Philosophy of Religion 30: 77-87, 1991. 9 1991 Kluwer Academic Publishers. Printed in the Netherlands.

Swinburne on natural evil from natural processes* DAVID O'CONNOR Department of Philosophy, Seton Hall University, South Orange, NJ 07079

Richard Swinbume has argued that the existence of natural evil is compatible with the existence of God as defined in traditional theism. At the core of that definition is the idea of God as the omniscient, omnipotent, morally perfect, benevolent, worshipful designer-creator of the universe. In particular Swinburne has argued that natural evil is necessary for human beings to develop moral responsibility and maturity. Among natural evils he is especially interested in the pain and suffering resulting from the operation of natural processes, henceforth NENP (natural evil from natural processes). In regard to that very common class of natural evils, Swinburne's argument is that, in a God-made world, NENP is logically necessary for human beings to become morally mature and to be able to choose their own destinies. 1 If he is right about this he will have established that natural evil is logically necessary in a God-made world, and thereby he will have succeeded in proving that natural evil is compatible with the existence of God. In this paper I will argue that Swinbume fails to justify the existence of NENP in a God-made universe. I will do this by showing that moral maturity and choice of destiny are possible without any NENP whatsoever. If I am right, the just-sketched chain of inferences will be broken and Swinbume's position on God and natural evil defeated. . Nowadays in the philosophical literature on theism and the problem of evil, it is customary to differentiate between two sort of problems of evil, the logical and the evidential respectively. It is the former alone that is at issue here, notwithstanding the fact that much of the present argument

* I am grateful to Richard Swinbume for his criticisms of an earlier draft of this paper, and also to Michael Martin for his comments upon that same draft.

78 revolves around epistemic notions. The logical problem may be further split into what we may call the quantified problem and the unquantified problem of evil. Restricting ourselves only to natural evil in the present context, by the former is meant theism's problem of reconciling all natural evil with the existence of God, while by the latter is meant theism's (comparatively) lesser problem of reconciling s o m e natural evil with the existence of God. 2 Borrowing Swinburne's understatement, the former is clearly more awkward for theism than the latter. 3 If Swinbume is right that NENP is logically necessary for human beings to learn to become morally mature, he will have solved the unquantified problem and, arguably, made creditable progress in the direction of solving the quantified problem too. But if I am right, both the quantified and unquantified problems of evil remain wholly unmitigated by Swinburne's argument. For purposes of our present discussion it will be sufficient for me to sketch only the bare bones of Swinbume's case for the place of NENP in a God-made universe. That case goes as follows:

(1)

let us agree that a God-made world reflects a divine plan,

(2)

let us accept that any such divine plan would include provision for human moral development.

(3)

moral develepment is possible only provided the world (including human nature) is a certain way.

(4)

moral development presupposes ability to make moral judgments

and

Now

But

which, in turn, presupposes

(5)

that human beings have the requisite knowledge - of the effects for good or ill of our actions, for instance - that makes such judgments possible.

(6)

the requisite knowledge can either be innate or acquired through experience.

Now

But

79 (7)

the knowledge that is needed could not possibly be innate.

Therefore

(8)

it must originate in experience.

But then

(9)

both good and evil - especially, in the present case, NENP must be available in the natural world as the sources of our knowledge of good and evil (that is, echoing a part of proposition (3) above, the world has to be a certain way).

Only thus could a first wrong-doer have acquired the necessary beliefs about actions and consequences. But (10)

knowledge is justified, true belief,

(11)

our beliefs about specific pieces of behavior and their consequences can only be justified through experience.

(12)

as it is the consequences for good or ill of our behavior that are in focus here, our beliefs about such behavior and its consequences can only be justified through experience of good and evil.

and

But

And, in particular, (13)

a first wrong-doer could become justified in such beliefs only by experience of NENP.

(14)

NENP is logically necessary as a justificatory grid in a Godmade world.

Thus

Thus, from (9) or (14), it follows that (15)

NENP is compatible with the existence of God.

(16)

natural evil is compatible with the existence of God. 4

Thus

80 In earlier papers I have suggested that the requisite knowledge need not come about in the way that Swinburne maintains. 5 The basic premise from which I have argued to that conclusion is that it would be logically possible for God to implant, incognito, the requisite knowledge of good and evil in our minds, so that later, upon our being in certain situations, the relevant information would come into our consciousness. In a recent paper Swinburne has responded to some features of my criticisms. 6 The key element in that response is a distinction between what he calls "moral knowledge" and "practical knowledge." By "moral knowledge" Swinbume means knowledge of principles or general propositions; for instance, "that causing pain is, other things being equal, wrong. '~ By "practical knowledge" he means knowing "what morally relevant features [a particular action] has (e.g. causing pain, being a lie, extending life, enriching experiences of life ...)" and so on. 8 Swinburne accepts that moral knowledge does not require experience for its acquisition or justification.9 With regard to that, then, he has no objection to my implant hypothesis, so let us not discuss moral knowledge further. But Swinburne goes on: With practical knowledge it is different, at any rate for agents, such as humans, who have a choice of destiny in that they can grow in knowledge of the nature and consequences of their actions. 1~ Let us focus here, then, upon practical knowledge. In regard to such knowledge, Swinburne's point is that it is impossible without the experience, thus the existence, of NENP. His reason is that, without justification through experience of NENP, our "practical" beliefs, even if true, will necessarily fail to meet the justification condition required for knowledge. 11 Contrary to Swinburne, my argument is that practical knowledge is possible without either the experience or the existence of any NENP whatsoever. My argument will take for granted that true practical beliefs could indeed be implanted incognito by God, and so the focus will be upon the question of whether such beliefs could be justified without experience of NENP. But, as Swinburne grants the possibility that true practical beliefs could be implanted or innate, 12 his position is not compromised by that assumption. Thus I will attack propositions (13) and (14) in the chain of Swinburne's reasoning, and conclude that NENP is not logically necessary for us to develop into full moral agents responsible for our own destinies.

81 .

Let us use the three following passages from Swinbume's paper to frame the issues concerning justification that we must discuss: If an agent is to have a free and responsible choice of destiny, his beliefs must be true and justified. 13 Why can't I just know what will happen if I do so-and-so tomorrow? I can indeed have such a belief, but it would not amount to knowledge, because it would not be justified. 14 God cannot give to men the sort of belief needed for a free and responsible choice of destiny without producing natural processes which bring about natural evils, and letting men observe and experience them ... The more and more varied are the evils observed and experienced, the surer is the knowledge of the consequences of natural processes gained, and so the greater is the opportunity for informed choice of action. The more that knowledge is made available to man by any other route than the observation of natural processes, the more his freedom and choice of destiny is reduced. 15 (italics added) In the present context there are two sorts of justifications involved and it will be best to separate them. They are epistemic justification and moral justification. Within the frame of this discussion epistemic justification is the more important. For, if I can show that we could have practical knowledge without experience of NENP, there will be no epistemic barrier to our being morally justified or unjustified in our decisions and actions, thus no such barrier to our developing moral maturity and coming to mastery of our own destinies. And it i s an epistemic barrier which, according to Swinburne, constitutes an absolute block here. Accordingly, separate discussion of moral justification will be very brief compared to my discussion of epistemic justification.

2.1 Epistemic justification With Swinburne let us accept that knowledge is justified true belief and that the distinction he has made between moral and practical knowledge is correct. With "knowledge" so defined, three questions come up in the present context: whether practical beliefs could be justified without reference to any experience whatsoever, whether practical beliefs could be justified without any experience of natural evil, and whether practical beliefs could be justified without reference to any experience of NENP? I think the answer to the first question is no. In this I quite agree with Swinbume. But nothing follows from this about the answer to either of the

82 two other questions. I shall not directly address the second question here, although my argument for an affirmative answer to the third question does point in the direction of a negative answer to the second. 16 Turning to that third question now, that is, whether practical beliefs could be justified without experience of NENP, let us pick up on a recent conjecture of Michael Martin's. 17 He argues, contrary to orthodox thinking on the problem of natural evil, that a mixed world of frequent, systematic miracles, together with natural laws governing natural processes, is possible and would be a better world, all things considered, than a world without such miracles. Let us accept that the actual world of our experience is a world of this latter sort. Adapting Martin's conjecture to our own purposes here, let us speculate about a possible world, which we may call worldP, that God could have made instead of the actual world. WorldP would b e so regulated that nobody (or any animal either) would ever experience pain or suffering due to natural processes. This does not mean that there would be no pain or suffering at all, though, for the pain we cause either through malice or negligence would remain unaffected. The regulation in question would involve a mixture of natural laws, as in the actual world, and patterns of miracles preventing the pain and suffering which, in the present world, comprises much of NENP, WorldP would be a world of regular orderly occurances, a world in which the future could be predicted on the basis of experience of the past with no less accuracy than in the present, actual, world. In worldP there would be both moral and natural evil, for instance, the painfulness of the pain that we maliciously cause, but there would be no NENP, for NENP would constantly be blocked by a constant series of miracles (that is, by regular, systematic, secret, divine intervention in the operation of natural processes). Three qualifications are needed before God's actualizing of worldP in preference to the actual world can be discussed. The first pertains to two different relations between natural evil and natural processes. We are considering natural evil from natural processes (that is~, NENP), but there is also natural evil that comes through natural processes. For instance, if I kick the cat, I am responsible for his pain, not certain processes operating in the natural world, even though he could not have felt the pain at all except through some such processes (for example, the operations of his own nervous system). In worldP there would be no NENP, but there would be natural evils through natural processes, that is, there would be pain and suffering. The second qualification concerns two senses of "negligence." The first sense of "negligence" is that involved in our failing to intervene to stop something which causes harm, while the second is the negligence that is

83 involved in our own reckless or thoughtless behavior. Let us call them passive and active negligence respectively. Now, insofar as NENP is concerned, there could be no passive negligence in worldP, for there would be no NENP there in the first place. We could, however, be passively negligent in regard to actions of other people that result in harm, and which we could justly have intervened to prevent. And, just as in the present, actual world, there would also be active negligence, recklessness for example. 18 In worldP, then, pain and suffering would all be due either to human malice or negligence, passive or active in the ways described. [On a side note, this possible world I am conjecturing seems compatible with one of the most deeply rooted portions of Christian mythology, namely: the story of Adam in the garden of Eden, with its explanation of subsequent pain, hardship, and so on as traceable to Adam's behavior. This is not evidence for my speculation about the possible world in question, of course, but it does suggest that worldP is not obviously out of step with Christian thinking.] The third point worth making before we go any further with discussion of worldP is that determinism would not be entailed by my implant conjecture. For it is information alone that would be implanted, not behavioral programs. Let us now examine some likely objections to the foregoing possibleworld hypothesis. In worldP, how could there be a first wrong-doer without NENP from which to learn about the harmful effects of his actions? That is, how could there ever come to be any moral evil in worldP? And, if moral evil is not possible in worldP, then worldP could not be a world that God, consistent with the divine plan for the universe sketched earlier, could actualize or permit. The answer is this: on the implant hypothesis, we would be born with true beliefs about the ill effects of certain actions (for instance, reckless actions), that is, with true practical beliefs. As noted before, Swinbume accepts the possibility of this. Of course we would not, just by having those true beliefs, know that such actions have certain consequences. But upon acting in those ways, or by experience of others so acting, we would see suffering result. Our true practical beliefs would then be justified by experience. And the knowledge we would have would be practical knowledge in Swinbume's sense of the term. Equipped with that knowledge, future behavior of those sorts would be negligent or intentionally evil, as the case may be. But the natural evils resulting from such negligent behavior, and which constitute the justifying base of the first wrong-doer's practical knowledge, would not be NENP, that is, natural evils from natural processes, although they would be natural evils through natural processes. The upshot is that we have provision for a first wrong-doer in worldP.

84 A second objection might be constructed as follows. Would it not be common knowledge in worldP that God existed, and would this not inhibit (through fear of certain punishment) our freedom of choice and action, and thereby moral evil? There are two lines of reply to this. First, moral evil would not be logically impossible under those conditions, and that is sufficient. 19 However, the second is the more substantive of the two replies and, for that reason, it is the one I shall sketch here. That second response denies the claim that God's existence would be known in worldP, and it goes like this. What would be known is that there was no literally accidental or morally neutral pain or suffering in that world. But it does not follow that it would also be known that this was because of divine intervention in nature. Perhaps there would be speculation to that end, just as there is, in the actual world of our experience, about the origin of this (actual) world, but there is no reason to suppose that such speculation about divine agency in worldP would be any more compelling or widespread than it is in the actual world. For, after all, God's systematic intervention would be unannounced, and, given God's omniscience, no telltale signs would be left. It is at least as plausible to suppose that scientists, philosophers, and others, would speculate that the absence of pain and suffering due to natural processes (thai is, NENP) was explicable in purely naturalistic terms as it is to suppose they would hazard explanations in terms of divine agency. And similarly, it is at least as plausible to suppose that the former sort of explanation, when hazarded, would be widely viewed as no less plausible than the latter. In short, in worldP, both science and ethics would remain open for business as usual. This opens the door to a third objection, namely: that my conjecture calls upon God to deceive us, thus compromising his moral perfection, thus making it impossible for God to do what is suggested. But this is false. All that is required is God's not informing us of the cause of certain patterns of events (or, more accurately, non-events) in worldP, much as God now fails to inform us of the ultimate origin or fate of the actual world. A fourth objection might be this. God could not actualize worldP in preference to the actual world because it would provide less opportunity for moral development than the actual world. The standard justification in theistic circles, and this is reflected in Swinburne's work, for the condition of the actual world, especially its NENP, is that it makes possible certain higher-order goods such as courage, tolerance, ingenuity, perseverance, thoughtfulness, compassion, and so on, that would otherwise be impossible. NENP is then supposedly justified as their unavoidable cost. But all of those higher-order goods are equally possible in worldP, and at less cost in evil. There would still be pain and suffering in the considerable

85 amounts due to our malice and negligence, both passive (in regard to the actions of others) and active. Thus there would always be an environment facilitating development of those higher-order goods mentioned. The difference is that, while there would be ample evil from which to learn about and develop those goods, there would be less evil than in the actual world. Prima facie, then, worldP would be more objectively desirable than the actual world. Equipped with the concept of this possible world, and with the objections raised met, let us go back to the issue of epistemic justification. If worldP were actual there would be no NENP. There would, though, be experience of pain and suffering, as well as experience of our ability to harm or help others. Why would this not be a sufficient experiential basis from which to learn about the consequences for good and ill of our actions and negligences? I can think of no good reason. We would have implanted practical beliefs about what does and would happen in certain situations. Those beliefs, let us say, are true. Those beliefs would also be experientially justified, for we would see the effects of our actions. To turn the issue around now, the question that goes backto Swinbume is that of why, in addition to such experience, we would also need experience of NENP in order to justify our true practical beliefs? I do not think that Swinburne has provided a good reason. But he must if his position is to survive. Thus, I conclude that he has not shown NENP necessary for humans to learn about the good or evil consequences of their behavior and thus to be justified in their beliefs about that behavior. Thus he has made headway in regard to neither the quantified nor unquantified problem of natural evil. The foregoing counter-argument to Swinburne's position is insufficient to establish that NENP is unjustified in a God-made world, for there could be justifications other than those proposed by Swinburne. It does establish, though, that NENP is unjustified as an educational device in a God-made world, and that is Swinburne's sole candidate-justification of NENP. Now, implicit in the argument I have formulated, if that argument is good, is a full and adequate defense of theism against one version of the logical argument from evil, namely: the version challenging theism to reconcile certain specific types or tokens of natural evil with God. For my argument here has allowed that certain natural evils - natural evils through natural processes - are compatible with the existence of God. But traditional theists may wish to look this particular gift horse in the mouth, suspecting it to be Trojan. For, if my argument is a good one, it seems to make another version of the logical argument even more difficult than before. That other version, the quantified version as I referred to it earlier, is one which challenges theism to reconcile all natural evil with the existence of God. And the foregoing argument both heightens theism's

86 need to justify NENP in a supposedly God-made universe and at the same time closes off one possible (and prima facie obvious and attractive) route to any such justification. Let us now briefly take up the issue of moral justification.

2.2 Moral justification I have accepted Swinbume's terms of discussion. On the ground for discussion that he has staked out, the key issue has been whether we could have practical knowledge without there having to be NENP from which we could learn of the good and evil effects of our actions and negligences. I have shown that we could. In worldP, then, there is no epistemic barrier to our developing into full moral agents with responsibility for our own lives and with full recognition of our moral responsibilities towards others. And so woddP is a world an omnipotent, omniscient, perfect being could bring about. The upshot is that, insofar as epistemic factors are concerned, we would be no less morally justified or unjustified in our practical beliefs and actions in worldP than we are in the actual world of our experience. This is because, in the context of the present discussion, although not just there, moral justification is a function of epistemic justification.

. Swinbume's natural theology is among the most ambitious projects in contemporary philosophy of religion. In the three books and various satellite papers where he has developed this natural theology, it is a rich tapestry of insight, clarification, and argument that Swinbume has woven. Yet, for all the light that his work has shed on many aspects of theistic philosophy, if I am right that he has failed to justify natural evil, or, more specifically in the context of the present paper, that he has failed to justify NENP, that large and ambitious project will be significantly reined in, and other philosophical tactics will have to be employed if NENP is to be shown justified in a God-made universe. Furthermore, and more generally, as Swinbume's is a version of "soul-making" theodicy, the failure of his attempt to explain and justify NENP in a God-made universe may have implications for that approach to theodicy beyond those it has for Swinburne's own particular version. However, that is a topic beyond the scope of this paper and so, apart from now pointing in the direction of that possibility, I shall not pursue the topic any further here.

87

Notes 1. Richard Swinburne, "Knowledge from Experience, and the Problem of Evil," The Rationality of Religious Belief, ed. W. J. Abraham and S. W. Holtzer (Oxford: Clarendon Press, 1987), pp. 150, 152-155: see also Swinburne's The Existence of God (New York: Oxford University Press, 1979), especially pp. 202-220. 2. David O'Connor, "On The Problem Of Evil's Not Being What It Seems," The Philosophical Quarterly 37, no.149 (October 1987): 441. 3. Swinburne, The Existence of God, 219. 4. This chain is a composite of Swinbume's thinking in the two works cited in fn. 2 above. 5. David O'Connor, "Swinbume On Natural Evil," Religious Studies 19. 1 (March 1983): pp. 65-73 and "A Variation On The Free Will Defense", Faith and Philosophy 4.2 (April 1987): 160--167. 6. Swinburne, "Knowledge from Experience, and the Problem of Evil," pp. 150--165. 7. Ibid., pp. 151-152. 8. Ibid., p. 151. 9. Ibid., p. 155. 10. Ibid., p. 156. 11. Ibid., pp. 158, 165. 12. Ibid., p. 164. 13. Ibid. 14. Ibid., p. 158. 15. Ibid., p. 165. 16. If this is right, that is, if the answer to the second question is indeed no, then I must trim back a proposition for which I have argued elsewhere, namely: that no experience of any natural evil is logically necessary for practical knowledge. See my "Swinbume On Natural Evil," pp. 72-73. 17. Michael Martin, "Reichenbach On Natural Evil," Religious Studies 24 (March 1988): 93-94. 18. I should tidy away a verbal point here. I referred to those actions from experience of which we would learn about the evil effects of our actions as reckless actions. But the term connotes culpability, whereas, when first done or observed they would not be culpable at all. The point, though, is this: that, upon observing the effects of certain actions which are of a sort with those which, in the actual world, we call reckless, we learn about the ill effects of such actions. 19. For discussion of this see my "Swinburne On Natural Evil," p. 69, and for Swinbume's response on this particular topic see his "Knowledge from Experience, and the Problem of Evil", p. 157, fn. 10