The Evil God Challenge

There is a thought experiment that has the rare quality of being both philosophically rigorous and immediately, viscerally obvious once you encounter it. It was developed in careful academic form by the British philosopher Stephen Law, but its underlying logic requires no specialist training to follow. You need only be willing to apply a simple test consistently: if an argument works in one direction, does it work equally well in the opposite direction? If it does, and if the opposite direction leads somewhere you find absurd, then you have a problem with the original. Not a minor technical problem requiring a footnote, but a deep structural problem requiring an honest reconsideration of the entire enterprise.

The thought experiment is called the Evil God Challenge, and it is directed squarely at the problem of evil, which is the oldest and most persistent difficulty in philosophical theology. The problem of evil, in its simplest form, asks how an all-powerful, all-knowing, and perfectly good God could permit the suffering that fills this world: the cancers consuming children, the earthquakes burying families, the famines hollowing out whole populations, the predatory violence woven into the fabric of biological life long before any human sinned against anything. Theologians have laboured for centuries over this question, producing a rich literature of responses called theodicies, each one attempting to explain why a genuinely good God would allow or even arrange such a world. The Evil God Challenge does not attack those theodicies directly. It does something more elegant and more devastating: it takes each theodicy in turn, flips its logic, and shows that the same style of reasoning would justify belief in an all-powerful, all-knowing, and perfectly evil God, one who arranged this world primarily to cause suffering, and who permits its pleasures only as instruments of cruelty. If the mirrored arguments are obviously absurd, as most people immediately recognise them to be, then the symmetry forces a question that deserves a plain answer: why are the originals any less absurd?

This essay walks through that symmetry in detail. It examines the major theodicies, constructs their evil mirrors, evaluates the standard responses to the challenge, and arrives at a conclusion that the problem of evil has not been solved by any of them. It has only been mirrored.

1. The Landscape Before the Mirror

Before the mirror is introduced, it is worth setting out the precise nature of the problem it is designed to illuminate. The problem of evil comes in two main varieties, and both matter here. The logical problem of evil, associated most powerfully with J. L. Mackie’s 1955 paper “Evil and Omnipotence,” argues that the simultaneous existence of an omnipotent, omniscient, omnibenevolent God and genuine evil is a flat logical contradiction. If God is all-powerful, he could prevent evil. If he is all-knowing, he is aware of every instance of it. If he is all-good, he would want to prevent it. Evil is a feature of this world, demonstrably and abundantly. Therefore, the God described by classical theism does not exist. Mackie’s argument was precise, and the philosophical debate it generated was productive, but most contemporary theologians regard the strictly logical version of the problem as having been at least partially addressed. They argue, following Alvin Plantinga’s free will defence, that it is not logically impossible for God to have good reasons to permit evil, and that freedom requires the genuine possibility of choosing wrongly.

The evidential problem of evil is considerably harder to dismiss. Associated with philosophers such as William Rowe, it does not claim that the existence of God and the existence of evil are logically incompatible. It argues instead that the sheer scale, distribution, and character of suffering in the world constitutes powerful evidence against the hypothesis of a perfectly good creator. This is not a deductive argument but a probabilistic one: the world looks far more like what we would expect from a universe with no guiding intelligence, or from one governed by an indifferent or actively hostile power, than it looks like what we would expect from the workshop of a being of infinite love and wisdom. The evidential problem is the one that bites hardest, because it cannot be defused by logical ingenuity alone. It requires a genuine account of why the evidence points where it does.

The theodicies produced in response to the evidential problem are numerous, but they cluster around a manageable set of families. The free will defence holds that God permits human moral evil because freedom, even freedom to do terrible things, is intrinsically valuable. The soul-making theodicy, developed most fully by the philosopher John Hick drawing on Irenaeus, holds that a world without difficulty and suffering could not produce the virtues, the character, and the spiritual depth that make human beings genuinely valuable in a moral sense. The greater-good theodicy argues that suffering, though bad in itself, serves some larger purpose that we cannot fully perceive from our limited vantage point, as a surgeon’s knife is harmful in isolation but beneficial in context. The sceptical theist response, associated with Michael Bergmann and others, argues that our cognitive limitations mean we cannot reliably judge whether or not God has sufficient reasons for permitting the evil we observe, and that our failure to see such reasons is no evidence that they do not exist. These are the main weapons in the theodicist’s armoury, and they have been refined over decades of serious philosophical engagement.

Stephen Law’s move, introduced in his 2010 paper “The Evil God Challenge” published in the journal Religious Studies, is to ask a question that, once asked, seems astonishing in its simplicity: do these arguments work symmetrically? Can the same style of reasoning justify belief in a God who is not perfectly good but perfectly evil?

2. Constructing the Evil God Hypothesis

The evil god hypothesis is not the claim that the universe is governed by a comic-book villain, cackling at human misery. It is a philosophically precise mirror of classical theism. Where classical theism posits a being of unlimited power, unlimited knowledge, and unlimited goodness, the evil god hypothesis posits a being of unlimited power, unlimited knowledge, and unlimited malice. This being created the universe with the intention of maximising suffering, not out of some limitation or competing interest, but as its fundamental purpose. It is all-powerful, so nothing occurs without its permission. It is all-knowing, so it is aware of every joy and every agony. It is all-evil, so its overriding disposition is the infliction of suffering and the frustration of flourishing wherever possible.

The immediate response from virtually every person who hears this hypothesis is that it is obviously wrong. The world contains too much genuine goodness, pleasure, love, laughter, beauty, and moral depth to be the product of a maximally evil creator. Sunsets are genuinely beautiful in a way that seems to serve no malicious agenda. Children laugh with real delight at things of no harmful significance whatsoever. Acts of extraordinary self-sacrifice occur in every culture and every period of history, serving no discernible purpose of cruelty. Science yields genuine understanding that improves real lives and alleviates real suffering. The objection seems powerful at first: an evil god would surely have made a worse world than this one.

But notice what has happened in making this objection. The person making it has just committed themselves to a principle: the character of the world counts as evidence about the character of its creator. If the world’s goodness counts against an evil creator, then the world’s evil counts against a good one. The evidential weight flows in both directions, or it flows in neither. You cannot selectively invoke the world’s features as evidence when they support your preferred hypothesis and dismiss them as irrelevant when they undermine it. That is not reasoning; it is motivated special pleading dressed in the language of argument.

The more interesting and philosophically serious response is to reach for the theodicies. If the theist can explain why a good God permits evil, perhaps the evil-god defender can explain why an evil God permits goodness. And here is where the symmetry becomes unavoidable, because the evil-god theodicies, which we might call reverse theodicies or anti-theodicies, map onto the standard ones with startling precision.

3. The Free Will Reverse Theodicy

The free will defence is probably the most widely deployed response to the problem of evil, and it has genuine philosophical substance. The argument runs approximately as follows. Freedom is among the highest goods a creator could bestow. A world populated by beings who genuinely choose their actions, including moral actions, is more valuable than a world of morally perfect automata who have no real choice in the matter. God, being perfectly good, desires the higher good. The higher good requires freedom. Freedom entails the genuine possibility of choosing evil. Therefore, God permits moral evil as a necessary condition of the freedom that makes genuine moral goodness possible. The suffering caused by human freely chosen wickedness is, on this account, the price of the gift of freedom, and it is a price worth paying.

Now consider the mirror. An evil God desires the highest evil, and the highest evil is not merely the mechanical infliction of pain but the genuine suffering produced by betrayal, cruelty, and moral failure chosen freely. A world in which beings freely choose to torture, to oppress, to deceive, and to destroy is more evil, in the sense that interests a malicious creator, than a world of suffering automatons grinding through preprogrammed misery with no agency of their own. The evil of a freely chosen betrayal is richer, more exquisite, and more satisfying to a malicious creator than the evil of a falling rock. Therefore, an evil God grants freedom, knowing that free beings will produce genuine moral evil, and accepts as the necessary price of this arrangement that some of those free beings will occasionally choose kindness, love, and generosity. The goodness in the world, on this account, is the regrettable but unavoidable cost of the freedom that makes genuine moral depravity possible.

The logical structure of this argument is identical to the standard free will defence. It uses the same move: posit a higher purpose that the observed contrary evidence serves as a necessary condition. If the free will defence is compelling in the theist’s hands, the reverse version is equally compelling in the evil-god defender’s hands. If the reverse version seems obviously absurd, the principle that condemns it, namely that the observed character of the world should constrain our hypotheses about its creator, equally condemns the original.

There are, of course, objections to the free will defence even in its standard form. Most significantly, it addresses moral evil, the suffering caused by human choices, but does nothing whatever to explain natural evil: the tsunamis, the childhood leukaemias, the parasites that blind millions, the plate tectonics that buried Lisbon in 1755. Free will is not implicated in the suffering of a deer burning alive in a forest fire started by lightning, nor in the intestinal worms that have tormented human children since before recorded history. The free will defence, taken alone, is therefore a partial response at best, and the reverse theodicy is equally partial in precisely the same direction: it explains why an evil God permits freely chosen goodness, but it does not explain why an evil God permits the natural pleasures of the world, the genuine warmth of sunlight, the satisfaction of a good meal, the biological reward of sleep.

4. The Soul-Making Reverse Theodicy

The soul-making theodicy, in its Hickian form, offers a response to a wider range of suffering than the free will defence manages. The argument holds that human beings are not created in a state of moral and spiritual perfection, but are instead created with the capacity for such perfection, which must be developed through struggle, adversity, and moral choice. A world of frictionless ease, in which every desire is immediately satisfied and every difficulty instantly removed, would be a world in which no genuine virtues could form. Genuine courage requires the presence of real danger. Genuine compassion requires that there be real suffering in others that we are called to alleviate. Genuine patience requires the friction of genuine frustration. Genuine integrity requires the reality of genuine temptation. The world’s difficulties and suffering are, on this account, the necessary gymnasium in which human souls are formed, and God, being perfectly good, arranges the gymnasium because the product, genuinely virtuous and spiritually mature souls, is worth the price of the training.

The mirror is immediate and exact. An evil God, concerned not with soul-making in the positive sense but with soul-breaking, or more precisely with the cultivation of cruelty, meanness, spite, and malice, would arrange a world in which genuine opportunities for the development of these qualities are available. A being cannot develop into a truly accomplished torturer in a world with nothing to hurt. It cannot refine its capacity for contempt in a world with nothing to despise. It cannot perfect its talent for betrayal in a world with no trusting relationships to violate. The goodness of the world, its beauty and its love and its moments of genuine human connection, serves as the raw material from which an evil God’s project of corruption is fashioned. The joy exists to be destroyed; the love exists to be betrayed; the beauty exists to be defiled. Far from being evidence against an evil creator, the goodness of the world is, on this account, its most essential resource. An evil-god world needs happiness as an ingredient in misery exactly as a soul-making world needs suffering as an ingredient in virtue.

Once again, the logical structure is identical. The theodicy works by identifying a higher purpose for which the observed contrary evidence serves as a necessary condition or raw material. The soul-making theodicy makes suffering the necessary precondition of virtue. The reverse makes goodness the necessary precondition of corruption and cruelty. If one argument is persuasive, the other should be equally so. If the reverse argument seems strained and grotesque, the question worth pressing is whether it is genuinely logically inferior, or whether it merely conflicts with our prior intuitions about the world’s creator.

It is worth pausing here to register just how unpleasant the reverse theodicy sounds. The suggestion that love exists primarily to be sharpened into an instrument of cruelty, that the laughter of children serves mainly to make their eventual suffering more acute, is morally repellent. But this reaction is itself evidence of something important. The reason the reverse theodicy sounds so bad is precisely that we have strong moral intuitions about the world’s character, intuitions that tell us it does not look like the product of a maximally malicious creator. Those same intuitions, operating in the other direction, should tell us something about the plausibility of the good-god hypothesis when confronted with the scale of the world’s unredeemed suffering.

5. The Greater Good and the Greater Evil

The greater-good theodicy is, in some ways, the most flexible and the least falsifiable of the standard responses. It holds that suffering, however inexplicable it appears from our limited perspective, serves some greater good that an omniscient God perceives and that we, with our finite and partial vision, cannot. The classic illustration is the parent who takes a young child for an injection. The child cannot understand why the person they trust most in the world is holding them still while a stranger pushes a needle into their arm. From the child’s perspective, this is inexplicable and apparently cruel. From the adult’s perspective, it is an act of love that prevents a far greater harm. The theodicy asks us to consider that our relationship to God is analogous: we are the child, unable to see the full picture, and the suffering we experience or witness is a small local harm in the service of a vastly larger good we cannot currently perceive.

The greater-evil reverse theodicy writes itself without any effort at all. A maximally evil God permits the goodness we observe because that goodness serves some greater evil that we cannot perceive. The momentary warmth of a human friendship is the setup for a more devastating betrayal. The years of health before a terminal diagnosis are the scaffolding that makes the final collapse more complete and more agonising. The love between a parent and a child intensifies the agony when that child is lost. Every joy is, in this framework, an investment in a larger portfolio of suffering, the returns on which will be collected in ways and at times that we, with our limited vision, cannot fully trace. The good in the world does not contradict the evil creator; on this account, it is that creator’s most sophisticated instrument, the long game of a being with unlimited time and unlimited malice.

Notice that both the greater-good and greater-evil theodicies share a structural characteristic that distinguishes them from the free will and soul-making responses. They are unfalsifiable in principle. No observation can count against them, because any apparently contrary evidence is explained away by appeal to inscrutable purposes beyond our comprehension. This is not a strength of the argument; it is its most serious weakness. An explanation that can accommodate any possible observation explains nothing. It is not a theodicy but an evasion dressed as one, and the tell is that its mirror image is equally unfalsifiable and equally accommodating. The sceptical theist response, which explicitly embraces this unfalsifiable structure, argues that our cognitive faculties are simply not equipped to judge whether God has sufficient reasons for the evils we observe. This may sound like intellectual humility, but it has a devastating implication: if our cognitive faculties cannot reliably assess God’s reasons for permitting evil, they also cannot reliably assess God’s reasons for permitting goodness. The sceptical move collapses both sides of the ledger simultaneously, which means the sceptical theist cannot coherently use the world’s goodness as evidence for a good God while simultaneously denying that the world’s evil is evidence against one.

6. Sceptical Theism and Its Perfectly Symmetrical Trap

The sceptical theist response deserves its own extended treatment, because it is the most philosophically sophisticated of the theodicies and because its interaction with the Evil God Challenge is particularly illuminating. Michael Bergmann and others have argued that human cognitive limitations mean we should not expect to be able to identify the goods that justify God’s permission of evil. We are, on this picture, like a child reading a complex chess manual and concluding, because no individual move seems to make sense in isolation, that no coherent strategy underlies the game. The inability to perceive the justifying reasons is not evidence that they do not exist; it is merely evidence that our perceptual and reasoning apparatus is inadequate to the task of evaluating an omniscient being’s decisions.

The argument has genuine philosophical content, and it is not simply dismissed by noting that it is convenient for the theist. The question is whether it is deployed consistently. And here is the trap: if the sceptical theist is right that our cognitive apparatus is insufficient to judge whether the world’s evil is compatible with a perfectly good creator, the same argument applies with equal force to the question of whether the world’s goodness is compatible with a perfectly evil one. An evil-god sceptical theist could argue with identical logical validity that our failure to identify the evil purposes served by love, laughter, and beauty is no evidence that such purposes do not exist. We are simply too cognitively limited to see how the world’s apparent goods serve the evil creator’s inscrutable malicious agenda. The sceptical move is perfectly reversible, which means it provides no net justification for preferring the good-god hypothesis over the evil-god hypothesis. It undermines the evidential value of the world’s features in both directions simultaneously, leaving the theist with no positive reason drawn from the character of the world to prefer their hypothesis over its mirror.

This is the heart of Law’s challenge, and it is worth stating it with full clarity. The Evil God Challenge does not require the atheist to prove that an evil God does not exist. It requires only that the theist explain why their evidence-based arguments for a good God survive the symmetry test while the corresponding arguments for an evil God do not. If both hypotheses are equally accommodated by the theodicist’s toolkit, neither hypothesis has any evidential advantage over the other, and neither has any evidential advantage over the null hypothesis, which is the hypothesis that there is no creator of any moral character at all. The burden of proof, as always, rests with the party making the positive claim, and it is not discharged by arguments that work equally well for the diametrically opposite claim.

7. The Asymmetry Objection Examined

The most common objection to the Evil God Challenge from theistic philosophers is the asymmetry objection. It holds that the world is not, in fact, evenly balanced between good and evil, and that the predominance of one over the other breaks the symmetry that Law’s challenge requires. There are two versions of this objection, and they pull in opposite directions, which is itself instructive.

The first version argues that the world contains more good than evil, that life, on balance, is worth living for most people and most creatures, that genuine happiness, love, and flourishing are more prevalent than their opposites, and that this tipping of the scales toward goodness is positive evidence for a good creator and against an evil one. This is a genuinely interesting argument, and it has some initial plausibility when considering human life in reasonably comfortable circumstances. But it faces several serious difficulties. The claim that life is predominantly good is empirically contestable once we expand our horizon beyond the lives of prosperous humans. The world’s biomass is largely composed of organisms whose existence consists primarily of predation, parasitism, disease, and violent death. The history of sentient life on earth, measured not by the pleasant evenings of prosperous modern humans but by the total experience of all sentient creatures across deep time, is a history in which suffering is not a marginal exception but a structural feature of the biological enterprise. Second, even if the net balance of human experience leans positive, this is not what we would expect from a maximally good creator who had unlimited power to arrange things differently. The question is not whether things could be worse, but whether they could be incomparably better, which they obviously could, and whether an omnipotent, all-loving creator would settle for the current arrangement when a vastly more benevolent one was available.

The second version of the asymmetry objection runs in the other direction and is sometimes deployed by people who have thought harder about the problem. It argues that the world contains so much suffering of such intensity and apparent gratuitousness that the balance actually tilts toward evil, which should on Law’s own terms be evidence against a good creator. This is essentially a concession to the evidential problem of evil dressed as an objection to the symmetry argument. The theist who makes this move is acknowledging that the world’s character is evidentially significant, which is precisely what the Evil God Challenge requires them to acknowledge. Having granted that principle, they cannot then selectively apply it.

Neither version of the asymmetry objection dissolves the challenge. The first version fails on empirical grounds and on the grounds that a maximally good God could clearly have arranged things far better. The second version concedes the central principle of the challenge while trying to use it against Law, which only works if one has already smuggled in the assumption that the world’s character should count as evidence in one direction but not the other. Both versions, examined carefully, leave the symmetry intact.

8. Natural Evil and the Particular Cruelty of Biology

The problem of natural evil deserves sustained attention because it is the domain in which the theodicies are weakest and where the Evil God Challenge lands with the most force. Natural evil refers to the suffering caused not by human choices but by the structure of the natural world: geological catastrophe, epidemic disease, predation, parasitism, and the biological machinery of pain and death that evolution has produced across hundreds of millions of years. No free will defence touches it. The child born with a neural tube defect did not suffer that fate because of anyone’s freely chosen wickedness. The antelope being disembowelled by a predator did not sin. The millions who have suffered and died from smallpox, tuberculosis, malaria, and intestinal parasites across human history were not the victims of human moral failure, but of organisms that had evolved efficient mechanisms of harm long before any theological category was available to explain them.

What makes natural evil theologically significant is not merely its scale, which is staggering, but its character. Evolution produces suffering not as a regrettable side effect of an otherwise elegant process but as a central mechanism. Pain exists because organisms that responded to tissue damage by withdrawing from its source survived better than those that did not. Fear exists because organisms that anticipated and avoided threats reproduced more successfully than those that were indifferent to them. The entire apparatus of sentient suffering is, from an evolutionary perspective, a highly refined and enormously successful adaptation. It is not a bug in the design; it is the design, and if it is the design of a creator, the design’s character says something rather direct about the creator’s character.

The standard response is to argue that evolution was God’s chosen mechanism for producing beings capable of genuine relationship with the divine, and that the suffering involved was either a necessary price or an outcome God permitted for reasons beyond our full comprehension. But this response, as we have already seen with the greater-good theodicy, is perfectly reversible. An evil creator could equally have chosen evolution as the mechanism for producing maximally suffering-capable organisms, beings complex and sensitive enough to experience genuine agony, betrayal, grief, and loss, rather than the simpler suffering of organisms lower on the ladder of sentience. The greater the complexity of the nervous system, the richer the capacity for suffering. If an evil God wanted to maximise the quality of suffering rather than merely its quantity, evolving organisms toward high sentience would be precisely the strategy to adopt. The process that produces beings capable of grief is also the process that produces beings capable of love; the same neurological structure supports both. An evil creator could accept the love as the price of the grief exactly as a good creator is asked to accept the grief as the price of the love.

There is something particularly uncomfortable about this implication when one considers the specific character of certain natural evils. The parasitic wasp that lays its eggs inside a living caterpillar, so that the larvae can eat it alive from the inside out, was famously cited by Darwin himself as making it very difficult to believe in a beneficent creator. The Ophiocordyceps fungi that infect ants and manipulate their behaviour before killing them in a precisely timed and choreographed manner, the botfly larvae that develop beneath the skin of living mammals, the nematode worms that cause river blindness by living in human eyes: these are not peripheral curiosities but structural features of the biosphere. If a good God designed them for purposes we cannot fathom, an evil God could equally have designed the world’s kindnesses as instruments of a malice we cannot perceive. The biological argument cuts both ways, with equal force, which means it resolves nothing in favour of either hypothesis.

9. The Moral Intuition Argument and Its Limits

Some theistic philosophers, recognising that the theodicies are symmetrically vulnerable to Law’s challenge, retreat to a different kind of argument altogether. They suggest that our strong moral intuitions about the world’s character, specifically the intuition that the world does not look like the product of a malicious creator, are themselves evidence that should be taken seriously and that breaks the symmetry. We do not merely fail to find an evil God plausible in some abstract logical sense; we find the hypothesis morally repellent in a way that seems to carry genuine evidential weight. The beauty of the world, the reality of love, the existence of genuine moral heroism: these things feel like testimony to something positive, and that felt sense should count for something in our epistemology.

The response to this is that the same argument cuts both ways, and in the relevant direction more forcefully. Our moral intuitions about the world’s character include our reactions to its suffering, not merely our responses to its beauty. The sight of a child dying from a preventable disease does not merely raise an intellectual puzzle; it produces in virtually every observer a reaction of moral horror that feels equally like testimony about the world’s character. The intuition that a loving God would not have arranged this is not merely a logical inference; it is an immediate and powerful moral response. If the positive intuitions generated by the world’s beauty count as evidence for a good creator, the negative intuitions generated by the world’s cruelty count as evidence against one. Both types of intuition are subject to evolutionary debunking arguments, cognitive biases, and the general limitations of human moral psychology. If we are going to invoke moral intuitions as evidence at all, we must invoke them consistently, and their consistent application does not obviously favour the good-god hypothesis over the null hypothesis of no creator.

There is also a deeper problem with the appeal to moral intuitions in this context. The theist who argues that our sense of the world’s goodness is evidence for a good creator is implicitly relying on moral faculties whose authority depends, on many theistic accounts, on God’s existence in the first place. There is a circularity here that careful theists should acknowledge. If God is the foundation of moral knowledge, then our moral intuitions about the world’s character are, at best, uncertain guides to the nature of their alleged source. If they are independent guides, as they must be if the argument is to have non-circular force, then they are operating autonomously of any theological framework, which raises urgent questions about why we need the theological framework to explain them.

10. The Challenge to Theodicy as a Project

Stepping back from the individual theodicies and their mirrors, it is worth considering what the Evil God Challenge implies about the theodicy project in general. Theodicy, as a discipline, aims to reconcile the existence of a perfectly good God with the existence of evil. The implicit methodology is to generate explanatory hypotheses about why a good God would permit the evils we observe, evaluate those hypotheses for logical consistency and evidential adequacy, and conclude that, if a sufficiently good explanation can be found, the existence of evil does not significantly undermine the good-god hypothesis. This methodology seems reasonable in isolation and has attracted genuinely first-rate philosophical minds over several centuries.

But the Evil God Challenge reveals a structural flaw in this methodology: the same standards of logical consistency and explanatory adequacy apply equally to the mirror hypotheses. If the bar for a successful theodicy is that it be logically consistent and not obviously refuted by the evidence, the evil-god reverse theodicies clear that bar as easily as the originals. The theodicy project, by its own standards, vindicates the evil-god hypothesis as thoroughly as it vindicates the good-god hypothesis. This is not a triumphant result for either position. It is a demonstration that the theodicy project, taken on its own terms, cannot distinguish between them, which means it cannot provide the evidential support for the good-god hypothesis that it is traditionally supposed to provide.

John Stuart Mill, writing long before the Evil God Challenge was formalised, recognised a related problem. In his posthumously published Three Essays on Religion, Mill argued that the evidence from the natural world for a benevolent creator was at best very limited, and that the honest assessment of nature’s character yielded a picture of something considerably more indifferent or even hostile to the welfare of sentient creatures. Mill was not making Law’s specific symmetry argument, but he was applying the same underlying principle: the character of the world should constrain our claims about the character of its creator, and the world’s character does not straightforwardly support the hypothesis of a maximally good one. Mill was willing to state plainly what the evidence suggested and what it did not, which is the intellectual honesty that the theodicy tradition has too often sacrificed in favour of maintaining a conclusion that has already been decided on other grounds.

The problem of evil has been the single most powerful argument against classical theism for as long as philosophers have examined it seriously. What the Evil God Challenge adds to the existing literature is not a new argument against theism but a new method of exposing the inadequacy of the existing defences. It is a symmetry test. It asks whether the tools of theodicy are genuinely doing explanatory work or whether they are merely providing verbal cover for a conclusion that has been reached by other means. The answer, when the test is applied systematically and without precommitment to either hypothesis, is that the tools of theodicy provide no net evidential support for the good-god hypothesis over the evil-god hypothesis, and that both hypotheses are significantly less supported by the world’s evidence than the straightforward null hypothesis that there is no morally characterised creator at all.

11. Why This Matters Beyond Philosophy Seminars

It would be tempting to treat the Evil God Challenge as a puzzle for professional philosophers, interesting within the confines of academic theology but of limited practical significance. That would be a serious mistake. The arguments that theodicies make are not confined to academic journals; they are deployed constantly in pastoral and public contexts to persuade people that their suffering is meaningful, divinely intended, or part of a plan that will eventually be revealed as good. When a cancer patient is told that God allowed their illness for a reason, when a grieving parent is assured that the death of their child serves a higher purpose, when a community devastated by natural disaster is encouraged to see the catastrophe as an opportunity for spiritual growth: these are theodicies in action. They are not merely intellectual propositions; they are responses to real human suffering, and they carry real consequences for how people process and respond to that suffering.

The Evil God Challenge has a direct implication for these pastoral deployments of theodicy. If the logic of soul-making applies symmetrically to an evil God, then the assurance that suffering serves soul-making purposes provides no actual evidence that the suffering is benevolently intended. It is, at best, logically consistent with either hypothesis. Telling a bereaved parent that their child’s death serves a good purpose is not a claim that the theodicy logic supports; it is a claim that requires independent evidence for the good-god hypothesis, evidence that the theodicy tradition, by Law’s argument, has not succeeded in providing. The intellectual honesty that the Evil God Challenge demands has real consequences for how religious consolation is framed and what epistemic commitments it asks of the people receiving it in their most vulnerable moments.

There is also a broader point about the culture of theological discourse. The problem of evil is regularly presented in popular religious apologetics as a solved problem, or at least as a problem that has been successfully defused by the available theodicies. The impression is given that philosophers of religion have developed satisfactory responses to the challenge, and that those who continue to find evil problematic for theism have simply not encountered or properly understood the literature. The Evil God Challenge is a useful corrective to this impression. It demonstrates, through an argument of genuine philosophical rigour that can be followed without specialist training, that the theodicy literature has not solved the problem. It has developed a set of moves that provide apparent cover for the good-god hypothesis while remaining perfectly available to the evil-god hypothesis. The symmetry is not evidence that the problem has been solved; it is evidence that the proposed solutions are not doing the work they are claimed to do.

Those who wish to pursue the broader context of theodicy’s failures may find the related discussions of why “God’s plan” arguments are a form of moral cowardice and of God as the silent observer of atrocity worth reading alongside this one. Together, they point toward a picture of classical theism that cannot be sustained under honest evidential scrutiny.

12. The Positive Case for the Null Hypothesis

The Evil God Challenge, as Law constructs it, is primarily a negative argument: it shows that the theodicy defences of the good-god hypothesis fail the symmetry test and therefore provide no net evidential support for that hypothesis over its mirror. But it is worth drawing out the positive implication, because it matters and because it is often elided in discussions that treat the Evil God Challenge purely as a polemical device.

If neither the good-god hypothesis nor the evil-god hypothesis is adequately supported by the available evidence, and if both hypotheses are rendered equally plausible and equally implausible by the theodicy literature, then the most epistemically responsible position is to withhold assent from both. The null hypothesis, that the universe has no creator with a moral character of any kind, is not merely the default position in the absence of positive evidence; it is the position that most naturally accounts for the world’s actual character without requiring any additional explanatory machinery that the evidence does not warrant.

A universe with no morally characterised creator is exactly what we would expect to look like if it were governed by the mindless processes that physics, chemistry, and biology describe. It would contain beauty as an emergent property of physical complexity, not as evidence of a designer’s aesthetic sensibility. It would contain suffering as a functional adaptation in nervous systems shaped by natural selection, not as evidence of either divine punishment or divine indifference. It would contain genuine moral goodness as an evolved social behaviour in highly intelligent, deeply social animals, not as evidence of divine inspiration or moral instruction from above. It would contain genuine love as a biological and psychological reality produced by the bonding mechanisms that successful social species require, not as a gift from a creator but as a feature of the creatures that natural processes happened to produce. All of these things would exist, as they clearly do, and none of them would require a moral explanation that appeals to a creator’s intentions. The null hypothesis does not impoverish the world; it describes it accurately without the added metaphysical baggage that neither the evidence confirms nor the theodicy tradition can adequately defend.

This is not a counsel of despair, and it is worth saying so plainly, because the objection that atheism drains the world of meaning is as tedious as it is unfounded. The absence of a morally characterised creator does not drain the world’s goodness of its reality or its significance. A sunset is genuinely beautiful whether or not it was arranged by a divine aesthete. Love between human beings is genuinely valuable whether or not it was installed by a creator who desired such love for its own purposes. The moral obligation to reduce suffering and increase flourishing is genuinely binding on reflective social beings whether or not it was decreed from above. As Carl Sagan put it, in words that bear quoting with their full weight: “The world is so exquisite, with so much love and moral depth, that there is no reason to deceive ourselves with pretty stories for which there’s little good evidence. Far better, it seems to me, in our vulnerability, is to look Death in the eye and to be grateful every day for the brief but magnificent opportunity that life provides.” What the null hypothesis removes is not the world’s value but the false comfort of believing that the world’s character has been adequately explained by an appeal to inscrutable divine purposes. That false comfort comes at a cost: it requires accepting arguments that, by their own standards, would equally support the proposition that an evil God made the world to suffer.

13. The Honest Answer the Theodicy Tradition Owes

The honest answer that the theodicy tradition owes, and that it has consistently failed to provide, is an account of why the standard theodicies are not symmetrically available to the evil-god hypothesis. Not an assertion that they are not, not a retreat to the inscrutable-purposes defence that undermines both hypotheses simultaneously, but a genuine argument that identifies a structural asymmetry between the good-god theodicies and their evil mirrors. Law has issued this challenge in the academic literature, and the responses that have been offered, from Mark Scott’s attempts to find asymmetries in the soul-making theodicy to various elaborations of the sceptical theist position, have not succeeded in demonstrating the required asymmetry. Each proposed asymmetry either rests on an assumption that begs the question in favour of the good-god hypothesis or applies with equal force to the evil-god mirror, leaving the symmetry intact.

The question is not whether the theodicy tradition has produced interesting and sophisticated arguments, because it clearly has. The question is whether those arguments survive the symmetry test, and the honest answer is that they do not. The free will defence survives it no better than the reverse free will defence. The soul-making theodicy survives it no better than the reverse soul-making argument. The greater-good theodicy survives it no better than the greater-evil reverse. The sceptical theist response undermines the evidential value of the world’s features in both directions simultaneously and so provides no net support to either hypothesis.

This is a significant conclusion, and it is worth stating what it does and does not entail. It does not prove that the good-god hypothesis is false, because Law’s argument does not require that conclusion. It demonstrates that the theodicy tradition has failed to establish what it set out to establish: that the existence of evil is compatible with a perfectly good creator in a way that preserves the good-god hypothesis’s evidential advantage over its competitors. The symmetry argument shows that the problem of evil is not a solved problem, not a defused problem, and not even a significantly mitigated problem. It remains exactly what it was before the theodicy tradition began its long and sophisticated attempt to neutralise it: the most powerful piece of evidence available to those who find the good-god hypothesis implausible, and one that has resisted every attempt to explain it away without simultaneously explaining away the hypothesis it was supposed to defend.

14. What Follows from the Symmetry

If the theodicies fail the symmetry test, several things follow that are worth stating plainly, because the implications extend well beyond the seminar room and into the public discourse about religion, suffering, and the intellectual honesty owed to people in distress.

First, apologists who claim that the problem of evil has been answered by the available literature are making a claim that the literature itself does not support. The existence of theodicies is not the same as the success of theodicies. The existence of sophisticated and logically consistent arguments for a position does not establish that position, particularly when equally sophisticated and equally logically consistent arguments are available for the diametrically opposite position. Sophistication and consistency are necessary conditions for a good argument, but they are not sufficient ones, and the theodicy tradition has too often substituted the former for the latter in the hope that the difference would not be noticed.

Second, those who deploy theodicies in pastoral contexts owe their audiences honesty about the epistemic status of those arguments. Telling a suffering person that their suffering serves a higher purpose is not a claim the theodicy literature establishes; it is a claim that requires independent support, and the theodicy tradition’s failure to survive the symmetry test means that independent support cannot be drawn from the standard arguments. This does not mean that religious frameworks are without comfort for those who find them helpful, or that the human need for meaning in the face of suffering is not genuine and important. It means that the comfort offered should be honest about what it is: a framework that some people find sustaining, not a conclusion secured by philosophical argument that has survived rigorous independent scrutiny.

Third, and most broadly, the Evil God Challenge is a demonstration of what rigorous philosophical argument can achieve when it is applied without precommitment to a conclusion. Law’s argument does not begin from atheism and work backward to undermine theism; it begins from the internal logic of the theodicy tradition itself and shows that the tradition’s tools are insufficient for the purpose they are designed to serve. This is the most powerful kind of critique, because it cannot be dismissed as the product of an alien methodology or a hostile starting point. It uses the theist’s own tools, accepts the theist’s own framework of evidential reasoning, and shows that the framework, applied consistently rather than selectively, does not deliver the theist’s desired conclusion.

The problem of evil has not been answered. The Evil God Challenge shows precisely why: every answer to it is available, in mirror form, to a position that every reasonable person immediately recognises as absurd. If we rightly and immediately reject an evil creator on the grounds that the world contains too much goodness to be compatible with such a being, we should apply that same standard to the good-god hypothesis with equal rigour and equal willingness to follow the argument where it leads. The world contains too much unredeemed, gratuitous, structurally embedded suffering to be straightforwardly compatible with a being of unlimited power and unlimited love. The theodicies have not changed this. They have only shown that the same style of motivated reasoning that defends a good God against the world’s evil can equally defend an evil God against the world’s good. That is not an achievement of the theodicy tradition. It is its indictment, clearly written and patiently waiting to be read.

For a related examination of the moral implications of divine silence in the face of historical atrocity, see the discussion of God as the silent observer of atrocity. For the specific failure of theodicy when applied to the doctrine of eternal punishment, the analysis at why eternal punishment fails develops many of these themes in a different but complementary direction.

The Evil God Challenge No Theodicy Survives