Sophia, Vol. 45, No. 1, May 2006. Copyright 9 2006 Ashgate Publishing Limited.

ROWE'S NEW EVIDENTIAL ARGUMENT FROM EVIL: PROBLEMS AND PROSPECTS N I C K TRAKAKIS Department o f Philosophy, School o f Philosophy & Bioethics Monash University, Clayton, Victoria, 3800, Australia Nick. [email protected]

This paper examines an evidential argument from evil recently defended by William Rowe, one that differs significantly from the kind of evidential argument Rowe has become renowned for defending. After providing a brief outline of Rowe's new argument, I contest its seemingly uncontestable premise that our world is not the best world God could have created. I then engage in a lengthier discussion of the other key premise in Rowe's argument, viz., the Leibnizian premise that any world created by God must be the best world God can create. In particular, I discuss the criticisms raised against this premise by William Wainwright as well as Rowe's attempt to meet these criticisms. The Wainwright-Rowe exchange, I argue, highlights some insuperable dijficulties in Rowe's challenge to theism.

William Rowe is perhaps best known for developing and defending for more than two decades a variety of evidential arguments from evil. Like most evidential arguments, Rowe's arguments rely on a crucial inductive step or 'noseeum inference' to the effect that, 'No goods we know of justify God's permission of evil; therefore, it is likely that no goods at all justify God's permission of evil'. This inference has come under much fire of late from 'sceptical theists' who question the underlying 'noseeum assumption' that, 'If there were goods justifying God's permission of evil, more likely than not we would discern or be cognizant of such goods'. Recently, however, Rowe has put forward a new evidential argument, one that is free from any such questionable 'noseeum assumptions', and relying instead on ideas or intuitions closely associated with Leibniz's Theodicy. In this paper I wish to examine Rowe's new argument, focusing in particular on the problems and prospects of its two key premises.

1. Rowe's New Evidential Argument Rowe sets out his new evidential argument in his 1999 paper, 'Evil and God's Freedom in Creation'. 2 In this paper he begins by noting that possible

58

NICK TRAKAKIS

worlds may be divided into three categories: those that are on the whole good, those that are on the whole bad, and those that are on the whole neutral, neither good nor bad. He then goes on to note that theists are committed to the claim that the actual world is on the whole good, since they hold that the actual world was created by a being that is omnipotent, omniscient, and wholly good, and surely such a being would only choose to create a good world. Rowe then proceeds to argue that theists are also committed to a much stronger claim, the claim that any world created by the theistic God would not only be overall good but also better than any other world ereatable by God. To support the attribution o f this stronger claim to the theist Rowe argues as follows: If there were some other creatable world that contained as much good as our world, but less evil, it would be a better world. But surely, God, being absolutely perfect, would then have created a world like that, rather than our world. Also, if there were another creatable world that contained no more evil than our world, but more good, it would be a better world. And surely a perfectly good being would have created a world like that, rather than the world that happens to be actual. So, if we are theists, it won't do to believe just that the actual world is a good world. We should also believe that no creatable world is better than the actual world. And we should also believe that no creatable world with more good than the actual world is better than the actual world. 3 We may therefore state the first premise o f Rowe's new evidential argument as follows: (1) If an omnipotent, omniscient, perfectly good being creates some world W, W must be the best world that being can create. A similar claim was, o f course, made by Leibniz, who infamously went on to argue that the actual world is the best o f all possible worlds. Premise (1) may therefore be called the Leibnizian premise. In the following step o f his argument, Rowe introduces the problem o f evil in the guise o f the question, Is our world the best God could do? To answer this question, Rowe distinguishes between two varieties o f the problem o f evil. First, there is the similar past worlds problem of evil. Consider the actual case o f a five-year-old girl who was brutally beaten, raped, and then strangled to death in Flint, Michigan on New Year's Day in 1986. 4 Following received tradition, call this girl 'Sue'. Consider, further, all those possible worlds which have an identical or very similar past to the actual world up until some time shortly before the aforementioned tragedy on New Year's Day in 1986. Let's assume that in these possible worlds

ROWE'S NEW EVIDENTIAL ARGUMENT FROM EVIL

59

which share a similar past to our world, God averts the tragedy by some way or other. For example, God may bring it about that Sue dies peacefully in her sleep, or he may bring it about that the girl's attacker is seized by a fatal heart attack moments after forming the intention of beating her. Now, even if some of these worlds in which God prevents the girl from being abused are worse overall than the actual world, it is surely the case that at least some of these worlds are better overall than the actual world9 And it is incredible to believe, according to Rowe, that an omnipotent being was powerless to create any of these better worlds instead of the actual world. 5 But the similar past worlds problem of evil is, according to Rowe, merely 'child's play' compared to the all worlds problem of evil. In the similar past worlds problem of evil, the issue is whether any worlds which share a similar past to our world (up to the time of the occurrence of the evil in question) can be judged to be better overall than the actual world. By contrast, the all worlds problem of evil concerns the question whether the actual world can be judged to be better than every other possible world, or every other world that could have been created by God. 6 In relation to this question, consider first the belief (defended by Leibniz) that no possible world is better than the actual world. Rowe takes this to be an 'absolutely absurd belief'. As he goes on to say, 9 the idea that our world with all its horrendous evil is just as good, if not better, than any other possible world is an idea that merits the ridicule Voltaire heaped upon it. 7 But consider next the slightly different view that no possible world creatable by God is better than the actual world. Rowe's attitude toward this belief is equally unfavourable: The idea that an omnipotent, omniscient being should be capable of creating the actual world but incapable of creating any of the possible worlds better than it is, I think, bound to strike us as absurd, at least when we seriously reflect on the horrific evils played out daily on the stage of the actual world. 8 Rowe's view, then, is that the belief that the actual world is the best possible world as well as the belief that the actual world is the best (or one of the best) of all worlds creatable by God cannot be taken seriously - both beliefs are absurd, deserving nothing less than ridicule. The second premise in Rowe's argument may therefore be put as follows:

60

NICK TRAKAKIS

(2) It is highly unlikely that the actual world is the best world an omnipotent, omniscient, perfectly good being can create. It is this premise that gives Rowe's argument its evidential character. For the claim here is that it is unlikely - but not (logically or metaphysically) impossible - that the actual world is the best world God could have created. But perhaps Rowe is making a stronger claim than merely that it is highly unlikely that the actual world is the best creatable world. 9 To borrow an example suggested by Rowe, it is logically possible that Saddam Hussein and Margaret Thatcher will marry this year and spend the rest of their lives on a farm in Texas growing peanuts - but this is not just highly unlikely, it is an 'absolute absurdity;. However, as this example indicates, Rowe does not identify the absolutely absurd with the logically impossible, and he is also committed to the claim that, necessarily, absolutely absurd beliefs are highly unlikely. Proposition (2), therefore, does not seem to be a misleading way of representing Rowe's views. The conjunction of premise (2) with the Leibnizian premise leads to the following conclusion: (3) (Therefore) It is highly unlikely that the actual world was created by an omnipotent, omniscient, perfectly good being. But since traditional forms of theism hold that the actual world is the creation of an omnipotent, omniscient, perfectly good being, it follows from (3) that it is highly unlikely that any traditional variety of theism is true.

2. Challenging Premise (2) In assessing Rowe's argument I will consider each of the argument's two premises, beginning with premise (2). Initially, at least, the task of mounting a successful case against (2) appears formidable. For to reject (2) it seems that one would need to place severe restrictions on God's power and knowledge without thereby rejecting his omnipotence and omniscience. Perhaps the recent move away from Molinism and towards open theism suggests a way of doing just this. 1~For if God lacks foreknowledge of the free choices of any creature, and if he chooses to create a world populated by free creatures, he could not know how the history of such a world would unfold. Even if, in such circumstances, God somehow knows that the world in question will turn out to be good overall, he might well be in the dark as to how much good and evil and what specific goods and evils will be instantiated. And so, on this view,

ROWE'S NEW EVIDENTIAL ARGUMENT FROM EVIL

61

the actual world - or something analogous to the actual world - is the best God could do. Whether the open theist can successfully run such an argument remains to be seen. Open theism has attracted much heated debate, and there is a perception amongst many critics that open theism diminishes God's goodness by turning him into a reckless risk-taker. 11 For if God's knowledge is as limited as the open theist holds, then he would not have the means to providentially govern the world, including the lives of each of his creatures (the 'open God', for example, would be unable to prevent the birth of people who freely choose a life of moral depravity and die unrepentant). But a God who cannot providentially govern the world and who nevertheless decides to proceed with creation is not so much playing dice as Russian roulette. I would like to suggest a more plausible line of response to premise (2). This premise, it will be recalled, states that, more likely than not, the actual world is not the best world a perfect being could create. Rowe attempts to support this view by claiming that a perfect being could simply replace the actual world with a better one in which such evils as that involving the abuse and murder of Sue do not occur. Rowe, as we have seen, thinks that to deny this is to be committed to the utterly absurd position that our world, littered as it is with horrific evils, could not have been improved upon in any way by God. But is it so obvious that this position is absurd or deserving of ridicule? It is undoubtedly true that there exists a better possible world than the actual world. To see this, simply imagine the actual world less one minor evil (e.g., the paper cut I experienced yesterday), or alternatively imagine the actual world with that evil resulting in slightly less physical pain. Even if Voltaire's famous depiction of Leibniz in Candide was merely a caricature, it is difficult not to be sceptical of Leibniz's view that the removal of even the smallest evil that comes to pass in our world would diminish the overall value of the worldJ 2 Indeed, as James Franklin has recently stated, 'Some think that the whole point of religion and ethical endeavour is to make the world better, which would seem to be impossible if it were already as good as it could be. '13 But the claim made in premise (2) of Rowe's argument is not that the actual world could have turned out better than it has, but that God could have ensured that the actual world turned out better than it has. And the theist - or, more precisely, the theodicist - may dispute this latter claim. For suppose that one were to offer the following theodicy with respect to the sufferings of Sue:

NICK TRAKAKIS

62

Following Hick's lead, we may postulate that the divine intention in relation to humankind is to bring forth morally and spiritually mature finite personal beings by means of a 'vale of soul-making' in which humans have the opportunity to transcend their natural self-centredness by freely developing the most desirable qualities of moral character and entering into a personal relationship with their Maker. Any world, however, that makes possible such personal growth must be one in which there is morally significant freedom, that is to say, the capacity to bring about not only good and evil, but also a range of actions that vary enormously in moral worth, from great and noble deeds to horrific evils. But given God's design of the world as a vale of soul-making, it is not (logically) possible for him to replace the actual world with a better one by intervening to prevent the murder of Sue. For if God were to prevent S's exercise of free will, then God would have to prevent every person's exercise of free will at all times, where 'S's exercise of free will' represents Sue's attacker's exercise of free will during his brutalization of the child. In support of this view, it might be argued that God would display a degree of unfairness or indiscriminate partiality if he were to intervene to prevent one instance of evil (viz., Sue's murder) but not intervene in other similar cases. But if God were to avert every misuse of free will, there would clearly be no freedom worthy of the name left for us to have. 14 The theodicist, then, is claiming that God permits evils such as child abuse for the sake o f goods such as soul-making and free will. This theodicist, I would add, must also claim that victims o f God's soul-making system such as Sue are adequately compensated by way o f a heavenly afterlife. Thus, the goods o f soul-making, free will, and heavenly bliss morally justify God's permission o f moral evil, even horrific moral evils. Call these goods God-

justifying goods. When are we justified in believing that a given good G qualifies as a God-justifying good with regard to God's permission o f some evil E? This is a difficult question, and a complete answer would require the development o f a meta-theodicy. But perhaps we can say that at least the following conditions must be satisfied before we can be justified in holding that some evil E serves a God-justifying good: (4) There is a good G that outweighs E, (5) G is such that God would have forfeited it had he prevented E. However, as Rowe points out, these two conditions are necessary but not sufficient for there being a God-justifying good: . . . it could be that permitting a particular evil is necessary for God to bring about a good that outweighs it; whereas preventing that very

ROWE'S NEW EVIDENTIAL ARGUMENT FROM EVIL

63

evil would enable God to bring about an even greater good. And if this were so, the outweighing good that God could not bring about without permitting that evil would not be a God-justifying good for his allowing that evil) 5 Rowe therefore seeks to identify a sufficient condition for there being a God-justifying good for God's allowing some evil to occur, and he proposes the following as a plausible candidate: (6) God's prevention of E would result in a world that is worse (or at least, no better) than the actual world. 16 Rowe is then led to consider 'the similar past worlds' and 'the all possible worlds' problems of evil. We can imagine some worlds whose history is identical to the history of the actual world but deviate from the actual world's history only by virtue of not including the sufferings of Sue (and any other events entailed by or supervenient upon the evil in question). We can also imagine worlds which have a very different history from the actual world, with one of the differences being that in the alternative worlds the sufferings of Sue never take place. At least initially, it seems obvious that at least some of these other worlds are better than the actual world, and that these better worlds could have been created by an omnipotent being. This initial appearance of obviousness, however, is deceptive. For as I have indicated by way of the theodicy sketched above, once God's purposes or goals for humanity are spelt out, it may turn out that God's prevention of a particular instance of evil would result in a world that is less good overall. And this, indeed, is the case when the evil in question happens to be an instance of moral evil. Thus, the truth of Rowe's claim in premise (2) that it is very likely that God could have created a better world is not as obvious or self-evident as Rowe makes it out to be. In fact, nothing less than a consideration of at least some of the major theodicies that have been developed by various theists over the centuries is required in order to arrive at a proper assessment of (2). Therefore, to dismiss the idea that this world could not have been bettered by God as 'an absurdity, fit for ridicule' is to fail to appreciate the complexity of the matter. 17

3. Challenging the Leibnizian Premise (1) If an omnipotent, omniscient, perfectly good being creates some world W, W must be the best world that being can create.

64

NICK TRAKAKIS

This premise has been heavily debated ever since the publication in 1972 of Robert Adams' seminal paper, 'Must God Create the Best?'. 18 Critics of premise (1) may be divided into two camps. Firstly, there are those who follow Thomas Aquinas in denying that there is a best world creatable by God. On this view, there is an infinite number of possible worlds, and for any world creatable by God there is a better world creatable by God. Alternatively, one may hold that, irrespective of whether there is a finite or infinite number of possible worlds, there is no single best treatable world but only several equally good creatable worlds. The second camp of critics consists of those who maintain that, even if there is a best world creatable by God, it is entirely consistent with God's perfect goodness that God creates a world that is less than the best. It is this latter and perhaps more adventurous line of thought that Adams sought to defend in his 1972 paper. 19 Each group of critics, it may be noted, is committed to the view that a perfect being (as represented by the theistic conception of God) may create a world W even when there is a world better than W that it could have created instead. However, in his recently published book, Can God Be Free?, Rowe argues that the idea of a perfect being creating a world when there is a better world it could create must be rejected as incoherent, z~ Rowe bases his case on the following very plausible principle: (7) If an omniscient being creates a world when there is a better world it could create, then it would be possible for there to be a being morally better than it. The idea here is that if some being A creates world W 1 even though A could have created a better world, W2, then it is possible for there to be another being B who creates W2. But then B, by virtue of deciding to create a world better than W 1, is morally better than A. And so, if being A creates W 1 when it could have created a better world, then it is possible for there to be a being that is morally better than A. Now, following Anselm, theists usually think of God as a maximally perfect being or the best being possible. On this view, it is not possible for there to be a being that is morally better than God. According to (7), however, if we assume that God creates some world when there is a better world he could have created instead, it follows that it is possible for there to be a morally better being than God. In other words, if we assume that God creates some world when there is a better world he could have created instead, it is possible for there to be a being better than the best possible being - and this, of course, is an incoherent view. Thus, the idea of a perfect being creating a world when there is a better world it could create ought to be rejected as (implicitly) incoherent, zl

ROWE'S NEW EVIDENTIAL ARGUMENT FROM EVIL

65

It is clear, then, that Rowe's defence o f his Leibnizian premise is heavily indebted to the principle expressed by proposition (7). In what follows, therefore, I will discuss what I take to be one o f the more forceful objections that has recently been raised against (7).

3.1 Wainwright's Dictum William Wainwright has argued that, given the assumption that there is no best creatable world, it is inappropriate to criticize God for failing to create a world better than the one he has in fact created. 22 Wainwright's case in support o f this position may be summarized as follows. Suppose that the number o f creatable worlds is infinite. It then follows that no matter what world God creates there would always be a better world that he could have created instead. But then 'no matter what God did, He would be exposed to the possibility o f a complaint o f this type. '23 That is to say, no matter which world God creates, he would be open to the criticism that he could have - and therefore should have - created a better world. But, according to Wainwright, 'a type o f complaint that is always in place is never in place. '24 In other words, a complaint - or the act o f censuring a person's moral behaviour - is appropriate only if the person whose conduct is being criticized could have acted in such a way that she would not be exposed to a complaint o f that type. The charge brought against God (that he should have created a better world) is thus shown to be illegitimate. Wainwright therefore concludes that, 'even though there are an infinite number o f possible created orders better than our own, God can't be faulted simply because He created an order inferior to other orders that He might have created in their place. '25 The key to Wainwright's case lies in his claim that a type of complaint that is always in place is never in place. Call this Wainwright's Dictum. What examples could be given in illustration o f it? Wainwright offers the following: For example, other things beings equal, it would be unfair to blame someone for the unfortunate consequences of his or her action if all of the available alternatives would have had equally unfortunate consequences. 26 Take, for example, the case o f Sophie in William Styron's novel, Sophie's Choice. She is forced by an SS officer to make the terrible choice of whether to send her son or her daughter to the gas chambers. If she refused to make a choice, both children would be killed. As it happens she chooses to save

66

NICK TRAKAK1S

her son, and so her daughter is taken away to be murdered. 27 The point to be made here is that whatever decision she made, the consequences would be extremely dire. It would be difficult to criticize her decision to not remain silent, for in that case she would lose both of her children, not just one of them (not that this would afford her much solace). But could she be criticized for choosing her son over her daughter? Well, if we assume that the altematives open to her are equally bad - i.e., if the overall consequences (for all parties involved) of choosing her son are just as bad as the overall consequences of choosing her daughter - then it would be wrong to criticize her on the grounds that she failed to save her daughter. And the reason this criticism is out of place is that the same type of complaint (viz., 'You failed to save one of your children') could be levelled at her no matter what she did. But a type of complaint that is always in place is never in place, as Wainwright puts it. In the light of such examples, Wainwright's Dictum has much to be said for it, and if his Dictum were accepted there would be little alternative but to reject Rowe's Leibnizian premise. It seems, then, that Wainwright has presented an ingenious response to Rowe. In fact, as Wainwright acknowledges, the line of response he has developed is simply a reformulation of a novel solution to the problem of evil proposed many years ago by George Schlesinger. Beginning with a paper published in 1964, and thereafter in a number of other places, Schlesinger made the astonishing claim to have resolved once and for all the seemingly intractable problem of evil. 28 His 'solution' is quite simple. It centers around a principle that he characterizes as 'one of the universal rules of ethics': Everything else being equal, increase the degree o f desirability of the state of a given person by as much as possible. 29 Schlesinger's notion of 'the degree of desirability of a state of existence' (or DDS) may need some clarification. The DDS of any given being is a function of two variables: (a) the kind of being it is, determined by the range of capacities for sensation and experience it has, and (b) the degree to which these capacities are fulfilled, thus yielding pleasure or pain, happiness or suffering. According to Schlesinger, the DDS of any sentient creature can be ordered and ranked on an objective scale of moral value. Socrates, for example, in virtue of his vastly greater capacities, has a higher DDS than a fool or a pig, no matter how content they may be. Schlesinger therefore endorses John Stuart Mill's well-known judgment that, 'It is better to be a human being dissatisfied than a pig satisfied; better to be Socrates dissatisfied than a fool satisfied. '3~ In short, Schlesinger's point is that the most fundamental kind of moral obligation any person has is not to maximize the happiness and minimize the suffering of others, but to endeavour to increase their DDS as much as possible.

ROWE'S NEW EVIDENTIAL ARGUMENT FROM EVIL

67

This, Schlesinger claims, allows us to resolve the problem o f evil in the following way. He points out first that there is no logical limit to the degree which the desirability o f a state may reach. And so the DDS of Socrates may be superseded by the DDS o f a super-Socrates 'who has a much higher intelligence and many more than five senses through which to enjoy the world and who stands to Socrates like the latter stands to the pig. '31 In like fashion, we can conceive o f a super-super-Socrates, and so on ad infinitum. The critic o f theism, however, may point to the pitiable state o f various creatures in our world, thus raising the complaint that God has failed to live up to his side o f the bargain by not fulfilling the aforementioned universal moral principle (viz., increase others' DDS as much as possible). But Schlesinger maintains that this complaint is empty. For as indicated earlier, there are no logical limits to the extent to which a creature's DDS can be raised. And so whatever God does - that is to say, whatever DDS he confers to a creature - the same type o f complaint will be forthcoming: Why didn't you give that creature a higher DDS? But as Wainwright's Dictum has it, a type o f complaint that is always in place is never in place. Schlesinger therefore concludes that There is no room for complaint, seeing that God has not fulfilled the ethical principle which mortals are bound by and has left his creatures in various states of low desirability. Thus the problem of evil vanishes. 32

3.2 Rowe's Rejoinder and Why It Is Inadequate Rowe has raised two objections to Wainwright's critique o f the Leibnizian premise, which would apply with equal force to Schlesinger's proposal. Rowe's first objection is that the kind o f example Wainwright employs to illustrate his Dictum is 'not germane to the case at hand'. 33 For in such cases as that involving Sophie's choice between her son and her daughter, the choice is between two outcomes that are presumed to be equal in value (or disvalue). But in the case involving God's choice as to which world to create, the choice is not between two or more worlds o f equal value. Rather, given the assumption that there is no best creatable world, God must choose from an infinite set o f worlds, where any given world is of greater or lesser value than any other. This, o f course, is a point o f difference between the two cases, but is it a relevant difference? Rowe thinks it is. For suppose that God, like Sophie, were limited to choosing between alternatives o f equal value. More specifically, suppose that God were limited to selecting from an infinity o f equally good worlds. In that case, whatever choice God were to make, his choice could not be faulted; at least, there would be no reason

68

NICK TRAKAKIS

to think, on the basis of the choice made by God, that he was less than perfectly good. On the other hand, if God were faced with an infinity of increasingly better creatable worlds, then the choice he does make would according to Rowe - give us reason to think that his degree of goodness is less than perfect. I think Rowe is right, at least to the extent that the kind of example given by Wainwright and illustrated further by my reference to Sophie's choice does not parallel the kind of choice we are presuming is faced by a worldcreator - that is to say, a choice from among increasingly better worlds, not a choice between equally good worlds. Rowe's task, therefore, is not to show that some being can be at fault in choosing a particular object from an infinity of equally good objects. Rather, Rowe's task is to show that a being will fail to be supremely perfect if it selects some world for creation from a set of increasingly better worlds, where none of these worlds can be labelled 'the best'. Whether Rowe succeeds in this respect is another matter. Rowe's second objection is that his position is entirely compatible with Wainwright's Dictum. 34 Rowe states that he need not disagree with Wainwright that no legitimate complaint or criticism can be made of a world-creator who selects one world for creation from an infinite array of increasingly better worlds. Rowe, therefore, states that he is not levelling some complaint or criticism against such a world-creator. Rather, he claims to be offering some reason for thinking that this creator's degree of goodness is surpassable. In Rowe's words, -

[G]iven an infinity of increasingly better worlds from a base starting point of a minimally good world, such a creator [i.e., an omniscient and omnipotent creator] presumably will select a world from the range of increasingly better worlds that is acceptable for creation given the creator's own degree of goodness . . . . But since there is no end to increasingly better worlds, there need be no end to the increasingly greater degrees of goodness in possible creators resultin~ in increasingly better sets of worlds judged as acceptable for creation.3J Rowe goes on to state that the view expressed here is underwritten by the following three assumptions: AI: A good, omnipotent, omniscient being will select a particular world for creation only if he judges that world to be acceptable for creation, given his own degree of goodness. A2: Given an infinity of increasingly good worlds and an omnipotent, omniscient being who creates one of these worlds, it is logically possible that there should be an omnipotent, omniscient being whose degree of goodness is such that he

ROWE'S NEW EVIDENTIAL ARGUMENT FROM EVIL

69

would judge that world, given its degree of goodness, to be unacceptable for creation. A3: The degree of goodness of a world that is acceptable for creation by an omnipotent, omniscient, good being is indicative of the degree of goodness of that being. 36 Given these three assumptions, it follows that no matter what world an omnipotent, omniscient being creates, its degree of goodness will be surpassable. 'But to point this out,' adds Rowe, 'is not to raise a complaint against this being. It is simply to note what follows from the three assumptions just noted.'37 And what follows from assumptions A1-A3 is Rowe's premise (7): If an omniscient being creates a world when there is a better world it could create, then it would be possible for there to be a being morally better than it. In this case, however, I think Rowe's reply to Wainwright is inadequate. The point behind Wainwright's Dictum is that a complaint of the form 'Person S did A when she ought to have done B' is not appropriate when a complaint of this form can be made against S no matter what she chooses to do. Similarly, if a reason R for thinking that a person S is not perfectly good is such that R cannot be defeated by anything S does, then R is a spurious reason. And so even if Rowe is Offering a reason for thinking that the creator's degree of goodness is surpassable, rather than raising a complaint against the creator, the reason offered by Rowe is baseless. More precisely, if Wainwright's Dictum is taken to be applicable to the case at hand, then there is no basis for thinking that a creator who decides to create some world W, when a world better than W was available for creation, is not perfectly good. For any reason to think that the creator's degree of goodness is not perfectly good (on the grounds that the creator selected world W for creation when he could have selected a world better than W) is voided by Wainwright's Dictum. Rowe, however, holds that his case in support of the moral surpassability of an omnipotent, omniscient being who chooses to create some world even though a better world is available, far from being baseless, is in fact based on assumptions A1-A3. As Rowe recognizes, the crucial and most problematic assumption here is A3. According to A3, the degree of a creator's goodness can be measured by the kind of world that creator judges to be acceptable for creation. Suppose, for example, that one possible world W1 is of far greater moral worth than another possible world W2. It would seem that a being who judges any world other than W2 to be unacceptable for creation is not as good as a being who judges any world other than W 1 to be unacceptable for creation. It is this seemingly plausible intuition that A3 is intended to capture.

70

NICK TRAKAKIS

But A3 presupposes that there is a best creatable world. If we remove this presupposition, whatever force lying behind A3 disappears. To see this, suppose that there is an infinity o f increasingly better creatable worlds, beginning with W1 and having no upper limit. Suppose also that an omnipotent, omniscient, good being called Ra (after the Egyptian sun-god) selects world W300 for creation. If the set o f creatable worlds were finite, with (say) world W500 representing the best creatable world, then we may have some reason to think that Ra's decision to create W300 indicates that his degree o f goodness is less than perfect. But if, as we are assuming, the set o f creatable worlds is infinitely large, so that for any world W one picks there is a world better than W~ then Ra's choice o f W300 cannot be taken as a reason for thinking that his character is not wholly good. For one could offer a reason o f this type in criticism o f Ra's character no matter what Ra decides to do. But given Wainwright's Dictum, a criticism that is always in force is never in force. A3 is therefore predicated on the assumption that there is a best creatable world, and it is from A3 in conjunction with A1 and A2 that Rowe builds his case in support o f the Leibnizian premise. 38 The structure o f Rowe's reasoning may be schematzed as follows: There is a best creatable world --) A3. A1 & A2 & A3 --) Premise (7) --) Leibnizian premise. O f course, if there is a best creatable world then Wainwright's Dictum is inapplicable in that it cannot be offered in defence o f God's decision to create our world. But as I noted at the beginning o f Section 3, some critics have sought to challenge the claim that God ought to create the best world by maintaining that there simply is no best world available for creation. If the position o f these critics is correct, then Rowe would not be entitled to assumption A3, and so his case in support o f the Leibnizian premise would collapse. But are these critics right to reject the idea o f a best creatable world? The answer to this question depends to a large extent on what we mean by 'best' when we speak o f 'the best creatable world'. What qualities must a (creatable) world have in order to count as 'the best'? I will consider several different responses to this question, each o f which (I will argue) fails to establish that there can be a best (creatable) world. 39 According to one line o f thought, a (creatable) world may be said to be the best only if it is inhabited by the greatest variety o f sentient and non-sentient things. This view is derived from what Arthur Lovejoy, in his now classic study The Great Chain of Being, calls the principle of plenitude, a principle which, as Lovejoy shows, was one o f the most enduring ideas in Western

ROWE'S NEW EVIDENTIAL ARGUMENT FROM EVIL

71

philosophy, going as far back as Plato and the Neoplatonists, and becoming a central element in the cosmology of medieval Christianity. 4~ Perhaps the chief reason the principle of plenitude found so much favour amongst theists was that it provided a neat solution to the following problem: Why did God create an enormously wide variety of living organisms, with the inequalities and conflicts this inevitably introduces amongst the different species, rather than restrict his creation to the most perfect of contingent beings? The principle of plenitude affords a simple answer: a world containing a rich variety of beings is more valuable than one containing only one sort of creature. The natural world is thus pictured as consisting of a great hierarchy of beings, beginning at the lowest end of the scale with inanimate objects (e.g., rocks, stars), then ascending to living things that lack sentience (e.g., trees, plants), and thence to living things that are sentient but not intelligent (e.g., animals), thence to living things that are both sentient and intelligent (e.g., humans), and finally to intelligent beings that are immortal (e.g., angels). At the summit of the hierarchy stands God. And it is usually added that God has chosen to create a world of this sort because it is intrinsically better that many kinds of beings exist, rather than just one kind of being or the best kind of beings. Given this value-judgment, it inevitably follows that the best (creatable) world will be one which displays the greatest variety of sentient and non-sentient things. But it is not difficult to see how this view fails to show that there is a best creatable world. For if the best world is one populated by the greatest variety of beings, then we must hold that there is an intrinsic maximum to the variety of things that can exist in the created order - that is to say, there is a degree of variety among created things than which there cannot be a greater degree of variety. But why should we accept that? Suppose, for example, that we follow Christian tradition in thinking of angels as the highest kind of creature in the great chain of created beings. Angels, like all other creatures, possess a suite of qualities, such as the ability to believe and know various things. The cognitive capacities of angels, however, are limited, for only the cognitive capacities of God are without (logical) limit. Let's assume, then, that one way in which the cognitive capacities of angels is limited consists in the fact that they can only entertain n number of propositions at any one time. Surely we can then conceive of another species who are like angels in all respects, but who have the gift of entertaining n+l propositions at any one time. To generalize, for any kind of created being K1 we can conceive of another kind of created being K2 which differs from K1 only by possessing to some higher degree some good quality had by K1 or by possessing some good quality not had by K1 at all. In this way, we can always add a better kind of creature to the hierarchy of created beings. As

72

NICK TRAKAKIS

long as the chain o f being consists o f creatures who possess various types o f characteristics in varying degrees, these characteristics and their degree o f realization can always be amplified. The chain o f being, therefore, must be potentially infinite, and so there cannot be an upper limit to the variety o f kinds o f creatures populating any given world. 41 A second way o f understanding the notion o f 'the best creatable world' would be in the following 'optimific' terms: a creatable world qualifies as the best only if it consists o f the maximum amount o f good, or the most favourable balance o f good over evil. 42 But which states o f affairs are to be counted as 'good'? Here we meet with a plethora o f ethical theories, ranging from the hedonist thesis that good states o f affairs are those that maximize pleasure over pain to the eudaemonist view that the good is that which maximizes happiness. But whichever ethical theory is accepted, the notion o f a best creatable world turns out to be just as problematic as on the previous interpretation o f this notion. For we can easily imagine, for any given optimific state o f affairs (e.g., one containing n units o f pleasure), another state o f affairs which is more optimific (e.g., one containing n+l units o f pleasure). Therefore, there cannot be a best creatable world, understood in optimific terms. 43 Up to this point, I have been following the standard practice o f speaking o f God as creating a world. But by 'world' I clearly do not mean 'possible world', for possible worlds are actualized, not created. It would therefore make better sense to speak o f God as creating a spatio-temporally distinct universe, thus giving rise to (or actualizing) a possible world containing both himself and his creation. Once this more perspicuous terminology is adopted, it is no longer a question o f God's having to choose one universe for creation out o f an infinite number o f universes. For God may simply choose to create no universe at all, and this - for all we know - may constitute the best option available to God. If so, then there is a best course o f action God may take, a best option, even if there is no best universe available for God to create. But is the decision to create no universe at all a genuinely better decision than the decision to create some universe or other? At the very least, creating a universe that has a favourable balance o f good over evil, and in which each creature o f that universe is guaranteed to enjoy a life that is good on the whole, would give rise to a possible world o f greater value than one consisting solely o f God (and whatever necessary abstract entities there may be). O f course, there is a sense in which the overall value o f a possible world consisting entirely o f an absolutely perfect being would remain the same regardless o f whether any contingent being were added to that world. For the value o f the world in question would be infinite both before and after

ROWE'S NEW EVIDENTIAL ARGUMENT FROM EVIL

73

the addition o f the contingent being. But there is also an intuitive sense in which it is possible to add something to an infinite collection (for example, one can add the number two to the set o f odd numbers, even though this set has an infinite number of members). And so it does not seem implausible to think o f the set {God, universe } as having a greater overall value than the set {God}. But there remains another possibility that may rescue the notion o f a best world or best option. In the spirit o f the principle o f plenitude, perhaps the best course o f action for God to take with respect to world-creation is to create everything that he can create. More precisely, the idea is that God creates every universe that passes some minimum standards of acceptability for creation (e.g., the universe must not contain any creatures whose lives are not worth living). Having created every overall good universe there is, God cannot be criticized on the grounds that he could have done better. For there would be no further good left for him to accomplish, as it were. And so when Rowe asks, 'Why didn't God create a better world, for surely he could have?' the theist may simply respond: 'He already has'. This line o f defence has, I think, a certain attractiveness, and it has even been proposed by some writers - for instance, Peter Forrest and Donald Albert Turner, J r - as a solution to the problem o f evil. 44There are a number difficulties, however, that such an approach needs-to address. For example, how plausible is it to think o f our universe as only one o f an infinite plurality o f universes? If these universes are spatio-temporally and causally disconnected, then is the 'multiverse' hypothesis open to falsification - and if not, can it rightly be thought o f as a 'hypothesis'? And does not the postulation o f an infinite number o f universes blunt the force o f theistic cosmological arguments grounded in the fine-tuning o f the fundamental physical constants? 45 There is, however, an even greater worry with this conception o f the best world as one containing an infinite plurality o f (overall good) universes. At least from the perspective o f Cantorian mathematics, not all infinite collections have the same cardinality. In fact, there are, according to Cantor, infinitely many infinities larger than 1~o, there being no largest infinity or infinite collection. And so even if some possible world was populated by a set o f universes whose cardinal number was ~ o, there remain other possible worlds populated by universes that have a cardinality of, say, R 1 or ~ 2 or R 3, etc. Thus, even the creation o f a world containing an infinity o f (overall good) universes can be bettered. The notion o f a best (creatable) world therefore remains in tatters. 46

74

NICK TRAKAKIS

4. Conclusion To sum up, each premise o f Rowe's new evidential argument encounters significant difficulties. The second, and perhaps initially more plausible, premise is not as obviously true as Rowe would have us believe, for its truth is largely dependent on the complex matter o f whether some theodicy for both moral evil and natural evil can be successfully developed. The first, or Leibnizian, premise is also problematic, for it relies either on the falsity o f a quite plausible moral principle, encapsulated by Wainwright's Dictum, or on the highly dubious assumption that there is a best creatable world. Given these problems, it may be concluded that the prospects for developing a compelling evidential argument from evil along Leibnizian lines are dim indeed. 47

Endnotes 1. I discuss and defend one of Rowe's 'noseeum' arguments from evil in my forthcoming book, The God Beyond Belief" In Defence of William Rowe's EvidentialArgumentfrom Evil (due to be published by Springer). 2. Rowe, 'Evil and God's Freedom in Creation', American Philosophical Quarterly 36 (1999): 101-13. 3. Ibid., p.101. 4. A fuller account of this tragedy is provided by Bruce Russell in 'The Persistent Problem of Evil', Faith and Philosophy 6 (1989): 123. 5. Rowe, 'Evil and God's Freedom in Creation', pp. 103-104. 6. It is important to not conflate the notion of 'best possible world' with that of 'best world creatable by God'. These notions correspond to what Bruce Langtry calls a 'maximal world' and a 'prime world', respectively ('God and the Best', Faith and Philosophy 13 [ 1996]: 313, 319). 7. Rowe, 'Evil and God's Freedom in Creation', p. 105. 8. Ibid. 9. Rowe suggested as much in personal correspondence dated 23 August 2003. The example in the main text that follows is drawn from the same correspondence. 10. Open theism (orfree-will theism) interprets God's omniscience in such a way that it does not include either foreknowledge (or, more specifically, knowledge of what free agents other than God will do) or middle knowledge (i,e., knowledge of what every possible free creature would freely choose to do in any possible situation in which that creature might find itself). On the other hand, Molinist theists (named after the sixteenth-century Jesuit theologian Luis de Molina, who developed the theory of middle knowledge) conceive of divine omniscience as encompassing both foreknowledge and middle knowledge. On open theism, see the movement's manifesto, The Openness of God."A Biblical Challenge to the Traditional Understanding of God (Downers Grove, IL: InterVarsity Press, 1994), authored by Clark Pinnock, Richard Rice,

ROWE'S NEW EVIDENTIALARGUMENT FROM EVIL

11. 12.

13.

14.

15. 16. 17.

18. 19.

20. 21.

75

John Sanders, William Hasker, and David Basinger. The most comprehensive defence of Molinist theism is provided by Thomas Flint, Divine Providence: The Molinist Account (Ithaca, NY: Cornell University Press, 1998). A good introduction to these and other conceptions of divine omniscience is provided by James K. Beilby and Paul R. Eddy (eds), Divine Foreknowledge: Four Views (Downers Grove, IL: InterVarsity Press, 2001). See, for example, Kenneth Perszyk, 'Molinism and Theodicy', International Journal for Philosophy of Religion 44 (1998): 163-84. See G.W Leibniz, Theodicy: Essays on the Goodness of God, the Freedom of Man, and the Origin of Evil, trans. E.M. Huggard, ed. Austin Farrer (London: Routledge and Kegan Paul, 1952), w pp.128-29. J. Franklin, 'Two Caricatures, II: Leibniz's Best World', International Journal for Philosophy of Religion 52 (2002): 46, emphasis his. In this paper, however, Franklin attempts to cast some doubt on the initially plausible claim that we can easily conceive of a better world. See also Bruce Reichenbach, Evil and a Good God (New York: Fordham University Press, 1982), pp.110-17 and Michael Peterson, Evil and the Christian God (Grand Rapids, MI: Baker Book House, 1982), pp. 113-16, both of whom argue that our cognitive limitations (among other factors) preclude us from 'designing' a superior world containing as much good but less evil than the actual world. I would like to make two brief points in response to some helpful remarks from an anonymous referee. Firstly, it should not be presumed that the theodicy sketched above presupposes that God relies on his foreknowledge in order to forestall the occurrence of some evil. This view, as many have pointed out, may well be incoherent. Rather, the above theodicy should be construed as grounding divine providence in either God's middle knowledge or God's knowledge of what is likely to take place were he not to intervene. Secondly, a policy of non-intervention motivated by considerations of fairness does not sit well with the traditional theistic view of miracles, according to which God does not evenly distribute his miraculous gifts. This, I suspect, highlights the importance of re-thinking the commonly accepted view of miracles. Rowe, 'Evil and God's Freedom in Creation', p. 104. Ibid. Bruce Langtry has similarly argued that 'accepting the a-premise [i.e., the premise that, if God exists then if a possible world is actual then God cannot replace it by a better possible world] will not make the task of defending theism much harder than it otherwise would be' ('Can God Replace the Actual World by a Better One?' Philosophical Papers 20 [ 1991 ]: 192). Robert M. Adams, 'Must God Create the Best?' Philosophical Review 81 (1972): 317-32. Another possible response to (1) would be to claim that most, ifnot all, possible worlds are incommensurable with respect to value, so that it is not possible to compare the moral value of any two worlds. Rowe elaborates this view and defends it from various recent criticisms in ch. 6 of Can God Be Free? (Oxford: Clarendon Press, 2004). Rowe offers a similar argument in 'The Problem of Divine Perfection and Freedom', in Eleonore Stump (ed.), Reasoned Faith: Essays in Philosophical Theology in Honor of Norman Kretzmann (Ithaca, NY: Cornell University

76

22. 23. 24. 25. 26. 27. 28.

29. 30. 31. 32. 33. 34. 35. 36. 37. 38.

39.

40. 41.

42.

NICK TRAKAKIS Press, 1993), p.230. Essentially the same argument is advanced by Philip Quinn in 'God, Moral Perfection, and Possible Worlds', in Frederick Sontag and M. Darrol Bryant (eds), God: The Contemporary Discussion (New York: The Rose of Sharon Press, 1982), pp. 199-215. William J. Wainwright, Philosophy of Religion, 2nd ed. (Belmont, CA: Wadsworth, 1999), pp.91-92. Ibid., p.92. Ibid. Ibid. Ibid. William Styron, Sophie's Choice (London: Corgi Books, 1980), ch. 15, pp.642-43. See Schlesinger, 'The Problem of Evil and the Problem of Suffering', American Philosophical Qitarterly 1 (1964): 244-47; Religion and Scientific Method (Dordrecht: D. Reidel Publishing Co., 1977), chs 9 and 10; and New Perspectives on Old-time Religion (Oxford: Clarendon Press, 1988), pp.5367. Schlesinger, Religion and Scientific Method, p.62. J.S. Mill, Utilitarianism, ed. Samuel Gorovitz (Indianapolis, IN: The BobbsMerrill Company, 1971 [originally published in 1863]), ch. 2, p.20. Schlesinger, Religion and Scientific Method, p.62. Ibid., pp.62-63. Rowe, Can God Be Free?, p.l18. Ibid., pp.l18-19. Ibid., pp.118-19, emphasis his. Ibid., p.119. Ibid., emphasis his. Although Rowe accepts A3, he hesitates to commit himself to the existence of a best world, whether it be a best creatable or a best possible world. Instead, he prefers to simply leave open the possibility that the series of good worlds may have an upper limit. See Rowe, Can God Be Free?, p.52. As pointed out in note 6 above, the question at issue is not whether there is a best possible world, or even whether it is possible for God to create a world than which none is better. Rather, the issue is whether it is possible for God to create a world which is better than any other world that some possible being or thing (God included) could create. See Arthur O. Lovejoy, The Great Chain of Being: A Study of the History of an Idea (Cambridge, MA: Harvard University Press, 1957 [originally published in 1936]). A further difficulty with the appeal to the principle of plenitude resides with the principle itself. For it is not necessarily true that the existence of many kinds of beings is better than the existence of only one kind of being. Consider, for example, worlds where the variety in species introduces a multitude of conflicts between them but no outweighing benefits, thus swiftly leading to the extinction of most of the species. The two conceptions o f ' b e s t creatable world' presented here are not mutually exclusive. Indeed, Leibniz took the best world to be one in which both variety and happiness are maximized - see David Blumenfeld, 'Perfection and Happiness in the Best Possible World', in Nicholas Jotley (ed.), The

ROWE'S NEW EVIDENTIALARGUMENT FROM EVIL

77

Cambridge Companion to Leibniz (Cambridge: Cambridge University Press, 1995), pp.382-410. 43. Bruce Reichenbach reaches a conclusion similar to the one advanced in these last two paragraphs - see his 'Must God Create the Best Possible World?' International Philosophical Quarterly 19 (1979): 205-208. See also Richard Swinburne, The Existence of God, rev. ed. (Oxford: Clarendon Press, 1991), pp.113-14. The same criticism can be made against the view that there is no best world but only a set of equally good worlds. For one can easily imagine, given some degree of variety of created beings or some optimal amount of goodness, an even greater degree of variety or an even greater degree of goodness. A very different line of argument against the intelligibility of the notion of a best possible world is suggested by Michael Martin (drawing on an unpublished paper by Patrick Grim) - see Martin's Atheism: A Philosophical Justification (Philadelphia, PA: Temple University Press, 1990), pp.440-41. 44 See Peter Forrest, God without the Supernatural: A Defense of Scientific Theism (Ithaca, NY: Coruell University Press, 1996), oh. 8, and Donald Albert Turner, Jr, 'The Many-Universes Solution to the Problem of Evil', in Richard M. Gale and Alexander R. Pruss (eds), The Existence of God (Aldershot: Ashgate Publishing, 2003), pp.143-59. See also John D. McHarry, 'A Theodicy', Analysis 38 (1978): 132-34, and R.K. Perkins, Jr, 'McHarry's Theodicy: A Reply', Analysis 40 (1980): 168-71. For an attempt to use the principle of plenitude to resolve the logical problem of evil, but without invoking the manyworlds hypothesis, see Robert Burch, 'The Defense from Plenitude against the Problem of Evil', International Journalfor Philosophy of Religion 12 (1981): 29-37. 45. Cf. Rodney Holder's list of problems facing the multiple universes hypothesis, in 'Fine-Tuning, Multiple Universes and Theism', Nofis 36 (2002): 310. 46. For similar sentiments, see Nicholas Everitt, The Non-Existence of God (London: Routledge, 2004), pp.241-42. Against the line of thought presented in the above paragraph, one may raise the following objection: Let's assume that the best world creatable by God is one in which he creates all (overall) good universes, and that there is an infinity of such universes. Even so, you need not think that all these universes can be grouped as a collection or set that has some cardinality. The infinite plurality of (good) worlds, in other words, may well have no cardinality, in which case it cannot be represented by, say, i or R 2" But this seems dubious, for it is difficult to see why any infinity of objects you choose cannot be grouped as a set, and if it can be grouped as a set then it must have some cardinality. 47. I am grateful to audiences at Monash University and the Research School of the Australian National University, as well as to William Rowe, Graham Oppy and an anonymous referee for this journal, for comments on earlier versions of this paper.