The Problem of Evil
1. Is There A Logical Argument From Evil?
Does the existence of pain and suffering pose a substantial objection to the existence of God?
This question is at the heart of the atheistic argument from evil, and it has become popular to distinguish two forms of this argument: logical and evidential. The logical (or deductive) argument claims that the fact of evil is logically incompatible with the existence of an omniscient, omnipotent, wholly good God. The evidential (or inductive) argument claims that, while at least some evil is logically compatible with the existence of God, the amount and diversity of evil that exists constitutes evidence against the existence of God, making theism improbable.
1.2 The Shift in the Literature
The debate in the recent literature about the problem of evil has shifted from preoccupation with the logical argument to an increased focus on the evidential argument. Thus William P. Alston (1996) claims that "it is now acknowledged on (almost) all sides that the logical argument is bankrupt, but the inductive argument is still very much alive and kicking" (97). He goes so far as to refer to the logical argument as the "late lamented deductive cousin" of the inductive argument (121). With respect to the contention that "the existence of evil is logically inconsistent with the existence of the theistic God," atheist William L. Rowe (1990) claims that: "No one, I think, has succeeded in establishing such an extravagant claim" (126 fn. 1). Stephen J. Wykstra (1990) concurs with this assessment, saying that Rowe's atheistic argument "exemplifies the recent turn away from 'logical' (or 'deductive', or 'demonstrative') formulations, construing the case instead as 'evidential' (or 'inductive', or 'probabilistic') in nature" (138).
Why this shift away from the logical argument from evil? The greatest catalyst for this development, acknowledged on all sides, is Alvin Plantinga's free will defence, set forth with rigorous precision in ch. IX of his The Nature of Necessity (1974a), and in a somewhat more popularised form in his God, Freedom, and Evil (1974b). On the basis of this defence Alston claims that, "even granting middle knowledge, Plantinga has established the possibility that God could not actualize a world containing free creatures that always do the right thing" (113). Rowe refers his readers to Plantinga for "a lucid statement" of the argument for the logical compatibility of God and evil (126 fn. 1). And Wykstra, commenting upon Philo's logical argument from evil, as Hume presents it in the Dialogues Concerning Natural Religion, claims that "in our day the work of Plantinga and others has made much clearer the import of such broadly logical constraints, making this talk by Philo (or, only a few decades ago, by Mackie) of 'decisive disproof' look like naïve bluster" (158).
One writer who does not concur with this assessment is Richard Swinburne (1998):
It seems to be generally agreed by atheists as well as theists that what is called 'the logical problem of evil' has been eliminated, and all that remains is 'the evidential problem'. See e.g. Paul Draper, who writes that he 'agrees with most philosophers of religion that theists face no serious logical problem of evil' ('Pain and Pleasure: An Evidential Problem for Theists', Nous, 23 (1989), 331-50: 349). But whether that is so depends on what we understand by 'the logical problem'. It has not been shown to the satisfaction of atheists that there is no valid deductive argument from the existence of certain evident bad states E (via some necessary moral truths) to the non-existence of God. It has been shown merely that there is no such valid deductive argument evident to theists, who dispute the validity of any such argument by disputing the necessity of the relevant purported necessary moral truths (20 fn. 13).
But, according to the recent literature, it is not 'merely' the theists who advance this claim. At least some atheists agree with the theists in "disputing the necessity of the relevant purported necessary moral truths." To be sure, whether the logical problem "has been eliminated" for an atheist will depend upon what set of beliefs that atheist (contingently) holds. But that is simply a consequence of the person-variable nature of rational persuasion, and is applicable to all arguments whatsoever.
1.3 The Failure of Plantinga's Defence
Although I happen to agree with the near-universal consensus that the logical argument from evil is a "late lamented deductive cousin" of the evidential argument, and thus will proceed in the remainder of this essay upon that basis, I do so for different reasons than the writers previously mentioned. I do not think that Plantinga has provided a refutation to the logical argument, at least a refutation that is open to most theists. To see this, let us briefly rehearse a crucial transition in Plantinga's argument (1974a).
According to Plantinga, the FWD must show that the following is possible:
To do this, the FWD "must demonstrate the possibility that among the worlds God could not have actualized are all the worlds containing moral good but no moral evil" (185). But, says Plantinga, the FWD can do precisely this by an appeal to the concept of 'transworld depravity': if an essence E suffers from transworld depravity,
then it was not within God's power to actualize a possible world W such that E contains the properties is significantly free in W and always does what is right in W. Hence it was not within God's power to create a world in which E's instantiation is significantly free but always does what is right (188).
Plantinga then advances the claim that "it is possible that every creaturely essence suffers from transworld depravity." The conclusion is that it was then "beyond the power of God himself to create a world containing moral good but no moral evil" (188-9). Without this latter claim, it would always be open to the atheist to ask, Why did God not create other people? "Instead of creating us, he could have created a world containing people all right, but not containing any of us. And perhaps if he had done that, he could have created a world containing moral good but no moral evil" (187).
Now the project of the individual Free Will Defender is presumably to defend, not the existence of a generic God in the abstract, but the existence of the God who actually exists, whose actual existence is asserted by the FWD. Let us concede (for the sake of argument) that Plantinga has successfully shown the compatibility of the following two propositions:
His argument also shows that (1) and (31) taken together are consistent with:
But what is to prevent the atheist from revising the FWD's initial claim to:
1*. God is omniscient, omnipotent, and wholly good, and Jesus was a sinless human being
Very few Christian theists in the history of the church would dispute (1*). Ecumenical orthodox doctrine, as defined by the first seven councils of the church, would require it. But now our (1*) is incompatible with (31), and so the FWD fails. In other words, Plantinga's FWD can only be successful at the expense of excluding a central truth of Christian theism. Perhaps it would be of some use to, say, Jewish theists. But despite appearances, it cannot consistently form part of a cumulative case for Christian theism.
To be sure, Plantinga's initial programme for his FWD was not to defend the sinlessness of Jesus, or even to assert that such a doctrine was relevant to theodicy. But surely it cannot escape the notice of the observant atheist that, just as (1) is merely an expansion of the Christianclaim that 'God is omniscient,' so (1*) is merely an expansion of the Christian claim that (1). After all, it was the specific claims about God - that he is omniscient, omnipotent, and wholly good - that generated the prima facie contradiction with the fact of evil in the first place. Presumably belief in the gods of the Greek and Roman pantheons, and perhaps in the deity of some forms of process theology, does not face the same sort of problem from the fact of evil.
Of course, the point is not to ask whether Plantinga has (1*) in his belief set. The point is, if a particular theist has (1*) in his belief set, then that theist cannot adopt Plantinga's Free Will Defence. He must find a defence or theodicy that is appropriate to the Christian scheme of which his theodicy forms a part. It must be person-variable (in the reverse of Mavrodes' sense), variable according to the convictions of the one offering (not necessarily receiving) the argument. Subjectively, this means the theodicy must be consistent with the theodicist's own theistic convictions. Objectively, this means the theodicy must be consistent with the kind of God on whose behalf the theodicy is being offered. Admittedly, this approach appears to introduce unnecessary complications into the coherence of the theism to be defended, since Christian doctrines can interact with and define one another in complex ways. But (assuming these complications can be addressed) it may also have the happy consequence of introducing metaphysical and epistemological resources not otherwise open to the theodicist. We shall find in due course that contextualising a theodicy within the particular claims of the Christian scheme is exactly what several recent philosophers of religion have done, to a greater or lesser degree. After responding to objections against the cogency of this approach, I shall ultimately recommend it (cf. § 3.3.2).
I regard the above considerations to be an application of Plantinga's own "Advice to Christian Philosophers" (his 1983 inaugural address as John A. O'Brien Professor of Philosophy at the University of Notre Dame). There he calls Christian thinkers to display more "autonomy and integrality" in their work. Part of what he means by 'autonomy' is that Christian philosophers "may have to reject widely accepted assumptions as to what are the proper starting points and procedures for philosophical endeavor." Part of what he means by 'integrality' is that Christian philosophers are to display throughout their work "integrity in the sense of integral wholeness, or oneness, or unity, being all of one piece," lest they "seriously compromise, or distort, or trivialize the claims of Christian theism."
1.4 The Greater Good Defence
But having argued that Plantinga's FWD fails against the logical argument from evil, I will now argue that this logical argument fails against a more general defence. For it is clear that FWDs are simply a species of a broader genus, the so-called 'greater good defence.' . A FWD forwards the specific claim that a world containing creatures possessing free will (in the strong libertarian sense) is more valuable, other things being equal, than a world without such creatures. Since it is a logical truth that not even God can determine the choices of creatures endowed with free will in this sense, the genuine possibility of evil is logically required for the greater good of the world. So what the FWD comes down to is the twofold assertion that:
For reasons of my own I doubt that (b) can function cogently in a greater good defence. Significantly, and consistent with the conclusion of this essay, some of these reasons are similar to the reason I earlier rejected Plantinga's specific formulation: an inability to successfully contextualise (b) within the larger scheme of Christian doctrines. But in any event, I find no reason why the theist needs (b) in order to successfully deploy a more general version of the greater good defence against the logical argument from evil. He can simply add (a) to make propositions (1)-(4) consistent:
This, I take it, is a sufficient response to the logical argument from evil. Of course, it may not satisfy all atheists, but as with Swinburne's remark about atheists who reject Plantinga's dismissal of the logical FWD, that is a matter of persuasion, not proof. Theists can always attempt to make (a) more persuasive by sketching scenarios that convey possible applications of the principle it embodies. But it is always open to the atheist to declare that such scenarios, even taken together, do not produce a morally compelling reason for the particular type of evil he is considering at the moment, and so the defence against the logical argument will fail for him.
1.5 The Connection Between the Two Forms of Atheistic Argument
And here, precisely, is where we see that both forms of the atheistic argument stand and fall together. Whereas Swinburne (if I have understood him correctly) might be inclined to argue that the first form of the problem has not really been vanquished, I would argue that the two forms of the problem are only superficially distinguished. For it is the question of the appropriateness of this more general greater good defence against the logical argument from evil, that is at the centre of the most recent controversies concerning the evidential argument from evil. The reason why the atheist finds so implausible the theist's simple assertion that God has a morally compelling reason for permitting the existence of evil, is that the atheist wants to know what that reason is, especially in the face of so much apparently pointless pain and suffering. But this leads us directly into the evidential argument from evil, from the fact of gratuitous evil to the disconfirmation (but not the strict incoherence) of theism. As we shall see, the strategy of recent theists has been to point out that, in light of the human cognitive condition and the complexities of the divine plan, the atheist's question is inappropriate. But if it is inappropriate in the context of the evidential argument, it is equally inappropriate in the context of the logical argument. Arguing from the fact of evil (logical argument), and from the amount and variety of evil (evidential argument) equally presuppose what recent theists in the literature are prepared to deny: that the atheist has grounds for knowing that God does not have a morally compelling reason for permitting the existence of evil. It is the fact that the atheist asserts he has these grounds that generates both 'forms' of the problem of evil.
2. The Wykstra-Alston Hypothesis Against the Evidential Argument From Evil
2.1 Rowe's Evidential Argument From Evil
With these concerns in mind, let us follow the shift to the evidential argument from evil, and the recent response to it, which I will ultimately formalise as the Wykstra-Alston hypothesis (cf. § 2.4).
Both Rowe (1990, 127-8) and Wykstra (1990, 140) start from the following version of the argument:
Given the validity of the above argument, and basic agreement between theists and atheists on the truth of (2), the crux of the debate is over the truth of (1), and whether we are in a position to give evidential support for it. While Rowe concedes that he cannot prove (1) with any kind of certainty, he believes that he can identify instances of apparently gratuitous suffering in the world that rationally support (1). Since 'it appears' to Rowe that there is gratuitous suffering in the world, this is evidence for the truth of (1), thereby making (1) reasonable to believe. And to that extent, the atheistic conclusion (3) is rationally supported and reasonable to believe.
2.2 Wykstra's ResponseWykstra counters this by finding in Hume's Dialogues "the main obstacle to any evidential case from suffering against 'theism'" (Wykstra, 1990, 139). This obstacle is the inevitable failure of the atheist's 'it appears' claims to satisfy what Wykstra formulates as CORNEA, the Condition of Reasonable Epistemic Access:
On the basis of a 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 (152).
one purposed by the Creator of all that is, whose vision and wisdom are therefore somewhat greater than ours... that we should discern most of them [the outweighing goods] seems about as likely as that a one-month old should discern most of his parents' purposes for those pains they allow him to suffer - which is to say, it is not likely at all (155-6).
claims only that if there is an omniscient and wholly good being, then there must be outweighing goods served by every instance of suffering he allows. This claim does not involve any presumption that we can know in any given case just what these goods are (141).
2.3 Alston's ResponseWilliam Alston has recently endorsed and extended a generalised version of the above reasoning. Joining Wykstra and others in their extensive pessimism concerning the human cognitive condition, he suggests "that the atheological argument from evil is vitiated by an unwarranted confidence in our ability to determine that God could have no sufficient reason for permitting some of the evils we find in the world" (102). Alston's article is unique because he produces two lists that are pertinent to the debate. The first list concerns the possible theodicies available to the theist:
The two lists are related in that the items on the second list make it quite improbable that the atheist is in a position to exclude the possible divine reasons on the first list, when examining any particular instance of evil in the world. Thus he is never warranted in asserting premise (1) of Rowe's argument. But by the same reasoning the theist, because subject to the same cognitive condition, can never confidently identify an item on the first list with any particular instance of evil in the world. He must rest content with possible theodicies. Alston's thesis is truly agnostic in this respect, which is why he repeatedly reminds his readers that he is not endorsing any specific theodicy, only a range of reasons which God may have for permitting any particular instance of evil.
2.4 The Wykstra-Alston Hypothesis
Combining the arguments of both Wykstra and Alston, let us formalise what I shall call the Wykstra-Alston Hypothesis. This response to the evidential argument from evil asserts:
It may seem that these two propositions are in considerable tension with one another; at first glance (2) seems to undermine the confidence inherent in (1). But it must be kept in mind that (1) refers only to possible divine reasons for any instance of evil. In its present form (1) actually functions to support (2), in that it undermines the atheist's confidence that he has exhausted the possible divine reasons for any instance of evil.
3. Assessing Swinburne's Recent Criticisms
In Providence and the Problem of Evil, Richard Swinburne (1998) has subjected part (2) of the Wykstra-Alston Hypothesis to a vigorous critique. In a section entitled, "Does the Theist Need to Prove his Case?" (14-29), Swinburne makes clear his difficulties with this particular response to the evidential argument from evil.
To be sure, Swinburne does adopt a more confident version of (1), and the remainder of his book is largely an expansion of that project. Following the general outlines of a greater good defence, Swinburne argues for the probability that 'the good goals of creation' (beauty, thought, feeling, action, worship) necessarily require the possibility of the moral and natural evils that actually exist. But one of the reasons that Swinburne sets forth his case at such length is that he believes, contrary to (2), that there are evils in the world which do have evidentiary force against theism, and this evidentiary force must be nullified through specific counterargument.
I shall not attempt to pursue the sufficiency of Swinburne's theodicy here, although I have briefly given some reasons why I doubt that a free will theodicy can be a cogent specification of a greater good defence for Christian theism (cf. fn. 10). Rather, in the remainder of this essay, I shall focus on the necessity of Swinburne's theodicy. This necessity hinges, of course, on the falsity of (2).
Swinburne's objections to (2) are twofold. With respect to (2a), Swinburne argues that Wykstra and Alston misuse the Principle of Credulity. With respect to (2b), Swinburne argues that Wykstra and Alston are guilty of begging the question against the atheist with respect to the existence of God and the complexities of his divine plan. I shall assess Swinburne's criticisms in turn.
3.2 Does the Wykstra-Alston Hypothesis Misuse the Principle of Credulity?
3.2.1 Assessing Wykstra's Contribution
Swinburne (1998) takes the Principle of Credulity to mean
that, other things being equal, it is probable and so rational to believe that things are as they seem to be (and the stronger the inclination, the more rational the belief). By 'seem' (or 'appear') I mean 'seem epistemically'; the way things seem epistemically is the way we are initially inclined to believe that they are. We find ourselves with involuntary inclinations to belief; in the absence of reasons against going along with such an inclination the rational person will do so... The Principle of Credulity applies not merely to the deliverances of sense but to the way things seem morally, mathematically, or logically (20-22).
Swinburne then argues that the Principle may be applied to the evidential argument from evil, saying that it follows from the Principle
that, if it seems to someone that there is some bad state incompatible with the existence of God, he ought so to believe, and so believe that there is no God - in the absence of counter-reasons... So, he must conclude, probably there is no God. All this, unless he has evidence that seems to him to tell against that conclusion... And if our understanding of possible reasons why anyone might allow suffering to occur provides us with no reason for supposing that a good God might allow certain suffering, we ought to believe that there is no God - unless we have contrary reason (22-23).
However, I question this application of the Principle to the evidential argument. In his fuller discussion of the Principle of Credulity in The Existence of God, Swinburne is clear that the Principle is formulated for the validation of perceptual claims based on mere sensory experiences. While I think the Principle is valid in this sphere, I share Wykstra's doubts (1990, 154) that it has the same cogency when applied to cognitive-epistemic appearings. Wykstra simply grants this application for the sake of argument with Rowe, but I would do so very reluctantly, for I do not see the argument in Swinburne for the legitimacy of this extension. Even in the discussion in Providence and the Problem of Evil, concerning the Principle's application to the problem of evil (20-22), the only intuitive justifications given for the Principle are strictly in the context of sensory experience.
This consideration by itself may be enough to block Swinburne's later 'ad infinitum' argument (1998, 26) against Wykstra's use of CORNEA. Although it was unwise for Wykstra to give CORNEA the appearance of an "unrestrictedly general epistemological principle," nevertheless it functions more narrowly in Wykstra's actual argument. Distinguishing between a 'sensory-epistemic appearing' and a 'cognitive-epistemic appearing', Wykstra holds that the latter is the proper description of 'it appears' claims that function in the context of the evidential argument from evil:
And this [the cognitive-epistemic sense] seems to be the way Rowe uses the term: in appealing to the fawn's suffering he is, after all, appealing not to a 'sensory experience', but to a 'cognized situation' (akin to the chess master's visualized board), which he takes to be uncontroversially instantiated in the real world, and which produces (via a range of other considerations) a strong inclination to believe a proposition about such suffering (viz., that it serves no God-justifying purpose) (1990, 154).
If this is the case, and if an (unmodified) Principle of Credulity doesn't apply to cognitive-epistemic appearings, then the Principle is irrelevant as a means of defusing Wykstra's argument.
But even granting (in the absence of supporting argument from Swinburne) that there is a legitimate application of the Principle of Credulity to those cognitive-epistemic appearings which are relevant to the evidential argument from evil, I question whether these appearings should be phrased in positive, rather than negative, terms. Swinburne, as we saw earlier, applied the Principle in this context to a positive seeming: "if it seems to someone that there is some bad state incompatible with the existence of God, he ought so to believe..." (22, my italics). And, to be sure, premise (1) of Rowe's argument is phrased in positive form:
But I follow Alston's interpretation, to the effect that what is being asserted in this premise is (despite phrasing) a negative existential statement:
To be sure, 1 is in the form of a positive existential statement. However, when we consider an instantiation of it with respect to a particular case of suffering, E, as Rowe does in arguing for it, it turns out to be a negative existential statement about E, that "there is no sufficient divine reason for permitting E." It is statements of this form that, I claim, no one can be justified in making (1996, 122 fn. 12).
If this is the case, then even if the Principle can be extended to cognitive-epistemic appearings, additional conditions need to be satisfied if the atheist's 'appears' claim is to be granted. Wykstra (1990, 150) notes the following common-sense limitation to the Principle of Credulity, given by Swinburne himself (1991, 254-55). I shall quote the paragraph in full:
Note that the principle is so phrased that how things seem positively to be is evidence of how they are, but how things seem not to be is not such evidence. If it seems to me that there is present a table in the room, or statue in the garden, then probably there is. But if it seems to me that there is no table in the room, then that is only reason for supposing that there is not, if there are good grounds for supposing that I have looked everywhere in the room and (having eyes in working order, being able to recognize a table when I see one, etc.) would have seen one if there was one there. An atheist's claim to have had an experience of its seeming to him that there is no God could only be evidence that there was no God if similar restrictions were satisfied. They could not be - for there are no good grounds for supposing that if there is a God, the atheist would have experienced him.
If this argument from The Existence of God is taken seriously, and thus if these restrictions upon the evidential import of 'negative seemings' are adopted, it will derail Swinburne's attempt to apply the Principle of Credulity to the problem of evil (and this is the case even if the Principle is relevant to cognitive-epistemic appearings). In particular it will decisively block Swinburne's ad infinitum argument (26) against Wykstra's amendment to the Principle (i.e. CORNEA), since the ad infinitum argument is formulated in terms of positive seemings.
Wykstra does attempt to offer Swinburne some help at this juncture. For Wykstra argues, with respect to the above-quoted passage from the Existence of God, that Swinburne's distinction between positive and negative seeming "depends upon rather arbitrary choices of formulation" (151). But Wykstra's counterexample irrelevantly focuses on the locution used to express the seeming, rather than focusing on the state of affairs to which the locution refers. Thus, on Wykstra's analysis, the 'negative' formulation "it seems that no table is in the room" could just as well be expressed positively: "it seems that the room is bare or empty." Since both locutions assert the same state of affairs, Wykstra charges Swinburne's distinction with arbitrariness.
But this misses Swinburne's original point in the above-quoted passage. Swinburne employed the positive/negative distinction among 'seemings' as a way of distinguishing the seeming of two distinct states of affairs, regardless of the language used to describe that seeming. For surely the state of affairs referred to by "it seems to me there is present a table in the room" is clearly different from the state of affairs referred to by "it seems to me that there is no table in the room." (In the first the existence of a table is asserted; in the second it is denied.) And it is in the case of this latter (negative) seeming that Swinburne proposes a qualified application of the Principle of Credulity: there must also be "good grounds for supposing that I have looked everywhere in the room, and ... would have seen one if there was one there." This is truly a common-sense qualification, and it is a pity Wykstra missed it, for the disparity Swinburne recommends in the two-fold application of the Principle of Credulity simply reflects the familiar dictum of inductive logic that (ceteris paribus) it is more difficult to prove a universal negative, than to prove an existential instantiation.
Let us then generalise from Swinburne's qualification to the Principle of Credulity (in the case of negative seemings). Generalising from Swinburne's example of the table in the room, the qualification is that one needs good grounds for supposing one both possesses and has executed the capacity to search the relevant vicinity before one is justified in asserting "it seems to me that there is no x in the relevant vicinity." First, one must possess the capacity to search the relevant vicinity; I would have seen a table if one were there. Second, one must have executed this capacity to search the relevant vicinity; I have looked everywhere in the room. When stated in this way (and I see no principial objection to this generalisation), it becomes clear that the heart of the Wykstra-Alston response to the evidential argument from evil trades upon this very qualification to the Principle of Credulity that Swinburne himself endorses. Since it is quite unlikely that we possess the capacity to bring the "outweighing good" served by the fawn's suffering "within our ken," then neither Rowe nor anyone else can offer any evidential support for premise (1) of Rowe's argument.
3.2.2 Assessing Alston's Contribution
But now let us suppose, again for the sake of argument, that Wykstra's application of CORNEA is an utter failure, an overly bold and therefore indefensible amendment to the Principle of Credulity. What of the weaker hypothesis that Alston proposes in its place? Rather than "proceeding on the basis of any such unrestricted epistemological principle," Alston says the more proper response
focuses on the peculiar difficulties we encounter in attempting to provide adequate support for a certain very ambitious negative existential claim, viz., that there is (can be) no sufficient divine reason for permitting a certain case of suffering, E. I will be appealing to the difficulties of defending a claim of this particular kind, rather than to more generalized human cognitive weaknesses (102).
That is, Alston (unlike Wykstra) is willing to initially grant the atheist his 'it appears' claims. But, Alston argues, a moment's reflection on the part of the atheist will rob premise (1) of Rowe's argument of any rational support. The atheist is perhaps initially entitled to say, "it appears that God could not have a morally compelling reason for permitting this evil." But when he is led to reflect upon, more particularly, his cognitive limitations vis-à-vis the complexity of God's plan, to reflect upon his inability to determine whether or not an omniscient God could have a sufficient reason for permitting this evil, he should realise that this very cognitive limitation is a reason that tells against the initial 'appears' claim. The common-sense qualification Swinburne makes to the Principle of Credulity has been fulfilled in the course of this very exercise.
Swinburne would disagree; he alleges that even if Alston is right that our moral and logical intuitions may be in error when examining an instance of evil, this is a worthless point, for Alston has still given us no "reason to suppose that our error is more likely to lie in the one direction rather than in the other" (1998, 28). That is, if Alston's argument is to be consistent, it must call into question all of our moral and logical intuitions. We need "positive argument for supposing that certain appearances rather than others are misleading" (1998, 28).
Three replies are in order here.
184.108.40.206 Alston Defines the Limits
First, this criticism overlooks the explicit restriction Alston places upon his scepticism, as well as his rationale for that restriction. The heart of Alston's argument is that
the critic is engaged in attempting to support a particularly difficult claim, a claim that there isn't something in a certain territory, while having a very sketchy idea of what is in that territory, and having no sufficient basis for an estimate of how much of the territory falls outside his knowledge. This is very different from our normal situation in which we are forming judgments and drawing conclusions about matters concerning which we antecedently know quite a lot, and the boundaries and parameters of which we have pretty well settled (120)... [The point] is that the judgments required by the inductive argument from evil are of a very special and enormously ambitious type and that our cognitive capacities that serve us well in more limited tasks are not equal to this one" (124 fn. 36) (my italics).
Alston is zealous to maintain this explicit restriction upon his scepticism. Thus,
having carefully examined my desk, I can infer 'Jones's letter is not on my desk.' But being ignorant of quantum mechanics, I cannot infer 'this treatise on quantum mechanics is well done' from 'so far as I can tell, this treatise on quantum mechanics is well done' (102).
If the Principle of Credulity cannot be fitted to obvious counterexamples like the above, then the Principle as it stands has no authority with respect to the judgements that are relevant to the evidential argument from evil, since these counterexamples form a precise parallel to those judgements. Even as ignorance of the complexities of quantum mechanics disqualifies me from passing judgement on how well done is a treatise on the subject, so ignorance of the complexities of God's plan disqualifies human beings (in their present cognitive condition) from passing judgement on how well done is a world God has made.
220.127.116.11 Alston Appeals to Discontinuity
Second, this criticism overlooks the fact that Alston's argument appeals to a clear principle of discontinuity between our knowledge of human affairs and our knowledge of God's affairs, to a discontinuity between our ability to conceptualise possibilities for sufficient reasons and God's ability to conceptualise them. Thus Alston asks his readers to
remember that our topic is not the possibilities for future human apprehensions, but rather what an omniscient being can grasp of modes of value and the condition of their realization. Surely it is eminently possible that there are real possibilities for the latter that exceed anything we can anticipate, or even conceptualize. It would be exceedingly strange if an omniscient being did not immeasurably exceed our grasp of such matters (109, my italics).
Such an appeal to discontinuity is familiar to those conversant with Joseph Butler's apologetic strategy in The Analogy of Religion Natural and Revealed to the Constitution and Course of Nature (1736). From the title alone it might be inferred that Butler only appealed to observed continuities between religious claims and the 'constitution and course of nature,' but this is not the case. While Butler does make great use of reasoning about unknown possibilities from the known constitution and course of nature, he recognises (and, like Alston, even exploits to his advantage) the limitations of this method of reasoning.
Thus, Butler argues for human immortality by pointing out that although we have in our lifetime undergone much change, we have still survived. Therefore, it is likely that we will survive death. This is Butler employing the principle of continuity. But, Butler also points out that we do not know enough about death to say (with the materialist) that it involves a loss of our present powers. This is Butler employing the principle of discontinuity. (For both of these points, see "Of a Future Life, " in ch. 1 of the Analogy.)
We can see Wykstra's emendation to the Principle of Credulity, and Alston's argument generally, as being in the spirit of Butler's appeal to discontinuity. There are areas concerning which human beings (in their present cognitive condition) are not qualified to speculate. If this is true for some of us with respect to human endeavours like quantum mechanics; why is it not all the more true for all of us with respect to divine endeavours?
As one commentator has expressed Butler's strategy:
Analogy tries to show that special revelation is analogous to natural revelation. Disanalogy, or the argument from ignorance, tries to show that the unlikeness of the supernatural invalidates unbelieving objections: the unbeliever does not know enough to object to scriptural teaching. This too is, in Butler's estimation, a practical kind of argument, like those by which we make decisions in everyday life. He presses the non-Christian to be consistent: admit in the religious debate the same continuities and discontinuities that you freely recognize elsewhere. When you do that, he tells the inquirer, you will see that special revelation has the same cogency that you accept in natural revelation. And if there are problems in special revelation, they are no greater than the problems of natural revelation. Since you are able to bear with the latter, you should be able equally to bear with the former.
18.104.22.168 Alston Provides 'Positive Argument'
Third, I take Alston as having, in any case, fulfilled Swinburne's requirement; he has given "positive argument for supposing that certain appearances rather than others are misleading" (Swinburne, 1998, 28). Namely, those 'appearances' into which the concept of God enters as part of its description are misleading, to the extent that we have not reflected upon the relationship between the cognitive abilities of that God, and our own cognitive limitations. We can interpret Alston as holding that 'it appears that p' (where p is 'God could not have a morally compelling reason for permitting this evil') would be misleading, if the aspect of 'God' being considered is merely that he is a perfectly good being. But if, in addition to his perfect goodness, his perfect wisdom and power are equally considered - that is, if it is truly the concept of God (and not a scaled-down substitute) which enters into the content of p in 'it appears that p' - then cause for scepticism as to the rational acceptability of p is due to enter in. It was the burden of Alston's article to argue this point at length. He initially concedes to the atheist his 'it appears that p' only because he is convinced that further reflection on the content of p will make p well-nigh indefensible.
To formalise a bit, if we take the following definitions:
h = 'God could not have a morally compelling reason for permitting this evil'
e = 'it seems to me that God could not have a morally compelling reason for permitting this evil'
k = irrelevant background knowledge
c = considerations of our human cognitive condition vis-à-vis God's omniscience and omnipotence
Then I am convinced that Alston's argument is meant to show that, even if it is a fact of human psychology that P(h/e.k)> 1/2, what is important with respect to the evidential argument from evil is that Alston's considerations lead to P(h/e.k.c)W 1/2.
3.3 Does the Wykstra-Alston Hypothesis Beg the Question Against the Atheist?
3.3.1. Is Wykstra's and Alston's method circular?
Finally, we must address Swinburne's charge that the Wykstra-Alston Hypothesis begs the question against the atheist. Swinburne takes the following to be a conditional essential to Wykstra's argument: "if there is a God, then there will not be any apparent good states which really subserve greater bad states" (27). But, says Swinburne, "we cannot assume at this stage without positive argument that there is a God and so that there are no such states; for that would be to beg the question against the atheist who has produced the argument under discussion in order to show that there is no God" (27). I have a few things to say about this.
First, it does not appear that Swinburne has properly represented the precise conditional which functions in Wykstra's argument. It is not:
(1) If there is a God, then there will not be any apparent good states which really subserve greater bad states.
(1*) If theism is true, then the outweighing goods served by instances of suffering "are, in a great many cases, nowhere within our ken" (cf. Wykstra 159).
(1) is a definitional comment (an unpacking of the definition of 'God'), but (1*) is an epistemological thesis.
It is only something like (1*) which functions to defuse Rowe's evidential argument from evil. Wykstra interprets its significance:
Rowe is correct in seeing that outweighing goods served by instances of suffering "are, in a great many cases, nowhere within our ken. The linchpin of my critique has been that if theism is true, this is just what one would expect: for if we think carefully about the sort of being theism proposes for our belief, it is entirely expectable - given what we know of our cognitive limits - that the goods by virtue of which this Being allows known suffering should very often be beyond our ken. Since this state of affairs is just what one should expect if theism were true, how can its obtaining be evidence against theism? (159).
That is to say, it does not beg the question to point out that, given c, e is just what one would expect. So how can e be evidence against c?
Second, Swinburne's counterargument overlooksthe use to which Wykstra and Alston are putting their respective theodicies. Wykstra and Alston are not appealing to the limitations of the human cognitive condition vis-à-vis the divine plan, as positive evidence one way or another, evidence that would lead the atheist to consistently interpret bad states as leading to good states, but not good states as leading to bad states (as if it were a piece of natural theology). The Wykstra-Alston Hypothesis is truly agnostic in that respect; it is a purely negative thesis that blocks inferences from states of any kind. And this is all they need to defuse the evidential argument from evil.
3.3.2 What to make of circularity?
I close with the observation that assessing the presence and significance of circularity in one's apologetic method is a subtle matter. My own view is that, in defusing the problem of evil, we are entitled to draw upon the resources of divine revelation, since the warrant for that revelation (not to mention its intelligibility) partly depends upon the successful vindication of the existence of God from the atheistic argument from evil. Thus in any sound case for Christian theism, positive statement and negative defence go hand in hand, each strategy drawing for its cogency upon the resources of the other. So if the atheist claims that the Christian view of men and things is either incoherent (deductive argument) or improbable (inductive argument), because of the fact and variety of evil, then he needs to be shown that, given certain doctrines implicit in theism and in the broader Christian tradition, evil doesn't pose that kind of a problem at all
At worst, what Wykstra, Alston, and others are doing is begging the question against the atheist, in hypothetical form, unpacking the definition of Christianity for their open inspection.
For Wykstra, Rowe simply
has not thought (in the light of our evident cognitive limitations) about the sort of Being theism proposes, and so has not realized what is 'contained' in the theistic proposal (1990, 159).
The observed sufferings in the world do require us to say that there are outweighing goods connected to them that are entirely outside our ken, but this is not an additional postulate: it was implicit in theism (taken with a little realism about our cognitive powers) all along. If we have realized the magnitude of the theistic proposal, cognizance of suffering thus should not in the least reduce our confidence that it is true (1990, 160).
Alston deems it prudent, in defusing the problem of evil, to appeal to the wider religious tradition in which belief in God is usually contextualised:
There are obvious advantages to thinking of the inductive argument from evil as directed against the belief in the existence of God as God is thought of in some full-blown theistic religion, rather than as directed against what we may call 'generic theism.' The main advantage is that the total system of beliefs in a religion gives us much more to go on in considering what reasons God might possibly have for permitting E. In other terms, it provides much more of a basis for distinguishing between plausible and implausible theodicies. I shall construe the argument as directed against the traditional Christian belief in God (1996, 100).
I might be accused of begging the question by dragging in Christian convictions to support my case. But that would be a misunderstanding. I am not seeking to prove, or give grounds for, theism or Christianity. I am countering a certain argument against Christian theism. I introduce these Christian doctrines only to spell out crucial features of what is being argued against. The Christian understanding of sin, human life, God's purposes, and so on, go into the determination of what the critic must be justified in denying if she is to be justified in the conclusion that Sam's suffering would not have been permitted by God (1996, 104).
And in similar fashion, Thomas Tracy appeals to a specific Christian doctrine (the afterlife) to make his case that God's permission of victimization does not necessarily violate a deontological conception of ethical obligation:
Because relationship with God extends beyond death, victims of evil need not be excluded from sharing in the highest good for the sake of which moral freedom, and consequently their suffering, has been permitted. Rather, God can guarantee that each person has the opportunity to participate in this good, notwithstanding what that person may suffer. Victimization, then, will never constitute the final or decisive word on a person's life (1992, 311).
If the above is to be condemned as vicious circularity, then it is interesting to find that even Swinburne's method would admit of a certain measure of it. This applies to Swinburne's books on argumentative theism (take as a whole), as well as to The Providence of God and the Problem of Evil (as a single book).
It applies to Swinburne's books taken as a whole:
A similar circularity also applies to The Providence of God and the Problem of Evil, for in that book Swinburne reveals that,
I seek only to explore the consequences of supposing certain doctrines to be true, for the central doctrine which is the concern of this book - the doctrine of Providence (1998, xi).
In The Existence of God, Swinburne assumed "that theodicy does not need to bring in doctrines peculiar to different religions... in order to show that the occurrence of evil does not count against the existence of God." But now the situation has changed:
I am not fully convinced about that any more. In any case most other contemporary humans are a lot more likely to be convinced if theodicy does bring in such doctrines. So this book invokes relevant Christian doctrines, such as those mentioned above, which make claims about what God has done and will do. I seek to show how the whole Christian doctrinal package taken together - the claim that there is a God together with the claim that he has done and will do certain things for his creatures - faces the difficulty of delay and suffering on the way to the good things God plans for us (1998, xi).
Inevitably, one must draw upon all of the resources of Christian theology to provide an adequate theodicy.
Perhaps here we could proceed [in our philosophical work] without appealing to what we believe as theists; but why should we, if these beliefs are useful and explanatory? I could probably get home this evening by hopping on one leg; and conceivably I could climb Devil's Tower with my feet tied together. But why should I want to? (Alvin Plantinga, "Advice to Christian Philosophers," 1983)
 Following William P. Alston (1996, 126-7), I shall define theism in the 'narrow sense': the affirmation of "an omnipotent, omniscient, eternal, supremely good being who created the world," and therefore I shall use the terms 'God,' 'theism,' 'theist,' 'atheist,' 'agnostic' with reference to this narrow sense of theism.
 'FWD' will hereafter be used as an abbreviation for both 'free will defence' and 'Free Will Defender'.
 One is reminded of Daniel Dennett's amusing definition: 'to alvin' is "to stimulate protracted dispute by advancing a bizarre claim." Entry in The Philosopher's Lexicon (APA, 1978).
 Thomas Tracy (1992) makes a related point in critiquing Plantinga's free will defence. He notes that "if the defense is to be relevant to the actual world, then a case must be made not only that R is possibly true, but also that ours is a world that God could be morally justified in creating" (303). Similarly, invoking the bare possibility of transworld depravity, for every possible creaturely essence that God could have created, is irrelevant to theodicy if such a possibility does not in fact obtain in the actual world. Plantinga's permissive description of R - that it "of course, need not be true or known to be true; it need not be so much as plausible. All that is required of it is that it be consistent with p" (1974a, 165) - sets up his theodicy for conflict with the actual world the Christian theodicist is seeking to make compatible with the existence of God. See also Adams (1987, 89-90), for additional argument as to why an account of the relations between God and evil must in general be credible, as well as logically possible.
 cf. Swinburne (1998, 30-31) on this latter claim.
 Indeed, Plantinga himself (1996) appears to take his own advice, when arguing that Paul Draper's atheistic arguments from pain (and pleasure) "are probably mistaken: if we take into account all that a typical theist believes, including (in the case of Christian theists) beliefs about sin, Incarnation, Atonement, eternal life, and the like, the alleged incoherence seems to disappear" (260).
 Both Paul Helm and Alston agree with this classification. "It is important to bear in mind that the free-will defence can also be thought of as a greater-good defence, though one of a peculiar kind. It is, in effect, a limiting case of such a defence" (Helm, 1994, 198). Alston describes the FWD as asserting that "the permission of such horrors is bound up with the decision to give human beings free choice in many areas, and that (the capacity to freely choose) is a great good, such a great good as to be worth all the suffering and other evils that it makes possible" (Alston, 1996, 112).
 Cf. Plantinga (1974b, 30).
 Unless God employs his 'middle-knowledge' in this context. See Helm (1994, 55-61), and Adams (1987, 77-93), for reasons why I believe this route to be incoherent.
 Although this should properly be the focus of another essay, I will very briefly list my concerns here, without further expansion:
Repeatedly, Scripture seems to represent God as determining the choices of human creatures, and yet holding them fully responsible for their choices, good and bad (cf. Ge 50:20; Isa 10:5-15; Lk 22:22; Ac 2:23, 4:27-28; 13:48; Php 2:12-13; 1Ki 8:58, 61; Ex 4:21, 7:3, 10:20, 10:27). In other words, it doesn't seem that Scripture shares the distinctive assumption of the FWD, that "indeterministic free will is a necessary condition of human personality and accountability" (Helm, 1994, 195). To give up a theme so deeply woven into the fabric of Scripture seems too high a price to pay for the Christian theist.
Scripture seems to represent heaven as eternal and immutable bliss, populated by "the souls of just men made perfect" (Heb 12:23). But if strong libertarian freedom is removed in the heavenly life, then (on the libertarian hypothesis), the spectre of God fellowshipping with mere 'robots' returns. And if libertarian free will is an intrinsic good, the possession of which makes a world more valuable, then it would appear (again, on the libertarian hypothesis) that heaven is a less valuable state than earth. This is surely a counterintuitive conclusion.
Moral concerns. I'm not at all confident that the FWD does save God's goodness. It's not clear that it answers the end for which it was designed: to absolve God from blame for permitting evil. It is a widely shared moral intuition that taking unnecessary risks (fraught with the potential for great suffering) is morally blameworthy. For example, think of a reckless chemist who endangers millions of city dwellers by mixing unknown chemicals in the open air, simply because of the possibility (under those conditions) of devising a universally-efficacious health serum. Wouldn't we hold the chemist morally blameworthy? Similarly, it may be argued that God was to blame for even taking the risk of creation in the first place, since the future good (as well as the future evil) was unknown to him.
Philosophical concerns. I agree with Alston (1996), that "theodicists often confidently assert, as something obvious on the face of it, that a world with free creatures, even free creatures who often misuse their freedom, is better than a world with no free creatures. But it seems to me that it is fearsomely difficult to make this comparison and that we should not be so airily confident that we can do so" (124 fn. 37).
 This may, in the end, amount to saying the same thing.
 "Our experience and knowledge of the variety and profusion of suffering in our world provides rational support for the first premise" (132). Rowe offers his example of a fawn suffering from a forest fire as a paradigm case (129-31).
 Swinburne refers the reader to his "much fuller and so more adequate discussion of the Principle of Credulity, as it applies to apparent perceptual experiences" in The Existence of God, ch. 13 (Swinburne, 1991). But notice here Swinburne's important qualifications to the Principle: "other things being equal, it is probable and so rational to believe..."; and "in the absence of reasons against going along with such an inclination," the rational person will go along with his involuntary inclinations to belief. Significantly, Swinburne allows that one sort of these "counter-reasons" may be "general reasons for doubting the force of this sort of inference to his [i.e. God's] non-existence" (23). Wykstra's argument from CORNEA, if sound, would be classified as just this sort of legitimate qualification to the Principle of Credulity.
 Here again, we should immediately notice the stated qualifications to the Principle: "in the absence of counter-reasons... unless he has evidence that seems to him to tell against that conclusion... unless we have contrary reason."
 As Alston charges. cf. Alston (1996, 102). For this reason, Swinburne may be at least partly right, in his 'ad infinitum' argument, that not just any 'appears' claim can credibly fall under CORNEA. I say this, even though I find it curious that nowhere in his response to Wykstra does Swinburne show how his own formulation of the Principle of Credulity could be adapted to cover the very paradigm cases cited by Wykstra as the intuitive basis of CORNEA. If Swinburne rejects Wykstra's particular emendation of the Principle, then what emendation will he make? I take this as further evidence that the Principle of Credulity, as it stands, really isn't meant to regulate cognitive-epistemic appearings. (Unless our philosophically-ignorant and linguistically-challenged ceramics professor is entitled to say, of a sentence of philosophy spoken in Greek, that 'it does not appear that this sentence has any meaning'!!)
 Alston is not asserting that this lack of rational support for premise (1) gives rational support to the denial of (1). Again, his is a truly agnostic thesis, meant only to block the evidential argument from evil, and not meant to function as a piece of natural theology in its own right. cf. Alston, 1996, 121 fn. 2, 122 fn. 4.
 John Frame, Cornelius Van Til: An Analysis of His Thought (Phillipsburg: Presbyterian and Reformed Publishing Company, 1995), pp. 279-80. In this passage Frame is defending Butler's methodology against what he takes to be Van Til's misdirected criticisms.
 Cf. Swinburne (1991, 260), under "Special Considerations which Limit the Principle of Credulity."
 It was a failure to keep this in mind that rendered Plantinga's FWD virtually unusable.
 Unless the arguments for the existence of God, while advanced independently of the concerns raised by the problem of evil, provide in advance a "super strong" disconfirmation of that problem. Cf. Wykstra (1996). "When evidence e strongly disconfirms claim c in a way that makes it reasonable to believe, 'in advance' as it were, that none of the arguments for c are likely to counterbalance e, let us say e disconfirms c in a 'super strong' sense" (pp. 145-6). But if that were the consequence of The Existence of God, then the production of such a tightly-argued tome as Providence and the Problem of Evil seems superfluous, to say the least.
Last Revised: 19 July 1999