Introduction
At a basic level, the argument from evil is that unjustified evil exists, so everything that exists is unwilling or unable to ensure that there is no unjustified evil. An omnipotent, omniscient, omnibenevolent being would be willing and able to prevent all unjustified evil, so no such being exists. It is a nice, neat argument. However, it depends on the claim that unjustified evil exists, not just that evil exists. Sometimes evil is justified, such as the suffering that a child goes through when getting life-saving surgery.Footnote 1 The pain is justified because it is the result of a procedure that saves the child's life.
Now, the fact that evil can be justified suggests an intriguing possibility: perhaps all of the apparent instances of unjustified evil in the world actually have justifications that are simply hidden from us. If this possibility is actual, then the argument from evil fails. If all suffering is justified, whether we can tell or not, then there very well could be an omnipotent, omniscient, omnibenevolent being. Sceptical theism is the view that our ability to distinguish between genuinely unjustified evil and justified evil is weak enough that we have to take seriously the possibility that all evil is justified. This is a neat response to the argument from evil.
However, this neat response has some unsavoury consequences. The very premise that allows sceptical theism to work against the problem of evil threatens to destroy moral decision-making. Sceptical theism claims that our moral judgements are prone to the illusion of seeing unjustified evil when in reality there is justified evil. However, we need to be able to identify unjustified evil in order to decide whether it is right to intervene to prevent evil. I will argue that the defender of sceptical theism faces a dilemma: either our judgements regarding unjustified evil are reliable enough that the problem of evil remains a problem, or our judgements regarding unjustified evil are so unreliable that we have no ground for using them in deciding what to do. The first horn undermines theism, while the second undermines our moral decision-making, leaving us to make decisions based on our more reliable judgements about self-interest.
Below I first explain how sceptical theism undermines the evidential argument from evil by claiming that our moral judgements are subject to illusion. Next, problems for sceptical theism are proposed, including the dilemma of accepting limited egoism or accepting the argument from evil. Finally some responses on behalf of sceptical theism are considered and found wanting, showing that sceptical theism is at best a problematic response to the problem of evil.
Basic sceptical theism
Traditional theism is the view that there exists an omnipotent, omniscient, omnibenevolent being (henceforth referred to as ‘God’). The evidential argument from evil is one of the most important arguments against traditional theism. The standard version of the evidential argument from evil, along the lines of Rowe (Reference Rowe1979), goes as follows:
1. There are instances of unjustified evil in the world.
2. God would not allow unjustified evil in the world.
3. Therefore, God does not exist.
‘Evil’ is used broadly to mean any bad event or state of affairs. Some examples include earthquakes, hurricanes, tsunamis, plagues, and genocides. Justified evil is evil that can only be prevented by sacrificing a greater or equal good or by permitting a greater or equal evil.Footnote 2 I do not presuppose that an instance of evil must be justified by local consequences of the evil. I also leave open what the greater goods secured or greater evils prevented might be. Autonomy might outweigh pain, for example, so a tremendous amount of pain might be justified if it is necessary to allow autonomy.
Premise 1 rests on the fact that there is an abundance of apparently unjustified evil in the world. To take but one example, it seems obvious that the existence of smallpox only caused suffering and death without providing any overriding benefit. Furthermore, smallpox has been eradicated, so the world can go on without it. This is but one example of something that seems to be completely unjustified. Premise 1 is justified by the fact that there are so many examples that it is unlikely that we are mistaken about every single one of them.
Defending premise 2, the important claim is that God would not permit an instance of evil without a justification. God's perfect goodness implies that He would only permit evil when faced with a dilemma such as, ‘Permit this earthquake, or permit something worse’. Assuming God exists, every evil was permitted by God, and thus there must be a good reason for each and every instance of evil. This is a good reason to accept premise 2.Footnote 3 Thus we have the evidential argument from evil. Traditional theism predicts a global lack of unjustified evil; there is unjustified evil; so traditional theism is false.
The inference used in the argument is unproblematic, so defenders of traditional theism must find a problem with one of the premises. The sceptical theist response defends theism by denying that we have grounds to accept premise 1. The reasoning to defend premise 1 depends on a noseeum inference, an inference from ‘I can find no justification for this evil’ to ‘There probably is no justification for this evil’.Footnote 4 Sceptical theists claim that this inference is flawed, and thus the evidential argument from evil cannot be compelling. Notice that the response does not undermine the truth of premise 1 directly by showing it is false (as a theodicy attempts to do), but undermines any confidence we might have in premise 1. Sceptical theists achieve this by defending the central claim of sceptical theism, ST for short.
ST: Our understanding of value is limited to such a degree that we cannot be confident that any apparently unjustified evil is in fact unjustified.
In other words, events that look like pointless evil to us might, as far as we can tell, turn out to be necessary and justified. We have such a limited perspective that we should never confidently conclude that an evil is completely unjustified. By accepting ST, one can undermine premise 1 above: given our limited capacity for distinguishing between justified and unjustified evil, we cannot be confident that there is any unjustified evil. Thus, we cannot be confident that premise 1 is true, and we cannot use the argument from evil as an argument against the existence of God.
The strategy draws upon the notion of reliable discrimination. For example, if Mike can't reliably discriminate between a grimace and a smile, we would be ill-advised to take Mike's word for it that a suspect was grimacing and not smiling. ST suggests that we are like Mike when it comes to distinguishing between justified and unjustified evil. We just are not good at it. Several reasons have been offered to support this claim. For example, questions about what the world would have been like given broad changes are difficult for us to answer. Thus, we cannot reliably tell what would have happened if some suffering (the existence of cancer, for example) had been prevented, so we cannot reliably say that the world would have been better off without that suffering. Another claim is that we may be unaware of the total range of values that could justify evil. Therefore, the fact that no known value justifies an instance of evil does not imply that nothing justifies it.Footnote 5 Together these considerations purport to show that we should not trust our own ‘testimony’ about the existence of unjustified evil. We cannot rely on the noseeum inference.
In summary, ST is the view that our capacity to recognize evil outstrips our capacity to recognize justifications, and thus our faculty for making value judgements is subject to the illusion of seeing unjustified evil where there is actually justified evil. This calls into question our grounds for accepting premise 1, the premise that unjustified evil exists. If we are that bad at identifying unjustified evil, how can we be confident that any actually exists? The sceptical theist argues that we cannot be at all confident, and thus we cannot accept the premise or the argument.
An initial worry about ST
In order for ST to rebut the problem of evil, we must assume an interpretation of ST that implies our judgements about unjustified evil are so unreliable that every instance of apparent unjustified evil might be justified. However, accepting the truth of that interpretation of ST undermines one's capacity for moral decision-making. In particular, a sceptical theist could have no basis for intervening to prevent evil.Footnote 6 Once sceptical theists accept that any evil they try to prevent could, as far as they know, be a justified evil, then they have no basis for concluding that they should intervene to prevent evil.
This line of reasoning starts with the fact that intervention to prevent evil is called for only if the evil is unjustified, or would be unjustified if allowed to happen. As a simple example, it would be horribly misguided to intervene to prevent surgery on the grounds that surgery is an evil. Even though surgery performed with no justification is an evil – it usually carries significant risks, causes damage to tissue, and causes a great deal of pain (and in America an extraordinary loss of wealth) – many surgeries are justified because of the nature of the diseases they treat. It would be much worse to stop the surgery than to allow it.Footnote 7
ST implies that we should distrust appearances regarding lack of justification for evil, because our ability to identify evil outstrips our ability to identify justifications for evil. We should expect to experience illusions of unjustified evil even if there is none.Footnote 8 The problem for ST regarding intervention derives from the practical consequences of distrusting appearances in general. In general when we learn that we are subject to illusions, the rational response is to discount appearances in the illusion-prone area. For example, there is a fairly well-known illusion among pilots called the somatogravic illusion or vertigo: when there is a lack of visual cues, it can appear that the plane is nose-up or nose-down when in fact it is level.Footnote 9 If the pilot tries to keep the plane level based on appearances, she will crash (which unfortunately has happened even to experienced pilots). Because pilots are taught about the illusion, they know that when their instruments disagree with appearances, they should discount the appearances and follow the instruments. The fact that one appears to be in a dive after take-off at night is not a good reason to think one is in a dive, because one is subject to illusions in this area.
Now return to ST. If our judgements about justification for evil are unreliable to the point of being illusions, then those judgements are like the perceptions of the pilot at night. The judgement that an evil had no justification would be an illusion providing little to no information about whether there was a justification for that evil. The judgement should be discounted. As a result, if one witnesses an evil occurring and wants to do the right thing, one's judgement that the evil is unjustified gives no help in deciding whether to intervene. One simply cannot tell whether the appearance of unjustified evil is an illusion or not.Footnote 10
An initial response
One might respond that when trying to decide whether to intervene, a person only has appearances to go on, so the best she can do is base her decisions on appearances. Since she is used to making decisions based on moral appearances, and since it is difficult to do otherwise, in the absence of an alternative she may as well go with appearances. If she is an oncologist dealing with a cancer patient, she may as well go with the appearance that cancer is unjustified and demands intervention when possible. It is as good a basis for decision as any other, and it has the added benefit of being familiar and minimally alienating. Ignoring one's own judgements is alienating, even in cases where they are known to be illusory. Thus, appearances of unjustified evil could still serve as a rational basis for action simply because we have to do something and there is nothing else to go on. One could go one step further and say that, while objectively these judgements do not guide one to what really ought to happen, subjectively they are still authoritative. The idea is that in a limited sense one ought to follow one's moral judgements; there is something wrong with feeling like something is wrong and doing it anyway. If this is right, one can accept ST yet still have a basis for moral decision-making.
Bergmann and Rea (Reference Bergmann and Rea2005, sec. III) make an even stronger claim. According to them, the appearance that an evil is unjustified is a pro tanto reason to intervene to prevent it. As they see it, ST only implies that we cannot assess the probability that unseen reasons override our pro tanto reason to intervene. The fact that we cannot tell whether invisible reasons are at work is not itself a reason to allow evil to occur. Since it is not a reason to allow the evil, it cannot override our pro tanto reason to intervene. Thus, sceptical theists can still find reasons to intervene to prevent evil.
Objection to ST, revised and expanded edition
At this point we need to deal with the interpretation of ST: how strong a claim does ST make? Possible interpretations lie on a spectrum. On one end of the spectrum is the interpretation according to which our judgements about unjustified evil are complete illusions giving no information at all about reality. On the other end is the interpretation according to which judgements about unjustified evil are a guide to reality, but a guide that can be overridden. We must consider weak and strong versions of ST. Each one leads to its own problems, creating a dilemma for sceptical theists.Footnote 11
Strong ST
Assume the strong interpretation, according to which our judgements that evils are unjustified are complete illusions. On this interpretation, Bergmann and Rea's claim that appearances give pro tanto reasons is untenable. Complete illusions do not provide any kind of reason. To illustrate, consider a pilot who knows that her sensations regarding pitch of the plane are complete illusions. She knows that appearances offer no basis for deciding how to fly the plane. The fact that it feels like the plane is nose up is not even pro tanto reason to think that it is.Footnote 12 Likewise, with the strong interpretation of ST, appearances of unjustified evil do not even provide pro tanto reason to intervene, since they are entirely illusory.
That leaves the response that appearances are all one has to go on, so maintaining familiarity and avoiding alienation make it reasonable to decide based on appearances. However, this response also fails for the pilot and for ST. Look at the case of the pilot again. The appearance of being nose up is not all she has to go on. Planes are equipped with instruments to measure the pitch of the craft. Even when the pilot is subject to an illusion of tipping, the instruments remain a basis for making decisions. Thus, the best strategy is to base decisions on information from the instruments, doing one's best to ignore the appearance of tilting. If there is no reason to think that the instruments are flawed, this is reasonable. Rather than being in a position where she must believe the appearances because there is no other option, she is in a position where she can ignore appearances and rely entirely on instruments.
Now return to Strong ST. Like the pilot, we are not entirely reliant on the illusion of unjustified evil to decide whether to intervene, since judgements about moral justification are only one aspect of our decision-making. When deciding what to do, in addition to moral considerations, we bring to bear considerations of personal preference, self-interest, and personal commitments. If we cannot trust ourselves to get certain moral value judgements right, we may at least be able to trust ourselves to know our own preferences and commitments. Certainly, we can be wrong about what we will prefer (I thought I would enjoy the film The Girl with the Dragon Tattoo, but was quite wrong), but as long as there is some degree of reliability in our sense of preference, then it is a much more solid basis for decision-making than judgements about justification, given Strong ST.
This is true as long as self-interest and preference are sources of reasons at all. Even if one adopts the view that moral value is infinitely more important than self-interest, so that the smallest difference in moral value swamps any difference in self-interest, self-interest can still play a role in reasonable decision-making when moral considerations do not settle matters. Thus, we have a relatively firm basis for deciding whether to intervene, without relying on moral illusions. If my wife is drowning and I want to know whether to intervene and save her, whether or not I trust my judgement that her suffering is not justified, I am going to intervene.
Whenever there is a conflict between an apparent obligation to intervene and what we prefer, we should do what we prefer, since there is little hope of getting morals right in such cases but we can do pretty well with preferences. This is not to say that one should ignore morality in these cases; it is to say that the conscientious agent who wants to be moral cannot use her own judgements about unjustified evil to figure out what is moral. It would be misguided to put appearances of unjustified evil over one's own preferences. It may seem like donating bone marrow to strangers is good because it prevents their suffering, but this is likely just an illusion. Thus, Strong ST leads to a decision strategy that is effectively a limited form of egoism.
Weak ST
Limited egoism is a result of accepting Strong ST. However, this outcome can be avoided with Weak ST, which says our judgements about justification are only somewhat unreliable, but still good enough to provide some basis for moral decision-making. As Bergmann and Rea claim, the appearances would give pro tanto reasons for thinking that an evil has no justification, and thus pro tanto reason to intervene. However, there is a trade-off involved for sceptical theists. If we grant better reliability to those judgements, then it is reasonable to conclude that some apparently unjustified evils are actually unjustified. This would provide the basis for a modified evidential argument from evil.
B1. As far as we can judge, there are numerous cases of seemingly unjustified evil in the world.
B2. Our judgements about lack of justification for evil are somewhat reliable.
B3. Thus, it is unlikely that we are wrong about every one of these cases of apparent unjustified evil.
B4. Thus, it is very likely that there is unjustified evil in the world.
B5. God would prevent all unjustified evil if He existed.
B6. Therefore, God does not exist.
Premise B1 is based on the prevalence of diseases such as cancer, malaria, AIDS, etc.; people dying in natural disasters like floods, landslides, hurricanes, tsunamis, and earthquakes; the prevalence of war, murder, torture, etc.; and so on. There may be some instances where many people would be inclined to say, ‘He really deserves to die of cancer. It is a good thing he has cancer’. However, according to the World Health Organization, about 7 million people die of cancer every year, from all over the world, of all different ages, all races, and all walks of life. It is of course possible that they all have done something so terrible that they deserve it, and the rest of us just don't know. However, it would hardly be much of a defence of theism to rely on the claim that every cancer patient is a horrible person who deserves cancer. Furthermore, it is unlikely that the millions of cancer patients, including children, have hidden so many terrible crimes. To cap it off, there does not appear to be any benefit from the existence of cancer. It is painful for those who have it and for those who know people who have it. It prevents people from living normal lives, including preventing them from working. It is expensive to research and treat, and it kills. Taking it all together, from consideration of cancer alone, we have millions of apparent cases of unjustified suffering.
Premise B2 is the proposed rebuttal of the claim that ST undermines moral decision-making. Without it the objection against Strong ST is back in force and the sceptical theist has no leg to stand on. So we have to accept that our sense of moral value is at least better than guessing.
Premise B5 follows from the fact that God is perfectly good. A perfectly good being would not allow suffering if there was no justification for it.
The inferences from B3 to B4 and from B4 and B5 to B6 are unproblematic. The part of the argument that leaves room for controversy is the inference from B1 and B2 to B3. The inference is defensible, however.
Line B3 follows from the first two premises. Assuming that traditional theism is true, all natural evils are ultimately justified, and thus every instance where we judge there is no justification is a false negative. Thus, our reliability at identifying justification is equal to the chance that we will recognize a justification when one is there (since there are no false positives). Let R be the chance that after thinking carefully about an instance of evil that in fact has a justification, we recognize that it has a justification. Consider an extreme case where, even though all evils are justified, we fail to recognize a justification for any one of them. The probability of this occurring given 1 million cases examined is thus (1 − R)1,000,000. Even if R is .001 so that we almost never correctly identify justifications for evil, the chances are near zero of getting them all wrong.
This is important, since this shows the argument does not rely on a noseeum inference. Recall, a noseeum inference is an inference from ‘I can find no justification after careful consideration’ to ‘There probably is no justification’. With millions of independent judgements, it turns out we only require an inference from ‘I can find no justification’ to ‘There is at least a 0.1% chance that there is no justification’. Unless a 0.1% chance renders an event probable, this is not a noseeum inference.
Now it must be admitted that the above probability calculation is only a first pass, for two reasons: first, it assumes that our judgements about millions of cases are independent, and second it assumes that every case examined appears unjustified. Let us try to use more realistic assumptions. None of us has individually examined each of the millions of cases of evil. Instead, we are familiar with some individual cases, and for the rest we make judgements about categories based on paradigm cases. This, combined with some idea of the prevalence of evils in the world, gives us our base of judgements. For example, we have a general idea about what different kinds of cancer are like, and the kinds of results that they lead to, and we have access to statistics about prevalence.
Next, we have to allow that some cases of evil appear to be justified. Assume that our judgements regarding most of the cases with which we have personal familiarity are independent. Second, judgements about general cases of unjustified suffering can be broken down into categories about which assessments are independent. Cases of hunger from drought might be one class, hunger from mismanagement of resources could be another. These two cases may be handled differently. Then cases of cancer differ based on how bad and the cause, whether genetic or from smoking. Undoubtedly, the number of individual cases plus the number of classes for which we have clear senses of unjustified evil is less than a million. However, the number should still be quite high. I cannot speak for everyone, but with personal experience, access to the news, the internet, movies, and from speaking to people, I have come across many cases and kinds of suffering that appear obviously unjustified to my lights. Many of them are quite different, the destruction from earthquakes being quite different from the pain of mental disorders or genocide. My grounds for thinking that each of these is unjustified are different. The diversity of grounds ensures some degree of independence among the judgements that these cases of suffering are unjustified.
If the number of independent judgements is as high as hundreds or thousands, we move into the range where the law of large numbers starts to apply. This means that, assuming traditional theism, the fraction of cases that seem to be justified should not deviate far from R, our chance of recognizing a justification when one exists. This means that if only a small fraction of evils appear to be ultimately justified, the sceptical theist is in trouble unless R is likewise small. It would be amazing for us to be quite good at recognizing justifications when they exist and yet only manage to find justifications for a few of the evils we have considered. Proponents of the problem of evil are in just such a state: unable to find justification for more than perhaps a few cases of evil. This leaves two choices: either accept that the problem of evil as captured in B1−B6 is sound, which is a problem for theism, or accept that their value for R is quite low, which would bring us back to Strong ST and the inability to intervene.
The difficulty for sceptical theism now takes the form of a dilemma: if ST is interpreted to be too strong, then it completely undermines moral decision-making and leaves us to pursue self-interest with no regard for morality. On the other hand, if ST is interpreted weakly then the problem of evil reasserts itself. Sceptical theism faces a tricky balancing act, requiring our value judgements to be good enough to allow moral reasoning, but flawed enough to make it plausible that we are just wrong about the existence of unjustified evil. I am doubtful that such a balance can be maintained. I suspect that the middle ground is not a land where our moral reasoning about intervention is secure and the problem of evil is averted, but rather the middle ground is the worst of both worlds where moral reasoning is undermined and the problem of evil is still a problem. If we are able to start trusting our judgement that there is no justification for an instance of evil when we are making decisions, we can start trusting it where it justifies the first premise of the argument from evil.
Final attempt for ST
A defender of ST could still mount a defence at this point. In the previous section, I argued that it would be unlikely for us to be wrong about every case of apparent evil. This argument depended implicitly on the assumption that we are just as likely to be wrong about any given case of evil. However, in any subject area, some questions are more difficult than others. Sceptical theists can respond to the above problems if they can show that it is easier to tell justified from unjustified evil in ordinary cases than in cases that give rise to the problem of evil, and that the difference in difficulty dissolves the problem of evil without leading to moral scepticism.
However, this can only save ordinary moral practice if we can reasonably claim that, in general, evils related to the problem of evil tend to be extraordinarily difficult while the ordinary moral practice cases tend to be rather easy. Without a relatively clear demarcation between easy and difficult questions, it would be extraordinarily unlikely that all and only (or almost only) the questions that give rise to the argument from evil are difficult. Furthermore, if we can rely on our judgement about which questions are easy, it would appear that the problem of evil includes some of the easy ones. To use an earlier example, smallpox looks like an easy call. Humans were able to eradicate smallpox, and that looks like a good thing. The absence of smallpox has not led to the collapse of society, and it has decreased suffering. It is almost obvious that God did not need to let it run rampant for so long before we finally took care of it. This seems like an easier case than deciding whether donating to Oxfam is a justified intervention to prevent hunger and disease.
Sceptical theism requires a principled reason to think that cases where we might be able to intervene usually tend to be easy ones, while those that are most problematic for theism are hard ones. I have suggested that this is not the case. It is reasonable to hold that there is no significant difference in difficulty between the cases where we are able to intervene and those where we are not, between those that we need to evaluate to make decisions and those that lead to the problem of evil. At the very least, this shows that it is reasonable to be find sceptical theism problematic.
Sceptical theists, however, have one last defence of the claim that ordinary cases are easy and problem of evil cases are difficult. This is the claim that the problem of evil deals with the reasons that could justify God in permitting evil, and we are poorly equipped to understand the kinds of reasons God can have.Footnote 13 Even given this difficulty, we can still be good at identifying reasons for and against humans permitting evil. Since our own decisions depend on our ability to identify reasons for or against us humans permitting evil, the response allows that we can be quite good at figuring out whether we should intervene, but miserable at determining what God would be justified in permitting. This would prevent the problem of evil from getting off the ground, while avoiding moral scepticism.
To have any hope of saving ST, this response must presuppose that God has reasons to permit evil that arise only because He occupies a special role, e.g. the role of creator. Just as parents can have reason to punish their son even though those reasons would not justify their daughter in punishing her brother, God could have reasons to allow evil by virtue of being the creator, even though His creatures could not have those reasons. If the response does not presuppose this, then the fact that God has a reason to permit an evil would still mean that there is a good reason to permit the evil. Uncertainty about whether there is good reason to permit evil is all it takes to force sceptical theism into the dilemma.
Assume, then, that God can have reasons to permit evil that derive from God's position and relation to us. The proposed defence of ST is to claim that we are not able to identify these reasons. Our initial judgement regarding an example of apparent evil might be that no good creator could allow such a thing to happen. ST would counsel us to show humility and admit that God, as the creator, might have good reason to allow that evil that we just cannot see owing to our limitations. In this case, we do not occupy the role of creator, so God's reasons to permit evil cannot be reasons for us.
However, sceptical theism does not escape the dilemma by this response. If God has a reason as creator to permit an evil, we would be wrong to interfere. Just as it would be wrong to interfere to prevent a police officer from justly forcibly detaining and imprisoning a suspect, even though it would be wrong for us ordinary citizens to imprison the person, so it would be wrong to interfere with God's actions as creator. If God has good reason as creator to let a falling branch hit Susan, we would be interfering with the goodness of God if we intervened to prevent the branch from hitting her. As long as we admit that we cannot determine when God has good reason to permit an evil, we cannot know whether it would be good to intervene to prevent evil since there would always be the possibility that we were interfering with God's plan. Moral paralysis is still the result of ST. The dilemma arises again: either we have a pretty good grasp of the kinds of reasons that God might have to permit evil, in which case the problem of evil is still a problem; or we do not have a good grasp and we are unable to decide whether to intervene based on moral concerns and forced to decide based on preferences.
Conclusion
We have seen that sceptical theism is problematic. Theists cannot successfully combat the argument from evil by defending ST. Furthermore, the discussion also reveals that the evidential argument from evil need not rely on controversial noseeum inferences. However, the arguments above fall short of conclusively demonstrating that it is irrational to accept ST and maintain ordinary moral practice. One could have strong independent grounds for believing in God and God's good commands that override the evidential considerations given above. Or one could have an enormous initial credence that certain interventions are justified, overwhelming the arguments above. However, these positions would be difficult to argue, which offers support to the atheist position.Footnote 14
