this ‘no human left behind’ policy, then all possible worlds in which humans exist are ones in which all humans are saved. Molinists like William Lane Craig object to universalism not because God can’t guarantee universal salvation without violating human freedom, but because they believe it is open to God not to adopt the ‘leave none behind’ policy needed to do so.Footnote ² That is, their objection is really to (1) of the universalist argument rather than (2) or (3).

My goal, however, will be to solve the problem of free guarantees without relying on any controversial assumptions about divine foreknowledge (including the assumption that God has foreknowledge). A quick overview of previous attempts will prove instructive.

Talbott (Reference Talbott1990, Reference Talbott2001, Reference Talbott2014, Ch. 11) argues that God can ensure universal salvation by placing any who resist God’s offer in a remedial version of hell. There the damned experience the full wretchedness of their spiritual condition. Inevitably, Talbott claims, those in hell will become so miserable that they will, like the prodigal son, come to their senses and accept God’s ongoing offer of salvation.

To avoid Pelagianism, Talbott (Reference Talbott1988, 17–18) adds that God’s grace enables the damned to accept this offer. Moreover, the first human movement in acceptance consists merely in the cessation of resistance to God, and is itself possible only through prevenient grace.Footnote ³ Taking advantage of their surrender, God bestows on them whatever positive attitudes (e.g., faith) are necessary for salvation.

Is it not possible for those in hell to hold out forever – to be ‘rebels to the end’ as C.S. Lewis (Reference Lewis2001, 130) puts it? Talbott says no. For a choice to be genuine, not some random spasm of the will, it must be rational in the minimal sense of being based on reasons. But anyone in hell who isn’t plagued by ignorance or disordered desire will have zero reason to continue rebelling.

How does God rid people of ignorance and disordered desire though? Talbott (Reference Talbott1990, 36–37) sometimes speaks as though God can accomplish this unilaterally,Footnote ⁴ but clarifies elsewhere that God educates the damned by allowing them to experience the natural misery that results from sinful resistance. Given enough time, personal experience convinces them that continued resistance is futile and they stop desiring to rebel (Talbott Reference Talbott, Parry and Partridge2003b, 261–262).

Jerry Walls (Reference Walls1992, Ch.5; Reference Walls, Parry and Partridge2003, 112; Reference Walls2015, Ch. 3) doubts that insight is inevitable. The misery of the damned might teach them the futility of resisting God … or they might construe it as further evidence of God’s malicious and tyrannical character. On the other hand, if the miseries of hell really are so unbearable as to render self-deception impossible, then repentance is determined and, assuming incompatibilism, thereby unfree.

In this way, Walls (Reference Walls2015, 81) presses Talbott with ‘A Hell of a Dilemma’. If the miseries of hell necessitate repentance, then that repentance isn’t free in a libertarian sense. If the miseries of hell do not necessitate repentance, then repentance isn’t assured.

As I read him, Talbott concedes the point (or at least declines to defend against it).Footnote ⁵ He writes,

Instead, Talbott turns his sights to what Eric Reitan (Reference Reitan2007, 415ff) calls ‘rational freedom’, which Reitan says ‘is [in Talbott’s estimation] the only kind of freedom really worth having, and hence the only kind of freedom that God would have a moral obligation to preserve’.

In short, Talbott doesn’t offer us a solution to the original puzzle that our acceptance of salvation cannot be both guaranteed and free in the libertarian sense.

Here is where Reitan takes up the case. While favorable to Talbott’s focus on rational freedom, Reitan (Reference Reitan2007, sect. IV, Reference Reitan, Parry and Partridge2003, 137–140) tries to show that those in hell might be guaranteed to repent even if they retain libertarian freedom. Offer those in hell enough chances to be saved while furnishing them with the grace that makes repentance possible, and it becomes ‘mathematically certain’ that they accept the offer eventually. Jeremy Gwiazda (Reference Gwiazda2011) makes the same point in terms of probability: so long as the chances of accepting each offer are greater than 0, then the probability that all offers are rejected approaches 0 the longer the offers continue.

Unfortunately, even if Reitan’s entire line of reasoning is granted, it doesn’t resolve the problem of free guarantees. This is because indefinite offers secure ‘what amounts to a guarantee’ but not an actual guarantee (Reitan Reference Reitan2007, 425). We can conclude that all will be saved but not that this outcome is necessary. Reitan (Reference Reitan, Parry and Partridge2003, 140) is aware of this:

Neither Talbott nor Reitan succeeds in solving the problem of free guarantees, then. Talbott gets a genuine guarantee but only by replacing libertarian freedom. Reitan maintains libertarian freedom but must settle for less than a full guarantee.

Source incompatibilism and rational determination

My own solution to the problem of free guarantees requires source incompatibilism. Since I cannot argue for the truth of source incompatibilism here, my thesis is conditional: if source incompatibilism is correct, then God can guarantee that all humans freely accept his offer of salvation.

Incompatibilism says that freedom is incompatible with causal determinism. Incompatibilism is generally motivated in one of two ways. Some identify alternative possibilities as central to freedom, then argue that determinism precludes alternative possibilities (e.g., van Inwagen Reference van Inwagen1983). Others identify ultimate sourcehood – ‘that the ultimate sources of our actions lie in us and not outside us’ (Kane Reference Kane2005, 6) – as the essence of freedom, and rule out determinism on the basis that determinism rules out sourcehood (Klein Reference Klein1990; Kane Reference Kane1996; Stump Reference Stump1999; Zagzebski Reference Zagzebski2000; Pereboom Reference Pereboom2001; Kane Reference Kane2005; Timpe Reference Timpe2012).Footnote ⁷ Source incompatibilism follows the latter path.

If causal determinism is true, then our choices are fully explained by factors external to us and outside our control – like the initial conditions of the universe in conjunction with the laws of nature (Kane Reference Kane2005, 122–123; Timpe Reference Timpe2012, 138–140). But in that case, the ultimate sources of our choices lie outside us.

Regarding alternative possibilities, some source incompatibilists are happy to relinquish them entirely (Stump Reference Stump1999; Zagzebski Reference Zagzebski2000). Others, myself included, think that we cannot be the ultimate source of our choices without having alternative possibilities at some point (Kane Reference Kane2005, 124–131; Timpe Reference Timpe2012, Ch. 9). Otherwise our choices are effectively in the same position they were in under causal determinism, being fully explicable by whatever external factors entail those choices. In this way, alternative possibilities are a necessary condition on freedom without being that in which freedom consists.

While alternative possibilities are necessary at some point, they are not necessary at every point. Given source incompatibilism, it’s possible that individual choices be determined (there is no possibility of doing otherwise in those circumstances) yet free.

To see how this is possible, start with the largely (but not entirely) uncontroversial observation that free choices must be made based on reasons.Footnote ⁸ Choices not based on reasons would not be genuine choices at all but something more like spasms of the will. So if an agent S must choose either to A or not-A, and she has reasons to A but no reasons to not-A, then S is determined to choose to A. When an agent is determined in this particular way, she is rationally determined.Footnote ⁹

While the clearest cases of rational determination are those where the agent lacks any reason to not-A, they are not the only ones. Eleonore Stump (Reference Stump1999, 323) offers the example of being offered a nickel to cut her daughter into little pieces. Here Stump has a reason to cut up her daughter (a nickel isn’t nothing!), but the reason is woefully inadequate. It is so inferior to Stump’s reasons for refraining that it is psychologically impossible for Stump to act on that reason. Thus, an agent can be rationally determined to choose to A not because she lacks any reason to not-A but because her reasons to A are overpowering.

From the standpoint of source incompatibilism, a choice can (in select instances) be both rationally determined and free.Footnote ¹⁰ To see this, begin with an ordinary choice that every libertarian concedes is free. Say there’s a cupcake in front of me, and I’m deciding whether to eat it. I have reasons on both sides, neither of which are overpowering: cupcakes taste good, but I don’t need the calories. I’m not causally determined to make either choice – the laws of nature and total state of the universe leave open what I do now. Add the most favorable story you can think of about how my choice comes about (e.g., via agent-causation). I choose to refrain from eating the cupcake.

Now change the situation slightly. Instead of this being just an ordinary cupcake, it’s my 4-year-old daughter’s special birthday cupcake – one that she’s been eagerly awaiting for days now. I notice that the cupcake looks tasty. If I scarf it down quickly and feign innocence, no one would know I ate it. Should I take it or not? Everything about this choice is the same as before with one exception: this time the reasons to refrain are overpowering. It’s not that I don’t have a reason to take the cupcake (cupcakes still taste good), but acting on that reason is unthinkable to me – I wouldn’t ever do it in those circumstances. I choose to refrain from eating the cupcake (and chuckle that the prospect even popped into my mind).

If the first choice is free, then why not the second? I am as much the ultimate source of the second choice as I am the first.

This reasoning seems persuasive save for one complication. I have the reasons I do, and those reasons have the strength that they have, because of my character. My character steers what things I take notice of, whether I recognize them as good, and how good I evaluate them as being. This means that coming to have a certain character can rationally determine me to make certain choices and not others. Well, what if stealing my daughter’s cupcake is unthinkable for me simply because I was born with an agreeable character, or, more extravagantly, because my neurologist wife did some rewiring the previous night so I wouldn’t ruin our daughter’s big day? Ultimate sourcehood is maintained only if the character which rationally determines my choice is formed in the right kind of way – or at least not formed in the wrong kind of way.

In particular, three things must be true about the character that rationally determines me if those rationally determined choices are to be free.

First, there cannot be any complete explanation of my character in factors external to me. If there were, then there would be a complete explanation of my choice in factors external to me (since my character fully explains my choice), and we would run into the same concerns we had with casual determinism. My choice could be explained without any mention of me.

This outcome is avoided if my present character can be partially explained (in a non-deviant way) by prior free choices that were themselves free and rationally undetermined. Following Robert Kane (Reference Kane1996, Reference Kane2005, 130–131), we can call these ‘self-forming choices’.Footnote ¹¹ If my character is the result of self-forming choices, then we cannot explain why my character is as it is now without mentioning me. But then we cannot explain my choice without mentioning me either.

Second, it must be that any self-forming choices were not made in conditions of coercion, manipulation, exploitation, or any others that eliminate my responsibility with respect to those choices.

Third, when I’m making these self-forming choices, it needs to be reasonably foreseeable that they might result in me coming to make choices of this further sort (the ones that end up being rationally determined) (Hartman Reference Hartman2020, 1426–1431). Let’s assume, following Timpe (Reference Timpe2011), that this requires me to have a justified belief to this effect (if only dispositional) or to be culpable for failing to have such a belief.

In sum, I propose that S’s choice to perform action A in accordance with character C can be both rationally determined and free in a libertarian sense if the following conditions are met:

1. (i) There is no complete explanation of C in terms of factors external to S;

2. (ii) No self-forming choices necessary to explain C were formed in conditions that removed S’s responsibility for them (such as coercion, manipulation, exploitation, etc.);

3. (iii) It was reasonably foreseeable to S that S’s self-forming choices (if there are any) might result in S’s choosing to do A.Footnote ¹²

When a rationally determined choice meets these conditions, let’s call it a ‘self-determined choice’.

From a source incompatibilist perspective, the claim that self-determined choices are free should be very appealing. There is no possibility of doing otherwise, granted, but freedom does not consist in alternative possibilities. While the lack of alternative possibilities is normally a reliable indicator of a failure vis-à-vis sourcehood, it is not so here. Everything we ordinarily want with respect to sourcehood is still present.

Precursors

The possibility of libertarian free actions being rationally determined has been explored in other contexts. Kane, in responding to Daniel Dennett (Reference Dennett1984), considers Martin Luther’s famous refusal to recant: ‘Here I stand. I can do no other.’ Kane (Reference Kane2005, 82), a libertarian, suggests that Luther might still have been responsible ‘even if Luther could not have done otherwise’ so long as ‘he was responsible by virtue of earlier choices and actions for being the sort of person he had become at that time’. Kane continues, ‘Often we act from a will already formed, but it is “our own free will” by virtue of the fact that we formed it by other choices and actions in our past for which we could have done otherwise.’Footnote ¹³

Building on an idea from James Sennett (Reference Sennett1999), Timothy Pawl and Kevin Timpe use this sort of reasoning to resolve the problem of heavenly freedom.Footnote ¹⁴ It is possible for the saints in heaven to be free yet determined not to sin so long as this determination results from ‘the character the redeemed have formed in their pre-heavenly existence’ (Pawl and Timpe Reference Pawl and Timpe2009, 410).

Finally, Timpe (Reference Timpe2014, Ch. 7; Reference Timpe, Timpe and Speak2016) argues that God can be rationally determined to make some choice (to keep a promise, for instance) yet still make that choice freely. This is because God ‘is not made to have his moral character’ by anything external to him. Nor is God’s character ‘something that He just happens to have’ (Timpe Reference Timpe2014, 109). It is necessary that his character be precisely as it is – that is, the divine nature precludes God from ever having anything other than an impeccable moral character. It follows that God’s decision to keep his promise meets all three conditions of a self-determined choice: the first is satisfied because God’s moral character is explained by his own nature and not by anything external, and the second and third because God makes no self-forming choices at all. That is why his choice remains free, even if determined.

The self-determining solution: The set-up

The possibility of free, rationally determined choices opens up a solution to the problem of free guarantees. What if those in hell inevitably become so miserable that they become rationally determined to repent? So long as they enter into this state of brokenness as a result of their own free choices, then their repentance is both free and guaranteed. This is ‘the self-determining solution’ to the problem of free guarantees.

Here are the central, non-trivial claims required for the self-determining solution to work. First, there is:

The term ‘hell’ is disposable here. By ‘hell’ I simply mean a place of remedial chastisement in which two things are true: one, it is increasingly miserable the longer people choose to rebel, and, two, this misery is the consequence of their own rebellion – that is to say the misery should not be externally imposed like lashes from a whip but consist in the inner turmoil that naturally results from rebellion against God.

I envision hell as an isolated environment where people undergo a vivid and unceasing experience of their true spiritual condition.Footnote ¹⁵ Since their spiritual condition is miserable, the experience of that condition would feel miserable, and would be accompanied by feelings of guilt, shame, rage, anxiety, longing, depression, despair, loneliness, and more – the full panoply of hellish miseries that plague the sinful even in this life. The rebels in hell would be free to deceive themselves about these experiences and the feelings they engender. They might tell themselves that God is a tyrant who demands slavish obeisance and that these experiences are illusions sent to torture them into submission. Still, they cannot escape the experiences themselves, which cannot be avoided even for a moment, and only get worse the more they double down on their rebellion.

Even if the miseries of hell are different than I’m envisioning, and start out less intense, they are certain to become horrifying when extended over a long enough period of time. Existence itself would become torturously boring if we did not have something as endlessly enrapturing as God to preoccupy us (Williams Reference Williams1973). Imagine, then, placing someone in a sinful condition in solitary confinement from which there is no relief. What solitary confinement does to people after just a few years is shocking. Can you imagine how desperate one’s situation would be after a billion years? It’s unspeakable. This is part of why mild views of hell aren’t plausible.Footnote ¹⁶ They not only fail to do justice to the Scriptural witness about the dreadfulness of hell, but they don’t properly take into account the length of eternity and how misery is magnified over time.

A common objection to the set-up is that some metaphysical obstacle renders postmortem repentance impossible. Most prominently, Aquinas holds that the disembodied soul cannot engage in the sorts of activities required to alter the soul’s fundamental disposition toward God – a concern shared by a number of early church fathers. This objection relies on controversial metaphysical commitments about the soul and the nature of moral and spiritual growth,Footnote ¹⁷ but even if those commitments are granted, the objection still fails. God can simply restore the damned to a mutable body. Walls (Reference Walls2011, 143) points out that Aquinas commits himself to this possibility in accepting the legend of Trajan, a Roman emperor restored to life by the prayers of Gregory the Great, who subsequently came to faith and was saved.Footnote ¹⁸ What’s more, the Scriptures themselves say that the damned will be restored to bodies prior to judgment (Jn 5:29; Dan 12:2). Accordingly, Walls (Reference Walls2011, 143) explains, ‘The more typical reasons for denying postmortem repentance are in the vein that God would not grant such repentance after death, rather than that he could not.’

Guaranteed repentance

The next step in the self-determining solution is to assert that, in the aforementioned conditions, repentance is inevitable:

The idea here is intuitive enough. The misery mounts. Eventually the reasons to stop resisting become overpowering (assuming one hasn’t thrown in the towel long before then). It needn’t be, as Talbott envisions, that people ever get to a point where they entirely lack reasons to keep resisting. It just needs to be the case that whatever they are getting from their rebellion pales in comparison to their overwhelming desire to escape. There are limits, remember, to just how irrational human choices can be – we cannot act for no reasons at all, nor for reasons that are overpowered. And given enough time, how could the scales not be tipped overwhelmingly in favor of surrender?

Avoiding guaranteed repentance requires that one’s commitment to rebellion can be so absolute that no matter how incomprehensibly miserable that rebellion makes her, she remains forever unwilling to cede even the slightest bit of territory to God. Such hard-heartedness strains credulity. Not only from a psychological standpoint but from a theological one.

If Augustine is correct that ‘You have formed us for Yourself, and our hearts are restless till they find rest in You,’Footnote ¹⁹ then we should expect anyone in a state of rebellion to be, well, restless, searching, longing, and so inherently unstable, capricious, liable to change. The invincible resoluteness, the fixed determination of will, necessary to hold out forever simply isn’t realizable within the sinful soul. We cannot be committed to anything so absolutely save God himself, for that would be to find our rest in something other than him.Footnote ²⁰ Reitan (Reference Reitan2022, 380) seems to be getting at something similar when he writes: ‘While we can, through sinful choices, make ourselves very broken indeed, we cannot reach such a degree of absolute depravity that there remains no lingering foothold for the good – because our nature, stamped with God’s image, is that foothold.’

Indeed, such a foothold seems built into the nature of rationality itself. All rational action is motivated by a desire for something construed as good, and thus grounded in a primordial desire for the transcendent Good as such. Even our rejection of God, therefore, arises from the inescapable pull we feel toward him. Accordingly, our rejection of God can never be absolute. We can never bring ourselves to a place where there is absolutely no desire for God to which he can appeal.

Furthermore, if we ever did reach a point of being completely closed off to God, why couldn’t God simply open us back up? Robert Hartman points out that, on widely held understandings of depravity, humans begin their earthly lives determined to reject God’s offer of salvation, yet God empowers us to accept his offer without any violation of our freedom. There is no reason, Hartman (Reference Hartman2023, 719–720) argues, God couldn’t offer this same prevenient grace to those in hell. Notably, the main objections to this reasoning contend that God wouldn’t soften the hearts of the damned, not that he couldn’t (Hartman Reference Hartman2023, 720-721). Such objections do not challenge God’s ability to guarantee repentance, only his willingness.

Even if these considerations are not decisive, surely the evidence weighs heavily in favor of guaranteed repentance. I suspect skepticism on this point usually arises from prior commitments about how to read Scripture,Footnote ²¹ or deference to the tradition,Footnote ²² rather than disagreement about the psychology of hell itself. In any case, I’ll say more in defense of this step when responding to objections (see objection 1).

Self-formed character

Assuming that repentance is guaranteed, does it meet the requirements of a self-determined choice?

Regarding condition (i):

This is pretty plainly true. The damned are given free rein. If their resistance makes them so miserable that it breaks them, then that brokenness is the result of their own choices.

The fact that they didn’t intend to become broken is irrelevant. Consider Rodion Raskolnikov, the protagonist of Dostoyevsky’s Crime and Punishment. Raskolnikov decides to murder a lecherous pawnbroker in a kind of Nietzschean attempt to live beyond good and evil. He intends for his conscience to die alongside his victim. The opposite occurs. Raskolnikov is so haunted that he is forced to acknowledge the moral law’s claim on him. Eventually, he is compelled to make a full confession, despite having the opportunity to get off scot-free. Raskolnikov’s repentant character was self-formed, even though he didn’t intend to become this sort of person and even though he intended to become the opposite.

I’ll consider another challenge to this step of the argument below (see objection 4).

No coercion

Condition (ii) on self-determined choices is that the relevant self-forming choices cannot be made under conditions of coercion (or manipulation, exploitation, etc.).Footnote ²³ The satisfaction of this condition is more controversial. Walls (Reference Walls2004, Reference Walls1992, 132) says this process seems an awful lot like conversion at the point of a sword. J.I. Packer (Reference Packer, Morgan and Peterson2004) says it comes uncomfortably close to brainwashing. The self-determining solution denies these characterizations:

I’ll defend this claim below (see objection 2).

Reasonably foreseeable

Condition (iii) for a self-determined choice is that it must be reasonably foreseeable to the damned that their rebellion ultimately results in repentance:

This seems true. It should be plain to anyone who considers it that those in hell might eventually make themselves so miserable that they reach a breaking point. The requirement, keep in mind, is not that the damned believe they will be broken, but only that they might be broken. If they truly lack this belief (even dispositionally), then such ignorance can only be culpable. God can easily ensure this by giving the damned evidence for all the key points. Such evidence can be distorted through an act of self-deception, but that only highlights the culpability of their ignorance. In any case, I argue below (see objection 1) that such acts of self-deception cannot last forever.

Genuine surrender

Two clarifications before moving on.

First, it may seem that, at best, hellish torment could guarantee only unwilling obeisance – conformance of outward action without inner change (e.g., McClymond Reference McClymond2018, 956). This is mistaken. If there is no genuine inner submission to the will of God, then there is no relief from the misery. Those in hell might try a million times to fake surrender, to ‘give God what he wants’ while remaining defiant in spirit, but God cannot be mocked. Only the genuine cessation of resistance can rescue them from the spiritual anguish their rebellion engenders. Eventually, they will see there is no recourse save true, on-their-knees, dear-God-please-help-me brokenness. And once that point is reached, it is God himself who cultivates the more positive aspects of repentance within them.Footnote ²⁴

As proof of concept, look at Alcoholics Anonymous and other 12-step programmes. Step 1 is admitting powerlessness and the inability to manage one’s own life. Step 2 is the recognition of a higher power that offers restoration. Step 3 is giving oneself over to this higher power. This is the foundation of recovery. And it is (apart from some miraculous intervention) achievable for addicts only after hitting rock bottom. What is rock bottom? It’s being so utterly broken by suffering that you stumble upon the gift of desperation – a willingness to do absolutely anything, even die, if that is what it takes to stop going on as before. This is the crack in one’s resistance that allows the grace of God in. And you know what – it works! The result of such recovery programmes is not mere behaviour modification but inner renewal, a completely new and abundant way of life.

Second, as any recovering addict will tell you, our initial surrender is almost always partial and imperfect. Resistance is relinquished bit by bit. Once the process gets started, however, there is no reason to doubt God can bring it to completion.Footnote ²⁵ After tasting relief, the person will have strong reasons to keep cooperating with God; but even if they reach a point where they refuse to surrender any further, the original process repeats itself. The misery mounts until it breaks their resistance. Even if they don’t run towards heaven, step-by-step they will back away from hell.

Objection 1: Ignorance and vice

Objectors like Walls (Reference Walls, Parry and Partridge2003) generally cite two obstacles to guaranteed repentance: ignorance and vice. The person in hell may have overwhelming motivation to escape but remain ignorant of how to do so, or the person may know how to escape but be so gripped by disordered desire that she refuses to do so. These are formidable obstacles, but they can and necessarily will be overcome given enough time (and the occupants of hell have no shortage of that).

However much the person desires not to repent, it’s inevitable that they come to desire relief even more. In saying this, I do not by any means downplay just how difficult disordered desire can be to overcome. Augustine writes:

I don’t disagree. My contention is simply that the damned will eventually come to realize the truth of Jesus’s words: ‘And if your right hand causes you to sin, cut it off and throw it away. For it is better that you lose one of your members than that your whole body go into hell’ (Matt 5:30).

But what if someone remains ignorant of how to escape her misery? What if she believes surrender will increase her misery, not lessen it? Such ignorance cannot be due to a lack of evidence. Personal experience will provide the damned with ample evidence that rebellion is the source of their troubles; and God is able to furnish supplemental evidence if necessary.Footnote ²⁷ The damned remain ignorant, then, only by skewing the evidence in an act of self-deception.

The thing about self-deception is that it must be motivated.Footnote ²⁸ If what I have said about the overwhelming misery of hell is correct, then the damned inevitably reach a point where they no longer have adequate motivation to deceive themselves. Their one, all-consuming priority is escaping their misery. Deception is a luxury they can no longer afford. The reasons not to self-deceive become overpowering.

It’s worth noting that even Walls (Reference Walls1992, 132–133) seems to concede this step in the argument. When he appeals to self-deception and disordered desire, his point isn’t that they prevent any guarantee of repentance. That much, Walls admits, is assured in conditions of ever-increasing misery. His point is that they prevent any free guarantee of repentance – that something about the conditions renders the choices made within them unfree.

I already addressed Walls’ concern that those choices are unfree simply because they are determined. Walls’ other concern is that those choices are unfree because coerced.

Objection 2: Turning the screws

Walls is not the only one who thinks this vision of hell sounds coercive. Michael McClymond (Reference McClymond2018, 956) describes it this way:

While surrender would be guaranteed in such conditions, the coercive nature of those conditions would prevent such surrender from being free.

The key difference between hell and torture is that the pain of torture is externally imposed, whereas the misery of hell is the result of one’s own sinful condition. The damned in hell are miserable because that is what it feels like to be separated from God, not because God is artificially visiting that suffering upon them.Footnote ²⁹ As the Scriptures say, ‘Your wickedness will punish you; your backsliding will rebuke you. Consider then and realize how evil and bitter it is for you when you forsake the Lord your God and have no awe of me’ (Jer 2:19).

This is not, however, to say that God plays no role in punishing the damned. On my view, God does two things which end up increasing their suffering compared to what they endure in this life: he withdraws any means of relief, and he draws near to his creatures. Let’s examine these to see if either is coercive.

Our present environment affords us temporary means of relief such as stimulants, entertainment, and sleep. We can use these to avoid awareness of our sinful conditions or suppress negative feelings. Such coping mechanisms never work perfectly (the misery always catches up to us), but temporary evasion is possible. In hell as I see it, God no longer affords us these means of relief with the intention of getting us to repent. Is this coercive?

Surely not. God’s role here is comparable to that of a father who, after due warning, withdraws financial support from his drug-addicted daughter in hopes that, by no longer shielding her from the full consequences of her addiction, she would choose to enter a programme of recovery.Footnote ³⁰ The daughter might accuse her father of coercing her (‘Why can’t you just let me live my own life!?’), but her accusation is baseless. He is letting her live her own life. The daughter might have a point if she is entitled to the father’s support, or if withdrawing that support leads to irrevocable harm, but nothing like that is true with God. God provides us temporary shelter from our sins out of mercy, not obligation. And if my thesis is correct, the harm brought about by this withdrawal is never irrevocable but inevitably leads to everlasting beatitude.

The second thing God does is reveal himself to all creatures. This is the traditional idea that all creatures will come to experience God’s presence and that this experience is simultaneously joy to the saints and torment to the wicked.Footnote ³¹ Zachary Manis (Reference Manis2019; Reference Manis2024) calls this ‘the divine presence model’ of hell. The unveiling of the divine presence enacts what Manis (Reference Manis2024, Ch. 19) calls ‘the judgment of transparency’, in which all deeds done in darkness, and all hidden purposes of the heart, are disclosed by the light of Christ – including to the sinful person herself. This is the vivid experience of our own sinful condition that I incorporated into my description of hell. The question before us is whether God’s revelation of himself, which also reveals us to ourselves, coerces repentance.

Right off the bat, it’s odd to construe felt awareness of our situation as a kind of external threat or pressure that might diminish freedom. Say I’m ridden by guilt and inner turmoil after secretly wronging my wife. These feelings become so unbearable that I confess and apologize. In what way was I coerced?

The difference, the objector might say, is that God brings about this felt awareness through his presence. Alright, let’s amend our example then. Say I’ve secretly wronged my wife, but I’m managing (with some success) to ignore or downplay this fact. When my wife enters the room, however, her presence cuts through my defenses and I’m struck by the enormity of my wrongdoing. Feelings of guilt and shame come pouring out. I confess and apologize. Did my wife coerce me? Obviously not. What if she had discovered the violation and entered the room intending for her presence to elicit a confession? Still, there is no coercion. How is the situation any different with God?

One potential difference is that my wife does nothing impermissible by entering the room. She has every right to be there, and I have no right to keep her away. There are, however, situations in which I might have a right to be left alone. The charge of coercion might gain some traction if it turns out that God’s revelation of himself violates this right to privacy.

We do not, however, have any right to be left alone by God. We are his creatures living in his creation. Is the Creator not allowed to dwell in his creation however he sees fit? The fact that God has so far chosen to remain hidden (Isa 45:15) should not deceive us into thinking we are entitled to it. And even if we did have such a right, everyone who’s privacy God ‘invades’ ultimately comes to an idealized standpoint from which she retroactively waives that right.

There is no reason, then, to think that God coerces us to repent by consigning us to hell.

Objection 3: Autonomy

Michael Murray (Reference Murray1999, 58) notes that theists tend to want more than mere free choices; they want ‘a freedom of choosing that is expressed in actions that influence the course of events in the world’. If I am given the freedom to choose, but those choices don’t make any impact on the world, then my freedom seems rather trivial. Murray labels the thing that is missing ‘autonomy’.

Murray argues that universalism undermines our autonomy with respect to one extremely important outcome: our final destinations. Consider how Talbott (Reference Talbott1990, 39) wraps up the initial presentation of his argument:

Murray’s complaint is that, while our choices make a difference to how we enter heaven, they don’t seem to have any influence over whether we enter heaven. The final destination is the same no matter what we do. Hence, we lack autonomy over one of, or rather the, most important thing in our lives.

The objection doesn’t contest whether the choices which lead us to heaven are free in the libertarian sense. So it isn’t challenging the possibility of a free guarantee. It’s asserting that a free guarantee is possible (if at all) only at the cost of autonomy.

The universalist could bite the bullet and concede that, when it comes to our final destinations, we are free but not autonomous. Why should we want such autonomy when it comes with the risk of eternal damnation? Perhaps a better response is that we do have autonomy in the universalist scenario, and that this becomes clear once autonomy is properly understood.Footnote ³²

Murray is assuming that we don’t have any real influence over our destinations unless there is some other destination at which we might end up. More precisely, autonomy with respect to heaven requires that there is a possible world where we end up in heaven and another possible world where we do not, and that our own free choices are the deciding factor between the two. Essentially, Murray is grounding our autonomy in alternative possibilities.

If we ground freedom in sourcehood rather than alternative possibilities, however, then we will almost certainly want to ground autonomy in a similar way. Influence over our ultimate destinations isn’t measured by whether there is a series of possible choices that result in a different outcome but by whether that outcome is brought about by us.

Accordingly, since we are operating from a sourcehood perspective, we should conclude that we are autonomous over our final destinations even on the universalist picture. We are autonomous not because there are some possible worlds where we end up in heaven and others where we end up in hell, but because we all arrive in heaven through our own free choices.

Objection 4: The limits of human nature

Finally, some objectors contend that the miseries of hell would pass beyond what any human (really, any free creature) is capable of enduring. Thus, they argue, we are determined to repent not by virtue of any particular human character or agential structure but by features inherent to human nature itself.Footnote ³³ The trouble with this is that we can explain why those in hell surrender simply by citing the limits of human nature – something over which those in hell have no control. It follows that their repentance cannot be attributed to any of their past self-forming choices. Worse yet, if the parameters of human nature are set by God, then we seem to have a complete explanation of our repentance in God’s prior activity.Footnote ³⁴

I will set aside this last worry: essences are independent of the divine will.Footnote ³⁵ God doesn’t determine the essential features of humans any more than he determines that squares have exactly four sides. But the prior worry remains: human nature isn’t shaped by our choices either.

An initial response is that we do need their past self-forming choices to explain why people repent in the time and manner they do. When exactly do they give in? Do they give up quickly or hold-out a long time? When they do repent, do they surrender fully or must they be dragged heavenward inch by inch? Human nature does not settle such things – the agents’ freely formed characters do. This seems enough to ensure that their brokenness ‘counterfactually depends on past directly free actions’ (Hartman Reference Hartman2020, 1426), which is generally all that source incompatibilists require here.

All this secures, the objector may retort, is that the manner in which the repentance occurs remains free. That the repentance occurs (at some point or other) is still required by human nature, and so the threat to freedom remains. I’m not sure this pushback succeeds. No particular choice to repent is ever determined by human nature but only ever by the specific characters that each person had a hand in freely shaping. I fail to see, then, how the freedom of those particular choices is threatened by the fact that, at some point or other, repentance is inevitable. Nevertheless, let’s follow the argument through on the assumption that there is a threat to freedom here.

There are two ways to respond depending on what kind of source incompatibilism we wish to endorse.

The first, more demanding kind says that we are the ultimate source of our action only if we are ‘responsible for anything that is a sufficient reason, cause, or motive for the action’s occurring’ (Kane Reference Kane2005, 121–122). Kane falls into this category. Kane (Reference Kane2002, 412) writes:

In short, the determining aspects of our characters must be freely formed.

The second, less demanding kind of source incompatibilism requires only that the determining aspects of our characters not be externally formed – i.e., that there be no ‘externally sufficient causal chain which results in that choice’ (Timpe Reference Timpe2012, 138–139; Reference Timpe2014, 104). Timpe’s treatment of divine freedom seems to place him in this category since God remains free despite the fact that God’s character does not result from any prior free choices. Sourcehood requires not that ‘her character is formed in the right way’ so much as that ‘it’s not formed in the wrong way’ (Timpe Reference Timpe, Timpe and Speak2016, 277).Footnote ³⁶

The difference between these two kinds of source incompatibilism is subtle and almost never relevant. In virtually all cases, if the determining aspects of our character aren’t formed through our own choices, then they come from something outside us; so there is no externally sufficient causal chain if and only if our character is shaped through self-forming choices. What pulls the two accounts apart are cases where the determining aspects of our characters are not formed at all – where those aspects result neither from any self-forming choices nor from any external source but are necessary features of all humans as such. Are we the ultimate source of the choices arising from these essential human features?

My own answer is yes. I find the divine case convincing here. To be the ultimate source of his choices, God doesn’t need to form the character from which those choices flow. It is sufficient that they stem from his essential features – features that define who and what he is – because in doing so they originate from him.Footnote ³⁷ The same, I think, is true of us. Our essences, or natures, should be understood as internal to us, not external influences. Thus, the choices that flow from those natures do not originate from without but from within.

Opposition to this position often misconstrues the sort of determination at issue. For instance, an objector might ask us to consider a human who was brought into existence determined to murder someone, or never to eat meat, or always to act morally. Shouldn’t we agree that such choices aren’t free? We should, but this poses no trouble to my position. Humans cannot be determined to make such choices by nature since clearly not all humans are determined in these same ways.Footnote ³⁸ The person in such scenarios is being determined by her initial agential endowment – an endowment that is simply given to her by external causes rather than belonging to her by nature. Very few choices, it turns out, are necessitated by human nature itself, and when we maintain focus on those rare few choices, I find little difficulty in saying that they remain free even though we did not get to pick the features of ourselves from which they flow.

If we adopt this less demanding form of source incompatibilism, then the overall objection is rebuffed completely. The fact that human nature entails eventual repentance is no threat to its freedom, for it still finds its origin in the essential features of these human agents and, hence, in the agents themselves.

If we adopt the more demanding form of source incompatibilism, then the objection might stick, though then again, maybe not. As stated earlier, even if repentance is bound to happen eventually, no particular act of repentance is ever entailed by human nature. How exactly, then, is that particular act of repentance not free?

Regardless, substantial progress has still been made. Worst case, God can guarantee that all repent without God himself violating our freedom. As a foil, consider the position of Marilyn McCord Adams. Adams (Reference Adams1999, 157) believes that God guarantees human repentance, but only because God stands willing to override human freedom. Many deem such a violation morally problematic. In contrast, by consigning people to a remedial hell, God can ensure that everyone repents without overriding anyone’s freedom. If anyone’s freedom is (partially) extinguished, it is not God’s doing but her own. Hence, God has the means to guarantee repentance while avoiding any morally questionable violations.