The Moral Argument from Evil (1999)
Theistic responses to the evidential argument from evil usually contend that God has morally justifying reason for allowing evil to occur. There are, of course, many cases where we cannot think what that morally justifying reason might be. But, it is argued, this at most reflects a limitation in our cognitive abilities. The fact that we mortals cannot think of any God-justifying reason gives us no reason whatsoever to suppose that there is no such reason. After all, perhaps God’s reason for permitting evil is simply beyond our ken.
Against this view, Bruce Russell has argued that "[if] we’ve conducted the relevant search for moral reasons to justify [God’s] allowing the relevant suffering … [then if we cannot find any such reasons,] we are justified in believing that there are no morally sufficient reasons for allowing that suffering" (Russell 1996:197). Call this thesis OV, since it may be thought to express a more "optimistic view" of human cognitive abilities.
Almost everyone will admit that there are some evils – the horrendous evils of the Holocaust, for example – that are totally inscrutable. We cannot think of any reason that would justify God in permitting them. If OV is correct, we are within our epistemic rights in believing that there really are no justifying reasons for allowing these evils. And since a perfectly good being would never willingly allow morally unjustified evil to occur, it would follow that we are also justified in believing that God does not exist. In this paper I will argue that Russell’s thesis is unsound, but that similar ideas to his can be used to construct a sound atheological argument. The moral argument from evil – so named because its crucial premise is a moral principle – claims that God’s existence is logically incompatible with certain actual instances of evil.
II. Critique of Russell’s Argument
Since Russell’s argument turns crucially on OV, the important question is whether OV is plausible. How, then, does Russell support it? He offers an indirect proof, arguing that denying OV leads to a conclusion the theist is unwilling to accept:
Why would it follow that we are not justified in believing that the human onlooker acted wrongly by failing to intervene? Because – and here is the crucial premise – "[t]he reason beyond our ken that would justify God in allowing [the evil act] could be the same reason which would justify the onlooker" (Russell 1996:198).
It will help if we put the argument a little more formally. We begin with an obvious consequence of the denial of OV:
What I have called the crucial premise is almost certainly an expression of a more general epistemological principle, which I take to be this:
It follows from R1 and R2 that:
R3 amounts to saying that we can never judge any person’s failure to stop or prevent a particular evil as morally wrong. But that conclusion is clearly false. Therefore, OV must true, since its denial leads to a false conclusion.
What can be said for this argument? Unfortunately, R2 is simply false. As William Alston points out, R2 fails to grasp that moral justification is partly a subjective matter:
In other words, from the fact that somebody is or could be justified in permitting a certain evil, it does not follow that everybody is justified in permitting that evil.
To illustrate this point, consider the following cases:
Jones’s case: Jones likes Doe, and wants the best for him. One day Jones notices that Doe is hooked up to a machine that is causing Doe substantial pain. (He also sees Smith standing by, laughing maliciously.) Jones knows three facts that Smith does not know. First, the machine that is causing Doe’s pain will soon cause Doe to experience intense pleasure that will last slightly longer than his present bout of pain. Second, the machine (which, incidentally, is the only one of its kind in the world) will render Doe immune to all earthly diseases – cancer, AIDS, influenza, and the like. Third, Doe freely consented to the entire process. "How nice," thinks Jones, "although Doe is now suffering, ultimately this process is for his own good." Like Smith, Jones can easily stop Doe’s suffering by pressing the OFF button on the machine. But he does not do so, because he is a nice person who likes Doe and wants the best for him.
What can we say about these cases with regard to the moral obligations of Jones and Smith? I think everyone would agree that Jones had no obligation to press the OFF button, and that he acted in a morally acceptable fashion by not pressing it. What about Smith? Did he too act in a morally acceptable fashion by not pressing the button? I think most people would agree that he did not. Smith’s failure to press the OFF button did by accident make the world a better place than it would otherwise have been. Nevertheless, surely we would say that given his particular circumstances – that is, given his knowledge and beliefs at the time in question, and, in particular, given that pressing the button was, so far as he knew, the only morally acceptable option – given all this, Smith did have an obligation to press the button. To say that he had an obligation not to press the button would be to require him to act on the basis of information he did not and does not possess, which would be manifestly unjust.
It seems, then, that whether someone is justified in permitting a certain evil depends on the level of information they have. Differing levels of information can give rise to differing moral obligations – as in this case, where, due to differing levels of information, Smith acted wrongly in permitting Doe’s suffering, whereas Jones acted correctly in permitting that suffering. In the given scenarios, we are not justified in concluding that one person (Jones) should have stopped or prevented a certain evil, but we are justified in concluding that another person (Smith) should have stopped or prevented that evil. It follows that R2 is false, hence that Russell’s proof of OV is unsound.
Although this objection to Russell’s argument is decisive, we might ask whether his ideas can be reconstructed in a more plausible fashion. In fact, I believe they can – not as a proof of OV, but as an atheological argument in its own right: the moral argument from evil (MAE). In developing MAE, I will first develop a moral principle specifying a sufficient condition for our being morally justified in allowing evil to occur. I will then argue that this principle, when applied to propositions accepted by theists, leads to the conclusion that God does not exist.
III. MAE’s Normative Premise
As Jones’s case showed, we are sometimes justified in allowing evil to occur. Although it would be an interesting task to find the conditions that are both individually necessary and jointly sufficient for our being so justified, for our purposes we need only specify a sufficient condition. To help us in developing this sufficient condition, consider a modified version of Jones’s case:
The crucial question is this: in Jones’s case #2, did Jones act wrongly by electing not to press the button? I suggest that he did not, and that, by not pressing the button, he acted in a manner equally acceptable to the way he acted in the previous scenario. For what difference could it make whether he knows the specific details of the outweighing good? Clearly none. What matters in terms of moral justification is surely the agent’s knowledge that an outweighing good will result, not the agent’s knowledge, or lack of knowledge, with regard to the particular form the good will take. So Jones’s non-intervention is equally acceptable in each case.
Doe’s suffering in the Jones cases is an example of what I will call "objectively justified evil". To say that some evil E is objectively justified is to say (i) that E is a member of a class of evils of equal magnitude the disjunction of which is either causally or logically necessary for the occurrence of an outweighing good, or the prevention of an outweighing evil, and (ii) either that the outweighing good has occurred (or will occur), or that the outweighing evil has been prevented (or will be prevented).
There seems to be some connection between, on one hand, an evil’s being objectively justified, and, on the other, our being justified in permitting that evil. Perhaps the connection is very direct, so that our moral intuitions are correctly expressed by the following principle:
This principle gives rise to two objections, one sound and one unsound. The unsound objection is that M1, based as it is on objective justification, seems inconsistent with what I previously wrote. After all, did I not previously argue that whether I am justified in permitting evil is partly a subjective rather than objective matter? Indeed I did. But M1 does not state that I am justified in allowing a particular evil purely by virtue of the fact that there is objective justification for that evil. Rather, what M1 says is that I am justified in allowing a particular evil by virtue of the fact that I know that there is objective justification for that evil. In other words, M1 says that my knowledge that the evil is objectively justified is sufficient to give me subjective justification for allowing that evil. Thus the objection is misguided.
However, M1 is open to a second objection. It may be pointed out that the objective justification for a particular evil might involve my instantiation of precisely those goods for which that evil is a necessary condition. And if that is so (one might continue), I am not justified in permitting that evil, since for all I know my failure to stop the evil may be preventing the occurrence of the outweighing goods.
Let us assume that this objection is sound. Where do we go from here? Well, we might note that the Jones cases are not cases where the onlooker’s failure to stop the evil may prevent the occurrence of the outweighing goods. The Jones cases are such that the evil in question will be objectively justified even if no onlooker stops or prevents (or attempts to stop or prevent) that evil. Let us reformulate M1 so that it refers only to this sort of evil.
It is this principle, I suggest, that operates in Jones-like cases.
Against the Jones principle, someone might argue that in cases where the objective justification for a given evil concerns goods or states of affairs that are beyond our ken, we are not justified in allowing that evil to occur. It would follow that in those cases where evil is permitted for reasons beyond our ken, only a being that can grasp those reasons (e.g., God) would be justified in permitting that evil.
I do not believe this objection succeeds. As I have argued, that we do not grasp the precise nature of the outweighing good is irrelevant to whether we are justified in permitting a given instance of evil. That we do not grasp the precise nature of the outweighing good only because we cannot do so is, I believe, equally irrelevant. To see this, we can easily consider a Jones’s case #3, in which the reason Jones does not know the nature of the relevant outweighing good is that (for some reason) he cannot do so: he is unable to comprehend the nature of prolonged pleasure, immunity from disease, or voluntary consent. Now this fact would surely not result in it being morally wrong for him to refrain from pressing the OFF button. As I have maintained, what is relevant is that Jones can grasp that there is an outweighing good. He does not have to know or comprehend, or even be able to know or comprehend, the precise form the good will take. Thus the objection fails.
Although the Jones principle is not self-evidently true, I take it to be a sound and well supported moral principle.
IV. MAE Formulated
We may now make an explicit formulation of the moral argument from evil. The first part of the argument is a conditional proof based on the following assumption:
By "the most rational theists", I have in mind those whose theistic belief was arrived at, and is sustained, no less rationally than the belief of any other extant theist.
MAE’s second premise requires some explanation, and is based on the following assumption:
It follows from A2a that:
Although MAE can be formulated without these assumptions, A2a and A2b are intuitively plausible. They are also fairly uncontroversial, since theistic writers on the evidential argument from evil are generally willing to grant that "if God exists, there is some [actual] outweighing good related…to every instance of suffering he allows" (Wykstra 1984:141).
The following argument, which takes the form of a conditional proof, uses A2b to prove premise A2 of MAE. Assume for the sake of argument that God exists. Now let E be any actual instance of evil for which there is an onlooker O, and let W be the actual world up to the point where O becomes aware of E. Given A2b, it follows that whatever possible successor-world to W becomes actual, there is in that world objective justification for E. Let PW1 be the set of all possible successor-worlds to W in which O intervenes; and let PW2 be the set of all possible successor-worlds to W in which O does not intervene. If O intervenes, then a member of PW1 becomes actual. If a member of PW1 becomes actual, then, since that member of PW1 is a world in which God exists, it follows that there is objective justification for E in that world. Thus, if O intervenes, then there is objective justification for E in the actual world.
Similarly, if O does not intervene, then a member of PW2 becomes actual. If a member of PW2 becomes actual, then, since that member of PW2 is a world in which God exists, it follows that there is objective justification for E in that world. Thus, if O does not intervene, then there is objective justification for E in the actual world. But either O intervenes, or he does not. So either way, E is objectively justified. Hence, there is in the actual world objective justification for E, justification that will occur even if O does not intervene to stop or prevent E. And since E is arbitrarily selected, it follows that:
The third premise of MAE concerns the knowledge of those whom I have called the "most rational" theists. Who is likely to be part of that group? The answer to this question depends in part on whether rationality is a matter of degree. Either it is legitimate to speak of certain people being "more rational" than others in holding certain beliefs, or it is not legitimate to speak in this way. If it is not legitimate to speak of degrees of rationality, then any given theist’s belief that God exists will either be "rational" (i.e., "completely rational", "not at all irrational"), or else "irrational" (i.e., "not at all rational"). Now, suppose a certain theist believes in God for reasons X, Y and Z. Whatever X, Y and Z happen to be – religious experiences, persuasion by reasoned argument, etc. – chances are that there are many other theists whose own belief in God is based on similar reasons to X, Y and Z. It is just a contingent truth about the world that theists tend to believe in God for reasons similar to those of other theists.
Clearly, X, Y and Z either provide, or do not provide, an adequate basis for belief in God. If they do provide an adequate basis for belief in God, then theists who believe in God for those reasons or for similar reasons will all most likely belong to the class of "rational" theists. On the other hand, if X, Y and Z do not provide an adequate basis for belief in God, then theists who believe in God for those reasons or for similar reasons will all most likely belong to the class of "irrational" theists. The result of this is that if any theists at all fit into a given category of rationality ("rational" or "irrational"), there will probably be many more theists who, due to having similar reasons for believing in God, fit into that category of rationality as well. Effectively, this leaves three live possibilities: either all theists are rational, or all theists are irrational, or, most likely, there are a substantial number of theists in each of the "rational" and "irrational" categories. Whichever of these possibilities is true, it follows obviously enough that a substantial number of theists belong to the class of most rational theists. And since A2 is (or seems to many to be) a reasonably obvious truth about the nature of God, it is highly likely that at least one of the substantial number of most rational theists will happen to know A2. This is especially so since the class of most rational theists will presumably include a number of theistic philosophers, who are the very sort of person likely to know A2. In other words, given the obviousness of A2, it would be very surprising if all the theists (and in particular, all the theistic philosophers) who belong to the class of most rational theists just happen to be theists who do not know A2. Thus, if rationality is not a matter of degree, then it is very likely that at least some of the most rational theists (as I have defined that term) know A2.
On the other hand, suppose rationality is a matter of degree. In that case, the most rational theists will presumably be those who have the most compelling (or at any rate "best") basis for believing in God, based on their access to, and evaluation of, philosophical and historical arguments for and against the existence of God, as well as their own religious experiences. What sort of characteristics would one expect to be possessed by the most rational theists? Well, just as those who are not in a good position to evaluate the evidence for and against the Big Bang theory can hardly be classed as some of the most rational believers in that theory, so those who are not in a good position to evaluate the evidence for and against God’s existence can hardly be classed as some of the most rational theists. Rather, the most rational theists will be among those who are best in a position to evaluate the evidence for and against God’s existence.
Evaluating the evidence for and against God’s existence consists largely in determining whether the predictions or consequences of God’s existence fit with reality. Necessarily, then, those who can best evaluate the evidence for and against God’s existence will be well informed about the consequences of God’s existence. Since the most rational theists are among those who can best evaluate the relevant evidence, and since those who can best evaluate the relevant evidence are well informed about the consequences of God’s existence, therefore the most rational theists will be well informed about the consequences of God’s existence. Since the most rational theists are well informed about the consequences of God’s existence, and since A2 is a reasonably clear consequence of God’s existence, it is highly likely that at least some of the most rational theists (particularly the theistic philosophers in that group) know A2. Again, it would be surprising if all the most rational theists just happen to be theists who do not know A2.
Regardless of whether rationality is or is not a matter of degree, then, it seems very likely that:
A3. Some members of the class of most rational theists (as I have defined that class) are theists who know A2.
At this point I want to assume – plausibly, I think – that if person A knows proposition p, and knows that p entails q, then A also knows q. If this asssumption is correct, it follows from A1, A2 and A3 that:
To this we add the moral principle developed in the previous section:
It follows from A4 and A5 that:
From A1 to A6, it follows by conditional proof that:
However, surely there exist instances of evil that theists, no matter how rational they are, and regardless of whether or not they know A2, have a moral obligation to stop or prevent. To say otherwise would be morally monstrous. Thus:
From A7 and A8, it follows by modus tollens that:
However, it is necessarily true that:
A10. If the most rational theists do not know that God exists, then no theist knows that God exists.
So from A9 and A10, it follows that:
A11. No theist knows that God exists.
Now, "All theists do know that God exists" is true if and only if, for any theist, (i) that theist believes that God exists, (ii) that theist is justified in believing that God exists, and (iii) it is true that God exists. So if no theist knows that God exists, that is, if all theists do not know that God exists, then, for any given theist (i), (ii) or (iii) must be false. It cannot be (i), since (i) is a necessary truth. So it must be either (ii) or (iii). That is:
I would argue, however, that if God exists, then at least some theists have a justified belief that God exists. After all, one of God’s foremost desiderata in world-making is the existence of rational creatures who seek a relationship with him. If belief in God is unjustified, then this objective would be difficult to achieve – in fact impossible to achieve, if the creatures are strictly rational. Thus, in order that he be able to achieve this desideratum, God, if he exists, would see to it that belief in him is justified for at least some theists. That is:
And of course:
A14. If God exists, then no theist has a false belief that God exists.
From A13 and A14, it follows that:
A15. If God exists, then some theists know (i.e., have a justified, true belief) that God exists.
However, it follows from A12 that:
A16. It is not the case that some theists know (i.e., have a justified and true belief) that God exists.
Finally, from A15 and A16, it follows by modus tollens that:
A17. God does not exist.
Although A13 seems to me correct, it might be challenged. As such, it is worth noting that there is another path from A12 to A17 that does not rely on A13. One can adopt a Rowe-style "friendly atheism" (Rowe 1979) and grant that at least some theists have a justified belief that God exists. Since (as A12 states) their belief that God exists is either false or unjustified, and since it is not unjustified, therefore it must be false. With or without A13, then, it follows that God does not exist.
V. MAE Without Middle Knowledge
In the preceding formulation, A2 assumes that God and objectively unjustified evil are not compossible. This assumption is correct only if it is also correct that God can guarantee the occurrence of outweighing goods. Why so? Well, assume that God exists and cannot always guarantee the occurrence of outweighing goods. In that case, there are possible worlds in which God exists and the outweighing goods do not occur. But in that case, God and objectively unjustified evil are compossible. So God and objectively unjustified evil are not compossible only if God can guarantee the occurrence of outweighing goods.
Presumably, however, some of those outweighing goods depend for their existence on free-willed creatures behaving in certain ways. God can guarantee that free-willed creatures will behave in certain ways only if he has middle knowledge of their actions. Thus A2’s assumption that God and objectively unjustified evil are not compossible is true only if God has middle knowledge. So A2 assumes that God has middle knowledge. Conversely, if God does not have middle knowledge, then he cannot guarantee the instantiation of all the relevant greater goods – in which case A2 is false.
The question therefore arises as to whether MAE poses a problem for those theists who deny that God has middle knowledge. On the above formulation it does not. But can it be restated in a way that is troubling for such theists? I believe it can.
To formulate the revised argument, let us first ask under what conditions God – without the luxury of middle knowledge – would allow evil to occur. A necessary condition for God’s allowing some evil E would surely be that E (or some evil of equal magnitude) is logically necessary for the occurrence of some outweighing good, or for the prevention of some outweighing evil. This is not, however, a sufficient condition. For suppose the scenario is such that there is objectively good reason to suppose that the greater good will not be attained, or that the greater evil will not be prevented. In that case God, if he allowed E to occur, would clearly be open to a charge of gross recklessness or negligence – which is not the sort of property one wants to attribute to a perfectly good being. This strongly suggests that God would not allow E to occur unless he has good reason to believe that the greater good in question will be instantiated, or that the greater evil in question will be prevented. Thus:
Let us now develop a possible scenario to which E1 and E2 can be applied. Let Scrooge be a theist whose theistic belief was arrived at, and is sustained, no less rationally than the belief of any other extant theist. It follows from this hypothesis that:
Now suppose that Scrooge one day encounters an instance of substantial suffering that he can stop or prevent with minimal effort and no risk to himself. Suppose further, however, that Scrooge has a history of failing to intervene to stop or prevent the sort of evil he is now witnessing. We know that the occurrence of G (or, as the case may be, the prevention of E’) either involves, or does not involve, Scrooge’s intervention. Given Scrooge’s history of non-intervention, God does not have good reason to suppose that Scrooge will intervene. But God does have good reason to believe that G will be attained (or E’ prevented). Therefore, God must have good reason to believe that G will be attained (or E’ prevented) without Scrooge’s intervention.
I now wish to add these assumptions to our account:
S3. If person A knows proposition p, and knows that p entails q, then A also knows q.
S4. Scrooge knows S5, where:
Now assume for conditional proof that Scrooge knows that God exists. From this, together with S2 and S3, it follows that:
By definition of knowledge as justified, true belief, S4 and S6 entail the true of S5 and E2 respectively. It follows from S5 and E2 that:
On the present assumptions, Scrooge knows S5 and E2. Assume he also knows that S5 and E2 together entail S7. From this and S3, it follows that:
Let us further assume that Scrooge is not irrational in this instance, hence that he comes to believe X if he knows that there is objectively good reason to believe X. From this and S8, it follows that:
This is a case, then, where the onlooker rationally believes that the evil in question will be outweighed by the occurrence of some greater good (or the prevention of some greater evil) even if he does not intervene. Is this sufficient to justify Scrooge in allowing the evil in question? Indeed it is. To see this, consider:
The fact that there is a small chance that the machine will not work does not, I take it, thereby give Jones an obligation to press the OFF button. And of course, that the machine will not in fact work cannot enter into our moral valuations, since this is something Jones neither knows nor believes. Once again, then, it seems that Jones acted in a morally acceptable fashion by not pressing the button. Cases such as this lend support to:
On the present assumptions, you will recall, Scrooge rationally believes that there is objective justification for the evil he observes, justification that will occur even if he does not intervene to stop or prevent that evil. From this and the Reduced Jones principle, it follows that:
Discharging the assumption that Scrooge knows that God exists, we get:
From S11 and S1, it follows that:
However, surely S10 is false:
From S12 and S13, it follow by modus tollens that:
And by the reasoning in the previous formulation of MAE, it follows from S14 that:
Of course, this conclusion only applies to those possible worlds in which a Scrooge-like scenario has occurred. If theists are not willing to grant that a Scrooge-like scenario has occurred in the actual world, I would be more than willing to arrange for such a scenario to take place. Since I do not anticipate that this scenario (or one relevantly similar to it) would be terribly difficult to arrange, we can safely treat the Scrooge scenario as having actually occurred. Thus the conclusion that God does not exist applies (or can be safely taken to apply) to the actual world.
VI. Comments and Conclusions
In this paper I have attempted to show that even if theists can successfully respond to the evidential argument from evil, there is a further difficulty to be faced in the moral argument from evil. If evil is merely the harbinger of greater good, why should we be opposed to its occurrence, and why, indeed, should we be expected to prevent it? These are questions that theists have not, to my knowledge, addressed in any detail. And yet they must be addressed, since to me they seem quite damaging.
It is these questions that motivate the moral argument from evil. The gist of MAE is that if God uses evil as a means to greater good in the way theists claim, then it would follow that some people have no obligation to stop or prevent evil. In other words, God’s existence is logically incompatible with the existence of evils that certain onlookers have a moral obligation to prevent. Since there clearly do exist evils that these onlookers have an obligation to prevent evil, it follows that God does not exist.
One way to avoid this conclusion is to deny that human onlookers have a positive duty to stop or prevent evil. While this would be a viable option for some forms of supernatural belief (some forms of deism, perhaps), it is certainly not an option for traditional theists. Christians, in particular, would agree with MAE’s contention that all of us do, on occasion, have an obligation to stop or prevent evil. Traditional theists, I think, would be more likely to avoid MAE’s conclusion by denying the Jones principles, which I take to be the most controversial aspect of MAE. However, since those principles are well supported (according, as they do, with our intuitions regarding Jones-like cases), one could not simply dismiss them out of hand. Rather, one would need to refute them by means of a counterexample. It remains to be seen whether this can be done. And until it is done, my tentative conclusion is that the moral argument from evil shows that the god of traditional theism does not exist.
 To say that God has morally sufficient reason for allowing some instance of evil is to say that God is responsible but not blameworthy for allowing that evil. Generally, this will occur only if allowing the evil in question (or one of equal magnitude) is logically necessary either for preventing some greater evil, or for attaining some greater good.
 To explain the notion of an outweighing good: where E is an instance of evil, and G is a good for which E is a causally or logically necessary condition, G outweighs E just in case the state of affairs involving the conjunction of G and E is better, all things considered, than the state of affairs that would have obtained if G had never occurred. There is, of course, a corresponding notion of an outweighing evil.
 I refer to a "human" person here, though it would be just as appropriate to refer to a "finite" person – that is, one who is not omnipotent, omniscient, etc. M1 includes a requirement of finitude because God would presumably not be justified in allowing a certain evil that is causally but not logically necessary for the attainment of some greater good. At any rate, I doubt theists would object to my speciesist phraseology.
 This sort of objection is pressed by Daniel Howard-Snyder (1996: 292-3) against Bruce Russell’s proof of OV. The objection is not applicable to my argument, however, since Howard-Snyder is not considering cases where the onlooker knows that there is objective justification for the evil in question (justification that will occur even if the onlooker does not intervene), whereas I am considering such cases.
 It should be noted that the present argument does not require that all members of PW1 and PW2 alike contain objective justification for the evil in question. Rather, all the present argument requires is that the evil is objectively justified in whatever world becomes actual.
 One objection to this line of argument might be that if rationality is a matter of degree, then for all we know there could exist one theist who is uniquely the most rational theist, and for all we know this theist might not know A2. In that case (one might continue), A3 is false and MAE fails. First, it is not clear that this objection brings up a live possibility, even from a theistic perspective. After all, theists generally – and in particular those who take belief in God to be properly basic – presumably hold that many theists are completely or perfectly rational in believing in God. Second, even if the objection did bring up a live possibility, this live possibility could easily be eliminated by having atheists embark on a large-scale information-dissemination campaign aimed at educating theists about the truth of A2. If such a campaign took place, A3 would effectively be established. Now I seriously doubt that theists would be willing to dismiss MAE on the possibility that such a campaign might not succeed in educating the most rational theist about the truth of A2. For theists to do that without finding any further flaw(s) in MAE would be to leave God’s existence dangling by a thread, subject to disproof the very instant the most rational theist comes to know A2. To neutralise MAE, then, theists will have to do more than cast doubt on the present truth of A3.
 It may be objected that it is unreasonable to have a premise such as A4 in an atheological argument. After all, atheists are presumably committed to the contention that there is not objective justification for every actual instance of evil. However, A4 is under the scope of an assumption in a conditional proof. Thus, what is being asserted is not that A4 is true, but that it follows from certain other assumptions.
Alston, William P. 1996. "Some (Temporarily) Final Thoughts on Evidential Arguments from Evil". In Daniel Howard-Snyder, ed. 1996. The Evidential Argument from Evil. USA: Indiana University Press.
Howard-Snyder, Daniel. 1996. "The Argument from Inscrutable Evil". In Daniel Howard-Snyder, ed. 1996. The Evidential Argument from Evil. USA: Indiana University Press.
Rowe, Willliam L. 1979. "The Problem of Evil and Some Varieties of Atheism". Reprinted in Daniel Howard-Snyder, ed. 1996. The Evidential Argument from Evil. USA: Indiana University Press.
Russell, Bruce. 1996. "Defenseless". In Daniel Howard-Snyder, ed. 1996. The Evidential Argument from Evil. USA: Indiana University Press.
Wykstra, Stephen J. 1984. "The Humean Obstacle to Evidential Arguments from Suffering: On Avoiding the Evils of ‘Appearance’". Reprinted in Marilyn McCord Adams and Robert Merrihew Adams, eds. 1990. The Problem of Evil. New York: Oxford University Press.