Evil Prevention Requirements and the God of Theism

Abstract

The central argument of James Sterba’s “Is a Good God Logically Possible?” relies crucially on the notion that a good God would have to abide by various evil prevention requirements. Because it appears that God has not done so, Sterba concludes that God does not exist. I challenge the notion that theists must accept the notion that God is bound by the particular set of evil prevention requirements Sterba’s argument presupposes. However, I argue that investigating ways God may in fact be required to prevent evils may serve as a helpful heuristic for theists as they seek further to understand God’s nature and purposes.

James Sterba’s Is a Good God Logically Possible? aims to demonstrate that the answer to its titular question is “no”, in the vein of J.L. Mackie’s 1955 “Evil and Omnipotence”, but utilizing what Sterba calls “yet untapped resources of ethics” (Sterba 2019, p. 5). Central to Sterba’s tapping of ethical resources is refining what Mackie had called the “quasi-logical” rule that “good is opposed to evil, in such a way that a good thing always eliminates evil as far as it can” (Mackie 1955, p. 200). In Sterba’s hands this rule branches into a set of “evil prevention requirements” that moral agents are obliged to obey, including three for moral evils, and no fewer than nine for natural evils. I will use the abbreviation EPRs throughout this article to refer to Sterba’s particular set of evil prevention requirements. The basic shape of Sterba’s main argument is that if God exists, he must operate according to the EPRs, but that if God did so, then we’d observe a great deal less evil in the actual world than we currently do. Hence, God does not exist.

Now theists might respond to Sterba’s argument by agreeing that God must obey the EPRs, but denying that the evils we observe in the world give us reason to doubt that he does so. My chief aim here, however, is to challenge Sterba’s argument in a different way, namely by questioning the claim that if God exists, he must operate according to the EPRs. I will argue that there are no good reasons why theists need to accept this claim. On the other hand, despite challenging Sterba’s central argument of the book in this way, I agree with him that there is a perfectly good sense in which theists can and should think of God as subject to certain evil prevention requirements. I will argue, furthermore, that it may well be profitable for them to explore the sense in which this is so. I will begin by examining in Section 1 of this paper the ways Sterba thinks God must prevent evil, if he exists, along with establishing a few things that I take it theists are committed to believing about God. In Section 2 I will argue that theists need not accept Sterba’s EPRs as constraints on God’s behavior. In Section 3 I will argue that while theists in my own Christian tradition will likely wish to reject Sterba’s EPRs as constraints on God’s behavior, they can still profitably discuss ways God is required to prevent evil as a heuristic for better understanding his nature and purposes.

1. Sterba’s EPRs and the God of Theism

Tacitly recognized in Mackie’s article is the recognition that good things eliminate evil as far as they can unless they have some good reason not to do so. Sterba’s EPRs can be thought of as ways of refining that recognition, utilizing the “resources of ethics” mentioned above, which include discussions of just political states and of the “Pauline Principle”: never do evil that good may come of it (see Rom. 3:8). Rather than list all twelve EPRs along with Sterba’s reasons for thinking morality requires them, it will suffice to describe generally how Sterba thinks a good God would have orchestrated things here on earth.

Sterba relies on various analogies to explore the ways God should have acted including a loving parent, a superhero, and a just political state. The latter is most pervasive. Just political states aim at the flourishing of their citizens. That means allowing their citizens freedom, but also curtailing freedoms at certain times when exercising them would impede the flourishing of others. Sterba thinks it clear that God should curtail the immoral activities of wrongdoers when they would have “significant and especially horrendous evil” effects on victims. He brings up the evil general from Ivan Karamazov’s story in the “Rebellion” chapter of Dostoevsky’s novel. A just political state would intervene if it could to make sure the general did not horrifically murder the serf boy as Ivan describes. A good God too would have intervened to make sure the consequences of the general’s horrific intent are not felt. Unlike the Grand Inquisitor Ivan introduces in the book’s next chapter, Sterba does not think it would be best to eliminate human freedom altogether. Sterba may (although I am not sure about this) agree that God and just political states can allow wrongdoers to ruin their own lives, just not those of others. Furthermore, he thinks God could allow wrongdoers to go ahead and “imagine, intend and even take initial steps toward” immoral deeds (Sterba 2019, p. 21). He might also have a good reason for allowing victims of wrongdoing to experience some harmful effects, so long as these are insignificant and non-horrendous, in cases when other citizens might have intervened but did not (Ibid., pp. 60–64). The idea is this. When you or I are in a position to intervene and prevent an instance of wrongdoing, and do so, God ensures that our efforts are fully successful. When we do not, he ensures that there are some noticeable and harmful, yet insignificant and non-horrendous, effects on the victim. He does this to make sure we still have ample motivation and opportunities for soul-making. Something similar is true in the case of natural evils. Just political states intervene to prevent in some degree the harmful effects of certain natural evils on humans and nonhuman animals alike, but their ability to do so is quite limited. A good God would also intervene to prevent needless suffering in sentient beings. Sterba does not think even God could prevent all significant consequences of natural evils (Ibid., p. 164). However, he thinks God could certainly limit them in similar fashion as just described: when we can intervene to limit the suffering of something like William Rowe’s fawn trapped in the path of a forest fire (say), and do so, God ensures that our intervention is successful, whereas when we do not God allows some noticeably harmful yet mitigated effects.1 He does this, again, so that our motivation to intervene is not undermined, and we retain ample opportunities for soul-making. Overall, Sterba thinks that by means of such limited interventions God could have given theodicists everything they typically say they want—morally significant freedom, opportunities for soul-making, etc.—without such horrors as the death of Ivan’s serf boy or Rowe’s fawn roasting in the forest fire. Again, the basic shape of his main argument in the book is that a just political state would act in these ways if it could, hence a just God would do so as well, and being omnipotent, could do so. Since God clearly has not done so, he does not exist.

Theists of any persuasion clearly will not agree. One way they might respond to Sterba’s main argument is by questioning the possibility of God intervening in the ways Sterba describes while preserving our morally significant freedom and opportunities for soul-making. Perhaps, even if God has to obey Sterba’s EPRs, certain avenues of theodicy or defense remain open. Alternatively, skeptical theists might maintain that God does indeed obey the EPRs, but that we’re unable to access the reasons he has for permitting all the evils we observe around us. Sterba devotes much of his book to combatting these responses—closing off avenues of theodicy/defense and explaining why skeptical theism cannot rescue theists either. For my part, as I mentioned previously, I am interested in pursuing a different type of response, namely rejecting the claim that God must act like a just political state in the ways Sterba believes—i.e., by obeying the EPRs—to begin with. I will argue this in the next section.

First, however, to scotch some possible misunderstandings, let me say briefly a few things I think theists are committed to believing about God. They must believe that God is the omnipotent and omniscient creator, and hence that he’s perfectly good—the best—in a metaphysical sense of goodness: he’s the most powerful being, the most perfectly actualized, etc. They must also believe, however, that God is perfectly good in the sense of perfectly possessing what we would think of in humans as good moral characteristics or virtues: he’s perfectly wise, just, loving, merciful, and so forth. Having said all this, I also take it theists are committed to God’s transcendence in some way or other, and hence that in some way or other God’s goodness, wisdom, justice, love and mercy differ from our own. Just how exactly this is so, and to what extent, are of course disputed matters.

I make the (hopefully) uncontroversial points above partly to sort through an unhelpful round of dialectic involving God’s moral agency. Sterba devotes a chapter to discussing Brian Davies’s Thomistically- inspired view that God is not in fact, a moral agent.2 I will say more about this view shortly, but for now it will be helpful simply to acknowledge that in my estimation theists—Davies included—are committed to God’s moral agency at least in the sense that he is an agent—i.e., acts for reasons—and is perfectly wise, just, loving, merciful, etc.3 With that sense of God’s moral agency, furthermore, I cannot imagine Davies disagreeing.

2. Against the Claim That God Must Obey Sterba’s EPRs

If God is a moral agent in the sense of being perfectly just, however, then here is how it looks to me like Sterba intends to convince theists that God must obey his EPRs:

(1) God is just.

(2) All just agents must obey the EPRs.

(3) God must obey the EPRs.

I am not sure Sterba ever states an argument like this explicitly, perhaps because it is too obvious to be worth making explicit. However, reasoning of this sort seems to underlie much of what he does say. The trouble with this argument stems from theism’s commitment to God’s transcendence. The reasoning in favor of (2) draws on considerations of human cases involving superheroes, just political states, the Pauline Principle and so forth, as mentioned above. However, if theists claim that God’s justice differs somehow or other from our own, then they can allege that the argument from (1) and (2) to (3) involves a fallacy of equivocation. Yes, God is just in his way, and yes, all just human agents must obey the EPRs, but since God’s justice differs from ours, it does not follow that he must obey them. I take it this response to the above argument is just as obvious as the argument itself, and I can think of two obvious routes by which Sterba might attempt to convince theists that God must obey the EPRs notwithstanding his transcendence. I do not think either of these routes succeeds, however, and I will explain why in this section, concluding my discussion of the second with a direct argument aimed at showing that no other route is likely to succeed either.

Here is the first route by which I think Sterba might try to convince theists that God, despite his transcendence, must obey the EPRs. God is rational. However, the rules of morality, including the EPRs, are binding on all rational agents as such. Hence, God must obey the EPRs. Sterba appears to have an argument like this in mind at one point in his chapter on Davies’s God-isn’t-a-moral-agent view, where he remarks as follows:

[T]he law of nature that God… implanted in our hearts is understood to apply to all rational beings including God himself. So it would… be contradictory for God to implant a law of nature in our hearts that applies to himself and then act contrary to that very law he promulgated.

(Sterba 2019, p. 116)

To set this in context, Davies’s view (following Aquinas) is that moral rules are simply ways of spelling out what right practical reasoning for rational animals such as ourselves involves, and some of them—the “natural law”—are “written on our hearts” (see Rom. 2:15) in such a way that we all implicitly know and acknowledge them whenever we do any practical reasoning at all.4 Sterba’s point is that if moral rules such as the natural law apply to practical reasoners as such, then they must apply equally to God, who is a rational agent too, after all. Elsewhere Sterba argues extensively for the view that rationality requires morality (Sterba 2013, especially chp. 3). He does not reiterate his argument here, but does seem to presuppose it.

Whether or not I am right that this is what Sterba has in mind, here is why I think the route I’ve just described will not succeed in persuading many theists that God has to obey the EPRs. For Thomas, and for Davies, it is important that the natural law spells out the rules for right practical reasoning among rational animals, such as ourselves—it tells us how we must behave in order to flourish as the kinds of things we are. God is not an animal, and is perfect no matter what he does. He is indeed rational in the sense of acting for reasons, but not in the sense of working discursively through a process of practical reasoning before settling on which option is best, or most conducive to his flourishing.5 This is why Davies wants to deny that God is a moral agent—because God is not bound by the rules directing our behavior toward flourishing as we are. It is the way Davies and Aquinas would claim God’s goodness transcends our own. For us, justice involves following certain rules of behavior, whereas for God it does not. What it does involve I will return to below. However, it seems fairly clear that the sort of argument that Sterba has given elsewhere for morality as a sort of rational compromise between egoism and altruism does involve settling on it after a discursive process, and hence will not apply to God.

Now some theists might indeed be persuaded that God’s rationality entails that he must obey the EPRs. Perhaps a sort of Platonist theist might think of moral rules as abstracta somehow binding both God and all other rational agents as such. Or perhaps a Kantian theist might think of moral rules as rules governing the will as a faculty of practical reasoning in the same way the law of non-contradiction governs theoretical reasoning. In that case, if the Kantian could be persuaded that the EPRs were among the moral rules, she might have to agree that God “has to” obey them, at least in the sense that they represent infallible descriptions of how he acts. If Sterba could show theists that they must embrace Platonism or Kantianism in one of these ways, then perhaps he could show that God must obey the EPRs. That would be an ambitious meta-ethical hurdle to clear, however, and might conflict with Sterba’s stated aim in the book of remaining neutral between a wide spectrum of different ethical outlooks.6 At any rate, for the route I’ve been discussing so far to succeed, Sterba would have to show that a natural law-style understanding of God and morality is either false or incoherent, and insofar as I doubt that could be done, I doubt this route is very promising.7

A second route by which Sterba might try to persuade theists that God must obey the EPRs is by targeting their claim about God’s transcendence. Theists think God’s justice differs from ours in some way or other. But just how different is it? Sterba agrees, presumably, that it is somewhat different. It is much larger in scale, for example, since God rules not just over a state but the whole world. And of course when it comes to preventing evil, an omnipotent and omniscient God would not face many of the limitations that we do. In these ways Sterba can agree that God’s justice transcends ours. However, suppose a theist wants to say that the transcendence extends further—to the point where God does not have to obey the EPRs. In that case, Sterba might ask, do theists really know what they’re talking about when they ascribe justice to God? Additionally, do they have any reasons for their ascription? Antony Flew argues in a widely-anthologized short article that theistic assertions like “God is just” die a death of a thousand qualifications, to the point that they no longer count as assertions at all (Flew 1971). Sterba might argue, similarly, that unless theists acknowledge that God is just in an EPR-obeying way, their belief in his justice is meaningless, groundless, or both.8 Hence, unless theists are willing to accept that their beliefs are meaningless and/or groundless, they must agree that God obeys the EPRs.

One way theists might respond to this argument is by biting the bullet, and agreeing that in a sense they don’t know what they’re talking about when they discuss divine characteristics like justice. Herbert McCabe, whose thought often aligns closely with Davies’s, suggests Aquinas might have been pleased with this sort of response:

Thomas Aquinas thought that theologians don’t know what they are talking about. They try to talk about God, but Aquinas was most insistent that they do not, and cannot know what God is. He was, I suppose, the most agnostic theologian in the Western Christian tradition… in the sense of being quite clear and certain that God is a mystery beyond any understanding we can have now.

(McCabe 2007, p. 96)

I am not sure this is right about Aquinas myself, although McCabe certainly is not alone in describing him this way (Cf. Hector 2007; Preller 2005). Nevertheless, I think there are many theists who would respond to the Flew-style argument I just sketched by denying that their acknowledged ignorance about God’s nature renders their belief meaningless or groundless in any problematic way. They might say that their beliefs are grounded in their conviction that God is telling them about himself in the metaphors, narratives, prophecies, etc. recorded in a book of scripture. They might claim that scripture allows them to understand God’s justice to some limited extent, even if they lack any univocal concept of justice under whose extension both God and creatures fall. They might claim that their limited, scripturally-based understanding of God’s nature is at any rate enough to render their belief meaningful in practical terms for them, in the sense of imbuing their lives with purpose, direction, etc. I can imagine theists from fideistic, apophatic or mystical traditions being attracted to some combination of these claims.

I can also imagine Sterba joining theists of a different persuasion who insist that we must possess univocal concepts that apply to God and creatures alike. Thomas Williams, for example, offers a Duns Scotus-inspired argument aimed at showing that the relationship between our concepts of God’s wisdom and human wisdom (say) not only cannot be one of equivocity, but also cannot be one of analogy either—i.e., different-but-related (Williams 2005). I find Williams’s reasoning persuasive myself. I think it’s worth noting than even Aquinas, known for claiming that “it is impossible that anything be predicated univocally about God and creatures”, nevertheless seems to agree that there are certain senses of terms like “justice”, “love”, “mercy” and so forth that apply both to God and to creatures.9 Regarding justice, for example, Aquinas denies that the notion of commutative justice—justice in transactions or exchanges—applies in any way to God. However, he thinks the notion of distributive justice—which involves a “governor or ruler giving to each thing according to its dignity”—does (Ibid., 1a.21.1). That is, he thinks that both God and a just human ruler are just in the sense of giving to each thing under their rule according to its dignity. Pending correction by other Thomists better informed about Aquinas’s views on the semantics of religious language, we might take this to mean the he thinks there is indeed a univocal concept of justice that applies both to God and creatures.10 Anyhow, whatever Aquinas’s view may have been, Williams’s conclusion is that unless we possess univocal concepts of characteristics like justice, wisdom, etc. under whose extension both God and creatures fall, we cannot meaningfully ascribe these characteristics to God.

Suppose Williams is right; the important question then for present purposes is whether this might give Sterba a way of convincing theists that God must obey the EPRs. I think the answer is certainly not. Theists can, it seems to me, meaningfully and with good grounds assert that God is good, wise, just, loving, merciful and so forth in univocal senses of these terms even if they deny any ability on their part to understand God’s purpose in creation. However, if we do not understand God’s purpose in creation, then for all we know it includes some good reason for not obeying the EPRs. Hence, even if theists agree that to speak meaningfully about God requires that we possess univocal concepts of characteristics like justice (that apply to both God and creatures), this will not suffice to convince them that God must obey the EPRs.

Let me unpack this bit of reasoning. Even if Davies is right that God is not bound by moral rules in the way we are—and hence not by Sterba’s EPRs either—I think we can still rule out at least some descriptions of ways God might have operated in creation as inconsistent with his nature. While some theists disagree, my view is that God’s nature did not require him to create anything at all. Hence, although it might be said that God is just by nature in the sense that if he creates anything he necessarily rules creation justly, God’s just rulership need not ever have been exercised. Having decided to create, however, God’s nature requires him to rule creation justly, and also wisely, lovingly, mercifully, and so forth. Given that, consider a world with lots of evil, none of which serves any purpose at all. I think theists should deny that God could create such a world. God’s wise rulership requires that he have a plan or purpose for what unfolds in creation. Nor could this purpose be just anything whatsoever. I think theists should agree, for example, that God’s purpose in allowing evil could not be his own amusement. It would be cruel and contrary to God’s perfect love and justice to allow sin and suffering, say, just to amuse himself.

Importantly, however, beyond a few negative restrictions like “not allowing sin and suffering just for amusement”, I do not think theists need to claim to know what God’s purpose in creation is. Suppose God creates a world with lots of evil, none of which serves any purpose that we can discern. (I am not sure such a world is conceivable, but suppose). Such a world, however much evil in includes, would also have to be good in a variety of respects, insofar as it includes creatures exercising their powers in various respects and thereby flourishing to some extent, however limited it may be. Theists might reasonably call the creator of such a world good insofar as he has created something good.11 They might call him loving insofar as he wills goods for his creatures.12 They might consider him merciful insofar as he mitigates the extent to which his creatures suffer evils.13 They might think him wise insofar as order in creation reflects rulership according to some plan or purpose, even if they do not claim to know what this involves. They might also call him just in the sense of distributive justice articulated above, namely insofar as he gives to the universe itself and to the creatures in it whatever they need to exist for some period of time and to flourish to some extent as members of their kind, however limited the extent may be. In these ways theists might meaningfully and with good grounds ascribe goodness, wisdom, justice, love and mercy to God even if the world contains lots of evil for which they can discern no purpose. Granted, the univocal concepts of these characteristics that I’ve just sketched are quite thin. They are certainly too thin to infer from any of them that God must obey the EPRs. However, they are thick enough, it seems to me, to give theists some idea what it means to ascribe them to God, along with some grounds for doing so.

I conclude from this reasoning that the second route I suggested by which Sterba might convince theists that God has to obey the EPRs will not, in fact, succeed. Furthermore, theists might advance a direct case against the possibility of his doing so by some other means:

(4) Any convincing argument why God must obey the EPRs will involve some robust assumptions about God’s purpose in creation.

(5) However, theists will always be in a position to deny the particular set of assumptions about God’s purpose in creation that such an argument involves.

(6) So no argument is going to convince theists that God must obey the EPRs.

Again, I think theists must make a limited range of assumptions about God’s purpose in creation, such as the fact that he does not allow sin and suffering for the sake of his own amusement. However, theists need not accept that God’s purpose in creation is the same as that of a just political state. Nor need they accept that God has to obey the EPRs.

I take it the argument I’ve just proposed might count as a sort of skeptical theist strategy insofar as it hinges on the theist’s ability to profess a degree of ignorance about God’s purpose in creation, so as to reject any particular robust set of assumptions about what this purpose might be. It differs, however, from the versions of skeptical theism Sterba discusses in the fifth chapter of his book insofar as these accept that God is bound by moral rules like the EPRs much as we are, but profess global skepticism about the range of possible goods that might justify God in permitting the evils we observe. On my proposal, in contrast, insofar as it is true to say that God is bound by moral rules at all—such as “do not murder”, “do not allow suffering for the sake of amusement”, etc.—the binding will stem from God’s nature together with his purposes, not some overarching set of rules binding both him and us alike. As a result, while Sterba may possibly be justified in arguing that the sort of skeptical theism he is considering “leads to moral skepticism and thus undermines morality”, I do not think the same could be said about argument from (4) to (6) (Sterba 2019, p. 76). At most it will involve a degree of skepticism about our knowledge about God, but not about human morality. Additionally, some skepticism about our knowledge of a transcendent God is something theists seem committed to anyway.

3. Should Theists Discuss Evil Prevention Requirements in Relation to God?

So far, my argument has been that since theists can claim that evil prevention requirements for God, insofar as there are any, stem solely from his nature and purposes, and that these remain at least partly hidden to us, they need not accept that God is bound by any set of requirements so specific as Sterba’s EPRs. A further question is whether it is open to theists to endorse Sterba’s EPRs as binding on God. Sterba himself, of course, would deny this, given that in his estimation the world as run by an EPR-obeying God would look very different from the actual world. However, as I mentioned above, theists who think God obeys the EPRs might pursue avenues of theodicy or defense within their constraints, or else might take issue with Sterba’s reasoning against the versions of skeptical theism that think of God’s morality as parallel to our own. Another further question is whether, even if it is theoretically open to theists to endorse Sterba’s EPRs, any of them ought to. Speaking solely for those within my own Christian tradition, I am doubtful that any of us should, for reasons I will explain below. If that is correct for at least some theists, however, then a third further question is whether it is in any way profitable for us to inquire about the ways God is or is not required to prevent evils. I think it can be, if approached correctly, and I will explain in this concluding section what I mean.

In an article offering “Advice to Christian Philosophers” Marilyn McCord Adams distinguishes between “aporetic” and “atheological” versions of the problem of evil, and counsels Christians to focus on the former instead of the latter (Adams 1988). When it comes to discussing evil prevention requirements for God, I agree with her. I argued at the end of the previous section that since theists can disavow knowledge in any but the most general terms about God’s purposes in creation, they will likely be in a position to reject any set of requirements so specific as Sterba’s EPRs as binding on God. However, if Sterba is correct that a successful atheological version of the problem of evil is best advanced by way of his EPRs or something like them, then it seems correspondingly unlikely that any such atheological argument will succeed. That in itself might give Christian philosophers a good reason for taking Adams’s advice and shifting our focus to aporetic versions of the problem of evil, by which she means efforts to resolve prima facie conflicts between commitments of the faith and the presence of various kinds of evils in the world. However, Adams herself suggests some other reasons for being leery of the kinds of theodicies and defenses that preoccupy many analytic philosophers of religion. These include reasons sometimes forwarded by proponents of the philosophical stance known as anti-theodicy, for example, that such responses to atheological arguments inevitably end up discounting the magnitude of the world’s ills in order to render them intelligible within a well-governed created order.14 Or that they distort God’s nature in their efforts to engage atheologians levelly on a neutral value-theoretic playing field. Christians are better off, Adams says, focusing on God’s “agent-centered goodness”, or as she puts it elsewhere, on “how God can be good enough to created persons despite their participation in horrors—by defeating them within the context of the individual’s life and by giving that individual a life that is a great good to him/her on the whole” (Ibid., p. 135 and Adams 1989, p. 306). I am not myself interested in casting aspersions on the endeavor of theodicy/defense in general terms. However, I think Adams is right that theists engaged in such practices as discussing evil prevention requirements for God ought to be wary of minimizing the reality of evil or compromising their commitments with respect to God’s nature, value theory, or whatever else. They ought to heed what Adams says about “genuine continuities between theoretical and practical problems of evil”, i.e., to keep firmly in mind the interplay between their theorizing and their faith and religious practice (Adams 1988, p. 140).

How might speculation about evil prevention requirements prove helpful within the ambit of the sort of inquiry Adams recommends? Consider the contrasting approaches of Adams and two other philosophers—Aquinas and Eleonore Stump—to one particular aporia all three face as Christians: the problem of hell. Many Christians think our scriptures commit us to believing that certain persons will not only sin grievously during their lives, but also as a result of these sins will end up eternally sundered from fellowship with God in hell. This seemingly evil state of affairs seems prima facie at odds not only with other scriptural passages—for instance, that God “desires all people to be saved and to come to a knowledge of the truth” (1 Tim. 2:4)—but also with other practices in our tradition, such as reciting weekly that God’s “character is always to have mercy”.15 How to reconcile God’s constant mercy and universal salvific will with the evil of eternal damnation in hell is puzzling indeed. What options are available to resolve the puzzle will depend in part on our understanding of God’s nature and purposes, as well as on other factors such as how we think God’s causality is related to human freedom. However, I think investigating what ways God may or may not be required to prevent evils can provide a helpful way of exploring and organizing our understanding of the ways these various factors interrelate.

To see this, a useful starting point is Adams’s remark about focusing on God’s “agent-centered goodness”, indicating that she accepts some agent-centered restrictions on God’s activity in creation. Eleonore Stump, who recognizes similar restrictions on God’s activity, puts them as follows “if a good God allows evil, it can only be because the evil in question produces a benefit for the sufferer and one that God could not provide without the suffering” (Stump 2003, pp. 461–62). Adams’s focus is a bit different, though similar: a good God must “defeat” any evils his creatures experience, as opposed to merely “balancing them off”.16 Both contrast these agent-centered restrictions on God’s goodness with “global goods” such as promoting free-will, soul-making, or whatever else considered in general.

Both Adams and Stump then, I take it, would be equally opposed to the way in which Aquinas, as I read him, addresses the puzzle concerning hell. Thomas agrees with Augustine that God “wouldn’t allow any evil to exist in His works, unless his omnipotence and goodness were such as to bring good even out of evil”.17 However, this evil prevention requirement stems just from God’s nature together with his general purpose in creation, namely “that his goodness might be communicated to creatures and might be represented by them”.18 For God’s goodness adequately to be represented, Aquinas thinks, God had to create a variety of different kinds of creatures, hierarchically arranged, including some that “can fail in goodness” and hence “sometimes do fail”.19 In short, Aquinas thinks, the “perfection of the universe” required that it contain evil. Furthermore, he argues, building on this theme, one respect in which evil is required for God’s goodness adequately to be represented is that his justice be shown by eternally punishing sinners in hell. That, Aquinas says, “is the reason God elects some and reprobates others”.20 For Thomas, then, God’s constant mercy means that he acts mercifully to all his creatures both by creating them in the first place and from sparing them from some suffering they could have undergone. This goes even for the damned.21 It certainly does not mean, though, that he spares humans from all the sin and suffering he could have, metaphysically speaking.22 Likewise God’s universal desire that all humans be saved is true only of his “antecedent will”, prescinding from facts about our sinfulness that his “consequent will” takes into account by damning some.23 As I read him, Aquinas does not think God’s goodness is agent-centered in the ways discussed above both insofar as he aims primarily at the perfection of the whole universe, and insofar as he allows instances of sin and suffering that are not good for those commit or undergo them.24

Stump’s reading of Aquinas is very different than the foregoing, as is her approach to the problem of hell. She reads Aquinas as a strong sort of libertarian who thinks that having created free creatures like us, God cannot prevent us from sinning when we make up our minds to do so. Our sinfulness infects us like a “spiritual cancer”, which will kill us, spiritually speaking, by resulting in our eternal separation from God in hell, unless God administers some pretty harsh spiritual chemotherapy in the form of suffering.25 God doses all of us as best as he is able, but again cannot prevent some of us from responding poorly to the treatment, and ending up eternally sundered from him in hell anyway. That is indeed very bad, but God still shows his constant mercy to those in hell by making things as good as they can possibly be.26 As Stump sees his, God wholeheartedly “desires all people to be saved” and does his very best to save everyone, but is thwarted in some cases by our wicked wills together with the constraints of libertarian freedom. Nonetheless, God abides by agent-centered restrictions in that he allows evil only insofar as doing so is truly good for those who commit or suffer from it—even for those in hell.

Perhaps, Adams would reply, but Stump’s God does not go far enough to abide by these restrictions insofar as the evils of hell remain forever undefeated.27 Adams makes no bones about her universalism. To the “pragmatic” question “whether or not a God who condemned some of His creatures to hell could be a logically appropriate object of worship”, her answer is no (Adams 1993, p. 302). This is not because God is bound by moral rules in the ways we are. Rather, as she sees it, C.S. Lewis is right to suppose that leaving humans to live out the effects of their own sinfulness in hell will necessarily result in a state of affairs quite like the traditional, Dante-esque hellishness.28 Additionally, in her view, “[a]ny person who suffers eternal punishment in the traditional hell will… be one within whose life good is engulfed and/or defeated by evils” (Adams 1993, p. 304). God’s goodness, as she understands it, requires him to prevent such a state of affairs, even if he would not necessarily be treating anyone unjustly in so allowing it.

Now Sterba does not think the measures Adams’s God takes to defeat evil are sufficient for him to count as good on the whole, according to the standards of his EPRs. Part of the trouble, as he sees it, is that the only kinds of goods with which God could possibly compensate our sufferings in this life are what he calls “consumer goods”, that is, “experiences and activities that are intensely pleasurable, completely fulfilling and all encompassing” (Sterba 2019, p. 36). Sterba reckons that the beatific vision, “which is said to involve ultimate communion or friendship with God” would be “the primary consumer good that would be experienced and enjoyed by those in the traditional heavenly afterlife” (Ibid.). However, it would be inappropriate, Sterba thinks, to reward persons who suffer in this life, or are deprived of goods such as opportunities for soul-making, with consumer goods in the afterlife. Such consumer goods would not only be unrelated to the goods lost through suffering, but also would be unearned—granted as a sort of consolation prize, as opposed to won through efforts at soul-making. Furthermore, they could not be chosen or accepted by the sufferers, and as Sterba sees it, morality “only justifies impositions that are reasonably acceptable to all those affected” (Ibid., p. 73). There is a sort of “informed consent requirement” for inflicting suffering. For reasons such as these Sterba dismisses Adams’s “attempt at theodicy”, even with its promise of universal salvation. Nor are the other responses to the problem of hell I’ve discussed so far any more likely to please him. Embedded in Stump’s theodicy is free-will defense of the sort Sterba rejects in his first chapter. Additionally, I take it Sterba would charge Aquinas’s God with blatantly violating the Pauline Principle (and probably others besides!) by allowing sin and suffering so as to better manifest his glory in the universe. Sterba, in fact, has proposed his own theistic solution to the problem of hell elsewhere, the gist of which is that the afterlife should look much the same as this one (Sterba 2020). Everyone enters the same afterlife and has the same options for choosing fellowship with God or rejecting it, except that God ensures the afterlife society meets the standards of basic justice that Sterba thinks God would be ensuring now if he actually existed.

These disagreements between Sterba and the three others I’ve discussed highlight, it seems to me, at least three important differences between the value theory presupposed by his EPRs and that of Christians. First, Christians would not want to see the beatific vision as simply the highest among many possible “consumer goods”. Communion with God is what we were made for, and it simply is the best thing for us—what we were made for—whether we like it or not. It is also of such infinite value as to swamp in comparison any evils that might be necessary for its sake. Second, without entering into thorny debates about grace and meritorious works, Christians of any sort deny that communion with God is something that can be earned. Hence, Sterba’s worries about needing to earn communion with God in successive “innings” of afterlives would strike them as strange.29 Third, even the most ardent theodicists among Christians typically deny that we’re able to understand God’s reasons for allowing particular evils in this life. Sterba writes that victims of suffering:

[W]ould have to be viewed to be incompetent throughout their entire earthly lives with respect to giving informed consent to significant and especially horrendous evil consequences that God would be permitting them to wrongfully suffer. Even so, if God’s permission of the infliction of such evil consequences is to be justified then at some point these victims of wrongdoing need to be able to give their informed consent to what was done to them.

(Sterba 2019, p. 74)

Christians would agree with the first statement here, about our incompetence to understand in anything more than general terms why God permits evils. They need not agree with the second statement, however, and might well prefer to maintain that we’ll never understand. Due to disagreements like these, I doubt that Christians ought to endorse Sterba’s EPRs. In general, recall, Sterba thinks God is required to prevent evils in basically the same ways a superpowered just political state would, equipping all of his creatures to pursue whatever aims they might have unimpeded by others and experiencing only whatever minimum of suffering is necessary for character formation in rational creatures. I argued in the previous section that theists need not agree with Sterba here, and I am doubtful that many Christians, at least, will wish to.

At the same time, however, it also seems to me worthwhile for theists to ask themselves in what ways God is or is not required to prevent evils, if they approach these questions in the aporetic way Adams recommends, and mindful of her warnings about not compromising their understanding of God’s nature and purposes or the commitments of a Christian value theory. Asking themselves why God is not, in fact, obliged to operate like a superpowered just political state (if that is in fact what they decide) will force them to clarify just how exactly their understanding of God’s nature and purposes differ from Sterba’s. When it comes to the particular puzzle I’ve focused on in this section regarding hell, asking questions like whether there are agent-centered restrictions on God’s goodness, or whether God’s goodness requires him to defeat the evils any of his creature experience, can likewise help us understand and organize our thoughts concerning God’s nature, purposes and related issues such as the relationship between divine causality and human freedom. In asking such questions theists must be careful not to put the cart before the horse. They must bear in mind that any obligations God has to prevent evil stem solely from his nature and purposes in creating. However, so long as they bear this in mind it seems to be that asking what obligations, exactly, stem from God’s nature and purposes can play a helpful heuristic role in understanding them.

For my part, I am inclined to agree with Aquinas (as I read him) that there’s no metaphysical reason why God could not have predestined everyone and reprobated no one. Hence, I disagree with Stump’s view that hell is a sort of backup plan on God’s part for those whom he could not prevent from rejecting fellowship with him.30 I think it must have been part of the plan from the beginning—that the fall was “truly necessary” as is said in the Exsultet, and that “sin is behovely”, as Julian of Norwich puts it.31 I am also inclined to believe, unlike Adams, that hell is populated. However, I have a very difficult time accepting Aquinas’s view that God’s purpose in allowing hell to be populated is to better represent his justice to the universe. I am inclined to agree that God’s goodness towards creatures counts as good only if it is good for them, in some way. Additionally, if God isn’t equally good and loving to everyone, as his allowing some to populate hell would appear to suggest, I am inclined to think God must base his decision about whom to save and whom to reprobate on something having to do with us, even if this cannot be a matter of earning salvation through meritorious works.32 I am not sure exactly how best to balance this set of intuitions, although I’ve made a few efforts to do so (Wood 2021, 2022). It may be that at the end of the day Julian—who I think shares most of these intuitions with me—is right that we will not truly understand how to balance our beliefs about God’s nature and purposes alongside our beliefs about evils like hell until God performs some “great deed” on the last day that will show us how in the end “all manner of things shall be well”.33 Perhaps we will not truly understand it even then. However, while I think there is much to admire in Julian’s willingness to hold in tension her view of God’s love and mercy with her acknowledgement of evils like hell, I think another sort of faithful response to such puzzles is to try to examine further what we truly believe about God’s nature and purposes. In doing so, sorting through the ways we believe God is required to prevent evils can play a useful role. Additionally, Sterba, who has done so in a comprehensive if questionable way, can prove a useful ally.