It is often assumed by all sides of the debate on the problem of evil that an omnipotent, omniscient, wholly good God could not permit any gratuitous evil, evil that is (roughly speaking) avoidable, pointless, or unnecessary with respect to the fulfillment of God’s purposes. In discussions of evidential arguments from evil, for example, any premise to the effect that God must have a morally sufficient reason for every evil he permits is usually conceded to be indisputable, self-evident, necessarily true, or something of that ilk. [1] The intuition here, as the Howard-Snyders explain, is that ‘on the face of it, the idea that God may well permit gratuitous evil is absurd. After all, if God can get what He wants without permitting some particular horror (or anything comparably bad), why on earth would He permit it?’ [2] The focus, then, tends to be placed on the atheologian’s claim that there does in fact exist pointless evil. Madden and Hare, for example, state that ‘the really interesting problem of evil is whether the apparent gratuity [of much evil] can be explained away by more ingenious measures or whether the gratuity is real and hence detrimental to religious belief.’ [3] That pointless evil is incompatible with theism is simply taken for granted.
An increasing number of theists, however, are beginning to question this assumption of incompatibility. This way of responding to the evidential problem of evil has been described by one prominent atheist as ‘radical, if not revolutionary,’ [4] but it is viewed by many theists as the only way to remain faithful to the common human experience of evil, according to which much of the evil we observe in the world strikes us as utterly pointless, absurd, or excessive. An ingenious attempt to square gratuitous evil with theistic belief has recently been made by Peter van Inwagen, and in what follows I will outline and then evaluate his proposal.
In his paper entitled ‘The Argument from Particular Horrendous Evils,’ van Inwagen considers an evidential argument from evil modelled on Rowe’s well-known argument but based on an instance of evil dubbed ‘the Mutilation.’ [5] Van Inwagen’s account of the Mutilation runs as follows:
Van Inwagen then devises the following Rowe-like argument:
Unlike most theists who have responded to such arguments, van Inwagen does not find fault with the first premise, but rather considers it to be ‘fairly plausible.’ [8] Instead, he attempts to debunk premise (3), which amounts to the claim that For any particular horrendous evil E permitted by God, there is a morally sufficient reason for God’s permission of E. This premise, states van Inwagen, appears to be predicated on a general moral principle of the following sort:
But is (MP) true? Van Inwagen thinks not, and to substantiate his position he draws attention to some scenarios of what would (or could) result if (MP) were put into practice by some moral agent. Each scenario functions in effect as a reductio ad absurdum of (MP). The first of these scenarios is described as follows:
Van Inwagen concludes that ‘this much, I think, is enough to show that the principle is wrong, for you have no such obligation.’ [11] But there is worse in store for (MP). By means of the reasoning employed in the above case, van Inwagen reveals that (MP) has further disastrous results. To paraphrase van Inwagen:
The above cases, in van Inwagen’s view, indicate that the problem with (MP) is that it ‘forbids the drawing of morally arbitrary lines.’ [13] (MP) does not countenance such things as a legislature’s right to set an arbitrary time span (e.g., ten years) as the minimum punishment for felonious assault. Although there is nothing intrinsically wrong with setting such arbitrary limits – indeed, this is a necessity in many facets of the real world – (MP) cannot make any provision for this practice and for this reason it has highly counterintuitive consequences.
The implausibility of (MP) is further highlighted by considering it in the light of God’s decision as to which evils to prevent and which to permit. Let’s assume that God’s plan for humanity includes the permission of a vast amount of horrendous evil for the sake of some future good that outweighs it—this future good may be called the ‘Summum Bonum Creatum’, though its nature or content need not detain us here. Consider, now, the case of the Mutilation. Should God have added this particular horror to his list of horrors to be prevented?
Suppose that God is guided in his deliberations about what to do by the principle expressed in (MP). In that case, God would prevent the Mutilation from taking place. For if the Mutilation were prevented, the world would not have been a significantly less horrible place, and so God’s plan for the realization of the Summum Bonum Creatum would not have been jeopardized. [14] But this point can be generalized as follows: For any n, if the existence of n horrors is consistent with God’s plan for the achievement of the Summum Bonum Creatum, the existence of n-1 horrors will be equally consistent with his plan. [15] Clearly, if this reasoning were applied to every instance of evil, God will end up creating a world totally devoid of evil. But even though God may prevent many horrors, he cannot (we have assumed) prevent all of them without thwarting his plan for achieving the Summum Bonum Creatum. (MP), therefore, would result in the frustration of God’s purposes.
To avoid this outcome, God must draw the line somewhere between horrors that are to be prevented and horrors that are to be permitted. But wherever he draws the line, it will be an arbitrary line. For assume that he draws the line so as to exclude the Mutilation from the set of actual horrors. In that case, no greater good would have been lost and no evil equally bad or worse would have resulted—thus, God would have no reason for drawing the line in such a way as to exclude the Mutilation. The line would have to be drawn arbitrarily. And it just so happens (we are assuming) that the Mutilation has fallen on the ‘actual horrors of history’ side of the line.
In short, if (MP) is rejected, the number of evils God must permit in order to fulfill his purposes becomes a purely arbitrary matter. What this means is that there is no minimum number of horrors that God must permit in order to realize the Summum Bonum Creatum. Let’s call this van Inwagen’s No Minimum Thesis:
As van Inwagen puts it, ‘to ask what the minimum number of horrors consistent with [God’s] plan is, is like asking, What is the minimum number of raindrops that could have fallen on England in the nineteenth century that is consistent with England’s having been a fertile country in the nineteenth century?’ [16] But if there is no minimum number of horrors, it is logically impossible for God to produce it and hence unreasonable for anyone to expect that he do so. [17]
To summarize, premise (3) is predicated on (MP), but (MP) has unwelcome consequences, as has been brought out by the examples involving prison sentencing and, especially, God’s permission of horrendous evil. (MP) therefore ought to be rejected and replaced by the No Minimum Thesis, which better reflects the realities and vagaries of our moral life. The No Minimum Thesis, however, entails the falsity of (MP) and, by extension, of premise (3). It seems, then, that proponents of Rowe-like arguments must find some reasonable grounds for rejecting the No Minimum Thesis. Daniel and Frances Howard-Snyder sum up the matter as follows:
I wish to argue, however, that the position advanced here by the Howard-Snyders is mistaken, for there need not be any incompatibility between the No Minimum Thesis and what the Howard-Snyders call ‘Rowe’s incompatibility claim.’
One way of understanding van Inwagen’s critique is as follows. If the principle expressed by (MP) were to be applied in a given situation, one would end up with a sorites-style chain of reasoning and hence an unpalatable conclusion (e.g., no criminal ought to spend even one day in prison). The problem is that sorites-style reasoning is flawed, even though it is infamously difficult to identify where precisely the flaw resides. As a result (MP) is also flawed, since any application of (MP) commits one to defective sorites-style arguments. The question, therefore, is whether there is any way of restating the evidential argument from evil that does not commit one to (MP) and thus does not oblige one to renounce the No Minimum Thesis.
To begin with, assume that van Inwagen’s No Minimum Thesis is correct, and so there is no minimum amount of horrific evil God had to permit in order to achieve his purposes. As some commentators have pointed out, it is entirely consistent with this view that God could have achieved those purposes with a lot less horrific evil than there is. [19] And if this is the case, then God should have set out to achieve his purposes by way of a lot less horrific evil. The point here is that even if there is no minimum amount of evil God had to permit, this does not entail that he would have permitted just any amount of evil. If, for example, we were to assume that there is no minimum prison sentence that would suffice as a sufficient deterrent for car theft, this would not justify a judge sentencing a convicted car thief to life in prison.
The claim, then, is that even though there may be no minimum number of evils permissible by God, there must at least be a threshold beyond which any further instantiations of evil are gratuitous. To illustrate this, consider first the following diagram:
![]() |
In the above diagram, evil is quantified in terms of ‘turps’ (to borrow Plantinga’s terminology). Worlds in which God permits anything less than, or anything more than, 385 turps of evil will not allow for the achievement of his purposes. Therefore, only worlds in which God permits 385 turps of evil are ones in which his purposes can be fulfilled. In this case, 385 turps is the minimum number of evils (as well as the maximum number of evils) that God can permit. [20] Now, compare this scenario with the following:
![]() |
In this diagram, in contrast to the previous one, there is no exact amount of evil God must permit in order to fulfill his ends. There is only a threshold, and any amount of evil falling within this threshold would be consistent with God’s plan. But since the boundaries of the threshold are ‘fuzzy’, there is no precisely specified minimum amount of evil that God must permit in order to achieve his purposes. This threshold, however, does not rule out the possibility that there are amounts of evil that are clearly excessive. For example, worlds where God permits 1,000 turps of evil would clearly thwart the realization of his plan. And so, even if there is no minimum number of evils that God must allow, this does not leave God free to allow just any amount of evil into the world. [21] In response, it may be held that it is impossible to determine whether our world is one in which the threshold, as represented in Figure 2, has been crossed. Of course, the amount of evil found in our world certainly appears excessive, but how can we tell whether appearances are in this case deceptive? For all we know, the amount of evil in our world falls somewhere within the threshold of ‘acceptable amounts of evil.’ Van Inwagen, therefore, can accept the claim that, even though there is no exact or ideal amount of evil that God must permit, not just any amount of evil would be permitted by God. But van Inwagen may question the further claim that our world contains excess evil, for it seems impossible to settle the truth-value of this claim. (He need not, of course, doubt our ability to tell, with respect to at least some possible worlds, whether they contain too much evil.)
Is this skepticism regarding our ability to determine whether there is excess evil in our world warranted? Consider Theodore Drange’s comment that, ‘I can know that I have too much or too little of something even if I do not know exactly what the perfect amount would be. For example, I can know that my food is too salty even if I do not know exactly how much salt would be ideal.’ [22] Or, to borrow another of Drange’s examples, I can tell whether my neighbour’s music is too loud, even though I do not (or cannot) know what the ideal volume would be. ‘In a similar way,’ Drange contends, ‘we can know that our present world contains too much suffering even if we cannot describe in detail just what the ideal amount of suffering should be.’ [23] But is Drange correct to draw this inference? If he only means to say that we can tell, with respect to at least some possible worlds, whether they contain too much evil, then it is difficult to dispute Drange’s contention. But Drange wants to say more than this, for he is claiming that we can just see that our world contains too much suffering. But how could anyone know this?
I think the most promising option here is to rely on a strategy used to great effect by Rowe, viz., the appeal to ‘noseeum inferences.’ That is to say, one can claim to know that our world contains excess – and hence gratuitous – evil on the basis of the failure of theodicy: despite thorough investigations on the part of many people over many years, no plausible justification for God’s permission of numerous evils has been uncovered. Since no goods we know of can justify God’s permission of these evils, we are entitled to infer that no goods whatever are served by the evils in question (in Wykstra’s terminology: we no see ’um [i.e., the goods in question], so they ain’t there!). The success of such a strategy will obviously depend on the degree of warrant that attaches to noseeum inferences. But if (as I believe) little can be said against, and indeed something can be adduced in support of, noseeum inferences from inscrutable to pointless evil, then there is a way to overcome skepticism regarding our ability to tell whether our world contains too much evil. [24] The upshot is that premise (3) withstands the objections levelled against it by van Inwagen via his No Minimum Thesis. But how is (MP) to be modified in order that it be immune from sorites-style paradoxes like those involving the sentencing of Blodgett? I propose the following reconstruction of (MP):
The threshold, however, is not precisely defined. Therefore, in some cases there may be no truth of the matter as to whether a particular amount of evil falls within the relevant threshold. Or, to venture into fuzzy logic, that a given amount of evil falls within the threshold may only be ‘partly true’. Such departures from classical logic are highly controversial, and there is no need to insist here that one must adopt a non-standard logic to dissolve sorites-style paradoxes. For even though the boundaries of the threshold are fuzzy, the important point is that God would only permit that amount of evil that clearly falls within the threshold (or that clearly falls within the bottom range of the threshold). This, of course, is not to deny that God’s choice of a specific quantity of evil, from amongst the various quantities of evil that fall squarely within the threshold, is entirely arbitrary. And so with this in mind, the earlier Rowean evidential argument can be recast as follows:
The move from (P) to (Q) represents Rowe’s ‘noseeum inference’, while (3*) is an amended version of premise (3). As long as (3*) is predicated on something like (MP*), the counterintuitive scenarios that plague (MP) will be avoided, since (MP*) allows for both the vagaries of moral life (e.g., there being no precise prison sentence sufficient to deter car theft) and for arbitrarily made decisions in response to these vagaries (e.g., setting twelve months exactly as the minimum prison sentence for car theft). But given that (MP*) does not entail the falsity of the No Minimum Thesis, there is no need – despite the suggestion made by the Howard-Snyders as quoted at the end of Section 2 above – to find some principled grounds for denying this Thesis before accepting Rowe’s incompatibility claim. In short, a commitment to the No Minimum Thesis does not exclude a commitment to the position that gratuitous evil is not compatible with the theistic world-view.
In sum, van Inwagen has failed in his attempt to reconcile theism with gratuitous evil. To be sure, a number of other attempts at reconciliation have been made, particularly by William Hasker and Michael Peterson, and these also merit attention. [27] The increasing number of criticisms voiced against the assumption that If there is a God there can’t be any pointless evil do at least indicate that this assumption can no longer be taken as self-evidently true and hence in no need of justification. Nevertheless, the foregoing discussion also indicates that the case against this assumption, as illustrated by the arguments developed by van Inwagen, is in need of further work if it is to have greater appeal. [28]