For several decades, discussions of the “inductive” or “probabilistic” or “evidential” argument from evil have been major fare in philosophy of religion. The target of this type of argument is standard theism—the view that there exists an omnipotent, omniscient, and perfectly good being who created the world. Of course, theistic religions call this being “God.” The argument purports to show that the facts of evil, though logically consistent with the existence of God, count against the truth of theism. The core argument goes as follows:
(1) There are gratuitous evils.
(2) If God exists, then no gratuitous evils exist.
Therefore,
(3) God does not exist.
Since the argument is clearly valid, if there are rational grounds to believe (1) and (2), then there are rational grounds for believing (3). (1) is grounded inductively; (2) is assumed as a necessary truth. Variations on the core argument as well as attempts at rebutting it occupy most of the literature on evil in the analytic philosophy of religion today, creating a complex discussion of the premises, their embedded assumptions, and their epistemic status.
My aim here is to assess William Rowe’s different formulations of the evidential argument from evil in connection with some important theistic responses. I find that both the Skeptical Theist Defense and Alvin Plantinga’s felix culpa theodicy (a new version of Greater Good Theodicy) make the common mistake of attacking the claim that there is gratuitous evil. Skeptical theism (appropriately for a defense) argues that there is no reason to think the claim is true; felix culpa theodicy (appropriately for a theodicy) argues that there is a reason to think it is false. I argue that both approaches are inadequate to show that Rowe is mistaken in claiming that it is more rational than not to believe that there is gratuitous evil. I further point out that both responses fail to question the other premise in Rowe’s argument: the claim that God must prevent gratuitous evil. I ground the case for rejecting this premise—and therefore improving prospects for effective response—in a certain understanding of God’s providence in light of distinctively Christian doctrines and themes. I contend that, in order to clear the path for developing this understanding, we must not accept various restrictions that Rowe wants to place on the debate.
THE EVOLUTION OF THE ARGUMENT
William Rowe has long been articulating, modifying, and defending the evidential argument from evil. His classic 1979 formulation became the progenitor of decades of debate:
1. There exist 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. [Factual Premise]
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. [Theological Premise]
Therefore,
3. There does not exist an omnipotent, omniscient, wholly good being. [Atheistic Conclusion]
Discussion naturally turns to the premises, which are labeled above to denote their functions in the argument.
The Theological Premise contains a necessary, but not a sufficient, condition for an omnipotent, omniscient, wholly good being to permit suffering that he could have prevented:
(i) there is some greater good that God can obtain only if he permits the intense evil in question, or
(ii) there is some greater good that God can obtain only if he permits the evil or some other equally bad or worse evil, or
(iii) God can prevent the evil only if he permits some other evil equally bad or worse.
The occurrence of an evil is gratuitous if it does not meet this condition, which Rowe believes accords with our basic moral principles.
Since most theists and nontheists agree that the Theological Premise is true, virtually all early discussions of the argument focused on the Factual Premise. Rowe explains that (1) is rooted in the intense suffering, both animal and human, which occurs daily in our world. He originally advanced a now famous hypothetical case of gratuitous evil: that of a fawn dying horribly in a forest fire. This example is entirely credible because we are familiar with myriad such cases; so, Rowe argues that it is more reasonable than not to believe that this evil is gratuitous. Although his 1979 argument turns on the reasonableness of affirming the Factual Premise in light of this specific case, Rowe admits that we could be mistaken about the gratuity of any particular case of such suffering. In some subsequent formulations of the argument, he maintains that it is not reasonable to believe that all instances of seemingly pointless animal and human suffering could not have been prevented by an omnipotent, omniscient, perfectly good being without thereby losing some greater good or allowing some equally bad or worse evil. So, it is highly likely that some, probably many, of the horrible evils in our world which appear to be gratuitous are genuinely gratuitous. Rowe has simply been a bit fluid in how he makes the point that (1) is a rational belief, more rational to believe than its denial.
As Rowe refined his thinking in response to criticisms of the grounds for (1), his argument evolved through 1988 and 1996 formulations. These formulations more precisely articulate the inductive logic involved and its bearing on the rationality of theistic belief. The 1988 version adds an instance of evil described by Bruce Russell: the case of a little girl beaten, raped, and murdered in Detroit. The dying fawn (E1) is evidence of natural evil and the victimized girl (E2) is evidence of moral evil. Rowe offers this statement of the argument:
P: No good state of affairs we know of is such that an omnipotent, omniscient, perfectly good being’s obtaining it would morally justify that being’s permitting E1 or E2.
P makes it likely that:
Q: No good is such that an omnipotent, omniscient, perfectly good being’s obtaining it would morally justify that being in permitting E1 or E2.
Rowe states, “Since Q, slightly qualified, is tantamount to (1) in the earlier argument for atheism, if we are justified in accepting P and justified in inferring Q from P, we are justified in accepting premise (1).” Then, from Q, the atheistic conclusion follows:
Not-G: There is no omnipotent, omniscient, perfectly good being.
Since the inference from P to Q recasts the case for (1), it became the topic of vigorous debate. Rowe writes, “If my project is to succeed, then the former fact must be a good reason for the latter assertion. And if the former fact is a good reason for the latter assertion, then my project succeeds.” As more theists began to accept P, the debate shifted to the inference from Q to not-G.
The inference from Q to not-G is not straightforward and requires an unstated premise very much like the original (2), but since Rowe believes that (2) is a “conceptual truth,” he tacitly employs it without extensive defense. But then this formulation of the argument still has both a factual and a theological premise, as all versions really must have. Rowe’s 1996 formulation is an even simpler argument: it avoids the inference from P to Q and reasons directly from P to not-G, assuming relevant background information. This argument is distinctive in its use of Bayesian probability theory, which means that “background information” will factor prominently into the debate.
Commentators accent the discontinuities between the 1979, 1988, and 1996 formulations, but Rowe’s more recent writings clearly frame the discussion around formulations of the argument that many thought he had abandoned. His 2001 exchange with Michael Bergmann and Dan Howard-Snyder revisits his 1979 formulation, with only slight verbal changes. In 2006, Rowe offered a statement of the argument that is virtually identical with the core argument at the beginning of this paper. Rowe just keeps working at the project of strengthening the argument—improving the inductive strategy for the Factual Premise, clarifying the assumptions it makes, and defending it against criticisms. But his central, driving insight remains the same: “[W]hen we consider horrendous evils or the sheer magnitude of human and animal suffering, the idea that an omnipotent, omniscient, perfectly good being is in control of the world may strike us as absolutely astonishing, something almost beyond belief.”
THE SKEPTICAL THEIST DEFENSE
The Skeptical Theist Defense has emerged in mutual interaction with Rowe’s evidential argument. Skeptical theists Stephen Wykstra, William Alston, Michael Bergmann, Dan Howard-Snyder, and others argue that Rowe has not provided a good reason to think that premise (1) is true or probably true because no human being is in an epistemic position to make such a judgment. Early on, Wykstra articulated a principle of rational belief—the Condition Of Reasonable Epistemic Access (CORNEA)—to show why it is not rational to believe the Factual Premise:
“On the basis of cognized situation s, human H is entitled to claim ‘It appears that p’ only if it is reasonable for H to believe that, given her cognitive faculties and the use she has made of them, if p were not the case, s would likely be different than it is in some way discernible by her.”
Essentially, if God exists, it is not rational to expect that we should be able to discern the goods for the sake of which he allows evils. Support for this point has ranged from emphasizing the limitations of human cognitive capacities in discerning complex relations between goods and evils to arguing that if God exists it is likely that there are goods “beyond our ken” which justify the evils in question. So, if we are not in an epistemic position to make the claim that there appears to be no greater good secured by God’s permission of some case of intense suffering, then (1) is not reasonable to believe.
Although Rowe offered credible support for his appearance-claim, his 1988 attempt clarifies the kind of appearance-claim being advanced: it is P, that we simply don’t know of any good that justifies God’s permission of certain evils. Although P here is meant to have prima facie evidential plausibility, leaving it open to discussion and even defeat, it gained acceptance from skeptical theists. So, Rowe’s support for (1) is the inference from P to Q, which has the form: So far as I can tell, there is no x; therefore, probably, there is no x. CORNEA and similar principles were modified to apply to this inference, dubbed a “noseeum” inference because it relies on a “noseeum assumption”: that if there were an x, it is likely that the person seeking it would discern it. Skeptical theism seeks to block this inference by denying that any human being has what it takes to discern God-justifying goods, while Rowe maintains that his claim that anchors the inference is based on careful contemplation.
Interestingly, Rowe maintains that his 1996 version, which argues directly from P to not-G in order to bypass Q altogether, does not require a noseeum assumption. In a 2001 piece, Bergmann disputes this and seeks to show that Rowe’s 1996 formulation of the evidential argument still depends on the denial of the skeptical thesis that we have no good reason for thinking that the possible goods we know of are representative of the goods there are. After all, as Bergmann explains, it is quite likely that reality far outstrips our comprehension of it. In his 2001 reply, Rowe argues that to suppose in regard to instances of terrible suffering that God exists and yet that the sufferers have no consciousness of his love or of the fact that there is some good which justifies their suffering is to assume unbelievable things about a perfectly good being of infinite wisdom and power. For Rowe, the supposition is inherently implausible, and its implausibility need not rest on an assumption about whether the goods we know are representative of the goods there are.
We cannot pursue here the details of this Bergmann-Rowe exchange. My take is that, in the final analysis, the Skeptical Theist Defense fails to undercut Rowe’s claim that the Factual Premise is more rational to believe than its denial. Rowe refers in various writings to some of the grounds that make it rational to believe the Factual Premise, including the fact that we have an admittedly modest grasp of some very great goods (such as beatitude in the presence of God) but that they are still goods we know and the fact that many theistic believers as well as unbelievers have the persistent intuition that in some sense inscrutable evil disconfirms theism. Skeptical theists have to maintain that people generally are simply mistaken about such things. An even more straightforward argument available to Rowe against this skeptical indictment is that the goods that most clearly matter in the dispute are those goods pertinent to our common human nature, which is reflected even in our dimly imagining fulfillment in God’s presence or in our widely shared sense that inexplicable suffering puts pressure on religious belief. Our common humanity, accessed by careful introspection and by acceptance of the wisdom of the race, grounds a generally reliable and widely agreed upon understanding of the types of goods and evils and their possible interrelations that are relevant to the kind of flourishing which is appropriate to our nature. We are not hopelessly in the dark about such matters. While it is logically possible that there are goods relevant to human nature that are beyond comprehension, and while it is surely highly probable that reality is far larger than our grasp of it, such considerations do not show that it is not reasonable to accept the Factual Premise. So, if we grant Rowe’s common sense support of the Factual Premise together with his use of a Theological Premise (explicitly or implicitly) in all versions of his argument, it would seem that the evidential argument from gratuitous evil provides a reason for atheism. At least it would seem so—as long as we do not question the various restrictions on the debate proposed by Rowe over time and accepted by many respondents.
One important restriction proposed by Rowe is based on the distinction between Restricted Standard Theism (RST: standard theism unaccompanied by other, independent religious claims) and Expanded Standard Theism (EST: standard theism conjoined with certain other religious claims). In an early reply to Wykstra, Rowe clarifies the structure of his approach: he is strictly evaluating RST solely on the evidence of evil. He correctly argues that Wykstra’s skeptical mode of defense unwittingly assumes a version of EST, one that adopts a certain construal of the wide gap between finite human knowledge and infinite divine knowledge. Of course, one irony here is that Rowe himself unwittingly imports into his understanding of RST assumptions from his own implicit version of EST. “[B]oth evidential arguer and skeptical defender,” William Hasker observes, “already claim to know quite a lot about God’s reason for permitting specific evils.” Both Rowe and skeptical critics claim to know, for example, “that God’s reasons must involve a specific good state of affairs that outweighs the evil, and moreover one that is related to the evil in question in a particularly intimate way: the good must be such that even God could not obtain it without permitting the evil in question.” After many years of debate, we should learn the lesson that it is extremely difficult, and perhaps not all that enlightening, to conduct the debate within the confines of RST. However, Rowe cautions that invoking some version of EST, say, drawn from Christian theology is actually counterproductive as a response to his argument. He admits that some particular version of EST (which entails RST) may not be rendered unlikely by the facts of suffering, but he maintains that it must nevertheless have a prior probability lower than that of RST:
“The reason this can be so, even though EST’s probability is not lowered by E, is that the probability of EST, given E, is a function not only of any tendency of E to disconfirm it, but also of the prior probability of EST, the probability of EST alone. Thus, even though E does not disconfirm EST, since EST commits us to much more than does RST, the probability of EST alone may be much lower than the probability of RST alone. In fact, given that EST accounts for E and entails RST, its prior probability must be much lower than RST’s, if the probability of RST on E is a good deal lower than the probability of RST alone.”
So, Rowe believes that appeal to any form of EST is problematic in terms of how he has structured the discussion. Skeptical theists accept restriction to RST but argue that Rowe fails to understand that RST still generates implications that block his noseeum inference. However, conducting debate in terms of RST virtually guarantees both that Skeptical Theist Defense will seem to many to be an attractive strategy and that the evidential argument, addressed on its own terms, will be virtually impregnable.
In his 1996 Bayesian argument (from P to not-G), Rowe employs another restriction, one related to the idea of “background information”: limitation of the background information (k) on which we rely in forming judgments about the probability of P and G (God exists) to the intersection of what the theist and nontheist know. This would include knowledge of evils, of various goods, and of how the world works. In this sort of Rawlsian “initial position,” no information that entails either theism or atheism is allowed. Rowe’s “level playing field assumption” governs the probability assignments to various pieces of information (with some assignments reflecting positive evidential impact on G and some of it negative) such that the totality of k leaves the probability value of G at 0.5 and of not-G at 0.5. Given this framework, Rowe’s claim is that, if we then acquire P as our sole new evidence, then, given the information and its values found in k, the probability of G is significantly lowered, making belief in atheism more reasonable than belief in theism. Now, if k is all we have to go on and the introduction of P makes atheism much more likely than not, then atheism is epistemically preferable to theism. Rowe admits that other evidence may make it reasonable to believe in theism but thinks that, in the absence of other theistic evidence, belief in theism is irrational.
Skeptical theists, of course, accept confinement to the “initial position” because they think that the atheist is in no position to conclude that the unavailability of God-justifying goods for the evils in question significantly lowers the probability of G, no matter how horrendous those evils are. But the ability of the skeptical move to block this inference must also be viewed as a function of the probability of skeptical theism itself, which is in dispute, partly because it may really be a version of EST with very definite views of God, his attributes, the finite-infinite knowledge gap, and so forth. Additionally, if the recognition of human cognitive limitations is leverage for skepticism regarding our ability to inquire successfully into God-justifying goods for the evils of the world, then the recognition of cognitive limitations would leverage skepticism in other areas of human inquiry as well. Yet various examples readily come to mind in which we do not allow the legitimate recognition of cognitive limitations to undercut the rationality of our judgments: regarding theories in quantum physics that have substantial evidential support but clearly do not capture the entirety of quantum reality, or historical claims based on reliable sources that are impossible to check directly, or even moral evaluations of persons and situations about which we cannot know all details or outcomes. Such examples accent the difficulty of understanding how recognition of our cognitive limitations successfully undermines the rationality of believing that not-G is more likely than G on P. Efforts by skeptical theists to confine their skepticism to the realm of God-justifying reasons do not seem effective; but I develop this point more fully in the next section where I consider some overlooked supports for the Factual Premise. So, it is looking as if confinement to the “initial position” all but guarantees that the theist will have irrational beliefs. Why on earth would a Christian theist who does not inhabit the initial position agree to be limited by it and the probability values it assigns, particularly if doing so makes theistic belief unnecessarily problematic? In reality, the Christian theist and nontheist simply believe and know quite different propositions—related to the deeper analysis of divine attributes, the significance of the human moral venture, goods and valuables, and other relevant elements of the Christian theist’s total version of EST.
Rowe’s artificial epistemic restraints set the stage for an invigorating academic exercise for both the theists and nontheists who accept them. However, the really important question is not whether the evidence of intense suffering counts against a limited form of theism considered against a highly controlled information set but how this evidence affects the evaluation of full-orbed versions of religion that involve theism. In contemporary analytic philosophy of religion, of course, it is typically the position of distinctively Christian theism that is at stake. Generally, the critic’s strategy is to argue against theism as a way of arguing against Christian theism; likewise Christian theists defend theism as an important way of rationally supporting Christian theism. However, unless we are dealing with an alleged contradiction in theism, we can no longer keep approaching matters this way. To make the complete position of Christian theism vulnerable to criticism of a subset of the position when that criticism is based on a piece of evidence, no matter how impressive, does not tell us nearly as much as critics claim. We should not avoid appeal to any form of EST because its prior probability is lower than that of RST. After all, many complex claims (some form of EST, Einstein’s theory of general relativity, etc.) about some complex reality (divine purposes in regard to evil, gravity as a geometrical property of spacetime, etc.) commit one to a lot more than do some simpler claims and therefore have low prior probability relative to simpler claims about the same reality. But we regularly find some complex claim—as opposed to other competing complex claims and even as opposed to competing simple claims—to be the best explanation given the total evidence.
PLANTINGA’S FELIX CULPA THEODICY
Theodicies within the Greater Good Tradition propose reasons to think that the Factual Premise is false. Rowe has critiqued many of these theodicies—notably, Augustinian free will theodicy, Leibnizian best of all possible worlds theodicy, and Hickian soul-making theodicy. Departing from his usual defensive approach, Alvin Plantinga has articulated a new greater good theodicy which identifies some good that God could not have achieved in the created world without permitting the evils it actually contains. Articulating what he calls felix culpa theodicy, he rightly refuses to be limited to RST and instead explicitly draws upon certain Christian doctrines.
Plantinga begins with the assertion that what God intended in creation was to weakly actualize a really good possible world. In considering the value or goodness of possible worlds, states of affairs such as John’s being in pain is bad while John’s suffering magnificently is good; there being many people acting in love toward each other is good while there being people who hate God and each other is bad; and so on. So, possible worlds can be evaluated in terms of the balance of good-making qualities—such as happiness, beauty, justice, and love of God, on the one hand—and bad-making characteristics—such as suffering, pain, sin, and rejection of God, on the other. But reflection on theistic belief extends our thinking about the value of possible worlds. If God’s existence is necessary, then he exists in all possible worlds; and theism entails that God himself is infinitely valuable. So, there are no possible worlds in which God does not exist; and any world God chooses for weak actualization will necessarily contain the great-making characteristic of God’s own infinite value. Given the truth of Christian belief, however, we can extend our thinking about the comparative value of worlds even further.
In developing his theodicy, Plantinga observes that, given the traditional concept of God as essentially unlimited in goodness, knowledge, and power, it follows that the world God has created is very good and that there are no worlds he would have created that are less than very good. Of course, we can in some sense imagine worlds in which all persons are always in excruciating pain; but no such worlds are in fact possible if God is a necessary being who essentially is perfect in goodness, knowledge, and power. So, all possible worlds are very good, including a possible world in which God alone exists. But this does not mean that even a world W in which God alone exists is of maximal value such that no possible worlds are better than W. A world that also contains free creatures who always do what is right would perhaps be a better world than W. This is true even if we grant that the good of God’s existence is incommensurable with both creaturely goods and evils. So, it still follows that every possible world is a very good world.
However, some possible worlds are much more valuable than others. Our world and some other possible worlds contain a contingent good-making feature: the towering good of divine Incarnation and Atonement. God, in the Second Person of the Trinity, became incarnate in Jesus of Nazareth, lived a holy and sinless life, and suffered and died for sinful human beings, providing reconciliation to the Father. Plantinga asks rhetorically, “Could there be a display of love to rival this?” Then he adds, “More to the present purpose, could there be a good-making feature of a world to rival this?” Plantinga maintains that this good is so clearly incomparably great, displaying the matchless beauty of self-sacrificing love, that any world with it is better than any without it (or at least without some similar divine initiative). So, any world with Incarnation and Atonement contains two infinite goods: the existence of God as well as Incarnation and Atonement. At this point, Plantinga articulates what he calls “the strong value assumption”:
A: There is a level L of excellence among possible worlds such that all the worlds at that level or above contain Incarnation and Atonement.
Now, if God intends to actualize a really good possible world, one whose value exceeds L, he will create a world containing Incarnation and Atonement. But all worlds containing Incarnation and Atonement also contain evil, since the presence of sin and evil is a necessary condition of Atonement. Hence, all of the really good worlds contain Incarnation, Atonement, sin, suffering, and evil.
We now have an explanation for why there is all the suffering and evil in the actual world. It is because God wanted to create a highly valuable world, one that contained Incarnation and Atonement; but all highly valuable worlds contain evil. O Felix Culpa! (O Fortunate Flaw!) Taken from the Exultet, the sentiment here is that Adam’s sin was fortunate because it necessitated Christ’s redeeming work. So, sin and evil occasion the highest possible good. This theodicy, then, like all greater good theodicies, purports to show why all evils are necessary to a greater good that God aims to achieve. What we have here is simply Plantinga’s particular construal of the relation of evils to the great good of Incarnation and Atonement which is drawn from the wider context of Christian beliefs.
But is the paradoxical lesson of felix culpa theodicy that the value of salvific relationship with God is so great that it is worth breaking so that God can restore it? Initially, we should be charitable in allowing the Easter Proclamation some poetic license to express profound gratitude and praise for what Christ has done. We cannot deny the beauty and love of God manifest in the whole Christ-event. Under scrutiny, however, there is something mistaken in this claim:
If humanity had not fallen, then this world would not be a really good world with a level of excellence of L or above.
To see why this is a mistake, we can consult the classical Christian doctrine of creation, which teaches that God’s original purpose is to invite finite personal creatures into intimate relationship with himself. Indeed, the historic Christian vision of the human telos is that we are meant for participation in the divine Trinitarian life. This entails that God works faithfully to bring our telos to fulfilment, such that the highest good for creation is available without creation’s descent into sin and evil.
Thus, Incarnation and Atonement are logically independent; their conjunction is contingent, not necessary. Even granting that the actual world and all other fallen possible worlds contain both Incarnation and Atonement, Incarnation is still possible in unfallen worlds without Atonement. Indeed, Incarnation is likely in unfallen worlds as God reveals himself and invites humanity into the divine life. Incarnation reveals God’s enduring nature to seek and draw created personal beings to himself, and it symbolizes that divine nature is forever bonded with human nature. Moreover, when God’s nature is revealed in unfallen worlds—through Incarnation or other means—it will be characterized by the same unrelenting, self-giving, self-sacrificing love demonstrated in Incarnation and Atonement in the actual world. Clearly, God could have carried out his original wonderful plan for humanity in worlds that do not contain sin and Atonement. So, it was always possible, and always more desirable, not to sin.
Plantinga anticipates the criticism that felix culpa theodicy portrays God as using created persons as means either to the divine end of a high degree of cosmic excellence or to his own glorification. This criticism gains traction in light of the widespread and intense suffering found in the actual world. Plantinga comments that his theodicy explains why there is evil, but he also recognizes that sin and evil cause much suffering. So, suffering is thought to be part of an excellent world. But a response at the world level does not address the point made by Marilyn Adams, Eleonore Stump, and William Rowe that God’s goodness must include goodness to individual persons such that their suffering must somehow benefit them and that this benefit must be included in their conscious experience in a positive, constructive way. God’s perfect love may mandate that he actualize a world in which those who suffer are benefited in such a way that their condition is better than it is in those worlds in which they do not suffer.
Addressing this Person Centered Requirement, as Rowe calls it, Plantinga explores the theme that persons who are redeemed from sin and persons who suffer can have greater intimacy with God than would otherwise be available in unfallen worlds. But should not strong caution be taken in presuming that the experience of God possessed by redeemed sinners and/or faithful sufferers is somehow of greater value than the experience of those who never sinned or perhaps never suffered? On the one hand, there may be a contingent truth in our world and in at least some other fallen possible worlds that experiencing redemption or suffering allows a particularly poignant sense of God’s presence and inner life. On the other hand, the previous points—about the capacity of Incarnation to reveal God’s nature without the necessity of sin and Atonement, and about God’s unchanging plan to bring humans into intimacy with himself—entail that there are no forms of intimacy with God that are fundamental to our humanity and yet attainable only by experiencing redemption from sin or experiencing suffering.
A NEW DIRECTION FOR THEODICY
The evidential argument from evil has proved remarkably difficult to dispose of—either by trying to undermine confidence in the evidence for its Factual Premise or by proposing reasons to think the premise is false. After sustaining these kinds of theistic assaults, the Factual Premise still seems more rational to believe than its denial. In reflecting on the severity and magnitude of evils in the world, it is entirely reasonable to think that there are evils such that the world would be better, or at least no worse, if God had prevented them. Even if we might be mistaken about whether a given instance of evil is genuinely gratuitous, it is still rational to conclude that many, if not most, apparently gratuitous evils are genuinely gratuitous.
There is weighty philosophical and theological support for the judgment that there are gratuitous evils. On philosophical grounds, there is strong presumption in favor of the general reliability of our rational and moral faculties. As I indicated earlier, excessively strong claims about our faculties being incomplete or incompetent or systematically mistaken in evaluating whether it is likely that horrendous evils have some God-justifying point portends an unacceptable epistemic and moral skepticism. Besides, if we are not entitled to make considered judgments about some of the most fundamental aspects of human life—such as the pointlessness of evil—the fact that we are creatures who naturally and quite regularly make epistemically illegitimate judgments about such matters would itself be a candidate for a completely pointless evil. For reasons given by skeptical theists, we literally have no idea whatsoever whether the life of Mother Teresa was better or worse on the whole, and made a better or worse contribution to the goodness of the world overall, than the life of Saddam Hussein.
On theological grounds, it is not clear that Christian theists who embrace classical orthodoxy should endorse any version of skepticism in order to find relief from the problem of evil. Christianity is an inherently revelatory tradition which affirms that it is God’s plan and good pleasure to reveal something of himself and his ways to us. This confidence is implicit in the Church’s historic formulation of key doctrines which assume that certain theological realities exist and that we do indeed know something about them: Trinity and Incarnation merely begin the list. Furthermore, the doctrine of humanity teaches that human persons are created “in God’s image” (in imago Dei)—and that God bestows upon human persons finite powers (such as reason and moral evaluation) which reflect powers he possesses in infinite measure. A balanced view of pertinent Christian doctrines, then, suggests that skeptical interpretations of the finite-infinite gap or of God’s intentional hiddenness and mystery are extreme. Whatever legitimate insights these interpretations contain, they need to be tempered by more positive themes drawn from the whole body of Christian doctrine—for example, related to the general reliability of the noetic powers and capabilities of finite created persons. So, something in the neighborhood of ontological, epistemological, and moral realism is implicit in classical theological orthodoxy. But this means that theology provides indirect support for the Factual Premise and the kinds of reasons Rowe gives for rationally believing it.
The continuing strength of the Factual Premise should motivate Christian theists to revisit the Theological Premise. The premise assumes the definition of a “gratuitous evil” as an instance of intense suffering that God could have prevented without thereby losing some greater good or permitting some evil equally bad or worse. The focus on suffering, of course, could be expanded to include what Marilyn Adams calls “all of the minuses of life.” But the core idea is that an evil is gratuitous if the world on balance would have been better had it never occurred. The Factual Premise also involves the assumption that God is required to prevent gratuitous evil so defined. The picture of God here embodies what I have called the Principle of Meticulous Providence: the requirement that God permits an evil only if it is necessary to achieving a greater good or avoiding an evil equally bad or worse.
Contemplation of such a requirement on God ushers us into the controversial territory of conceptions of divine providence, that is, of the nature of God’s purposes for and interaction with the world. Theists who accept theological determinism or Molinism typically embrace Meticulous Providence, but theists who reject these models are often drawn to views that envision some degree of openness in God and the world, and thus endorse general providence. Without rehearsing here the details of this complex debate, I believe that prospects for answering the evidential argument from gratuitous evil are much better for views that envision God, his ways, creation, and the human venture as “open” in significant respects, not “closed.”
The fact that RST can be interpreted to support various views of providence further accents the need to resist its artificial limitations and develop distinctively Christian intellectual resources to address the problem of evil. Plantinga’s felix culpa theodicy displays his willingness to start thinking along these lines; Marilyn Adams has been working in this vein for a long time. We should not, however, envision including claims drawn from Christian doctrinal understanding as simply conjoining them with RST, since the body of Christian theistic belief forms an organic whole in which a more nuanced and profound interpretation of the concept of God is provided, including interpretations of the theistic attributes. This clearly affects the implications we can draw from theism considered as a component of Christian belief. Appeal to the rich version of EST I have in mind will not settle all disputes over models of providence, but it will provide important insights for rebutting the evidential argument.
As long as the Principle of Meticulous Providence is accepted, responses to the evidential argument will be driven in the predictable directions of either Skeptical Theist Defense or Greater Good Theodicy, neither of which are prepared to admit that the world would have been better without the evil in question. Since I have always thought that such responses are doomed to failure, I have long recommended rejecting the Principle of Meticulous Providence in order to reject the Theological Premise which assumes it. The principle is neither necessary nor essential to theism, and is certainly not essential to classical Christian orthodoxy. So, the Theological Premise cannot be assumed as a necessary truth or as essential to the theistic conception of God, and certainly not as essential to classical Christian orthodoxy. Showing that this is so involves drawing important implications about God and his purposes from certain major Christian doctrines and themes because bare theism simply does not contain enough information about God to do this.
We can sketch here only the outlines of the relevant richer information from the Christian theological perspective: The doctrine of the Trinity implies that God is an inherently personal, inter-personal, relational, self-living Life whose nature is mutual self-giving, self-sacrificing love. This Three-Personal Life graciously created a finite reality with an evolutionary trajectory that would eventually bring forth personal-rational-moral beings that are intimately related to the physical realm and can relate to God. These finite personal beings are invited to find their proper flourishing by participation in the divine Trinitarian Life. This mainstream construal of the doctrines of Trinity, Creation, and Humanity entails that this is a deeply relational universe. Only in a truly relational reality can certain great goods—such as love, self-giving, free choice of the good, interpersonal communion, and voluntary connection with God—be realized by finite rational-moral creatures. For the world to be genuinely relational, it cannot be ready-made—or “closed”—but must contain real contingency. This contingency, whether in the natural or personal realms, occurs within lawful structure. Contingency is required as a source of novelty so that there can be an actual history—of both nature and human affairs—and a meaningful future. Lawful structure provides stability so that both natural events and human choices can have meaningful consequences over time. This general description resonates with everything we know from both the physical sciences and the human experience of moral and personal life.
Of course, God does not actualize any possible world but rather creates an initial world state, with inherent propensities (deterministic and non-deterministic) which set the stage for the development of successive world states. In so doing, God sovereignly limits himself and leaves some of the details of the future unspecified. We might say that God actualizes a world type, a certain kind of world, with certain beings and certain structural features forming the fabric of their existence, and then allows the world—in its physical and personal dimensions—to develop without his ensuring that only those evils occur which are necessary to some greater good. In Christian theology, kenosis (Greek; Phil. 2:7) refers to the divine act of “self-emptying” which is involved in the Christ event. But a common application of this concept pertains to God’s refraining from overwhelming creation with his power by allowing an “open space,” so to speak, in which genuine relationship between God and finite persons can occur.
A world structured with a significant range of openness makes possible the great goods that can come only through proper relationship with God and others. This kind of world is indeed a very great good. But the opportunity for such great goods is commensurate with the possibility of great evils: the high degree of freedom in creation makes for real risk. Clearly, Christian theological teachings about sin and the fall are woven into the Christian story to account for the fact that all of creation, with its obvious and abundant fundamental and residual goods, is now damaged, wounded, and warped. Things are not as they should be: the world could certainly be better if events had not unfolded as they did. The Christian theist who holds this perspective can agree with the atheistic critic that not every bad thing that happens is connected to a greater good.
There is indeed a strangeness to the conditions of our existence: opportunity and risk linked as part of the divine plan for governing the world. God’s providence, at one level, then, amounts to following certain intrinsically valuable policies regarding how human beings (rational-soulish animals) may attain the highest types of goods and what scope must be given to freely willed actions that are successfully executed. Created reality involves a physical environment that must run by natural laws that form a relatively stable framework in which personal decision and action can take place. But this means that there will be occasions in which the regular operations of nature interfere with human interests and aims—giving rise to what we often call chance or contingency. God cannot frequently contravene the operation of natural laws without creating, as Peter van Inwagen terms it, “massive irregularity,” which is a defect in a world, indeed, a great evil. In the personal realm, in which the power of libertarian freedom is exercised, there is likewise the contingency of choice and its consequences. God cannot generally be interfering in this sphere and still maintain meaningful freedom. All of this entails that God’s providence is general rather than meticulous; it does not entail that God is aloof or unavailable for intimate relationship, or that nature and humanity are entirely on their own.
Thus, the possibility of both natural and moral evils is inherent in the structure of creation, including those evils that the world would be better off without. So, the possibility of gratuitous evil is necessary in a good type of world. Christian theists should not attempt any answer to the problem of evil that entails that every particular evil is necessary to a greater good or that God has specifically allowed every case of suffering. Instead they should focus on articulating what sorts of reasons the God of the historic Christian faith might have for allowing His creatures to live in a world with the structural features outlined here. Articulation of God’s general policies for the governance of the universe should not minimize the suffering of individual persons as a result of how contingencies play out. This accents what is right in the intuition underlying the Person Centered Requirement: that a good God would allow an individual to undergo serious suffering only if it is necessary to achieve some benefit for the individual that will be included in his or her conscious experience. Such a requirement on God is perhaps still too much under the spell of Meticulous Providence. It is better to admit that, in a contingent and chancy world, benefits and burdens, pains and pleasures, will be distributed in ways that defy detailed explanation from an ultimate point of view. This lays the groundwork for developing the intuition in a more helpful direction: that the Triune God’s infinite goodness, mercy, love, and justice—as revealed in the sweep of biblical revelation and most particularly in the life, death, and resurrection of Jesus—entail that, given the contingencies of evil, God works redemptively with what happens to us, that no ultimate harm comes to any individual for reasons that are outside of his or her control, and that God’s plan of seeking all who will participate in the Divine Life will not be thwarted. We don’t have a formula for how all of this will go; but we have a confidence rooted in Christian knowledge and experience which gives rise to hope.
RE-EVALUATING ROWE’S ARGUMENT
As more Christian philosophers reject Rowe’s restrictions and systematically mine the intellectual resources of the classical Christian theological vision, it will become increasingly obvious that the effectiveness of the evidential argument from evil has been overestimated. In “Friendly Atheism, Skeptical Theism, and the Problem of Evil,” Rowe continues this overestimation. “In an age of faith,” he observes, “before the growth of the scientific knowledge that produced alternative, credible explanations of the emergence of human life,” traditional theism struggled with the problems of evil and divine hiddenness but persisted. The idea of God continued to play an essential explanatory role because it was met by no credible alternative. However, “in this age of reason and science,” Rowe continues, “the idea of God no longer plays an essential, rational role in explaining the world and human existence.” With these comments about reason and science as prelude, he then summarizes his own reasoning: given horrendous evils, divine absence is evidence that there is no God. For him, the best explanation of reality, all things considered, is atheism. Or, since atheism is not much of an explanation of anything, Rowe presumably means atheism’s default worldview in secular Western culture: philosophical naturalism.
Indeed, Rowe simplistically characterizes the “age of faith” as premodern and dominated by theistic thinking while describing the modern alternative as an intellectually sophisticated combination of atheism, science, and naturalism. It would be fairer to pit Restricted Standard Theism (RST) against something like Restricted Standard Atheism (RSA) as opposed to the “expanded” version of atheism smuggled into this passage. But this game of restrictions just repeats old mistakes. It would be more enlightening to allow Christian understanding to interpret and contextualize theism as well as to allow atheism to invoke the fuller intellectual resources it needs to construct a competitive comprehensive naturalistic worldview alternative. This larger debate transcends the academic exercise of arguing that evil lowers the probability of minimal theism and revolves around the more serious questions of confirmation and disconfirmation of worldviews.
An intellectually robust Christian theism will display the explanatory power of the Trinitarian Godhead as ultimate reality and of all that Christian doctrine tells us about his ways and his world. Atheistic naturalism can likewise explain the way things are based on positing brute nature, material stuff, as ultimate reality and deploying other intellectual resources in its orbit. In this larger debate, one wonders how Rowe can support quite a number of important realities to which he is committed—free will, agent causation, and objective moral values—within a purely naturalistic framework. Such realities are, of course, very much at home in a universe described by Christian theism. And then there are the ultimate questions which must be answered: Why is there anything at all? Is rational thought appropriately reducible to material processes? Why is nature accessible to rational thought? Is the existence of finite personal-moral beings fully explicable by reference to impersonal matter? Christian theism certainly does not take a back seat to naturalism in credibly addressing such questions. In short, it is not difficult to see that Christian theism could be argued to have a high degree of confirmation relative to naturalism.
Along these very lines, Rowe’s statement that science helps tip the rational scales toward atheism is curious in light of his Bayesian restriction on k to knowledge that, on balance, leaves the probability of God’s existence at 0.5. Among the things k includes, Rowe explains, is our “common knowledge of how the world works,” which would presumably include our knowledge of science. And he does say that k may include some items that raise the probability of theism and some that lower it. Apparently, Rowe thinks that science in k favors atheism. It is hard to understand why he believes this. It could be for the Ockhamistic reason that science explains things with fewer kinds of entities. But this commits the category mistake of thinking that Christian theism and science explain the same kinds of things when in fact they provide very different kinds of explanations of very different kinds of things. Indeed, in worldview comparisons, science itself and all of the conditions that make it possible—from the existence and lawful order of nature to the reliability of human reason which investigates it—become data for metaphysical explanation. One relevant consideration would be the respective likelihoods, the probabilities of science as we know it coming to be by way of these two rival metaphysical hypotheses. Let S be science, CT be Christian theism, and N be naturalism. The question is, then, which is greater: P(S/CT) or P(S/N)? Although it is difficult to make precise judgments about such matters, Trinitarian Christian theism, once again, does not obviously have low probability and can be easily be argued to have higher probability than naturalism here.
Without minimizing the problem that evil presents for religious belief at many levels or discounting Rowe’s careful work articulating the problem as he sees it, the debate must now be put on different footing. Christian theism explains much about evil and much about other key realities. Whether or not some pieces of evidence taken in isolation lower the probability of theism is irrelevant to whether Christian theism is confirmed by the total evidence. Given that Christian theism has a respectable degree of confirmation on the total evidence, its explanation of evil in the world must be taken more seriously. And our estimation of the strength of the evidential argument from evil must be significantly moderated.
Michael L. Peterson, “Christian Theism and the Evidential Argument from Evil,” in Philosophy and the Christian Worldview: Analysis, Assessment and Development, ed. David Werther and Mark D. Linville (New York: Continuum, 2012), 175-95.