The Problem of Evil

An objection a student has never felt is an objection they cannot answer. What follows states the case against Christian faith as strongly as its best advocates state it, and no reply will be offered until the next section. If you find yourself reaching for a solution while you read, stop and finish the section first. Preachers who have never let the argument land produce congregations who collapse the first time it lands on them.

The oldest form is attributed to Epicurus and is preserved in a Christian source, the treatise on the anger of God by Lactantius in the early fourth century. Is God willing to prevent evil but unable? Then he is not omnipotent. Is he able but unwilling? Then he is malevolent. Is he both able and willing? Then where does evil come from? Is he neither able nor willing? Then why call him God? The Christian answer has never been that the questions are impertinent.

The modern logical version was put by the Australian philosopher J. L. Mackie in a 1955 article in the journal Mind. His claim was strong and precise: the propositions that God is omnipotent, that God is wholly good, and that evil exists are not merely hard to hold together but form an inconsistent set, so that the believer is in the position of someone holding a contradiction. He got there by adding two premises he took to be obvious, that a good thing always eliminates evil as far as it can, and that there are no limits to what an omnipotent thing can do. If both of those hold, then a good omnipotent God and any evil at all cannot both exist, and the believer is not merely unsupported but refuted.

The evidential version replaced it and is now the serious form of the argument. William Rowe, in a 1979 article, granted that there might be no formal contradiction and pressed a different case. Take, he said, a fawn trapped in a forest fire started by lightning, badly burned, lying in agony for days before dying, unseen by any human being. Rowe's argument is not that we can prove no good comes of it. It is that we have no reason to think any does, that the world is full of such cases, and that in every other area of life we take the absence of any discoverable reason as evidence, defeasible but real, that there is none. Paul Draper sharpened this a decade later by arguing that the actual distribution of pain and pleasure among sentient creatures, which tracks biological usefulness far more closely than it tracks anyone's moral deserts, is much more what you would expect if the universe were indifferent than if it were governed by a loving God. Notice that neither argument requires the atheist to prove anything. It puts the believer in the position of someone with an unexplained anomaly.

Then there is the form that is not an argument at all and is more dangerous than either. In Dostoevsky's The Brothers Karamazov, in the chapter called Rebellion, Ivan tells his believing brother a series of true stories he has collected from newspapers about the deliberate torture of small children, and then says something that no theodicy has yet answered. He is not denying God's existence. He says that if the final harmony of the world requires the unavenged tears of one tortured child as part of its price, then the harmony is not worth that price, and he most respectfully returns his ticket of admission. That is a moral refusal rather than a logical objection, and it is the form in which the problem actually reaches most people. A pastor who can refute Mackie and cannot sit in the same room as Ivan has learned the wrong half of this subject.

One distinction before we move. Moral evil is what free agents do: cruelty, betrayal, murder, systems built to grind people down. Natural evil is what happens without an agent: earthquakes, cancers, birth defects, the fawn in the fire. Several of the replies that follow work well on the first and poorly on the second, and you should watch for which is which. A third category, horrendous evil, will be defined in the last section, and it is the one that breaks the most theories.

The free will defence is the oldest Christian reply and its classical form is Augustine's: God made rational creatures with wills genuinely their own, a will that cannot turn away is not a will, and the goods that depend on freedom, love, worship, loyalty and virtue, cannot be produced by a mechanism. Evil came in by the free turning of created wills, angelic and then human, away from God towards lesser goods. God did not make it and did not need to.

Alvin Plantinga gave this argument its technical form in the 1970s, in The Nature of Necessity and the shorter God, Freedom, and Evil. His target was specifically Mackie's claim of inconsistency, and this matters for judging whether he succeeded. Mackie had said a good omnipotent God would have made free creatures who always freely choose the good. Plantinga answers that if a creature is genuinely free in the libertarian sense, then what it would freely do in a given situation is not up to God, so it is not within God's power to bring about that a free creature freely does right. He then introduces the possibility he calls transworld depravity: it might be that every possible free creature God could have created is such that, in any situation in which it is free, it would go wrong at least once. If that is so much as possible, then it is possible that God could not have created a world containing moral good without moral evil, and the alleged contradiction is gone. The argument does not claim that transworld depravity is true or even likely. Possibility is all a consistency proof needs.

By the standards it set itself the argument worked, and this is one of the few places in philosophy of religion where something approaching consensus exists. Mackie himself, in The Miracle of Theism in 1982, conceded that the free will defence shows that the theist's position is not internally inconsistent, while going on to press the evidential case instead. Rowe granted the same. The logical problem of evil is now generally regarded as answered, and a Christian student should know that and should also know exactly how little it establishes: it shows that the believer is not holding a contradiction. It offers no account of why anything actually happened.

Three costs. First, it is a defence and not an explanation, which the next-to-last section takes up. Second, it does nothing obvious about natural evil, and Plantinga's own suggestion, that natural evils might be the work of fallen angelic agents, was offered explicitly as a logical possibility sufficient for the consistency argument rather than as a claim he was asserting; quoting it as his view of earthquakes misrepresents him. Third, and most important for this module, the whole argument runs on libertarian freedom, and two of the accounts of providence you met in the last lesson deny it. A Reformed or Thomist theologian cannot use the free will defence in this form, and has to say something else, which section 3 supplies.

John Hick proposed a different theodicy in Evil and the God of Love in 1966, and he proposed it as a genuine theodicy rather than a defence, which is what makes it more ambitious and more vulnerable. He distinguished two types in the tradition. The Augustinian type looks backwards, explaining evil by a fall from an original perfection. The type he called Irenaean, after the second-century bishop, looks forwards: human beings were not created finished but immature, made in God's image and destined to grow into his likeness, and the world is not a paradise but an environment for that growth. Hick borrowed a phrase from the poet Keats and called it a vale of soul-making. On this account a world with real dangers, real consequences, real suffering and real moral stakes is precisely the kind of world in which courage, compassion, patience and self-giving can be formed, and a pain-free world would be a nursery in which no one ever became anything. Hick added that God must keep an epistemic distance, not making his presence overwhelming, because a creature coerced by the sheer obviousness of God could not freely come to love him.

The objections are serious and Hick knew it. The distribution objection is the sharpest: suffering very often destroys people rather than forming them, and the cases are not marginal. A child starved into permanent cognitive damage, a woman broken by years of abuse, a man whose dementia erases him: it is grotesque to describe these as soul-making, and Hick does not try. His answer is that the account only works if no one is finally lost, so that the growth can be completed beyond death, and he therefore commits to universal salvation, which is a substantial theological price that many who like the rest of his account will not pay. There is also the instrumentalising objection, that the account risks treating the victim's agony as raw material for someone else's virtue, which is Ivan Karamazov's complaint in philosophical dress. And animal suffering long before there were any souls to make is untouched by it. Finally, a historical caution: patristic scholars have questioned whether Irenaeus held the position Hick attributes to him, so treat the label as Hick's construction rather than as a report.

Augustine spent nine years as a Manichaean, holding that good and evil are two eternal powers and that evil is a substance. What broke the position for him was a metaphysical argument he sets out in the Confessions and again in the Enchiridion: everything that exists is made by God and is therefore good so far as it exists, so evil cannot be a thing. It is a privation, privatio boni, the absence of a good that ought to be present. Blindness is not a substance added to an eye; it is sight missing where sight belongs. Cruelty is not a positive quality; it is love and justice absent from a will that should have had them. Aquinas takes the analysis over and it remains the standard Catholic account, and it is held by most Reformed and Lutheran theologians too.

It does real work. It kills dualism, which the first lesson of this module showed is the alternative on offer, because if evil is not a thing then it is not an uncreated rival thing. It explains how God can be the cause of everything that exists without being the cause of evil, which is the move the doctrine of providence needed in the last lesson. And it accounts for the parasitic character of evil, which is genuinely observable: a lie depends on the value of truth, a betrayal depends on the good of loyalty, and no one is corrupted by anything except a good used wrongly.

The objections are not frivolous. Hick pressed the first: an account on which evil is a lack and the whole shows a beauty invisible from inside can end by saying that evil is not really so bad from God's point of view, which is false to the thing itself. A second objection is phenomenological: pain does not feel like an absence, and torture is not experienced as a gap in a good. A third is moral, and it is the one that has been pressed hardest since the twentieth century: that a metaphysics of deficiency looks thin beside industrialised murder, and that a doctrine which cannot name Auschwitz as a positive horror has failed. The defenders' reply, argued carefully by Brian Davies among others, is that privation is a claim about what evil is metaphysically and not a measurement of how bad it is, that privations can be the worst things there are, since death is a privation and so is the loss of a child, and that nobody has produced a coherent account of evil as a positive substance that does not end in dualism. Whether that reply is sufficient is a live question, and it is worth noticing that the argument is between people who agree entirely about how bad the evils are.

Then permission, where the Catholic and Reformed accounts genuinely differ and both deserve their own statement. The Catholic account, stated in the Catechism at paragraphs 309 to 314 and resting on Aquinas, is that God permits evil without in any way causing it, and permits it only because his power and goodness are such that he can draw good out of it. The Catechism is emphatic at 311: 'God is in no way, directly or indirectly, the cause of moral evil.' It then points to the case that proves the principle, the killing of Christ, the greatest moral evil ever committed, out of which came the greatest good. The Easter liturgy goes further still with the phrase felix culpa, the happy fault of Adam that merited so great a redeemer, which the Catechism quotes at 412 while insisting that this is a reason for awe and not a justification of sin. The hardest question put to this account is whether permission is a real category for a God who could have prevented the act, knew what would follow, and did not: a bystander who could stop a murder without cost and does not is not excused by saying he permitted rather than committed it. The standard answer is that the analogy fails because God is not a bystander among others but the author of the whole, that preventing all evil would mean preventing the free creatures who commit it, and that the good drawn out is not a compensation added afterwards but the reason the permission was not wrong. The sentence the Catechism reaches for at this point is Augustine's, from the Enchiridion, which it quotes at 311: God, being supremely good, would never allow any evil whatever to exist in his works unless he were powerful and good enough to bring good out of the evil itself. Aquinas states the same maxim, and the Catechism quotes him too at 412. On this account permission is never bare and never final. Notice also what the account refuses to say. It does not say that God wills the evil act under any description, and paragraph 311 is deliberately absolute about that, which is exactly where it parts company with the next paragraph.

The Reformed account refuses to rest on permission in that sense, and refuses deliberately. The Westminster Confession says at 5.4 that God's providence extends to the first fall and to all other sins of angels and men, 'and that not by a bare permission, but such as hath joined with it a most wise and powerful bounding, and otherwise ordering, and governing of them, in a manifold dispensation, to his own holy ends', and then in the same sentence that 'the sinfulness thereof proceedeth only from the creature, and not from God, who, being most holy and righteous, neither is nor can be the author or approver of sin.' Calvin argued the same way, holding that a merely permissive God is a God standing idle while the world runs, which is neither the God of Genesis 50:20 nor of Acts 4:28. The positive case is textual: if God's hand and will had decided beforehand what Herod and Pilate did (Acts 4:28), then something stronger than permission is being described, and the same chapter treats this as a ground for prayer rather than a scandal. The hardest question put to this account is the obvious one, and the confession raises it itself by inserting the guard in the same sentence: if God ordains the act, in what sense is he not the author of the sinfulness? The answer given is that the decree ordains the event while the sinfulness arises wholly from the creature's own motive, and that God's end in ordaining and the agent's end in acting are different, which is precisely the structure Genesis 50:20 asserts when it says that what the brothers intended for evil, God intended for good. Critics reply that the distinction between ordaining an act and ordaining its sinfulness cannot be made to stand up, and the debate has run for four centuries without resolution.

A third voice belongs here because it is the one most often missing from Western textbooks. Eastern Orthodoxy has produced relatively little systematic theodicy, and that is a position rather than an omission: the East has tended to treat evil as an intruder to be healed and defeated rather than a problem to be justified, and to locate the answer in the incarnation, the harrowing of hell and the final restoration rather than in an explanation. David Bentley Hart argued this forcefully in The Doors of the Sea, written after the Indian Ocean tsunami of 2004, insisting that the New Testament presents this present age as under the dominion of powers God is at war with, that Christians are not required to see each particular horror as a direct instrument of divine purpose, and that the God revealed in Christ is one who weeps at the tomb of Lazarus rather than one who explains it. The Reformed reply is direct and fair: what, then, of Acts 4:28, and of the two intentions in Genesis 50:20, which say more than that God defeats evil afterwards. Hart's reply is that the cross is precisely where God's purpose and the world's malice coincide, and that a single unique case does not license a general theory. It is worth adding that the Orthodox instinct here is liturgical before it is theoretical: the funeral service laments death as an enemy rather than explaining it, and the Paschal proclamation that Christ has trampled down death by death is offered as the answer, which is an answer of a different kind from the ones this section has been weighing. Those are the real positions, and a student who can state all three of them has more than most graduates.

Rowe's evidential argument has a hinge, and skeptical theism attacks the hinge. The argument moves from we can see no justifying reason for this suffering to there probably is none. Stephen Wykstra, in a 1984 article, argued that this inference is only as good as our grounds for thinking that if there were such a reason we would likely be able to see it. He formulated a principle for when an appearance counts as evidence, which he called CORNEA, the condition of reasonable epistemic access, and argued that the gap between God's understanding and ours, if there is a God, is large enough that our failure to spot a reason is about what we should expect either way. William Alston pressed the same point in 1991 with a survey of how badly placed we are to make such judgements, and Michael Bergmann has developed it since. The comparison usually offered is a novice looking at a chess position between grandmasters and seeing no reason for a move.

It is a strong reply and it has a bill attached. If our grasp of the goods and evils at stake is that unreliable, several other things wobble. We normally think we can tell that a particular suffering is gratuitous, and the position says we cannot. We normally think we can judge that a state of affairs is bad enough that we ought to intervene, and if the same skepticism applies, our reasons for intervening are undercut too, which is the objection pressed by Bruce Russell and by Michael Almeida and Graham Oppy in a well-known exchange. Some have added a worry about divine deception: if God may have reasons beyond our grasp for permitting suffering, he may have reasons beyond our grasp for allowing us to be deceived, and the reliability of our beliefs about him is in question. Skeptical theists answer that the skepticism is deliberately narrow, restricted to claims about the total set of goods and evils and their entailments, and that it does not touch ordinary moral knowledge or our duties. Whether it can be kept that narrow is the argument.

Now a distinction that will stop you making a common mistake in public. Plantinga drew it and it has held. A defence attempts to show only that the existence of God and the existence of evil are compatible, by describing a possible state of affairs in which both hold; it does not claim that the described state of affairs is actual and it does not claim to know God's reasons. A theodicy is more ambitious: it claims to give the reasons, or at least the kind of reasons, for which God actually permits evil. The word itself was coined by Leibniz for his book of 1710, from the Greek for God and justice, and it means literally the justification of God.

Almost all the serious philosophical work of the last fifty years has been defence rather than theodicy, and almost all popular Christian speech about suffering is theodicy without knowing it. When someone at a funeral says that God had a purpose in this, they are not offering a consistency proof; they are claiming to know what God was doing, which is a claim about a particular death that nobody has the standing to make. The distinction gives you something honest and usable to say: you can tell a grieving person, truthfully, that there is no contradiction in trusting a good God while what has happened is genuinely evil, and you can decline, truthfully, to say what the reason was. That is not evasion. It is the difference between the two projects, and Job's friends did not know it.

Add up the ledger honestly. The logical problem is answered; almost everyone in the field agrees. The evidential problem is not answered, it is contested, and the contest is genuine rather than a stalemate anyone can declare in their own favour. The free will defence handles moral evil and leaves natural evil largely untouched. Soul making explains some suffering and is grotesque applied to the rest. Privation tells you what evil is and not why there is any of it here. Permission, in either of its forms, names the relation and leaves the question of why this permission in this case. Skeptical theism blocks the inference to gratuitous evil at the cost of admitting how little we can judge. Not one of them tells a particular person why the particular thing happened to them, and no theory in Christian history has ever done so.

Marilyn McCord Adams pressed exactly this point in Horrendous Evils and the Goodness of God in 1999, and her argument is the most important recent development in the subject. She defines a horrendous evil as one whose participation gives the person prima facie reason to doubt whether their life could be a great good to them on the whole: the paradigm cases are things like the murder of one's own child, betrayal by the person who was supposed to protect you, the corruption of a person into an agent of atrocity. Her claim is that generic theodicies miss the target entirely, because even a successful global explanation of why a world contains evils would do nothing for the individual whose life has been ruined by one. What is needed is not a reason but a defeat, and the defeat has to happen within the individual's own life: God must be good to that person, and Adams argues that the resources for this are Christian ones, the incarnation and the cross, where God himself participates in horror from the inside, so that the sufferer's experience is integrated into an intimacy with God rather than compensated for by something else. You may find that persuasive or not. Notice what it is doing: it moves the question from why to what God does about it, and from the general to the particular person, which is where the question is actually asked.

And then Job, which is the datum that governs the whole subject. The book takes forty chapters. It gives the reader the scene in heaven and then withholds it from Job permanently. It has three friends deliver theodicies of increasing confidence, many of whose individual sentences would pass any doctrinal examination. It lets Job say things about God that would get a man disciplined in most congregations. God then speaks from the whirlwind for four chapters and does not explain anything. He asks, 'Where were you when I laid the foundations of the earth?', and then walks Job through wild goats and ostriches and rain falling on land where no one lives, and asks in 40:8 whether Job will condemn him in order to be in the right himself. Job's response in 42:3 is that he had spoken of things he did not understand, things too wonderful for him to know. He is not given a reason; he is given God. And in 42:7 the explainers, not the complainer, are the ones God is angry with, and Job is told to pray for them. If the inspired book about this question ends without an explanation and rebukes the men who supplied one, a teacher who supplies one has not merely overreached. He has taken the part the book assigns to Eliphaz.

So how do you teach this to someone in the middle of it, which is where most of this school's students will have to do it, rather than in a classroom? Four things, in this order. First, do not explain. Not because explanation is forbidden in principle, but because you do not have the information, and the person can tell. Every sentence beginning God is teaching you or everything happens for a reason or he needed another angel is a claim you cannot back, and the last of those is not Christian doctrine in any tradition. Second, give them the Bible's own words for what they are feeling, because the Bible has more of them than your tradition probably uses. Roughly a third of the Psalms are laments. Psalm 88 ends in darkness with no turn. Psalm 22, which Jesus quoted from the cross, opens by asking God why he has abandoned the speaker. Habakkuk asks how long. Permission to say those things out loud, in church, is one of the most useful things you can give a sufferer, and many congregations have quietly withdrawn it.

Third, say the small true things rather than the large uncertain ones. That God is not absent from this. That he is not indifferent to it, and Jesus wept at a grave he was about to open, which means the tears were not a failure of faith. That the Son has been in the place of forsakenness himself and did not stay there. That the world is not as it will be, since Romans 8:22 says the whole creation is groaning as in childbirth and Revelation 21:4 promises that he will wipe away every tear. And that they are not going to be left alone in this, which is a promise you can actually keep. Fourth, stay. Nicholas Wolterstorff, writing after his son was killed in a climbing accident, said in Lament for a Son that what he wanted was not someone to explain but someone to sit beside him on the mourning bench. Eleonore Stump's Wandering in Darkness makes the philosophical case for the same instinct, that what is offered to sufferers in scripture comes in narratives and in the presence of persons rather than in propositions. Job's friends, remember, did one thing right: they sat with him on the ground for seven days and said nothing. The trouble started when they opened their mouths.

Last, the reason this lesson is in a seminary course at all. Teach the arguments to people while they are well. The time to learn that the logical problem is answered, that a defence is not a theodicy, that Psalm 88 has no happy ending, and that God rebuked the men who explained Job's suffering, is years before the phone call comes. A person who has that in place will not have to construct a doctrine of God at three in the morning in a hospital corridor. And when it is your turn, and it will be, you will find that what holds is not the argument you found most convincing but the God the argument was about, which is precisely what happened to Job.