On the Significance of Assumptions about Divine Goodness and Divine Ontology for ‘Logical’ Arguments from Evil

Abstract

Sterba’s Is a Good God Logically Possible? (2019) draws attention to the importance of ethical assumptions in ‘logical’ arguments from evil (LAfEs) to the effect that the existence of (certain types) of evil is incompatible with the existence of a God who is all-powerful and morally perfect. I argue, first, that such arguments are likely to succeed only when ‘normatively relativized’—that is, when based on assumptions about divine goodness that may be subject to deep disagreement. I then argue that these arguments for atheism are also, and more fundamentally, conditioned by assumptions about the ontology of the divine. I criticise Sterba’s consideration of the implications for his own novel LAfE of the possibility that God is not a moral agent, arguing that Sterba fails to recognize the radical nature of this claim. I argue that, if we accept the ‘classical theist’ account that Brian Davies provides (interpreting Aquinas), then God does not count as ‘an’ agent at all, and the usual contemporary formulation of ‘the problem of evil’ falls away. I conclude by noting that the question of the logical compatibility of evil’s existence with divine goodness is settled in the affirmative by classical theism by appeal to its doctrine that evil is always the privation in something that exists of the good that ought to be.

1. Introduction: Mackie’s ‘Logical’ Argument from Evil

James Sterba’s Is a Good God Logically Possible? (2019) draws attention to the importance of ethical assumptions about what divine goodness entails in any plausible argument for the conclusion that the existence of evil (or, of certain types and/or degrees of evil) is incompatible with the existence of a theist God who is both omnipotent and perfectly good. Taking God’s providence over creation to be analogous to the care an ideally just state takes of its citizens, Sterba argues that such a God would need to meet a set of ‘evil prevention’ moral requirements which, given the facts about horrendous evils experienced in history, have (he maintains) evidently not been met. We may thus draw the conclusion, Sterba argues, that no such all-powerful and perfectly good God exists.

J. L. Mackie’s ‘Evil and Omnipotence’ (Mackie 1955) is widely taken as making the opening moves in the contemporary debate about ‘the problem of evil.’ However, it needs to be acknowledged, I believe, that this debate is, historically speaking, a relatively parochial one conducted by philosophers in the Anglophone analytic tradition. In particular, as I shall argue in what follows, the contemporary debate amongst analytic philosophers is limited by key assumptions, not only about divine goodness but also, even more fundamentally, about divine ontology.

Mackie argued that ‘the propositions that a good omnipotent thing exists, and that evil exists, are [logically] incompatible’ and the argument as he presented it has become known as a ‘logical’ argument from evil (LAfE).1 Mackie clearly acknowledged, however, that these propositions do not, purely by themselves, entail a contradiction. To arrive at a contradiction certain ‘additional principles’ are needed—and these crucially include the assumption that ‘a good thing always eliminates evil as far as it can’ (Mackie 1955, p. 201). But that is, of course, an ethical assumption—or, at least, it becomes an ethical assumption if the ‘thing’ concerned is understood to be a morally responsible personal agent.

It was thus clear from the start of this contemporary debate on ‘the problem of evil’ that judgments about what divine goodness requires were essentially implicated in any argument for the incompatibility of God’s existence with the existence of (certain kinds of) evil. The fact that this is so is reinforced by the strategy pursued by philosophers in response to Mackie’s LAfE. For, that strategy—and I am speaking here at the level of broad overview—was to deny the ethical assumption required to produce Mackie’s contradiction by holding that a perfectly good agent may have a morally adequate reason for permitting, or even directly causing, evil. Divine goodness, it may be argued, does not require eliminating evil whenever one can, since sometimes evil may be necessary, either as a means for achieving an outweighing ‘higher’ good or avoiding a worse evil, or as an unavoidable side-effect of bringing about and sustaining some significantly outweighing good.

Now, it is true that a pertinent doubt arises from the thought that morally adequate reasons for permitting or causing evil arise for us typically from our human limitations in knowledge and power, and so would not apply to omnipotent and omniscient God. Defenders of the coherence of theism sought to assuage that doubt, however, by arguing that even omnipotent God is subject to logical limitations, so that morally adequate reasons could arise for God to permit evils of a kind which were logically implicated in outweighing higher goods. Speculative theodicies were then produced to show how specific impressively important goods might indeed be logically unobtainable without permitting certain kinds of evil.

Thus, for example—to continue my ‘broad overview’ approach—it may be argued that evil is logically unavoidable if significant moral freedom is to be exercised, and those further important goods achieved (such as the good of inter-personal loving relationships) which presuppose the exercise of significant moral freedom. A world with significantly morally free beings (arguably) needs to be a world where natural law operates consistently and without constant supernatural intervention, and suffering by sentient beings (‘natural’ evil) will thus be an unavoidable feature. As well, free beings will use their freedom to make morally wrong choices which cause them and others harm (‘moral’ evils). Besides, having to cope with evil promotes the development of various important virtues (courage, sympathy) which would otherwise not emerge. Accordingly, it may be argued that, for all we know, omnipotent God could have morally adequate reasons to permit evils (both ‘moral’ and ‘natural’) based on logical limitations on his power—and this is all we need to rebut Mackie’s LAfE. Finally, if appeal is made to the (plausible) existence of evils which seem not to be clearly ‘covered’ by any speculative theodicy, a ‘sceptical theist’ move may then be made by arguing that God’s logical-limitation-grounded reasons for allowing evils may extend beyond our ability to discern, not simply what they are (for who would dare claim to know that?), but even what they could, for all we know, possibly be. Given some success in making intelligible the existence of some kinds of evils in a world created by a perfectly good omnipotent creator, it may be urged that we may be satisfied that there’s no overall incompatibility with God’s existence generated by the existence of any evils, since we may consistently believe that God has morally adequate reason for allowing even those that are egregiously inexplicable from our own perspective—it is just that that adequate reason is quite beyond our imagining. Of course, these responses to Mackie’s argument are open to contestation and further debate: my focus in this paper, however, is not on the detailed dialectic but on certain features of the framework within which it has been conducted.

2. Normatively Relativized’ ‘Logical’ Arguments from Evil

We may observe, then, that unless one has some grip on what divine perfect goodness would imply for a world that is the creation of an all-powerful and all-good creator, there is no prospect of arguing that some features of the world of our experience do not fit with its status as such a creation. And one can have such a grip only by taking a stance on the normative ethics of divine goodness. Thus, questions about what ethical principles would apply to the creator if the creator is understood to be an all-powerful moral agent have understandably been at the heart of debate about Mackie’s LAfE, as I have outlined in Section 1.

A point of the first importance now emerges: differences and disagreements about what the normative ethics of divine goodness actually are will potentially produce differences and disagreements over whether some actual feature of the world of our experience (such as the existence of certain types of evil, and the lack of clearly disambiguating evidence of God’s caring presence) is or is not consistent with its being the creation of an omnipotent and perfectly good creator. Sterba bases his version of a LAfE on some quite specific claims about the moral principles a perfectly good agent would adhere to—principles suggested by Sterba’s original use of a ‘just state’ analogy.2 Objectors might thus reply by disputing whether adherence to these, or to some of these, principles does indeed characterize divine goodness.3 And it might ensue—whether in this particular case, or more generally with other disputes over what ethical principles characterize the goodness of that-than-which-a-greater-cannot-be-thought—that, from our necessarily limited perspective, we find it moot, or a matter of ‘deep’ disagreement, whether or not some ethical principle that would seem to have been violated by an omnipotent agent that permits actual evils of a certain specific kind is, or is not, essential to perfect divine goodness.

When such deep disagreement does ensue, we will get what Ken Perszyk and I have called a ‘normatively relativized logical argument from evil’ (NRLAfE) (Bishop and Perszyk 2011). That is to say, it may be evident that, if one endorses such-and-such an ethical principle as essential to divine goodness, then certain actual evils logically could not have occurred had the world’s creator been an all-powerful and morally perfectly good intentional agent, yet no such incompatibility obtains if that principle is held not to be entailed by divine goodness. If, furthermore, the ethical principle in question is the subject of a deep disagreement, then one party may have a successful deductive argument from the existence of the evils in question to the non-existence of a perfectly good omnipotent personal creator, while the other party may be logically consistent in accepting that such a creator exists despite agreeing that the world does contain those very same evils.

As I have emphasized in Section 1, it was evident from the start of the contemporary debate that any LAfE must make ethical assumptions about what divine goodness entails. In that sense, one might say, any such argument might be described as ‘normatively relativized.’ What Perszyk and I meant to draw attention to, however, is the contestability of claims about what divine goodness entails, with the result that a claim that a certain evil or kind of evils logically could not obtain in the presence of divine goodness if divine power could have prevented it might stand only relative to a particular normative stance. Such a NRLAfE then justifies rejecting the existence of an omnipotent perfectly good personal creator only for those who find it reasonable to take that particular normative stance, where the context may be one of deep, irresolvable, disagreement in which an opposed normative stance cannot be shown to be unreasonable. Then, the argument in question does not merely rest essentially on ethical assumptions—it rests on assumptions over which there is disagreement, and potentially irresolvable disagreement. It may thus be that the closest an attempted ‘logical’ argument from evil gets to its goal is to succeed only relative to accepting a certain contestable normative stance (that is, that a successful LAfE would inevitably be a NRLAfE).

3. Ivan Karamazov’s ‘Rebellion’: A Key Example

In case my key claim in Section 2 seems unduly abstract, it will be useful to give examples of ‘logical’ arguments from evil that are relativized to a contestable normative stance. One example may be drawn from Ivan Karamazov’s ‘rebellion’ in Dostoevsky’s The Brothers Karamazov. Ivan relates harrowing tales of the abuse and torture of children, including one example of a child, beaten, birched and kicked by her own parents, who smear her face with her excrement and shut her up in the outdoor privy, where she ‘beats her little aching chest in that vile place, in the dark and cold, with her tiny fist and weeps searing, unresentful and gentle tears to “dear, kind God” to protect her.’ (Dostoevsky 1958, vol. I, p. 283) Finally, Ivan asks his brother, Alyosha, a novice monk:

… imagine that it is you yourself who are erecting the edifice of human destiny with the aim of making men happy in the end, of giving them peace and contentment at last, but to do that it is absolutely necessary … to torture to death only one tiny creature, the little girl who beat her breast with her little fist, and to found the edifice on her unavenged tears—would you consent to be the architect on those conditions? (op. cit., p. 287)

Ivan’s question, which expects—and in the novel’s narrative does indeed receive—the answer ‘no,’ envisages a normatively relativized ‘logical’ argument from evil. Relative to a stance that holds that any agent could not be morally perfect if he could have prevented, but, for whatever reason, did not prevent that little girl’s torture, the fact that she was tortured shows that, if the creator exists as an all-powerful agent, that creator logically cannot possess perfect goodness. But that normative stance can be contested. Analytic philosophers who defend theism are inclined to view such an argument as naively reliant on tugging at heartstrings. If, as Ivan envisages, the overall happy destiny of humanity really was at stake, would not a God with an eye on the overall goals of creation need to steel himself against compassionately intervening to prevent or cut short the child’s sufferings? As speculative theodicy suggests, for all we know, it may be necessary for achieving creation’s ultimate purposes to allow even the torture of children in order to preserve morally significant free will, to secure the historical order as a ‘vale of soul-making,’ and to achieve further essential goods whose very existence—or, at least, whose logical connexion with such horrifying evils—is necessarily beyond our ken.

The normative stance to which the LAfE suggested by Ivan’s question is relativized is thus contestable: though there is a strong moral obligation on agents able to prevent the torture of children to do so, it may be held that this is not an absolutely uniform obligation on all moral agents, but might be overridden in God’s case by a yet higher moral obligation unique to God’s role as creator. Yet there may be occasion for deep disagreement here. Maybe it is reasonable to hold—as Ivan’s question prompts us to imagine—that an all-powerful agent deliberating over creating a natural Universe and finding that achieving certain higher goods would require standing by while even one child was tortured to death would, if perfectly virtuous, regretfully choose to create a ‘safer’ and less ambitious Universe, or even decide not to create at all. Yet regarding such an outcome as a failure of divine courage may also seem reasonable: what’s the use of a God who balks at carrying out his key role as creator or settles for a creation in which the highest forms of good cannot emerge? Thus, deep disagreement about what divine goodness requires might yield, on one side of that disagreement, a NRLAfE: someone who is convinced that a perfectly good God could not have any reason for refraining from exercising his power to stop the most harrowing kinds of evils, such as the torture of children, will rightly find the actual existence of those evils logically inconsistent with the existence of an all-powerful agent who is also perfectly morally good.4

4. Assumptions about Divine Goodness

Such a strong conviction that some evils are so bad that there could not be any morally valid reason for allowing (let alone sustaining) them, no matter what ‘higher good’ was at stake, may be seen as a protest against the generally consequentialist ethical approach of much speculative theodicy. It often seems to be assumed that the normative ethics of divine goodness are straightforwardly consequentialist—that is, that the logical necessity for evils of certain kinds if certain important goods are to obtain is by itself enough to reconcile the existence of those evils with an omnipotent agent’s perfect goodness. Yet theist ethics have typically rejected the general principle that ‘the end justifies the means,’ preferring Kant’s categorical imperative (in one of its formulations) that we should never treat humanity, whether in ourselves or in others, as a means only but always as an end in itself. And Sterba emphasizes what he calls the ‘Pauline Principle’ that ‘we should never do evil that good may come of it,’ which, as he observes, ‘seems to be in direct conflict with God’s permitting evil and making up for it later’ (Sterba 2019, p. 49).

It might be argued, of course, that a baldly consequentialist normative ethics does apply to the unique case of the creator, even if a suitably action-guiding ethics for finite agents like ourselves requires respecting the Pauline Principle and Kant’s formula of humanity as an end in itself.5 Nevertheless, thinking of God as a supremely good personal agent, and reflecting on what is held to be revealed about God’s goodness in the Abrahamic religious traditions, it must surely seem that the Creator would not be concerned solely with an overall ultimate good outcome, but also with the care of each and every one of his creatures. Marilyn Adams argues that God’s perfect goodness makes him good to every created person, including those implicated in ‘horrendous evils,’ which are, on the face of it, so severe as to render their lives not worth living (see Adams 1999, pp. 20–22, 26, 31). Accordingly, the need to allow such evils for the sake of outweighing goods does not by itself vindicate God’s perfect goodness; God will also ensure that participants in horrors—perpetrators as well as victims—are eventually brought into the joy of eternal relationship with him (see Adams 1999, pp. 49–55 and chp. 8). Adams here places the emphasis of a theist response to a ‘horrendous-evils-based’ LAfE on the resources that God has to deal with evil, rather than on speculatively justifying God’s permitting evils in the first place—although she certainly does not deny God’s ultimate responsibility for all the evil that blights the creation. On her account, then, even the worst horrors are consistent with the existence of a personal God who is both all-powerful and morally perfect, provided the overall context is one where all victims are immeasurably recompensed and all perpetrators of horrors eventually reconciled in the ultimate bliss of eternal communion with the God of love.

May we appeal, as Sterba may suggest, to the ‘exceptionless minimal components’ of the ‘Pauline Principle’ to rule out God’s possibly ‘making up later’ for horrors in the way that Adams, speaking as a Christian philosophical theologian, maintains he will? Adams would, I believe, regard such a reply as completely underestimating the vastness of the love of God, and its incommensurably outweighing even the worst horrors of human history. Philosophers might, however, raise objections to the view, essential to Adams’ account, that mortal human history is not all there is for individual personal existence—objections, that is, to the assumption that there can be a meaningful post-mortem existence in which important developments can take place amongst persons, including reconciliation between evil-doers and God, between the perpetrators of horrors and their victims, and between those who have suffered evil and the God who let them suffer. Those objections will include concerns about whether the conditions for diachronic personal identity with historical persons could be satisfied in such an ‘after-life,’ whether never-ending existence as a finite person would eventually be inescapably tedious, and whether human mortal history would be rendered less significant, even otiose, as a mere preliminary to the ‘real thing.’

If all those concerns could be met, a further question to which Perszyk and I have drawn attention still needs considering: could a God who first allows people to suffer horrors and then ultimately brings them into eternal relationship with him have acted so as to form the best kind of overall inter-personal relationship with those persons? We have envisaged a NRALfE according to which, granted a certain stance in relationship ethics, an omnipotent person who presides over the whole suffering-and-redemption scenario as described by Adams would fall short of perfect goodness in relationship with others (Bishop and Perszyk 2011). Adams would reject our intuition that God would not be placing himself in right relationship with created persons if he acted in relation to participants in horrors as she thinks he does—after all, this is a ‘logical’ argument from evil which is relativized to moral assumptions about which deep disagreement is possible. Nevertheless, Adams has agreed that the issue is an important one, needing more exploration.6

5. Assumptions about Divine Ontology

‘Logical’ arguments from evil in the ‘Mackie tradition’ are limited, then, by their dependence on ethical assumptions, about which there may be deep disagreements, so that it may turn out that the most compelling versions of the charge that certain evils are inconsistent with the existence of an all-powerful and morally perfect God are convincing relative only to a specific ethical stance which others may reasonably reject. But these arguments are subject to a further, more fundamental, limitation. This further limitation is the ontological assumption that, if God exists, then God is—at the level of how reality is, fundamentally and ultimately—a personal intentional agent whose agency is all-powerful and whose goodness is personal, moral, goodness. Thus, for example, Sterba’s ‘Moral Evil’ and ‘Natural Evil Prevention Requirements’ apply to God only if God is a moral agent, morally responsible for his actions in creating and within creation. Sterba’s own LAfE, then, fails if God is not a moral agent. Now, Sterba is well aware of this limitation, devoting his Chapter 6 to the question ‘What if God is not a moral agent?’, and I’ll now consider how Sterba deals with that question.

I have already noted that God might be a moral agent, and yet—given God’s unique situation as a supreme individual being on which all other beings depend for their existence—actions may be permissible for him which would be impermissible for any finite, created, agent. Making allowances for God’s unique moral situation should not be allowed to go too far, however. For example, it might be suggested that, as our maker, God is within his moral rights in treating us in any way he wishes.7 On that view God’s goodness as an all-powerful supreme agent would be straightforwardly compatible with the existence of any variety and amount of evil. However, God’s goodness would then be rendered unintelligible as such from a human perspective, and worship could only be submission to God’s power and not praise for God’s goodness.8

No, God as a moral agent may be justified in allowing preventable evils up to a certain point, but—as is a key theme in Sterba’s argument—God must surely, if perfectly virtuous, keep matters in proportion and not remain aloof in situations where, for example, respecting the significant moral freedom of a genocidal tyrant will result in extinguishing the moral freedom of millions. Human history features events of just that kind, however, which may thus plausibly be regarded as showing that, if a supremely powerful moral agent does exist, he has not respected any such principle of proportionality. An LAfE based on that observation may, however, be, at best, a normatively relativized one of the kind to which I’ve already drawn attention—decisive only for those who take a certain normative stance on one side of a potentially deep disagreement about what divine goodness entails.

Marilyn Adams notes that her ‘favourite five’ scholastics (Anselm, Bonaventure, Aquinas, Scotus and Ockham) ‘were unanimous that God is too big to be networked to us by rights and obligations’ (Adams 2013, p. 22). She herself endorses that view, yet retains a conviction that God is a personal agent whose ‘morality’ is similar to the honour code that bound a feudal lord to his serfs and vassals. It’s unclear why being honour-bound would not be a matter of having duties (namely, those duties proper to one’s privileged position); but, in any case, even if God’s goodness should not be understood in terms of God’s duties to creatures, God’s goodness surely would still pertain to the way God relates to his creatures. In that case the concern raised by Perszyk and myself (mentioned above in Section 4) will still arise even if Adams is right in holding that God is not subject to moral obligations. That concern (as argued) is the source of another potential LAfE which goes so far as to concede the reality of an unlimited post-mortem existence for human persons. Once again, though, that argument is a normatively relativized one—yet none the less decisive for those for whom the normative stance invoked is a compelling one.

Could it be, though, that God is not a moral agent at all—not merely a moral agent whose special position makes him subject to morality in a unique way? Sterba attributes the claim that God is not a moral agent to Brian Davies, or, more precisely, to Davies as interpreting Thomas Aquinas (Davies 2011). Sterba’s main argument in response is that if God is a rational agent, then the moral law for all rational agents, which itself arises from God’s goodness, must apply to God himself (see Sterba 2019, p. 116). This reply seems cogent: if God is a (personal) rational agent, then God surely has to be also a moral agent. Sterba argues further, however, that, even if it is conceded that God somehow escapes being a moral agent, then, if God is the agent who creates the world, God still remains causally responsible for ‘far more evil than that [which] has been produced by all the great villains among us’ (Sterba 2019, p. 117)—and that surely is a serious obstacle to ascribing worship-worthy goodness to God.

6. Divine Agency, Divine Simplicity, and Analogous Predication

Sterba seems to have missed, however, the radical nature of the claim that God is not a moral agent. This ‘Davies/Aquinas claim’ is not that God is a rational agent, only just not a moral one (Sterba is surely correct to find that proposal unstable); rather the claim is that God is not a moral agent, because God is not ‘an’ agent at all—nor, indeed, any kind of ‘being amongst beings.’ According to this claim, then, God’s power is not an individual agent’s power, neither is God’s goodness an individual agent’s virtue. Thus, if—as ‘Mackie-tradition’ LAfEs claim—certain actual evils would not have existed had a supreme all-powerful and perfectly virtuous individual also existed, there is no implication that God does not exist, since the existence of God is not a matter of the existence of such an individual.

Davies has drawn attention to the importance of recognizing two opposing ways of interpreting the theism of the Abrahamic religious traditions, which he has called classical theism and theistic personalism.9 Theistic personalism is familiar from most contemporary discussion of theism in analytic philosophy; it holds that God’s existence is the existence of a supernatural, immaterial, person with the omni-properties (‘the personal omniGod’). However, according to the classical theism of Aquinas, for example (and, for example, Maimonides in the Jewish tradition, and Ibn Sina in the Islamic), though personal language is revealingly used in conveying truths about God in scripture, and properly used in addressing God in prayer and worship, when faith seeks understanding it recognizes that God’s reality is not that of a supreme individual person or mind like a human person or mind only vastly greater. As source and ultimate end of all that exists, God’s reality absolutely transcends the reality of any particular being or entity amongst other beings or entities. God is not ‘a’ thing, ‘a’ substance, or ‘a’ being at all—not because God is unreal, but precisely because God’s reality is complete reality (‘pure act’ to use one of Aquinas’s descriptions). God may be said to be ‘no-thing,’ yet this emphatically does not entail being nothing in the sense of not-being. To the contrary, God’s reality is so great—to recall Anselm (of course, another classical theist)—that nothing with greater reality could even be conceived.

We may conclude, then, that if Sterba’s LAfE succeeds in showing that there is no personal omniGod, it will nevertheless fail to support atheism if a classical theist divine ontology is correct. This is an important limitation—and one which Sterba comes close to recognizing when he acknowledges the objection that the natural law grounded in God’s goodness may not apply to God because ‘natural law only applies to beings in virtue of their belonging to a certain kind or kinds, and God does not belong to any kind of being, and so natural law does not belong to him’ (Sterba 2019, endnote 9, p. 137). This claim that God does not belong to any kind of being is a consequence of the teaching that God is ‘simple,’ in the sense that there is no ‘composition’ in God. This teaching is central to classical theism: in particular, with reference to the Aristotelian metaphysics it deploys, God lacks the ‘composition’ of essence and existence. Creatures exist only because God ‘composes’ each individual’s existence with the essence of the kind of creature that they are. But God, who is absolutely a se (‘from himself’), does not depend on anything else to give him existence, and is thus necessarily not an instantiation of any kind of thing (a fortiori, then, God is not an individual person or mind, however supreme and exalted).

But Sterba dismisses the claim that God does not belong to any kind of being: ‘Nevertheless,’ Sterba writes, ‘God is said to be rational, and it is in virtue of his being rational that the same (moral) natural law applies to God as to ourselves’ (ibid.). Classical theists will indeed agree that ‘God is said to be rational’—and they will speak of God as knowing and willing—but they will understand these ways of speaking as an analogous extension from the human personal context in which they are at home and where their meaning is understood. They will not accept that this language is transparent to an underlying metaphysics of God as ‘a’ human-like (though supremely exalted) rational being with human-like (though infallible) knowledge and a human-like (though always finally unimpeded) will. On a classical theist view, God may be said, and said truly, to be rational, and we understand what this means on the basis of our understanding what it is for us to be rational; yet we cannot comprehend what it positively amounts to, in reality, for God to be rational. Our comprehension is limited to the negative recognition that divine rationality (and knowledge, and will) are not just further instances, possessed by a supreme being and appropriately ‘extended’, of these properties as possessed by finite humans. Thus, classical theists block the very inference Sterba makes: though we may truly say that God is rational, the moral law grounded in God’s goodness which applies to all rational beings does not apply to God as to one rational being amongst all others. That God is rational involves an analogous predication, which has readily intelligible consequences—in particular, it excludes the possibility that the divine purposes in creation could be at odds with one another (we may be sure, for instance, that God does not hate anything that he has made). However, what it is in divine reality that makes it true that God may be said to be rational is beyond our full comprehension, though we may clearly understand that the truth-maker for this claim is not the existence of a supreme individual personal being possessing the property of rationality.

7. The ‘Classical’ Alternative to ‘Personal-omniGod Theism’

Sterba has not, then, closed off a ‘classical theist’ reply to his LAfE. It might be that God is not a moral agent because God is not a rational agent at all, even though God may rightly be said to be rational by an analogous predication. Admittedly, this shortcoming is not much of a handicap in the contemporary analytic philosophers’ debate, since classical theism—with its doctrine of divine simplicity—is widely rejected by analytic philosophers as unintelligible or absurd.10 Furthermore, some philosophers expert in mediaeval philosophy maintain that classical theism—in the normative sense of the authentically transmitted tradition—is not the non-personalist ‘classical theism’ defined by Davies, but, rather, the personal-omniGod theism familiar from the ‘analytic’ debate.11 Still, within the context of the ‘standard’ assumption that God is (or is, near enough) the personal-omniGod,12 Sterba’s argument poses a significant challenge. As I have suggested above, Sterba’s argument gives decisive grounds for personal-omniGod-atheism to those who endorse the argument’s normative assumptions. At the same time, however, there may be good prospects for personal-omniGod theists to defend the coherence of their worldview by rejecting or suitably amending those assumptions. Arguably, the underlying ethical debate here is not rationally resolvable, so that neither side can claim that its position is ‘the’ only rational stance to take. Unless one thinks that there can in principle be a completed rationally grounded normative ethics, it must be allowed that such an inherent evaluative rational impasse is possible—and then it is a matter for judgment whether impasses of this kind do indeed make an appearance in normative-assumption-based arguments from evil.

Those who do find a logical argument from evil compelling (because they are committed to the values that, according to that argument, render an all-powerful agent morally flawed for permitting certain actual evils or kinds of evil) need not rationally commit themselves to atheism if they can endorse a ‘non-personalist’ understanding of theism, such as is provided, according to Davies and others, by classical theism with its key doctrine of divine simplicity.13 The possibility that Sterba dismisses, then, may be of considerable intellectual and existential importance to many people.14

Analytic philosophers’ tendency to reject as incoherent the idea that God is not ‘a’ personal being, nor, indeed, ‘a’ being of any kind, might be seen as a refusal to allow that ultimate reality could be ‘incomprehensible’ in the sense that it transcends all the categories limited human minds have available for comprehending anything.15 But what could justify such a refusal? Physical reality’s (relative) comprehensibility seems compatible with the ‘incomprehensibility’ of the ultimate meaningfulness and purposiveness of reality which religion may affirm. Our physical scientific understanding of reality in its most abstract and general nature is subject to certain inherent limits; how much more limited, then, might our understanding be of reality at its most concrete, namely in its working out of an ultimate purposiveness such as theism posits?

Nevertheless, even if the essence of divinity is beyond our comprehension in the sense that it is beyond our intellectual mastery (in some supposed theoretical metaphysical science), commitment to theism has to be intelligible as a particular and distinct stance on ‘how things ultimately are.’ The metaphysics of the classical theism of the mediaevals is largely apophatic—affirming that God is atemporal, immutable, impassible (not such as to undergo any process or experience), necessary (not contingent, not able to not exist), and simple (not in any way ‘composed,’ and so, not a thing of any kind, nor a thing which ‘possesses’ its attributes). Some account of what the God who is none of these things positively is would be welcome—though we must recall that mediaeval apophaticists had no doubt that the atemporal, immutable, etc. One was revealed in their various traditions as (for example), the One who brought Israel out of Egypt and gave the law to Moses, the One who is Father of Our Lord Jesus Christ crucified and risen, or the One whose messenger, Muhammed, received the final revelation of the Qur’an—and these, of course, are distinctly positive descriptions. Still, philosophers will surely be keen to seek a positive metaphysical account from within the theist perspective of what it is for God to exist if God is ‘no-thing’—an account that might satisfy fides quaerens intellectum even when such faith humbly accepts that full comprehension is beyond the created mind.16

8. Classical Theism and the Problem of Evil

Sterba quotes Davies (or, Davies interpreting Aquinas) as claiming that the problem of evil as usually understood by contemporary philosophers of religion ‘is not a serious problem at all but rather the result of a confused way of thinking about God’ (Sterba 2019, p. 111). In response to this claim, Sterba’s strategy is to argue that the clarifications Davies/Aquinas makes to correct this ‘confused way of thinking’ do not in fact render the problem of evil as usually understood less ‘serious.’ According to Sterba, even if these ‘classical theist’ clarifications are accepted, a ‘logical’ argument from evil (the detailed one he himself proposes) remains standing. In the previous Section, I have criticised Sterba’s strategy, arguing that he fails to recognize just how radical the Davies/Aquinas clarification about divine ontology actually is. It is not just that God is not a moral agent, God is not a rational agent either, nor any kind of agent, nor ‘a’ being of any kind at all. Though talk of God as exercising rational agency is apt, and (from the perspective of faith) may convey revealed truth, it does so through an analogous extension of the language of agency and of beings and their properties. Indeed—to extend a little my earlier remarks—even our referring to God as something that can be the subject of (analogous) predication is itself a significant piece of analogising and thus not transparent to, nor grounded in, an ontology of a (divine) substance ‘having’ attributes.

This ‘Davies/Aquinas’ clarification of how to think about God needs to be seen, I believe, as an honouring of the absolute transcendence and incomprehensibility of divine reality—transcendence, that is, over the created world of our experience and the categories that make that world intelligible for us. As I’ve noted, analytic philosophers typically find that such an ‘honouring of transcendence’ descends into unintelligibility—and there clearly is an important issue here as to whether showing Abrahamic theism to be an intelligible worldview requires a properly anthropomorphic understanding of God’s reality or else demands that we disown it. I will not attempt to address that issue any further here. However, there is one final item of business for the present discussion, concerning how ‘the problem of evil’ stands if one does concede to Davies/Aquinas that thinking of God as—literally and metaphysically—a supreme all-powerful and morally perfect being is ‘a confused way of thinking about God.’

If one does agree that God is not any kind of agent, nor any kind of being at all, then it is clear that the problem of evil as usually understood by contemporary philosophers ceases to be a serious problem. In fact, the problem then does not arise at all, since, as usually understood, ‘the problem’ is the intellectual one posed by a ‘logical’ argument from evil—and ‘logical’ arguments from evil in the Mackie-tradition simply assume the ontology of God as an all-powerful and morally perfect personal agent.

It is important to emphasize, however, that a classical theist ontology in which God is not ‘a’ being does not side-line evil as a problem altogether. Evil presents an existential problem, which theist faith purports to help us resolve. Faith assures us that, through God’s grace and mercy, we may overcome evil and persist in the pursuit of the good in the well-founded hope that we may thereby find our own fulfilment as the kind of beings we were created to be. A soteriology is essential to any theist worldview—which could hardly be so if theism did not hold that evil presents a serious problem. However, an intellectual problem of evil still remains for classical theism—even though it is no longer the specific intellectual problem of how an ultimately all-controlling person who sustains evil can possibly be fully virtuous. For classical theism, the world is God’s creation and God is perfectly good. How could this possibly be so given the horrendous evil that mars this (supposedly) good creation? One might suspect that the Davies/Aquinas clarification about how we should think about God tips us out of the frying pan into the fire so far as understanding how evil can exist in the first place, if—as theist soteriology assures us—God can ultimately deliver us from it.

In fact, however, classical theism offers a way of understanding what evil is—namely as privatio boni, the privation of the good—which (as I shall shortly explain) makes it clear that evil’s existence is logically compatible with God’s goodness. This understanding of evil is, of course, an understanding from within a theist perspective—but this is enough to dispel the charge that theist commitment is irrational because it is commitment to a worldview that cannot possibly be true on account of its internal logical incoherence. (However, it is important to recognize that the charge that theist commitment is irrational because it goes against the weight of all our available evidence may yet remain.)

Philosophers often take the idea that evil is privation of the good as implying that evil is illusory.17 But that is a mistake. This is because the claim is that evil is the privation, or lack, of the good that there ought to be: accordingly, where there is evil, this lack in what does exist is anything but illusory and can amount to something quite horrendous.18 Where creation is understood to exist for a supremely good purpose—into which fit the particular, naturally discernible, ends (telē) of the diversity of creatures—what ‘ought to be’ is, of course, the good that God wills. Thus, as Aquinas argues, there can be no evil in this sense unless God exists, so that evil, so understood, is clearly logically compatible with God’s existence.19 However, it may still seem puzzling how it could actually turn out that God’s creation does contain privations of the good that God wills, even though this is a logically coherent possibility. Furthermore, the worst horrendous evils seem to weigh strongly as evidence against the actual existence of a sovereign divine goodness. Indeed, it may seem that the worst horrors need classifying as inherent evils, rather than as privations of the good that ought to be. Much more needs to be said, then, to respond fully to evil-generated doubts about God’s existence. One thing theists will need to maintain, I believe, is that privative evils are inherent in the concrete realization of the highest forms of good in any finite, material and historical world. For a full response to these evil-generated doubts, theists will no doubt need to deploy theological resources specific to their particular religious traditions (such as, in the Christian case, the soteriological significance of the Incarnation). However, it is beyond my present focus on the question of theism’s internal logical coherence to attempt any elaboration along these lines here.

Within my present focus, then, I believe I have said enough to show that we may give an affirmative answer to Sterba’s specific title-question, ‘Is a good God logically possible?’ There is no internal contradiction in a theist worldview that accepts that evils exist. A proviso is needed, however. There is no internal contradiction, provided we accept a classical theist account of evil as the privation of the good and the classical doctrine of divine simplicity. If it is insisted, however, (contrary to divine simplicity) that God (if God does exist) is a supernatural personal being who is both all-powerful and morally perfect, then the threat of internal contradiction in the theist worldview remains. For, if such a theist metaphysics is insisted upon, one may well find oneself, with Sterba, committed to moral values which do indeed entail that an all-powerful personal being could not be fully virtuous in sustaining certain actual kinds of evils.