SCEPTICAL THEISM AND EVIDENTIAL ARGUMENTS FROM EVIL

Sceptical theists (William Alston and Michael Bergmann) have claimed that considerations concerning human cognitive limitations are alone sufficient to undermine evidential arguments from evil. We argue that, if the considerations deployed by sceptical theists are sufficient to undermine evidential arguments from evil, then those considerations are also sufficient to undermine inferences that play a crucial role in ordinary moral reasoning. If cogent, our argument suffices to discredit sceptical theist responses to evidential arguments from evil.

a number of philosophers endorsed skeptical theism [Alston 1991, 1996; Bergmann 2001; Fitzpatrick 1981; Howard-Snyder 1996a, 1996c; van Inwagen 1991; Plantinga 1979, 1988; Wykstra 1984, 1996]

The core idea behind these responses to evidential arguments from evil is that considerations of human cognitive limitations are alone sufficient to undermine those arguments. This core idea is developed in different ways:

Some skeptical theists' [Bergmann 2001; Howard-Snyder 1996a, 1996c] human cognitive limitations in the realm of value are alone sufficient to undermine evidential arguments from evil.
Other sceptical theists [Alston 1991, 1996; van Inwagen 1991; Plantinga 1979, 1988; Wykstra 1984, 1996] human cognitive limitations in various spheres including the realm of value are alone sufficient to undermine evidential arguments from evil.

Three sections:

First, we argue against [Bergmann 2001, et al.]and
then argue against the second group

What is crucial is that the kind of evidential argument which we consider and which is the target of sceptical theist' critique contains a key inference which moves from a premise of the form We have found no reasons why . . .' to a conclusion of the form There are no reasons why .. The burden of the sceptical theist' critique of the argument is to attack this inference by appeal to nothing more than certain (alleged) human cognitive limitations; we shall argue in reply, that the inference in question cannot be undermined merely by appeal to considerations about human cognitive limitations by the first group or the second group.

Secondly, stephen Wykstra's claim that the premise in the key inference does not even weakly support the conclusion of that inference, i.e., the claim that our failure to find reasons of a certain kind does not raise the probability that there are no reasons of that kind at all. We suggest that this claim is massively implausible: it would require very special circumstances in order for a claim of the form We have found no reasons why . . .' to fail to provide any support for the claim that There are no reasons why . . .'; and it is not credible to suppose that our evidential argument from evil provides such a case.

thirdly, Michael Bergmann's sceptical theist' argument based on a general scepticism about our knowledge of the realm of value' that the premise in the key inference does not provide substantial support for the conclusion of that inference [Bergmann 2001].3 We claim that, if the inference from We have found no reasons why . . .' to There are no reasons why . . .' is blocked by the considerations to which Bergmann adverts in the case of evidential arguments from evil, then similar inferences will be blocked in cases of ordinary moral reasoning which we all have reason to endorse. Thus, we claim, Bergmann faces a dilemma: either his general scepticism about our knowledge of the realm of value' is too benign to save theism from the evidential argument from evil; or else his general scepticism about our knowledge of the realm of value' is so strong that it threatens to disrupt our ordinary patterns of moral reasoning.

In the fourth and fifth sections of our paper, we turn to consider possible objections to our argument against Bergmann. In particular, we consider the suggestion that our argument fails to pay due attention to the fact that it makes a huge difference that it is not we but rather a perfect being who is the subject of the key inference in our evidential argument from evil; and we also consider the suggestion that we are wrong to think that the moral scepticism which is mandated by the principles that Bergmann endorses is anything other than benign.

In the sixth section of our paper, we consider the prospects for extending our critique of Bergmann [2001] to other kinds of sceptical theist' responses to evidential argument from evil. In particular, we focus on Alston [1991, 1996]. Unlike Bergmann, Alston does not claim that considerations concerning human cognitive limitations in the sphere of value are alone suf®cient to undermine evidential arguments from evil. However, it seems to us that Alston is committed to the claim that considerations concerning human cognitive limitations in a range of spheres including the realm of values are alone suf®cient to undermine evidential arguments from evil.

if the objection that we develop against Bergmann is good, then it carries over to this view as well. Moreover, if we suppose as we think we should that it is constitutive of sceptical theist' responses to evidential arguments from evil to claim that considerations concerning human cognitive limitations in a range of spheres including the realm of values are alone suf®cient to undermine evidential arguments from evil, then our argument carries over to all versions of sceptical theist' responses to evidential arguments from evil.

However, it is important to note that the sceptical theist' response is distinctive because, were it correct, it would establish that no-one theist or non-theist should make the crucial probability judgement that undergirds evidential arguments from evil. The direction of the sceptical theist' argument is from considerations concerning our cognitive limitations and nothing but our cognitive limitations to the conclusion that a certain probability judgement is out of bounds. And it is this that we claim to be able to show is wrong.

section 1

We begin, then, with the formulation of a more or less canonical version of an evidential argument from evil. The main sub-conclusion of the evidential argument from evil that we shall discuss is that there are evils for which it is true that, were a perfect being to prevent those evils, then the world would be non-arbitrarily improved thereby. It follows from this sub-conclusion that, if there are such evils, then there is no perfect being. It is tempting to call such evils gratuitous'; however, it seems doubtful that this usage is in accord with the standard de®nition(s) in the literature.5 So let us instead call them problematic'.

Evidential arguments from evil of the kind which we are investigating here aim to establish that there are problematic evils, or that there is most reason to think that there are problematic evils, or that there is at least some reason to think that there are problematic evils. (Later, we shall fuss more about the strength of the conclusion which is to be drawn in arguments of this kind; for now, we leave the matter open.)

we should explain what we mean when we insist that problematic evils are evils that are such that, if they were prevented by the actions of a perfect being, then the world would be non-arbitrarily improved.

**it should be noted that it would not suffice for the purposes of the argument to claim that there are evils for which it is true that, were a perfect being to prevent those evils, then the world would not thereby be made worse.

Suppose that there are two worlds, W1and W2 which are in all respects alike except that W1 contains evil E1 and W2 contains E2, where E1 and E2 are of equal (dis)value. If W1 is actual, then it casts no doubt on the claim that there is a perfect being to point out that a perfect being could have prevented the occurrence of E1 by actualising world W2 instead without thereby making the world worse. (We assume what some have denied that, if a perfect being must choose between W1 and W2, then it can do so.)

should be noted, too, that it would not suf®ce for the purposes of the argument to claim that there are evils for which it is true that, were a perfect being to prevent those evils, then the world would be improved.

Suppose that there is a sequence of worlds W1, W2, . . ., Wn, . . . which contain sequences of evils {E1, E2, . . ., En, . . .}, {E2, E3, . . ., En, . . .}, . . ., {Ek, Ek 1, . . ., En, . . .}, . . ., and which are otherwise identical. (So, by hypothesis, W1 is worse than W2, which is worse than W3, etc.) Suppose further that any world which contains only a ®nite number of the Ei is worse than any of the worlds W1, W2, . . ., Wn, . . ., because a ®nal in®nite segment of the sequence of evils E1, E2, . . ., En, . . ., is required in order to ensure some massive `outweighing' good. If Wk is actual, it casts no doubt on the claim that there is a perfect being to point out that a perfect being could have prevented the obtaining of EkÐby actualising world Wk1, sayÐand thereby made the world better. (Again, we assumeÐwhat some may denyÐthat if a perfect being has to choose one ®nal in®nite section of the sequence E1, E2, . . ., En, . . ., then it can do so.)

The point here is that it is possible that a perfect being may need to make an arbitrary choice, from amongst a set of worlds which it can make, if it is to make any world at all. If, for example, for any world that a perfect being can make, there is a better world, then, if the perfect being is to make any world at all, it must arbitrarily choose one which is good enough. But, in these circumstances, it is then no criticism of the activity of the perfect being that it could have made a better world: in the circumstances, the mooted improvement of the world is arbitrary. The case that we have described in the previous paragraph is simply meant to show that improvements of the world by the prevention of evils could be arbitrary in exactly the same sense: there may be evils which are not required for the obtaining of greater goods, or for the non-obtaining of greater evils, and yet which a perfect being does not have reason to prevent.

However, we do assume what some may perhaps deny that, if the world would be non-arbitrarily improved by the prevention of some evil then, a fortiori, a perfect being would prevent that evil; in other words, if there are problematic evils, then there is no perfect being

With these preliminaries out of the way, we can now turn to our formulation of an evidential argument from evil. Let E be some candidate i.e., prima facie problematic evil which has occurred, e.g., the rape, beating, and murder by strangulation of a five year old girl; or the prolonged and painful death of a fawn which has been trapped in a forest ®re; or the like. We shall use the following as our representative evidential argument from evil:

(1) We have been unable to find even pro tanto reasons why the world would not be non-arbitrarily improved if a perfect being prevented E. (Premise) (2) (Therefore) There are not even pro tanto reasons why the world would not be nonarbitrarily improved if a perfect being prevented E. (From 1) (3) There are at least pro tanto reasons why the world would be non-arbitrarily improved if a perfect being prevented E. (Premise) (4) (Therefore) There is all-things-considered reason why the world would be nonarbitrarily improved if a perfect being prevented E. (From 2, 3) (5) (Therefore) The world would be non-arbitrarily improved if a perfect being prevented E. (From 4) (6) (Therefore) There is no perfect being. (From 5)

In support of (1), we note that we can point to no greater good which would be lost if E were prevented by a perfect being; no greater evil which would ensue if E were prevented by a perfect being; nothing which suggests that there must be some events like E violations of young children if there are to be greater goods which would otherwise be lost; nothing which suggests that there must be some events like E violations of young children if there are not to be greater evils which would otherwise ensue .

In support of (3), we note that we all agree that the world could have been nonarbitrarily improved if one of us had intervened to prevent E.

Indeed, we note that we all agree that, other circumstances permitting, we would have a moral obligation to intervene: if we could, without risk to ourselves and others, and without extravagant use of resources, prevent E, then that is what we are required to do. (No doubt, our moral obligation runs further than this; however, all we need is the uncontroversial claim that our obligations run at least this far.) The awfulness of E is enough to establish that a perfect being has at least a pro tanto reason to prevent it.

The inference from (4) to (5) is uncontroversial. The inference from (2) and (3) to (4) also looks solid: if there are not even pro tanto reasons why the world would not be non-arbitrarily improved if a perfect being prevented E, and there are pro tanto reasons why the world would be non-arbitrarily improved if a perfect being prevented E, then it surely follows that there is all-things-considered reason why the world would be nonarbitrarily improved if a perfect being prevented E.

That leaves the inference from (1) to (2); and here there are many philosophers who will wantÐand have wantedÐto raise objections. In what follows, we shall consider some of the objections that have been made to the inference from (1) to (2).

II

The inference from (1) to (2) moves from It is not the case that we have found reasons of such-and-such a kind' to There are no reasons of such-and-such a kind'.9 Plainly, this is not in general a good deductive inference: it is perfectly possible for claims of the former kind to be true while the corresponding claims of the latter kind are false. However, it does not follow from this claim alone that the inference from (1) to (2) is not a good inference; it may be that there is some other kind of evidential or probabilistic support which (1) lends to (2). Moreover, it seems naturalÐat least initiallyÐto suppose that this is the case: surely (1) does lend some kind of evidential or probabilistic support to (2). That I have failed to ®nd reasons of such-and-such kind may well be evidence for the claimÐmay well make it more likely that there are no reasons of such-and-such kind

We need to distinguish at least two different possible claims here. One claim is that (1) provides strong evidential support for (2): given (1), we have substantial reasonÐ perhaps even more reason than notÐto believe (2). A weaker claim is that (1) provides weak evidential support for (2): given (1), we have more reason to believe (2) than we had before we took (1) into account. We shall ®rst brie¯y consider views that dispute even the weaker claim, before turning to consideration of arguments against the stronger claim.

A clear example of someone who disputes even the weaker claim is Wykstra [1984]. In his view: `Cognisance of suffering . . .should not in the least reduce our con®dence that [perfect being theism] is true. When cognizance of suffering does have this effect, it is perhaps because we had not understood [what perfect being theism] proposes for belief in the ®rst place' [Wykstra 1984: 91, our italics]. Of course, this kind of position is very strong. If cognizance of suffering should not in the least reduce our con®dence that perfect being theism is true, then it seems that observations of evils in the world must be completely irrelevant for the question of the assessment of the truth of the claim that there is a perfect god

Suppose, for example, that we were to discover that there are a billion other inhabited planets in our galaxy, and that the trillions of intelligent inhabitants of all those other planets live lives of unrelieved misery. Suppose that we extend our search, and ®nd the same ®gures reproduced for the millions of other galaxies. Suppose, ®nally, that we are utterly unable to think of any way in which all of this misery could subserve some greater good. The view in question entails that none of this would provide any evidence at all against perfect being theism. It is very tempting to suppose that this is a reductio of the view in question.