Theodicies and Defences

The previous article ended with two arguments and a warning. The logical version claims theism and evil are inconsistent, and a bare possibility refutes it. The evidential version claims the pattern of suffering is strong evidence against theism, and a bare possibility does nothing against it.

Read the responses below with that distinction in hand. Several of them succeed completely against one version and are close to irrelevant against the other, and the literature is easier to follow once you stop expecting any single answer to do both jobs.

The other kind of response

The word is Leibniz's, assembled from the Greek for god and justice, and the ambition it names is much larger than the one a defence carries. A defence describes a way things could be. A theodicy claims to say how they are.

The difference in ambition produces a difference in vulnerability. A defence cannot be refuted by showing that its proposal is improbable, because it never claimed the proposal was probable. A theodicy can be refuted that way, and can also be refuted by showing that the reason it identifies fails to cover some particular case. It has to hold for every instance, since the commitment it defends is a commitment about every instance.

Not everyone thinks the more ambitious project should be attempted at all, and some of the people who think that are theists. The advanced tier takes up the position that constructing justifications for particular sufferings is itself objectionable.

Evil as a lack

The view is Neoplatonic in origin. Plotinus held that evil is not a substance but a deficiency, and Augustine took it over as the answer to the question that had held him among the Manichees, who held that good and evil were two competing powers. Augustine's discussion runs through the seventh book of the Confessions and the Enchiridion.

What it accomplishes is narrower than it is usually credited with. If evil is a privation, then God did not create it, because there was nothing there to create. That removes one specific charge: that a good creator made a bad thing. It also fits a metaphysics on which everything that exists is good insofar as it exists, which the tradition wanted for independent reasons.

What it does not accomplish is anything about the problem as stated in the previous article. The problem is not that evil is a substance. The problem is that a fawn burns to death and no good comes of it, and describing the fawn's condition as a privation of health leaves the fawn in exactly the same condition. A creator who allows a great deal of privation is in the same position as one who allows a great deal of positive badness, because on this account the thing we objected to was never the metaphysical category.

The privation view also strains at the point where the pain is most acute. A toothache is not naturally described as a lack of anything. It is a presence, and describing it as the absence of proper functioning captures the cause without capturing what is bad about it. The view survives in Thomist writing, where it does real work in an account of the will, and it has largely stopped being offered as a response to the argument from suffering.

The free will defence

The name says defence, and it is used as one. Alvin Plantinga developed the version that settled the logical problem, in chapter nine of The Nature of Necessity, Clarendon Press (1974), with a less technical presentation the same year in God, Freedom, and Evil.

Recall Mackie's objection from the previous article. If a person can freely do the right thing on any given occasion, there is no contradiction in a person freely doing the right thing on every occasion, so an omnipotent being could have created agents who always freely choose well.

Plantinga's answer turns on a distinction Mackie's argument does not observe. Grant that for each free agent there is a possible world in which that agent always chooses well. It does not follow that God could bring about such a world, because whether an agent freely does something is not up to God. God can create the agent and the circumstances. What the agent then does with them is precisely what freedom means God does not settle.

So the question is not what worlds are possible, but which of the possible worlds God is able to actualise. And that depends on facts about what each agent would freely do in each circumstance, facts God knows but does not determine. Whether such facts exist at all, and what determines them, is the subject of a substantial dispute that the advanced tier takes up.

The defence needs only this: it is possible that every agent God could have created suffers from transworld depravity. If that is possible, then it is possible that God could not actualise a world containing free agents and no moral evil, and Mackie's claim of inconsistency fails.

Note carefully what is being claimed, because this is where the argument is most often misreported. Plantinga does not assert that every agent is transworld depraved. He does not think it is likely. He needs only that it is not impossible, because his opponent's claim was that theism and evil cannot both be true, and a single coherent possibility defeats a claim of that form.

Natural evil is the obvious gap, and the defence addresses it in the same key. Suffering not chosen by any human might be chosen by some non-human free agent, and for the purposes of a defence that possibility does not need to be believed, only to be coherent. The proposal strikes most readers as desperate, and Plantinga's reply is that it is not offered as an account of what happened. It is offered as a way things could be, which is all a defence trades in.

The result is generally accepted. What is generally accepted is also narrow: the logical problem is answered, and the evidential problem is untouched, since the reply's whole strategy is to trade in bare possibilities, and bare possibilities do not lower probabilities.

The defence also has a cost in the other direction that its users rarely count. If the value of freedom is what justifies permitting moral evil, then freedom has to be worth what is actually done with it, and the price includes every atrocity in the historical record. Some philosophers accept that. It is not obvious that a good being would.

Goods that need evils

Almost every theodicy has this structure, and stating it in the abstract makes the shared weak points visible.

The first is that the good has to be worth the price, judged from a standpoint that includes the person who paid it. Courage requires danger, and compassion requires suffering to respond to, and it is a real question whether a world of courage and compassion is better than a world where neither was needed. But the person who supplies the danger is not the same person as the one who develops the courage, and an account on which one person's agony is justified by another person's moral improvement has to answer for that.

The second is that omnipotence puts pressure on the word "requires". If God is unlimited, why does this good require that evil? A being who can do anything logically possible cannot say the link was a matter of convenience. It has to be a genuine necessity, and for most proposed pairings it is not obvious that it is.

The third is quantity. Even where the link holds, the argument has to account for the amount. If some suffering is needed for compassion, that does not explain the amount there is, and the evidential argument was always about the amount rather than the existence.

The soul-making theodicy

John Hick set out the modern version in Evil and the God of Love (1966). He labelled it Irenaean, after the second-century bishop of Lyon, and contrasted it with the Augustinian tradition on which the world was made perfect and then broken by a fall. Hick traced the alternative type through Irenaeus, Origen, and Schleiermacher, and the honest description is that he constructed a type rather than reporting a doctrine any of them held in his form.

The phrase is Keats's. Writing to his brother George and his sister-in-law Georgiana in 1819, Keats proposed calling the world a vale of soul-making, on the ground that a world of pains and troubles is what schools an intelligence into a soul.

The core claim is that we were not made finished. We were made capable of becoming something, and the process requires an environment with real stakes, real consequences, and real difficulty. A world engineered so that no action ever caused harm would be a world in which no action mattered, and the beings in it would remain what they started as.

The theodicy has genuine explanatory reach, which is why it survives. It gives an account of why the world contains obstacles rather than a permanent easy provision, and it accommodates natural evil directly, which the free will defence does only by a manoeuvre nobody finds comfortable.

Three difficulties are standard.

Distribution is the worst. If suffering is for growth, it is allocated with no discernible relation to that purpose. Some people are given exactly the amount that shapes them, and some are destroyed by theirs, and some receive almost none. A curriculum that kills a substantial fraction of its students, at random, is not defensible by reference to what the survivors learned.

Animal suffering is untouched. Whatever the fawn's death accomplished, it did not build the fawn's character, and hundreds of millions of years of animal pain preceded any creature capable of moral development at all.

And the deaths of small children fall entirely outside it. There is no soul-making in a life that ends before the capacity for it develops.

Hick saw the shape of this and gave the consistent answer, which is worth stating because it shows what the theodicy costs. If the process fails as often as it succeeds, the suffering is not justified by it, so the process cannot be allowed to fail. Hick therefore held that it continues beyond death and is completed for everyone, and accepted universal salvation as a requirement of his account rather than an optional extra.

That is intellectually honest and it is expensive. The theodicy now depends on a substantial eschatological claim that most of the traditions it was meant to defend do not accept, and it purchases its answer to the distribution objection by moving the accounting to a place where no one can check it.

Skeptical theism

This is not a theodicy and does not try to be. It offers no reason at all. It attacks the inference at the centre of the evidential argument, which is the step from "we can see no justifying good" to "there is none".

Stephen Wykstra gave the first influential statement in "The Humean Obstacle to Evidential Arguments from Suffering: On Avoiding the Evils of 'Appearance'", International Journal for Philosophy of Religion 16, no. 2 (1984), pp. 73-93, where the governing principle is that appearances only license a belief when, had things been otherwise, they would likely have appeared otherwise. His claim is that this condition fails here. If there were a justifying good beyond our ken, things would look to us exactly as they do.

Michael Bergmann, in "Skeptical Theism and Rowe's New Evidential Argument from Evil", Noûs 35 (2001), pp. 278-296, formulated the position as a set of theses about representativeness: we have no good reason to think the goods we know of are representative of the goods there are, nor the evils, nor the entailment relations between them.

The analogy usually offered is the difference between a novice and an expert. Seeing no reason for a move in a game you barely understand is not evidence that there is no reason. The gap between us and an omniscient being is larger than the gap between a novice and a grandmaster by an amount that cannot be stated.

Rowe's counter, from the previous article, is that this inference pattern is the ordinary one and that suspending it selectively is not a coherent policy. Three developments of that counter are the live objections.

The first is that the scepticism does not stay put. If we cannot judge whether the goods we know of are representative, we cannot judge that any action's likely consequences are good on the whole, and ordinary moral deliberation depends on exactly that judgement. Preventing a child's suffering might, for all we know, forestall some greater good. Skeptical theists deny that their theses have this consequence, and the argument over whether the denial works is not settled.

The second is that it generalises past morality. Divine reasons beyond our grasp could as easily justify permitting us to be systematically deceived, and the ordinary reply that a good God would not deceive us is unavailable to someone who has just said we cannot assess what a good God would have reason to do.

The third is the sharpest for the theist. The position undercuts revelation and religious testimony, which require confidence that God has no reason beyond our grasp for allowing the record to be misleading. A defence of theism that leaves you unable to trust what you have been told about God is not a defence anyone wanted.

The fair summary is that skeptical theism is the strongest response to the evidential argument and that its costs are borne in areas its users would rather not spend. It also, by design, leaves the theist with nothing positive to say. It does not explain any suffering. It says only that we are not in a position to conclude that no explanation exists.

Next

The problem of evil is a tension between attributes and the world. The next article turns to a tension inside the attributes themselves, and asks whether the properties classical theism ascribes to God can all be held at once even before any facts about the world are brought in.