REVIEW: The Evidential Argument from Evil

REVIEW: The Evidential Argument from Evil

Howard-Snyder, Daniel, ed. The Evidential Argument from Evil. Bloomington, IN: Indiana University Press, 1996.

This paper was originally presented during the PhD seminar “Problem of Evil” (2023) at New Orleans Baptist Theological Seminary.


Evidential Arguments from Evil

To understand evidential arguments from evil it is best to start with the logical problem of evil. The argument takes a deductive form. If its premises are true then its conclusion must necessarily follow. Due to the work of Alvin Plantinga and others, most consider the logical problem now to be no longer valid.[1] The evidential (or inductive) argument, on the other hand, is mostly about probability, often taking an inductive form. If its premises are true then its conclusion likely follows.[2] Two of the biggest proponents of evidential arguments are William L. Rowe and Paul Draper, both of which have papers and responses included here. Bruce Russell and Richard M. Gale have also contributed.

The rest of the volume is comprised of theodicies and defenses from theists: Richard Swinburne, Eleonore Stump, Alvin Plantinga, William P. Alston, Stephen John Wykstra, Peter van Inwagen, and Daniel Howard-Snyder. This is a dense volume with most writers presenting nuanced proofs to make their case. The book has the shape of claim-counterclaim-claim-counterclaim. Rather than tracing that, the following review will first look at the case against theism and then case for it. Additionally, I have not reproduced the probability calculous many have used to make their cases.

A Summary and Critique of the Case Against Theism

Chapter one is William L. Rowe’s 1979 article, “The Problem of Evil and Some Varieties of Atheism.” He looks at three questions: “Is there an argument for atheism based on the existence of evil that may rationally justify someone in being an atheist?” “How can the theist best defend his position against the argument for atheism based on the existence of evil?” And “What position should the informed atheist take concerning the rationality of theist belief?” (1). The argument he puts forth is:

  1. There exists instances of intense suffering which an omnipotent, omniscient being could have prevented without thereby losing some greater good or permitting some evil equally bad or worse.
  2. An omniscient, wholly good being would prevent the occurrence of any intense suffering it could, unless it could not do so without thereby losing some greater good or permitting some evil equally bad or worse.
  3. There does not exist an omnipotent, omniscient, wholly good being.

Rowe introduces two kinds of evil that define the conversation: natural evil (the fawn who burns to death alone in a forrest fire, later called “Bambi”) and moral evil (the real-life story of the rape, beating, and strangulation of a five-year-old girl, later called “Sue”). He focuses his time on premise one, noting that certain evils do “not enable us to establish the truth of (1) . . . But it is one thing to know or prove that (1) is true and quite another thing to have rational grounds for believing (1) to be true” (4). Rowe believes it is “quite unlikely that all” the evil happening in the world is “intimately related” to the work of a good God (5). In responding to some objections, such as the process of soul-making, he notes: “it is reasonably clear that suffering often occurs in a degree far beyond what is required for character development” (6).

But how can he possibly know this? It is only on the naturalistic assumption (that there is no life after death) that such a position is as “reasonably clear” as he assumes. However, the critique he is bringing is not against a naturalistic worldview, but against a theist one. So on what theistic basis would this be reasonable? This criticism is a recurring theme in the atheistic/naturalistic arguments that follow throughout the volume.

To Rowe’s credit, he argues for what he calls the position of “friendly atheism,” where “the atheist may believe that some theists are rationally justified in believing that the theist God exist” (8), noting that “it is perhaps as absurd to think that no one was ever rationally justified in believing that the theist God exists” as it was to believe that men did not walk on the moon (9).

In chapter fourteen, Rowe provides a “second look” to his original argument, responding to criticisms and “strengthening” his argument. In a similar spirit as his first, he beings by admitting that “the whole issue of whether what we know about evil in our world makes it likely that the theistic God does not exist is quite complex and cannot be satisfactorily addressed in any single essay” (263, emphasis added). Two important factors that his updated article rest on are the “conjunctive good” and the background information which can shift the probability positively or negatively. Looking at the first, he believes we should likely “expect” a single good that overpowers both the natural and moral evil.

But he never states why this restriction should be here. Why can the theist not simply have two answers, one for each? The issue of background knowledge is a curious one. The point in question is: What other knowledge do we know about the world that would make theism or atheism more likely? Rowe adds a stipulation to this. The information must be “shared by most theists and nontheists” familiar with the problem of evil (265). In all of the essays by theists (and atheists) in this volume, only Plantinga responds to this with what seems to be a somewhat obvious response: the significant background knowledge comes from the life of Jesus. Gary Habermas’ “minimal facts” for the resurrection would qualify for Rowe’s stipulation. It seems incredulous to think that if that (the resurrection) happened, then would theism not be at least more likely than not?

The last questionable issue Rowe discusses is the response to the skeptical theist (more on this in the section below). While Rowe accepts that the “complete felicity” of experiencing God in eternity would outweigh the “horrendous evil” Sue had to go through in this life, “it strains credulity to think that it is beyond the power of God to realize this good without having to permit” the evil that befell Sue (277). Creaturely freedom is no acceptable response for Rowe. But why? This again betrays his naturalist worldview. As awful and unquestionably evil as is what happened to Sue, for Rowe to take this position is to assume a certain knowledge to be able to value what comes in life after death. That’s an odd position for an atheist to take. 

As a theist myself, Rowe’s atheistic challenge is both thoughtful and welcome. Welcome, for the charity he displays. And thoughtful for the challenge he presents. As Howard-Snyder summarizes in the Introduction, theism is still “quite likely” and “sufficiently reasonable to believe” (xvi). Perhaps, one may say, upon answering the challenges posted by Rowe, et al, that theism is even more likely and more reasonable to believe now.

The second major challenger is Paul Draper. In chapter two, he agues that “pain and pleasure creates an epistemic problem for theists” (12). He presents the Hypothesis of Indifference (HI) as a better explanation for evil than theism. HI states that “neither the nature nor the condition of sentient beings on earth is the result of benevolent or malevolent actions performed by nonhuman persons” (13). Taking his cues from Hume, Draper believes an alternative hypothesis (such as HI) is the key to arguing against theism. Draper turns his attention to two main points in favor of HI: the biological utility of pain and pleasure and the moral value of pain and pleasure. On the first, he believes that HI is stronger than theism. His reason: God “could create goal-directed organic systems (including humans) without biologically useful pain and pleasure” (17). And while he does not directly say anything more than “a biological explanation of pain and pleasure is just the sort of explanation that one would expect on HI” (17), I assume he’s referring to the standard picture of evolutionary development as justification. (This brings up a point that seems a common feature of Draper. Throughout his essay he will make unsupported value statements like this.[3])

Looking at his second point, moral value, he makes the claim: “few would deny that most theodicies are rather obvious failures” (19). I would imagine quite a few theists would, in fact, deny that. As Draper advances his argument, it seems, in his world, that pain and pleasure are the greatest virtues this life has to offer. He later writes, on the assumption of libertarian freedom, “one would expect that God would behave like a good parent, giving humans great responsibility only when we are worthy of it” (23–24). Potshots aside, what leads Draper to conclude this? One of the deepest and most important tenets of Christianity is that we can not merit God’s favor. It is only our freewill that allows such scenarios as grace to play out.[4] In building up the problem against theism, he expounds: 

traditional and contemporary arguments for theism are far from compelling—that is, they are far from being so persuasive as to coerce the acceptance of all or even most rational theists. Thus, even if some such arguments were sound, most theists, including many philosophically sophisticated ones, would not recognize this, and hence the argument would not provide them with evidence favoring theism over HI (25–26).

Draper’s second chapter (nine) is a response to skeptical theistic claims. There are two important points to make about this. First, he concludes his chapter “with a challenge to the skeptical theist to produce some good, aprobable,[5] and undefeated theistic stories. In the absence of such stories, skepticism about my claim that [the probability of observations on HI is “much greater” than the same on theism] is not warranted” (187). I again present Habermas’ minimal facts around the resurrection of Jesus.

Second, Draper seems to feel confident in his position due to the constraints of the skeptical theists’ own requirements. Consistency and intellectual honesty requires a skeptical theist to extend their skepticism even to special revelation, as a result they “must treat such claims with skepticism” (188). But this is not what skeptical theist is claiming. Rather, the claim is that if the biblical God exists, then it is not unreasonable to believe that there are reasons and facts beyond our ken when it comes to understanding why he does or does not interact in the world. The skeptical theist is not claiming, as Draper believes, that we should be skeptical of all knowledge about God.

The final two contributors to the argument against theism are Bruce Russell (chapter ten) and Richard M. Gale (chapter eleven). Assessing background knowledge, Russell writes that “we need to take account of goods beyond our ken only if there is reason to believe there are any” (194). Such a statement seems reasonable. But it will hardly make for an objective treatment between atheists and theists. Regarding what one is justified to believe, he soon concedes that “it is very difficult to specify in any detail what makes one explanation better than another” (195). Throughout, he maintains the need for relevant evidence. “The issue,” he notes, “turns on the question of which of those two competing hypotheses is more likely to be true on background evidence” (196).

Bringing in a positive case for theism (such as my previous note about the resurrection) would be damning enough against his case. But he goes on to make some strange claims. One is the comparison of God’s moral obligation to ours: “I am arguing that if we are not justified in believing that no reason would justify God in allowing the brutal rape and murder, then we are not justified in believing that no reason would justify the onlooker in allowing that same act” (198). It is hard to imagine he is being serious with this statement.[6] At no point does the theist claim that any of the onlookers are omniscient, omnipotent, or omnibenevolent. Such a comparison just seems silly.

Gale’s chapter was the only one that made me laugh out loud, and not for ironic reasons. As Peter van Inwagen notes in a subsequent chapter, Gale is funny and charming. To avoid similar criticisms as above, I’ll focus on just the unique parts. In his section, “axiological laziness,” he criticizes the doctrine of original sin as a way of holding back “the whole class for the misdeeds of a single person” (215). This may be the teaching of some theistic traditions, but it is not the God of Judaism or Christianity. The biblical God holds no one back on behalf of any other, but, as Paul wrote, “all have sinned” (Rom 3:23), and it is due to this that all are guilty. He seems to be so pleased with this assessment that he uses it as his finisher. As noted above Russell said some strange things that one could take him as being serious. Gale, on the other hand, is much harder to read. Is he making outrageous claims to make a point? Or is just making outrageous claims?[7]

One the whole, the case against theism is not thin. This volume is a testament to serious philosophers who have presented serious claims and objections. The following section will survey and assess the responses.

A Summary and Critique of the Case for Theism

Richard Swinburne (chapter three) begins with the claim that “theodicy is the enterprise of showing that appearances are misleading” (30). He goes on: “I accept that an omnipotent being can prevent any evil he chooses, but I deny that a perfectly good being will always try to do so,” which will result in “some good” (30–31). Swinburne is arguing for a greater good theodicy. Without naming him, he immediately corrects Draper’s “very narrow conception of good and evil,” which is limited simply by sensory pleasures and pains (31). The case he builds throughout his essay is that there are greater goods all around us in life. Such a thing is not unreasonable. And if this is the case, then it is not unreasonable to believe that an omnipotent, omnibenevolent God is operating within a greater good framework. But what of unmet desires? “If God wants to make creatures sensitive to all that is good, He will allow them to have desires which are permanently frustrated” (35). Or what about Bambi? The evil is lessened “if others know about this and respond with compassion” (36). “A world without any suffering or in consequence any compassion,” he notes, “would not be better than a world with little suffering and its proper response” (36). And “blessed is the man or woman whose life is of use” (43). Unlike some of the other denser chapters, Swinburne’s chapter is filled with aphorisms like this. Just about all are thought provoking. Though, I do wonder if his overall case would not be stronger without a few of the weaker ones.

Of all the chapters, Eleonore Stump’s (chapter four) felt the most like an appendix. Not for its quality—it was interesting and well done—rather, for its subject matter. She used Aquinas’ view of suffering as a lens into a medieval worldview, with its pain-and-suffering-as-medicine approach to evil. It was interesting and thought-provoking, but it did not seem terribly relevant to today.

Alvin Plantinga has two chapters. In chapter five he tackles epistemic probability. Given the nature of the argument against theism (mostly inductive) and the response most theists in this volume give (often some form of skeptical theism), epistemic probability is deeply related to both sides. “There is no reason to think that if God did have a reason for permitting the evil in question, we would be the first to know” (70). This here is the divide between the theists and the atheists represented in this book. A naturalist’s perspective would, by definition, see the world as both knowable and complete. That is (for most) there is no such thing as the spiritual world. The theists, on the other hand, believes in both an entirely different world (spiritual) as well as life after death. It is for this reason that Plantinga proposes that God’s reasons for allowing evil may “involve a good for the sufferer, a good that lies in the future life” (73). For this reason and others, he takes Rowe to task for supposing God would have no reason for allowing evil. Epistemically, Rowe “can’t sensibly suppose” this (74). Upon examining the claim that a good God could create the same good with less evil, Plantinga replies,

How would we know what would have happened if God had created a world quite different from this one . . . We know far too little about the world that is in fact actual, far too little about the sorts of creatures it contains and the counterfactuals of freedom that characterize them, to be justly confident of opinions about God’s alternatives (78). 

For Plantinga, the “real question here concerns the nonpropositional warrant, if any, enjoyed” by belief in God (70). “Perhaps belief in God resembles certain perceptual beliefs, memory beliefs, certain a priori beliefs and others in being properly basic (in the right circumstances); if so, belief in God, like these others, will (under the right circumstances) have nonpropositional warrant” (89). Plantinga goes on to argue that this entire debate is misunderstood as being founded on some kind of epistemic neutrality, but is rather an “ontological or theological dispute” (93). Plantinga’s second chapter (thirteen) is largely a response to Draper’s “Pain and Pleasure” essay, discussing the “evidentially challenged” position of theism. Plantinga questions this, noting: “For any proposition I believe (so long as it is not logically impossible) I can find an evidential challenge” (248). The same is true for “every scientific hypothesis” (248).

William P. Alston’s chapter (six) looks at skeptical theism from the limitation of human cognition.[8] Responding to Rowe’s claim—see premise (1) above—“our epistemic situation is such that we are unable to make a sufficiently well grounded determination that 1 is the case” (98). Ultimately, his position is to attack the label of gratuitous evil for any specific case. I take his point here, and it is consistent with my theistic beliefs, but, I’ll admit, it is hard to swallow. The case of Sue does seem gratuitous. But if gratuitous means something like “too bad for any good to justify it” then I do agree with him, even if, as the skeptical theist says, I cannot understand it.[9] Chapter sixteen is meant to “clean up” and finish the book. He moves around, touching on various points that have previously been presented. I’ll note one that was helpful. “The main trouble with Rowe 1988 is that . . . it encourages the supposition that we can tell whether known goods could provide God with a justifying reason for permitting certain evils” (315).

Stephen John Wykstra (chapter seven) applies his CORNEA method (Conditions of Reasonable Epistemic Access) to what he calls “Rowe’s Noseeum,” in which Rowe believes the evil observed has “no such goods” (126). Specifically Wykstra focuses on “levering evidence,” which, as its name implies, is what can “move one from one rational square belief-state to another” (131). Wykstra concludes that accepting “Core Theism . . . gives us a great deal more than ‘no reason whatsoever’” to take a skeptical theist position (145). Peter van Inwagen’s two chapters (eight and twelve) are responses to Draper (mostly) but also Russell and Gale. Contra to Rowe’s “conjunctive good,” van Inwagen does not believe there need to be a single reason encompassing both natural evil and moral evil (157). Arguing for the “consequences of modal and moral skepticism,” so as to avoid a “massively irregular world,” van Inwagen believes that evil “gives us no reason to prefer HI to theism” (163). His second article (chapter twelve) is to clarify some of his previous claims by presenting them more “systematically.”

Daniel Howard-Snyder begins his chapter (fifteen) with a bold claim, that “God need not have a reason to permit that [specific] evil” (287). Then, arguing similarly to Plantinga, he notes: “What reason is given to think that by thinking, talking, and reading about the matter, we will very likely comprehend God-justifying goods, if there are any? None” (292). The rest of his chapter feels like a conclusion (though that job was apparently given to Alston) touching lightly on some of the high points throughout the book.

In conclusion, this was a dense but important book. There are strong inductive arguments from evil against theism, as Rowe, Draper and others make clear. But as the theist responses have shown, there is nothing intellectually shameful about holding to a certain level of epistemic humility. Mostly, what stood out are the value of examining one’s presuppositions. Ultimately, this seems to be what anchors belief to one side or the other.


[1] Richard M. Gale (chapter eleven) claims that he’s shown that “no version of this defense works” (206). Nonetheless, his view seems to be in the minority. In chapter nine, Paul Draper comments, “Logical arguments from evil are a dying (dead?) breed” (176).

[2] Rowe’s latest formulation (chapter fourteen) combines inductive with deductive.

[3] For example: “we obviously have much more reasons on theism & O1 & O2 than we have on HI & O1 & O2 to expect sentient beings (especially nonhuman animals) to be happy” (18). But why? What reasons? As a theists writer in this volume notes: given sin and the state of injustice and the expectation for the way things should be, the theist has perhaps greater reason for frustration in this life. (This, to note, may be distinguished from hope.) Additionally, there are various degrees of personal maturity among adherents of theism, and so a genuine theist may not be happier than a naturalist who believes there are no consequences after this life.

[4] This holds true for both compatibalists and libertarian views of freedom. 

[5] A theistic story that satisfies the condition: observational data that cannot be assigned values like “low” or “high” (180).

[6] Other strange claims are his no-blue-crows search which aligns “similarly” with his no-moral-God search. This is an epistemic equivocation. We as people are quite capable of exploring the natural world to determine if blue crows exist. We know where to look, and we know the limits of where blue crows would be (e.g. not in outer space, not underwater). On the other hand, we are far from capable for exploring the supernatural world. Especially those of us committed to naturalism.

[7] For one example: “Defensive skepticism is an ivory tower invention of the detached epistemologist of religion that is completely out of touch with the grimy realities of everyday religious faith and experience” (214).

[8] To note, he is not referring to general human cognition, but human cognition as it relates to evaluating the issue of gratuitous evil (102).

[9] Alston makes many good points in his essay, such as “Rowe’s generalization is more like inferring, from the fact that no one has yet produced a physical theory that will unify relativity and quantum mechanics, the prediction that no one will ever do so. . .” (110). But for the sake of space, I must leave out many of his other points.