You are viewing a free preview of this lesson.
Subscribe to unlock all 10 lessons in this course and every other course on LearningBro.
Where the logical problem (Lesson 1) claims that God and evil are strictly logically incompatible, the evidential problem makes a more modest but, in the view of many philosophers, more dangerous claim: the sheer amount, distribution and apparent pointlessness of much suffering makes the existence of an omnipotent, omniscient and wholly good God highly improbable. This is an inductive argument — it reasons from observed evidence to a probable conclusion, rather than claiming to prove a contradiction. Because it survives the free will defence that defeats the logical problem, the evidential problem is widely regarded as the most formidable contemporary objection to theism. For AQA A-Level Religious Studies (Component 1A), the central figure is William Rowe, and the central reply is sceptical theism.
Most philosophers now accept that Plantinga's free will defence shows there is no demonstrable contradiction between God and evil (Lesson 5). Rowe himself granted this. His strategy was therefore to concede the logical point and press a probabilistic one: even if God's existence is logically possible alongside evil, the specific character of the evil we actually observe gives us good reason to think God probably does not exist. The argument does not need certainty; it needs only to make atheism more reasonable than theism in light of the evidence.
The contrast is best grasped through the standard of success each argument sets itself. The logical problem aims at a proof: it tries to derive a contradiction, and so can be defeated by exhibiting a single coherent possibility (which is what the free will defence does). The evidential problem aims only at rational support: it offers evil as evidence that tells against theism, in the way that a smoking gun and a motive tell against a suspect without strictly proving guilt. Because of this, the evidential problem cannot be defeated merely by describing a logically possible scenario in which God and the evil coexist; the theist must show that such a scenario is not improbable, or must undercut our ability to assess the probabilities at all. This is why sceptical theism, rather than the free will defence, is the natural reply to Rowe: the issue is no longer possibility but reasonable expectation.
It is important to be precise about what kind of inductive argument this is. Rowe is not arguing by simple enumeration ("many evils seem pointless, therefore all evil is pointless"). His structure is subtler. He selects a single, carefully chosen instance (the fawn) about which the claim "this is gratuitous" is maximally plausible, and points out that one genuine instance of gratuitous evil is logically sufficient to refute theism (given Premise 2). The induction is concentrated, not diffuse: it asks us to judge of a particular case that no greater good plausibly justifies it, and then notes that the world contains countless such cases. The strength of the argument therefore depends not on the number of apparently pointless evils but on whether our judgement about even the clearest single case is trustworthy — which is exactly the point sceptical theism contests.
Rowe later refined his presentation by distinguishing two cases of seemingly gratuitous suffering: "Bambi", the fawn dying unseen in the forest fire (natural evil, no human involvement), and "Sue", a case of a child subjected to appalling abuse and murder (moral evil of the most extreme kind). The pairing is deliberate: Bambi removes any appeal to human free will, while Sue removes any suggestion that the suffering is trivial or that the victim's own moral development could be at stake. Between them, the two cases are meant to close off the standard theodicies — the free will defence (which cannot touch Bambi) and the soul-making theodicy (which is obscene when applied to Sue, as Phillips argues in Lesson 4). The evidential problem is thus tailored to survive precisely the responses that answer the logical problem.
Key term: Gratuitous evil is suffering that an omnipotent, omniscient being could have prevented without thereby losing a greater good or permitting an evil equally bad or worse. The whole evidential argument turns on whether any gratuitous evil actually exists.
William Rowe (1931–2015) gave the canonical version of the evidential argument in his 1979 paper "The Problem of Evil and Some Varieties of Atheism". He set it out as a valid syllogism whose second premise is a piece of moral common sense and whose first premise is an evidence-based generalisation:
Premise 2 is almost undeniable: it simply states what perfect goodness combined with perfect power would imply. The argument therefore stands or falls on Premise 1 — the claim that at least one instance of gratuitous evil genuinely exists. Rowe is careful: he does not claim to prove Premise 1 with certainty. He claims it is reasonable to believe on the evidence. If even one instance of truly pointless suffering exists, the wholly good and omnipotent God of theism does not.
To make Premise 1 vivid, Rowe offered an example designed to strip away every theistic consolation:
In some distant forest, lightning strikes a dead tree and starts a fire. A fawn is trapped, badly burned, and lies in terrible agony for several days before death relieves its suffering. No human observes it; no lesson is learned; no character is built; no soul is made.
The example is carefully constructed. The suffering is natural, so it cannot be blamed on human free will. It is unobserved, so it cannot serve as a warning or build human compassion. It is an animal, so it cannot be soul-making for the sufferer in Hick's sense. Rowe's point is that, however hard we try, we can identify no greater good for which this particular suffering seems necessary. It looks, as far as we can tell, simply pointless. And if it is pointless, Premise 1 is true.
Rowe's reasoning rests on an inference philosophers nickname the "noseeum" inference (from "no-see-um"): if, after careful reflection, we can see no justifying reason for an instance of suffering, it is reasonable to conclude that there probably is none. This applies a familiar epistemic principle — in the right conditions, the absence of evidence is evidence of absence. If you search a room thoroughly and see no elephant, you are entitled to conclude there is probably no elephant. Likewise, having canvassed every greater good a theist might propose and found none that fits the fawn's case, we are entitled to conclude there probably is no such good.
The most powerful theistic response is sceptical theism, which attacks the noseeum inference at its root. Its core claim is epistemic: we are simply not in a position to tell whether any given evil is genuinely gratuitous, because our grasp of the realm of possible goods, evils and their necessary connections is radically limited. From "I cannot see a justifying reason" it does not follow that "there is no justifying reason", unless we have grounds to think that, if such a reason existed, we would probably see it.
Rowe and others argue that sceptical theism, taken seriously, proves too much. If our moral discernment is so unreliable that we cannot tell whether the fawn's agony is pointless, then for all we know any atrocity — even the Holocaust — might serve some hidden divine good. That conclusion threatens to paralyse moral reasoning altogether: it would undermine our confidence that we should prevent suffering when we can, since perhaps God is permitting it for reasons we cannot fathom. A position that corrodes ordinary moral judgement and the practice of petitionary or protesting prayer pays, Rowe suggests, too high a price to count as a satisfactory answer.
Rowe also distinguished three attitudes an atheist might take. He called himself a "friendly atheist": he held that the evidence makes God's existence improbable, yet he accepted that a thoughtful theist could be rational in continuing to believe. This is contrasted with "unfriendly (hostile) atheism", which judges theistic belief outright irrational, and "indifferent atheism", which takes no view on the theist's rationality. The friendliness matters philosophically: it signals that the evidential argument trades in probabilities, not proofs.
The evidential approach has a distinguished ancestor in David Hume's Dialogues Concerning Natural Religion (1779). In Parts X and XI, the sceptical character Philo presses the problem of evil against Cleanthes, the defender of the design argument. Philo's strategy is precisely evidential rather than logical. He is willing to grant, for the sake of argument, that the world's mixture of good and evil is not strictly incompatible with a benevolent deity; his point is that the world, observed without prior assumptions, gives us no good reason to infer such a deity. As Philo puts it, the question is whether, looking at the phenomena, we can conclude that the cause of the universe is wholly good. He argues we cannot: the data are at best ambiguous and at worst suggest indifference.
Philo dramatises this with what he calls the "four circumstances" that aggravate the world's suffering — features of nature, such as the role of pain (rather than mere loss of pleasure) in motivating creatures, the rigid operation of general laws, and the niggardly way the world is provisioned, none of which seem necessary for a good God to have included. Even if each could be explained, Philo argues, their combination tells against benevolent design. Crucially, he draws the famous conclusion that, confronted with the world's mixed character, the most reasonable inference is not that its author is good, nor evil, but morally indifferent — a conclusion Paul Draper would later make precise as the "Hypothesis of Indifference". Hume thus stands as the bridge between Epicurus's ancient dilemma and the modern evidential argument: he reframes the problem as a challenge to the inference from world to benevolent God, which is exactly the form Rowe and Draper give it.
Paul Draper (b. 1957) gave the argument a more rigorous, broadly Bayesian shape in a 1989 paper. Rather than focusing on isolated horrors, Draper compares two hypotheses against the overall pattern of pleasure and pain in the world:
Draper argues that the biological facts — that pleasure and pain are largely keyed to survival rather than to moral desert, that creatures suffer in ways unrelated to any soul-making, that suffering pervaded the animal world for hundreds of millions of years before humans existed — are far more probable on HI than on T. He does not claim to disprove God; he claims that the data give us a strong prima facie reason to prefer indifference to benevolent design. This is the evidential argument at its most disciplined: a comparison of explanatory probabilities.
Sceptical theism is the most distinctively evidential reply, but theists also deploy positive theodicies against Rowe, arguing that the apparently gratuitous evils do, after all, serve goods we can partly identify. Three are worth noting.
Subscribe to continue reading
Get full access to this lesson and all 10 lessons in this course.