Divine Hiddenness: What is the Problem?

The problem of evil centers on the claim that some evil is so bad that God would not allow it. If evil has such a character, it might be claimed that evil constitutes evidence against the existence of God. Even more, it might be held that evil constitutes an epistemic problem for theism, that the presence of evil has the power to alter the epistemic status of theism. For even if there are reasons which in themselves might confirm theism, factoring in the presence of evil might decrease the status of theism from, say, being rational to believe to being counterbalanced. Though not universal, this account of the epistemic implications of evil is widely shared, an account on which evil constitutes evidence with sufficient epistemic force to alter the epistemic status of theism. On this conception, then, evil constitutes an epistemic problem for theism.

This account of the nature of an epistemic problem for a point of view focuses on the concept of change (that some piece of information has the power to alter the epistemic status of that point of view), and that isn’t quite right. The real issue is whether the new information has the power to make the epistemic status of a point of view different from what it would have been otherwise, different from what it would have been had that information not been present or available. Information can do the latter without ever changing the epistemic status of any point of view. For ease of expression, however, I will continue to speak with the vulgar on this matter by saying that epistemic problems are created by information with the power to change the epistemic status of some point of view.

In addition to this epistemic problem of evil, some have maintained recently that there is an additional epistemic problem regarding the existence of God, the problem of divine hiddenness. The most detailed investigation of this problem is that of J.L. Schellenberg who, in Divine Hiddenness and Human Reason, argues that the epistemic hiddenness of God presents a strong argument for atheism; in particular, that it yields an argument with the power to change the epistemic status of theism from counterbalanced to disconfirmed. ^{^[i]}

I believe there is a mistake underlying any such project, for I do not think that there is an epistemic problem of hiddenness. There are ways in which God is hidden, some of them evidential in character. But there is no epistemic problem of divine hiddenness. Nothing about the hiddenness of God has the power to change the epistemic status of theism.

Before arguing for this claim, I want to take it back, but only a little. For there is an epistemic problem of hiddenness if you agree with me about matters epistemic. Regarding such matters, I am a subjectivist, inclined toward coherentism, save for an articulated account of the possibility of justified inconsistent beliefs.^{^[ii]} I believe that what a person is justified in believing is not determined by what is true or objectively likely to be true nor by what the evidence actually confirms. Such facts can give way to a person’s subjective conception of the world and the

evidential connections perceived in that conception. The heroes of my epistemological faith are Keith Lehrer and Richard Foley.^{^[iii]} On their theories, and on any subjective theory, there clearly can be an epistemic problem of hiddenness. For on such theories, nearly any piece of information can be embedded in a cognitive structure in which that piece of information counts against the existence of God, and the epistemic status of theism could be diminished by coming to learn of that information and seeing its significance. All that is needed on such subjective conceptions, to repeat, is a cognitive structure that ties the two pieces of information together in an appropriate way. So on such conceptions, the existence of a pair of scissors on my desk can pose an epistemic problem for theism.

Just as true, however, is that solutions to such epistemic problems come in a wide variety for subjectivists. For almost any set of beliefs, there is some belief system in which those beliefs would have good subjective epistemic standing. Of course, there are some beliefs for which this may not be true--for example, it is not possible to be justified in believing (in the same way, at the same time) both p and not-p. But the set {God exists, God is hidden} is surely such that there are many ways of embedding it in a system of beliefs that has universally high subjective epistemic status. And the same point holds even for those who have read and fully grasped attempts to argue against the consistency of this set, such as the one Schellenberg develops: there will or might be individuals for whom it is subjectively more reasonable, or rational, or justified, or warranted, to deny the premises of such arguments than to deny either that God loves us or that God is hidden.

Such is a project familiar in recent religious epistemology: to resolve an epistemic dilemma, find a belief system in which it is rational to hold both purportedly inconsistent claims --God is hidden and God is perfectly loving, or evil exists and God is perfectly good, all-powerful and all-knowing. In one way, this game is boring because the outcome is sure, at least if you lean toward subjective conceptions of the key epistemic terms. For there is always some belief system adequate to the task of epistemic absolution. The only question is how long it might take to find it.

Some of these belief systems will even be quite familiar. Regarding the problem of hiddenness, for most traditional Christians such as I, the explanation of how God is perfectly loving and yet hidden lies in the doctrine of the Fall. Schellenberg considers, all too briefly, this

line of response to the problem he raises. He notes that much of Pascal’s approach to the hiddenness of God hinges on Pascal’s assumption that “a historical “Fall” actually took place.” In response he says, “But in light of the findings of disciplines like evolutionary biology and biblical criticism, it is hard to see how such an assumption could be successfully defended.”^{^[iv]}

There are several things to complain about here. First, I think Schellenberg must be thinking that a historical doctrine of the Fall must treat the Adam and Eve story as literally true. Without this assumption, I don’t see how the findings of evolutionary biology and biblical criticism could undermine the view. But that assumption is false. Adopting a literal interpretation of the Adam and Eve story is only one way of understanding the Fall historically. The story of Adam and Eve might be representative of the fall of peoples, or of the fall of each individual person, or of some cosmic catastrophe that needs to be redressed through the process of redemption. Second, it is far from clear that a historical doctrine of the Fall is required to resolve the difficulty Schellenberg presents. The story of Adam and Eve might be a mythic account of a fallenness endemic to the created order, perhaps a fallenness that results from the inclusion of freedom in the created order, the rectification or removal or cleansing or purification of which God has undertaken and is in the process of completing. Whether historical or not, the doctrine of the Fall points to a need, both cognitive and conative, that can be addressed only by

the intervention of the divine. We are more used to the phenomenological data of conative fallenness, but there are data on the cognitive side as well. For example, there is a logical and empirical gap between recognized sufficient warrant and actual belief formation very much like

the fallenness of humanity displayed in cases of weakness of the will. There are also other aspects of cognitive fallenness: the possibility of self-deception and the Jamesian possibility of such paranoia over the prospect of false belief that one sets the standards for warranted belief too high, for two other examples. A doctrine of the Fall is but a theological interpretation of such fallenness, a doctrine according to which all aspects of fallenness, including both cognitive and moral aspects, are somehow fleeting and regrettable aspects of or earthly existence which God is in the process of removing. The traditional form of such explanation, whether through a literal story about Adam and Eve or some mythic interpretation of this story, ties fallenness in all its aspects to a desire (or purpose) to be God-like, to a desire for complete autonomy, or self-control, culminating for some desire (or purpose) to be rid of God, perhaps even to spite him or show him up. The common element of such explanations is that an account of fallenness is tied

directly to a rejection of God, either through wanting to take his place (by being the most

important thing there is), or not caring about his existence, perhaps wanting him to exist in order to defy him, perhaps even denying his existence altogether.

There are many questions that remain about this account of divine hiddenness, including questions concerning the competitor explanations that might be proffered for the data of fallenness and questions about the epistemic effects of such a doctrine, including the question of

whether such a doctrine defeats any reason we could have for believing it. Recall, however, the context of our discussion here. I am considering how an epistemic subjectivist will look at the possibilities for reconciling any perceived tension between the claims that God is hidden and that God loves us, and have claimed that one such possibility is found in the doctrine of the Fall. Schellenberg demurs, but my point has been that the epistemic utility of this doctrine simply can’t be undermined by some passing remark about what evolutionary biology and biblical criticism shows.

In the present context, there is another aspect of Schellenberg’s rejection of an appeal to the doctrine of the Fall that is puzzling, for he claims, not that the doctrine is false, but that it cannot be “successfully defended.” For any subjectivist about justification, such a remark is completely irrelevant. What is relevant are the connections, partly explanatory ones, between the various beliefs a person has. The doctrine of the Fall, whether conceived historically or not, provides an explanation both of fallenness and of the hiddenness of God. We are, according to this doctrine, damaged goods both conatively and cognitively, and the damage in question is so intimately linked to the nature and existence of God that there is a strong incentive for self-deception and confusion regarding the truth or falsity of theism. Given these explanatory connections, one will be hard-pressed to find subjective grounds that undermine the power these explanatory connections possess to generate warrant for belief. Most important, such power won’t be affected in the slightest by the issue of whether the beliefs can be defended successfully. That is a completely different matter, having to do with the capacity of a person to answer criticisms of his or her view. But that is irrelevant to epistemic questions subjectivistically construed.

So the caveat I must make to my initial remarks questioning the existence of an epistemic problem of hiddenness is that we must focus on more objective construals of key epistemic terms in order for the project of finding a troubling epistemic problem of hiddenness to get off the ground. I think such a construal is just what Schellenberg needs to make his worries about whether certain claims can be successfully defended to the point. For a complaint that some claim cannot be successfully defended might also mean that the totality of the evidence, objectively conceived, does not support the claim. What I want to defend is that there is no epistemic problem of hiddenness, no factor involved in the hiddenness of God that can change the epistemic standing of theism, when such an objective understanding is assumed.

In order to explore this issue, I first want to begin by looking more carefully at Schellenberg’s argument. The value of this approach is that the failures of his argument give us insight into what must hold in order for there to be an epistemic problem of hiddenness. So after looking at a particular attempt to find such a problem, we will be in a position to look at the issue more generally, a perspective from which we will be able to see that there is no such problem to be found.

I. Schellenberg on Divine Hiddenness

Schellenberg begins his work by making several assumptions, one of which he formulates as follows: “I will further assume that the relevant evidence..does not clearly favor either the conclusion that God exists or its denial” (pp 11-2). I think Schellenberg means to assume that the evidence is counterbalanced for and against theism, though the claim does not quite say this. Instead, it notes a lack of clarity on the issue (the evidence doesn’t “clearly favor” either view, he says), which is not a point about what the evidence supports, but rather one about our understanding, our grasp of what the evidence shows. Throughout the work, however, Schellenberg takes the significance of his project to be that of altering the epistemic status of

theism, claiming that “the weakness of evidence for theism...is evidence against it” (p. 2). So, in spite of language not quite endorsing this conclusion,^{^[v]} I conclude that an initial assumption

that Schellenberg makes is that evidence for theism is counterbalanced, favoring neither theism nor atheism.

Schellenberg argues for this assumption by saying that it seems warranted by recent discussion in philosophy of religion which often ends in deadlock, and it is widely granted by

those inside and outside the academic study of religion (p. 12). The former claim, that disputes in philosophy of religion end in deadlock, yields a bad argument for the counterbalanced claim. All the point about ending in deadlock shows is that the issues are controversial and consensus is not likely to be achieved soon. Nothing follows about the quality of the evidence for a claim just

because people can’t agree on what it shows. Positing the counterbalancing of the evidence might be thought to explain the disputes and their irresolvability, but it doesn’t; it merely sides with some of the disputants over others. So the only argument left in Schellenberg’s claims is that the assumption is widely granted, and that is not much of an argument at al. At the very

least, it is not a method by which we ought to decide matters philosophical. (I might point out here that the confusion between two assumptions, the first being that the evidence does not favor theism, and the second that it is not obvious that the evidence favors theism, may incline Schellenberg toward this argument, since it is a good argument for the claim that the evidence doesn’t obviously favor theism, but not for the assumption that the evidence doesn’t favor theism.) So Schellenberg does not have a good argument for the assumption that the evidence is counterbalanced. More important, I want to argue that the assumption can’t be granted unless we

employ a different conception of evidence and belief from the ones that Schellenberg uses.

After assuming the evidence is counterbalanced, Schellenberg argues that a loving God would put us all in a strong epistemic situation regarding his existence. That is, a loving God would provide evidence adequate to make us believe that He exists. In defending this position, Schellenberg responds to two attacks. One questions whether any evidence at all is needed.

Why not just require God to make us believed that exists? The other questions whether

adequate evidence is good enough. Perhaps what God should give us is absolutely conclusive evidence that He exists, so that those who examine the evidence won’t see it as less than conclusive and thereby fail to believe.

Neither response is adequate, I want to argue, and I’ll take up the response to the latter objection first. In response to the objection that God should provide conclusive evidence, Schellenberg licenses inferring belief from taking the probability of the claim to be greater than .5. He says, “Of course, if I consider the evidence to be evenly balanced, I will be uncertain

whether G [the claim that God exists] is true. But if I see G’s probability as greater than .5, it seems natural to suppose that I believe it to some extent” (p. 37).

One problem with this response is that it implies that no reflective theist will be able to agree with the Schellenberg’s initial assumption that the evidence is counterbalanced as to

whether God exists. For Schellenberg says that if you consider the evidence to be evenly balanced, you will be uncertain whether God exists, “uncertain,” in the psychological sense of

failing to be convinced one way or the other on the question. He further identifies belief with having at least a weak disposition to feel the claim to be true, a disposition clearly lacked by

those who are uncertain in the psychological sense of the existence of God. Hence, believing that the likelihood of theism is no greater than .5 is incompatible with believing that God exists, on Schellenberg’s account. So every reflective theist thinks there is adequate evidence for the existence of God, given Schellenberg’s conceptions of evidence and belief.

There are, of course, reflective theists who think that the existence of God is not probable

at all. Here come to mind Pascal and, even more starkly, Kierkegaard. But such reflective theists don’t conceive of the relationship between belief and probability the way Schellenberg does. My point is not that every reflective theist must think that there is good evidence for the existence of God, but rather that this claim follows when we adopt the accounts of evidence and belief that Schellenberg employs.

More important, however, is the perplexity we should experience upon hearing Schellenberg’s reply to the objection. The objection, recall, suggested that God ought to give us absolutely conclusive evidence, not merely evidence adequate to show that He exists. Schellenberg’s reply attempts to answer this objection by weakening the notion of belief, and tying it to our assessments of the probability of a claim. The idea, I take it, is that if we weaken the notion of belief sufficiently, it will be harder to imagine cases where a person has adequate evidence for a claim but misjudges the quality of evidence. But I don’t see how the reply accomplishes that. I don’t see how it gives us any reason at all to think that mistakes about quality of evidence are less likely than we might have supposed. Furthermore, even if such mistakes are less likely than we might have thought, they are not impossible, and if they are not, it is hard to see how Schellenberg’s reply is to the point at all.

There is nonetheless an adequate reply to the objection, I think. If mistakes about the quality of evidence are possible, I see no reason to think that such mistakes are not possible about even absolutely conclusive evidence as well. No matter how good the evidence is, it still appears possible to mistake its quality, and if so, you can’t argue that God should have provided absolutely conclusive evidence so as to eliminate the possibility of people seeing it as less than conclusive and thereby failing to believe.

The other objection has no such simple confusion in it, however. That objection questioned why the discussion is phrased employing the concept of evidence at all. Why, it asked, doesn’t God simply force belief on us? To stave off this criticism, Schellenberg endorses a broad understanding of evidence that includes “anything that can serve as a ground of belief” (p. 33), so that if it merely seems strongly to a person that a claim is true, that very experience of it seeming to him or her to be true is evidence for the belief (p. 33). The idea here is that if we broaden the notion of evidence so that more counts as evidence than we thought, and tighten the connection between belief and evidence so that it is more difficult to find cases in which belief and evidence come apart, there will be less opportunity to raise the possibility of belief forced on us without evidence.

There are two problems with this reply. The first is that it commits Schellenberg to a fairly strong and subjective epistemic conservatism principle in conflict with a more

objective conception of evidence and rational belief required for the central arguments of the book. In the passage just quoted, Schellenberg only infers the existence of evidence where a proposition seems strongly to a person to be true. Yet, he also claims that anything that can serve as a ground of belief can be evidence for a belief, so there is no reason whatsoever to count only strong seemings as evidence. For moderately strong seemings, and even weak seemings, ground belief as well. It would be appropriate, perhaps, to hold that the evidence these seemings provide is not as good evidence as that involved in cases of strong seemings, but it would be unprincipled to count only strong seemings when one’s understanding of evidence is supposed to include anything that can serve as a ground of belief. It is a truism, however, that everything we believe seems to us to be true, so in order to answer the worry that evidence for belief is not required, Schellenberg endorses a conception of evidence broad enough to imply that we have evidence for every single one of our beliefs.

Schellenberg is thus committed to a strongly subjective conservatism principle, according to which merely believing a claim constitutes evidence for it. I do not believe that any such conservatism is adequate,^{^[vi]} but I will bypass that point here to focus on the relationship between Schellenberg’s conservatism and important claims made throughout the rest of the book, claims that require a strongly objective notion of evidence that are in tension with the subjectivity involved in conservatism. I shall look quickly here at only one. Near the end of part I of his book, Schellenberg considers Stephen Wykstra’s claim that if God exists, we should expect some goods to be inscrutable, and that the apparent pointlessness of evil can be rebutted by that fact. Schellenberg wishes to attack the strategy of applying Wykstra’s position to the particular evil of reasonable nonbelief. He says,

Suppose that it is likely as not...that if there is a god, at any rate some evils of human

experience serve inscrutable goods. How can we know, with regard to the occurrence of

reasonable nonbelief, that it is a member of this class? Without independent information

bearing on the question, we would have to say, “Well, it might belong to that group, but

then again it might not.” Hence the probability that reasonable nonbelief will be

apparently pointless if God exists would seem to be at most half that of the proposition

that some instances of evil will be so. But then, given that we can surely say no more

than that the latter proposition is as probable as not, the probability of the claim that

reasonable nonbelief will be apparently pointless if God exists is clearly too low for

rational acceptance. (Pp. 90-1)

There are several mistakes in this passage, but let me focus on two central ones. First, the probabilistic assessment of reasonable nonbelief serving inscrutable goods is a complete wash. It endorses the faulty principle that when you have no idea what the probability of a claim is, you should assign a probability of .5 to it. That’s obviously wrong, however. When someone shows you a deck of cards, revealing it to be a normal deck of cards, one has a basis on which to consider it a fair bet at 3:1 odds that the next card drawn will be a spade. If all you know is that there is a deck of 52 cards, with no information whether it is normal, the principle in question would require considering it a fair bet at even odds that the next card drawn will be a spade, but that is clearly mistaken. For it confuses uncertainty about the fairness of the deck with a precise understanding of a particular bias of the deck. The correct attitude in the latter case would be to hold no opinion whatsoever about which bets would be fair bets.^{^[vii]}

The more important point, however, concerns Schellenberg’s last claim. He performs his untoward probabilities calculations, and then announces that the results are “too low for rational acceptance.” But this conclusion is in tension with his conservatism. Whatever the actual probability, Schellenberg’s conservatism principle implies that if I mistakenly judge the

probability to be high, then I have evidence that the probability is high. The mere fact that the actual probability is low doesn’t imply anything at all about what is rationally acceptable for a person, once Schellenberg’s subjective conservatism is accepted, for a person can acquire additional, relevant evidence just by making a probabilistic mistake.

I don’t claim, of course, that there is a straightforward inconsistency here, but the tension cannot be dismissed lightly, either. Schellenberg wants to let in just enough subjectivity into his theory of evidence and rationality to avoid certain objections to his view, but you can’t do that responsibly without endorsing some subjectivist arguments about how a theory of rationality ought to be constrained by actual probability.

There is a worse problem, however, with Schellenberg’s reply to the claim that God should simply force belief on us. To see it, note the seeming irrelevance of Schellenberg’s reply to this objection. The objection quarrels with evidence playing any part in the discussion; why doesn’t God just build us so we believe that He exists? Schellenberg’s response is that there’s a lot more evidence for any belief than you might initially have thought. But how is that relevant? The only way I can see to make it relevant is to hold that the question of God making us believe

is not logically distinct from the issue of having adequate evidence, so that if God were to make us believe, He could only do so by providing us with adequate evidence. Otherwise, no matter how much evidence exists in virtue of our beliefs seeming to us to be true, the question can be raised again--namely, why this focus on evidence when God could simply have placed the belief in us? In order for Schellenberg’s discussion of this question to be to the point, belief and evidence would seem to need to go hand-in-hand to such an extent that mere belief carries its own implication of adequate evidence. Such a viewpoint endorses an irresponsibly strong epistemic conservatism principle, irresponsible because it is inconsistent with the obvious possibility of irrational belief.

What we are left with, then, is that if we travel with Schellenberg we end up construing the problem of hiddenness as a doxastic issue about why God doesn’t force belief on us. And if God’s merely forcing us to believe that he exists solves a problem of hiddenness, the problem that it solves is not an epistemic problem of hiddenness. It is not a problem concerned with the concepts of evidence or rationality at all.

One might think that the solution is simply to abandon Schellenberg by adopting better construals of the key epistemic terms, and an epistemic problem of divine hiddenness will be easily formulable. Let us see if that is so.

II. Is there an Epistemic Problem of Hiddenness?

To find an epistemic problem of hiddenness, we need to establish some parameters. A first reasonable parameter is something like Schellenberg’s assumption that the evidence for theism is counterbalanced, for if there is adequate evidence for the existence of God, there is no

epistemic problem of hiddenness. We should also note that there is no such problem, if the evidence is adequate, regardless of whether anyone believes that there is a God. There may be some other problem lurking here, but it is not because of God’s hiddenness. The problem would rather be one of human density or resistence to what the evidence shows.

Again, Schellenberg holds a contrary position, one that comes to grief fairly quickly and underlies the host of difficulties uncovered in the last section. He holds that a loving God would provide “evidence that is sufficient to produce belief” (p. 33). This requirement involves two aspects: first, the evidence must be adequate, and second, it must be efficacious. The tension between these two elements turned out to be intolerable, as we saw in the last section, for Schellenberg ends up with no good answer as to why there should be any focus at all on evidence once we introduce the question about efficacy. Once that question was introduced, Schellenberg had no good answer to the question of why mere forced belief wouldn’t be good enough. Avoiding this problem is simple, however. We need only note that the question about the efficacy of evidence is not a problem concerning the hiddenness of God. If you can’t see me standing in front of you in broad daylight, even though your eyes are open and directed at me, the problem is not that I’m somehow mysteriously hidden from you. The problem, instead, is that you just can’t see. Now, we might want to investigate the problem of the epistemic defects of human beings if it should turn out that there is adequate evidence for the existence of God and many people do not recognize it as such. We might ask under what circumstances we should expect adequate evidence to fix belief and when we shouldn’t. Perhaps we could even do a bit of scientific investigation of the issue, perhaps positing that we should expect belief fixation by evidence when it has survival value and not otherwise. Then Schellenberg’s concern for the efficaciousness of evidence would receive the scientific answer that God should have limited the survival and reproductive capacities of non-theists.

But I mock; perhaps I mock a bit too much. Perhaps evolutionary explanations themselves raise problems, so that the problem of the epistemic defects of human beings is not answered by such evolutionary explanations, but only raises the further question of why a good God would have chosen a set of laws for the actual universe that leaves room for such defects. Such is not my topic here, however; suffice for the present to repeat that this issue is not one of divine hiddenness, even if it does raise issues that are important in their own right.

So, in order for there to be an epistemic problem of hiddenness, we must assume that the evidence is counterbalanced between theism and non-theism, or at least not favorable to theism. For the sake of simplicity, I will talk of the needed assumption in terms of a counterbalancing of the evidence, for the same points I will make will apply if we take a bit broader interpretation of the assumption. The implications of this assumption, however, undermine the attempt to find an epistemic problem of hiddenness.

If the evidence is counterbalanced between theism and atheism, one of two accounts of the quality of the evidence must be accepted. On one picture, what is proffered as evidence for

or against the existence of God is not really evidence at all. On this picture, the proofs for the

existence of God are not even defeasible pieces of evidence; they don’t make for the objective rationality of accepting theism even an iota. The same would be said on this picture about other defenses of the rationality of religious belief, defenses that appeal to the nature of religious experience or the grounds which are sometimes claimed to support the proper basicality of certain claims.^{^[viii]} In no such case is there anything to the claim that there is any positive evidence or grounds in support of theism. But equally true on this picture is that there is no evidence in the other direction, either. The problem of evil is a mere pseudo-problem; the existence of evil counts no more against the existence of God than does the existence of grass, or stars, or trees. There is just no objective epistemic connection at all between the claim that God exists and the claim that evil exists.

It is not hard to see that on this picture, there will be no epistemic problem of divine hiddenness. For the purported problem of hiddenness is just a special case of the general problem of evil. If hiddenness is an epistemic problem, it is because the inscrutability of divine existence is a bad thing that a perfectly loving God would not or could not allow. But that, of course, is simply a special case of the problem of evil, and on the picture above of how the evidence is counterbalanced between theism and atheism, no instances of evil are relevant at all to the truth or falsity of theism.

So we must turn to the other picture of how the evidence is counterbalanced. On this picture, there is good, objective evidence for the existence of God. Of course, none of it is indefeasibly good evidence, for then it could not be counterbalanced by evidence against the existence of God. So the good evidence must be defeasible, capable of being counterbalanced by evidence against the existence of God. The proofs, religious experience, and the grounds which might be cited on behalf of the proper basicality of theistic belief can provide strong evidence for theism, but the evidence must not be conclusive. For its positive impact on the claim that God exists is offset by other information.

This other information is in the form of defeaters. Defeaters come in many forms, but the relevant form here is that of a rebutter. A rebutter of evidence e for p is some evidence r for ~p, such that e&r leaves p less than justified. As we saw above, there will be no epistemic problem of divine hiddenness if no instance of evil counts against the existence of God, so we must assume here that evil is evidence against theism and for atheism. In order for the total evidence to be counterbalanced, evil must provide sufficient evidence for atheism to offset, to counterbalance the evidence available for theism in the form of arguments for the existence of God, appeals to religious experience, and grounds for the claim that belief in God is properly basic. So the picture we get is one for which the metaphor of the scales of justice is appropriate. Each pan of the scale has something in it, and the two sides balance each other out.

Yet, if this is the picture we adopt of the claim that the evidence is counterbalanced, what room is left for the claim that there is some further evidence provided by the fact that God is

hidden, that God’s existence is not as obvious as it should be, which will tip the scales on behalf of atheism? We ought to find the suggestion initially quite puzzling, for such a problem can only arise after evaluating the evidence for and against theism, and finding that evidence to be counterbalanced. Yet, it is also obvious that the problem of hiddenness is but a special instance of the problem of evil, the epistemic weight of which has already been factored in. So how could hiddenness be a further problem, a further piece of information to be weighed in the balance?

One might wonder why a defender of the epistemic problem of hiddenness can’t reply as follows: “I’m imagining the scales to be balanced when all is taken into account except for the hiddenness of God. Once hiddenness is added, the scales tip toward atheism.” Such a claim is precisely one that Schellenberg makes. He qualifies his remarks about the counterbalanced nature of evidence for theism by formulating the epistemic indifference point “exclusive of the evidence adduced in this book,” the book he writes (pp. 11-12).

But such a reply cannot work. Since the problem of evil is already present in the balance, adding a new kind of evil to the scales will have no more effect on the balance than would adding some new religious experiences to the other side of the scale. The theist can say, “I’ll see your hiddenness problem and raise it five religious experiences.” That is, if adding new evils to one side counts, certainly adding new proofs, or new grounds, or new religious experiences (I don’t mean new kinds of religious experiences, but I don’t rule that out either) has to be allowed on the other side. So citing a special evil won’t have the required effect, if we are assuming that all evil counts against the existence of God.

Moreover, the scales must include all known information relevant to theism, if it is to be an adequate representation of the epistemic question in mind. And the fact that God is hidden is

not unknown, not even to theists--they have discussed it for centuries. The most that could be claimed is that there are some implications of the claim that have not been appreciated by theists, and it is these implications that constitute the epistemic problem of divine hiddenness.

There is another point worth making as well. We are imagining that the epistemic balance is equally weighted for and against the existence of God. We have noted that for this to be so, there must be information on the negative side to counteract the positive evidence for the existence of God on the other side. We have focused on rebutters above, but there is an important presupposition to be noticed here regarding the possibility of such rebutters. In order for there to be rebutters, there must be defeaters present of a different sort. These defeaters must delimit the evidential force of the information on each side of the scales. These defeaters, that is, must operate so as to prevent, on the one side, any information from compellingly establishing the existence of God. In some cases, the evidence itself implicitly reveals this defeating information, such as in the case of probabilistic arguments derived from sampling. In other cases, the presence of such defeaters is signalled by the author of an argument when he or she draws a conclusion preceded by a phrase such as “So it is probable that”. Regarding the question of the existence of God, the delimiting defeaters are best identified with the objections to the traditional proofs and arguments against appeals to religious experience and proper basicality. These objections and arguments must be effective enough to prevent the positive evidence from conclusively establishing theism. Equally so, there must be delimiting defeaters to prevent the problem of evil from decisively establishing atheism. Presumably, the standard responses to the problem of evil, whether in terms of necessity for a greater good, soul-making, the importance of freedom, or simply the fact that we do not know what all the good things are, provide this kind of defeater. Most generally, these defeaters must have the power to neutralize sufficiently the evidential force of both sides of the scale so that such force can be rebutted by the information on the other side of the scale to result in counterbalancing. If so, however, it is hard to see what evidential force could be added by the problem of hiddenness. This problem could tip the scales toward atheism if the delimiting defeaters for the problem of evil failed to delimit the implications of hiddenness. In this way, hiddenness would constitute a different kind of problem for theism than the general problem of evil. But if it does not constitute a different kind of problem in this way, then all it does is add further examples of a problem already weighed in the scales, and we have already seen the failure of this attempt to argue that hiddenness constitutes an epistemic problem for theism. So all that is left is to argue that it constitutes a different kind of problem than the general problem of evil, and for that to be true, the delimiting defeaters regarding the problem of evil will have to fail to delimit the evidential power of hiddenness. No one has argued such, and it is hard to see how such an argument could be successful. If we consider the plausible candidates for such delimiting defeaters-the value of freedom, necessity for a greater good, the importance of soul-making, cognitive limitations, and the like-there is no particular reason to think that such responses succeed only for the general problem of evil but not for the specific problem of divine hiddenness. If not, however, there simply is nothing for hiddenness to add to the scales that won’t be neutralized sufficiently by information already present.

I might encode the view I am defending by saying that the evidential value of hiddenness is merely epiphenomenal. I am not arguing that hiddenness does not constitute evidence against the existence of God; instead, I am merely claiming that this evidence, if it be such, cannot do any epistemic work. It is akin to my testimonial evidence to you that you exist. What I say is evidence that you exist, but it can’t do any epistemic work, for whatever the evidential status of your existence for yourself, it won’t be changed by such testimony.

I believe that there are other examples of epiphenomenal evidence as well. For example, our awareness of the lack of complete reliability in our sources of information and in our own cognitive abilities is awareness of evidence against most everything we believe. But it is merely epiphenomenal evidence, for in the absence of more specific concerns about reliability or our fallibility regarding a particular subject matter, it doesn’t alter the epistemic status of our beliefs: factoring in the information that we are not infallible does not make any of our beliefs less rational than they would have been, had that information been ignored.^{^[ix]} And when there are more specific concerns present, it is these concerns that are responsible for the epistemic change, not the awareness of general facts about fallibility.

My interest, however, is not to defend or articulate any general theory of such epiphenomenal evidence, but merely to point out the existence of such, and to use that concept to articulate my thesis about the nature of the hiddenness of God. If we grant that such hiddenness is evident against the existence of God, it is merely epiphenomenal evidence: it does no work regarding the overall epistemic status of belief in God.

[i].J.L. Schellenberg, Divine Hiddenness and Human Reason, (Ithaca: Cornell University Press, 1993). Page references in the text are to this work.

[ii].For more on the relationship between coherentism and inconsistency, see Richard Foley, The Theory of Epistemic Rationality, (Cambridge, Mass: Harvard University Press, 1986).

[iii].Keith Lehrer, Knowledge, (Oxford: Clarendon Press, 1977); Richard Foley, The Theory of Epistemic Rationality, (Cambridge Mass.: Harvard University Press, 1986).

[iv].Schellenberg, p. 146.

[v].I must note, however, that I am engaging in charitable interpretation of remarks that recur throughout the work. As another example, Schellenberg asks in the Conclusion of the work, “Consider...the situation in which S takes the relevant (independent) evidence to strongly favor the claim that there is a God...Must S, to be rational, deny or come to doubt the existence of God because of the undefeated argument [from hiddenness]?” (p. 209) If we construe the epistemic notions here objectively, the supposition preceding the question is irrelevant, for what makes for rationality is not what a person takes the evidence to favor, but what in fact it favors. I interpret the actual language as unfortunate. What Schellenberg meant to ask, I believe, was what his purportedly undefeated argument would imply epistemically, on the supposition that the evidence strongly favors theism.

[vi].See my “Conservatism and its Virtues,” Synthese 79 (1989), pp. 143-163.

[vii].For further discussion on the failure of this principle, see, e.g., Bruce Reichenbach, “The Inductive Argument from Evil,” American Philosophical Quarterly 17 (1980), pp. 221-227.

[viii].Claimed most famously by Alvin Plantinga. See his “Reason and Belief in God,” in Faith and Rationality: Reason and Belief in God, Alvin Plantinga and Nicholas Wolterstorff, eds., (Notre Dame, Ind.: University of Notre Dame Press, 1983), pp. 16-93.

[ix].To prevent a possible misunderstanding here, I note that failing to factor in the information that one is fallible is not to be identified with factoring in the claim that one is infallible. The middle ground is that of making neither assumption.