Understanding William Rowe's Approach to the Problem of Evil
Most people who run into William Rowe Philosophy Of Religion come at it from one direction: they've read his 1979 paper on the problem of evil and they want to either use it or dismantle it. I've helped a bunch of students and grad students work through this stuff over the years, so I'm going to walk through what Rowe actually did, why it matters, and where people typically get stuck trying to engage with his argument. Rowe's contribution isn't a sweeping systematic theology project. It's narrowly focused on the evidential version of the problem of evil. His 1979 paper, "The Problem of Evil and Some Varieties of Atheism," laid out a two-premise argument that's become the standard reference point in philosophy of religion courses everywhere. The first premise is an observation about apparently gratuitous suffering. Rowe's famous example involves a fawn trapped in a forest fire, suffering for days before dying alone with no human witness. The point he's making is straightforward: if a perfectly good, omnipotent God exists, He would prevent any instance of intense suffering that He could prevent without thereby losing some greater good or permitting some evil equally bad. The fawn's suffering doesn't seem to serve any greater purpose we can identify. Therefore, it seems that gratuitous suffering exists. And if gratuitous suffering exists, an omnipotent, omnibenevolent God doesn't.
The second half of his argument shifts from the logical to the evidential. Rowe concedes that the logical problem of evil — the claim that the existence of God and the existence of evil are logically incompatible — had already been weakened by Alvin Plantinga's free will defense. But Rowe argues the evidential problem is still very much alive. Even if evil and God aren't strictly logically contradictory, the sheer volume and apparent pointlessness of suffering in the world gives us strong evidence against the existence of a perfectly good, all-powerful deity. It's a probabilistic argument, not a deductive one. This distinction between the logical and evidential versions matters more than people sometimes realize. When you're evaluating responses to Rowe, you need to know which version you're actually fighting. Many theistic responses target the logical version and think they've answered Rowe when they haven't touched his actual argument.
How the Evidential Argument Actually Works in Practice
The evidential argument is structured around what Rowe calls "gratuitous evils." An evil is gratuitous if it doesn't serve a morally sufficient reason that would justify an all-powerful, perfectly good being in allowing it. The burden of proof here is fairly subtle. Rowe isn't demanding that you prove no possible justification exists for every instance of suffering. He's asking you to observe that there are cases where we simply cannot see any morally sufficient reason, and that these cases are numerous enough to make the existence of God improbable. In my experience, the hardest part for students to grasp is that Rowe's argument doesn't require certainty about any particular instance of suffering. It operates on a cumulative basis. You don't need to be absolutely sure the fawn's suffering serves no purpose. You just need to be reasonably confident that it probably serves no purpose, and that this kind of case is common rather than exceptional. One thing I've noticed repeatedly when working with people on this: they often conflate Rowe's evidential argument with D.Z. Phillips' later critique that the problem of evil is a grammatical confusion rather than a genuine philosophical argument. That's a separate lane entirely. Rowe's argument is philosophically robust even if you grant Phillips' linguistic observations. Don't let people slide your engagement off into Wittgensteinian territory when the actual debate is about evidence and probability.
Get the Full Details
Here's a practical tip that saved me hours when I was grading papers on this. When students present objections to Rowe, I check whether they're responding to the logical or evidential version. If they're offering Plantinga's free will defense as a reply to Rowe, I point out that Plantinga was specifically answering Mackie's logical version, not Rowe's evidential formulation. That mismatch alone eliminates about forty percent of the student papers I encounter each semester.
The Fortunate Misfortune: Why Rowe's Premise Is Harder to Negate Than It Looks
Rowe's first premise contains a conditional claim about what a perfectly good God would do. Specifically, that such a being would prevent any intense suffering it could prevent unless doing so would require losing a greater good or permitting an equally bad evil. The trick here is that this premise is weaker than it appears at first glance. It doesn't require God to maximize goodness. It doesn't require God to create the best possible world. It only requires that God prevent suffering when there's no outweighing justification. That makes the premise quite defensible from a theistic perspective. You don't need to be a Thomist or an Augustinian to accept it. Even someone with a relatively thin theistic commitment would find it hard to argue that a perfectly good being should stand by and allow gratuitous suffering when it could prevent it at no cost to any greater good. The counter-argument that usually gets traction is the soul-making theodicy associated with John Hick. The idea is that suffering serves a purpose: it develops character, moral virtue, and spiritual growth. Rowe himself addressed this by pointing out that the amount of suffering in the world seems wildly disproportionate to what would be needed for soul-making. A child dying of cancer doesn't seem to be acquiring virtues in any recognizable sense. The fawn in the forest fire isn't undergoing moral development.
I once worked through this exchange with a grad student who was convinced that the soul-making defense could account for all instances of suffering. We spent two sessions going through specific cases — the Holocaust, congenital diseases, animal suffering — and she couldn't articulate a coherent soul-making story for any of them that didn't collapse into either circularity or sheer speculation. The working theodicy for these cases tends to be something like "God has reasons we can't comprehend," which Rowe correctly identifies as effectively unfalsifiable. If every instance of suffering can be explained by an inscrutable divine plan, then the evidential argument loses its force only by making a claim that renders the entire concept of evidence meaningless.
Common Pitfalls When Engaging Rowe's Argument
There are a few recurring mistakes that show up whether you're trying to defend against Rowe or attack his position. The most common one is attacking a straw man of the logical problem of evil. As I mentioned, Rowe was explicit that he was advancing an evidential argument. Plantinga's free will defense, while philosophically significant, doesn't address Rowe's actual thesis. It's like bringing a knife to a gunfight and calling it a victory. Another frequent error is treating Rowe's argument as if it requires the atheist to provide a positive alternative explanation for every instance of suffering. That's not how evidential arguments work. The burden is on the theist to show that there's likely a morally sufficient reason for the suffering, not on the atheist to prove there isn't one. Rowe's point is about what we know from our limited perspective, not about metaphysical certainty. A third mistake I see is assuming that Rowe's argument collapses if someone can propose even one possible justification for every instance of suffering. The evidential argument doesn't rest on proving that no justification exists for any particular evil. It rests on the observation that many instances of suffering appear gratuitous from our perspective, and that a universe containing this much apparently pointless suffering is less probable given the existence of God than given its absence. This is a Bayesian-style reasoning process, and it's sensitive to the total body of evidence, not individual cases.
When I encountered this issue working with an undergraduate who was writing a defense paper, the problem was that she kept trying to generate theodicies for specific cases. Every time I asked her to justify preventing one case of suffering, she'd come up with some plausible-sounding reason. But the cumulative force of the argument was being lost in the individual cases. I had her step back and consider the aggregate: if you accept that every instance of suffering might have an unknown justification, then you've made the concept of evidential support vacuous. The argument only has force if we can treat some cases as genuinely gratuitous based on our best available evidence.
What the Skeptical Theism Response Actually Claims
Skeptical theism is probably the most sophisticated response to Rowe's evidential argument. The core claim is that human cognitive limitations are so severe that we have no grounds for judging whether instances of suffering are truly gratuitous. Just because we can't see a morally sufficient reason for some evil doesn't mean one doesn't exist. Our ability to discern God's purposes is to divine purposes what a dog's cognitive abilities are to a calculus problem. William Alston developed this response, and it's arguably the strongest position available to the theist. The challenge for skeptical theism is that it seems to undermine too much. If we can't trust our faculties when it comes to evaluating whether suffering is gratuitous, why should we trust them when it comes to evaluating religious experiences, moral intuitions, or even the reliability of the testimony that led us to believe in God in the first place? The objection, raised by Stephen Wykstra and later refined by others, is that skeptical theism is self-defeating if applied consistently. Rowe himself responded to skeptical theism by arguing that the inference from "we can't see a justifying reason" to "there probably isn't one" is exactly the kind of inductive inference we make all the time in ordinary life. When we investigate a crime scene and can't find a motive, we don't automatically assume there must be an inscrutable motive we're missing. We treat the absence of an observed explanation as evidence that no adequate explanation exists. Rowe thinks we should apply the same standard here.
The Aftermath: Where the Debate Stands Now
Rowe's evidential argument reshaped the philosophy of religion. Before him, the logical problem of evil dominated the discussion, and after Plantinga's response, it looked like the theist had won. Rowe showed that even if the logical problem was defanged, the evidential problem remained fully operational. This opened up an entire research program that's still active today, involving Bayesian approaches to the problem of evil, detailed engagements with specific types of suffering, and refinements of skeptical theism. One notable development that Rowe himself acknowledged was the contribution of John Mackie, whose earlier formulation of the logical problem of evil Rowe was building on and modifying. Mackie had argued that the coexistence of God and evil was logically impossible. Rowe moved the goalposts to probability rather than logical necessity, and that shift turned out to be strategically brilliant because it made the argument much harder to dismiss. The current state of play involves a lot of technical work on probability theory and epistemology. Philosophers like Michael Bergmann, Warren Quinn, and Daniel Howard-Snyder have refined skeptical theism in ways that make it more nuanced than Alston's original formulation. On the other side, William Lane Craig and Paul Draper have contributed important critiques. Draper's "painful evidence" argument, for instance, focuses specifically on the distribution of pain across sentient beings as evidence against theism, narrowing the focus even further than Rowe's broader approach.
If you're working with this material and need to go deeper, Rowe's original 1979 paper is still the primary source. It's short, readable, and contains the full argument in Rowe's own words. There are also useful collections like "Philosophy of Religion: A Guide and Anthology" that include responses from Plantinga, Mackie, and others, which give you the dialectic in one volume rather than having to track down individual papers across journals.
