Skeptical Theism: A Powerful Response to the Problem of Evil

Context and Background

Long ago there was once an argument called the logical problem of evil, and it was defeated. It has a quite simple form:

  1. If God exists, then there does not exist any evil
  2. There exists some evil
  3. God does not exist

These days, even the most ardent sceptical philosophers do not defend the logical problem of evil. To poorly summarize centuries of fairly technical dialectic, the theist rejects 1 by saying that God is justified in permitting evil if God uses that evil to bring about some greater good that depends on the existence of the evil. This seems pretty plausible, and so most atheists do not try to put forward 1 any longer.

But this doesn’t bother many of them, because they attempt to put forward an evidential formulation of the argument. Here is what modern evidential problem of evil arguments looks like:

  1. If God exists, then there are no evils which do not bring about some greater good
  2. There exist some evils for which we cannot see any corresponding greater goods
  3. For some of these evils, if such a corresponding good did exist, we would see it
  4. Therefore there exist evils which do not bring about some greater good
  5. Therefore God does not exist

This argument is valid, and the first premise is pretty plausible. So the theist must proceed by attacking or undermining premises 2 or 3. In general, an attack on premise 2 is a theodicy, and an attack on premise 3 is called skeptical theism.

I think theodicies are too heavy a burden for us to bear, at least given an evidential argument like the one above. In this case, the Christian has to give a sufficient reason for the existence of every evil that exists. We have to explain specifically how a child with a painful terminal cancer produces some higher-order good. I do not think that we can do that, and I think attempts to do so often end up being grotesque.

I think the wiser route, and the more biblical route, is to question premise 3. That is, argue that we are not in a position to see such reasons if they do exist, so the inference from “we do not see them” to “they do not exist” is not justified. As Josh Rasmussen helpfully puts it, we can make two observations here. We can either 1) fail to see a reason why God would allow it, or 2) see that God would have no reason to allow it. And are we in a position to differentiate between 1 and 2? The atheist has to put forward an argument that we are in situation 2. But instead of just sitting and waiting to hear that argument, I think we can do better.

What I aim to do here is to put forward some reasons that the theist has for thinking that we are in case 1 rather than case 2. If we have such reasons, this will undermine the atheist’s claim that we are in case 2.

Let’s introduce some terminology that will make our life a bit easier. We will call an evil Gratuitous (G) if it is an evil for which there is no resulting greater good that justifies the existence of that evil. We will call an evil a Horror (H) if we cannot find such a justifying reason. A G evil is the kind of evil that would be incompatible with the existence of God, and the evidential argument for evil is essentially a move from H to G, and from G to the nonexistence of God.

Epistemology and Skeptical Theism

We can develop some reasons to put forward the skeptical theist position from a consideration of some very plausible principles in epistemology. The main example of such a principle is something called CORNEA, put forward by Wykstra and Perrine.

On the basis of cognized situation s, human H is entitled to claim ‘It appears that p’ only if it is reasonable for H to believe that, given her cognitive faculties and the use she has made of them, if p were not the case, s would likely be different than it is in some way discernible by her.


CORNEA is a powerful principle which, if true, would undermine many conceptions of the problem of evil. Even “commonsense” problems of evil such as this one. I think it is more intuitively phrased in Bayesian terms, and is better understood as simply a restatement of what Bayesian evidence is. That is:

Observation E is evidence of some proposition H if P(E|H) > P(E| ¬H)


To get a better intuitive handle on this, let’s consider a real-world example. Suppose I get home and notice that my grass is wet, and I form two competing hypotheses as to why:

  1. My wife watered the grass
  2. It rained

Now suppose that I know that if it rained, the grass is always wet, but that if my wife waters it, she does a bad job, so it only gets wet 50% of the time. Knowing these things, my conclusion should be that the grass being wet is evidence of rain. Because the rain hypothesis does a better job at explaining the observation, the observation is evidence of the rain hypothesis.

If Bayesian reasoning is new to you, I suggest spending some time learning about it. It’s very useful. The point here is that if some observation is equally likely given two competing hypotheses (or we are unable to calculate either of those likelihoods), then that observation can’t be evidence one way or the other.

The point of this kind of idea is this: if both theism and atheism predict H with a similar degree of confidence, then H simply can’t be evidence either way. Being generous to the atheist and granting that atheism does predict H, the atheist now has to argue that theism does not predict H. And while we could wait for the atheist to make such arguments and then respond to them, I think it is also worthwhile putting forward some positive arguments for the claim that theism predicts H with some high degree of certainty.

One such argument, given by Howard-Snyder, is the Progress Argument. I will rephrase it here:

  1. Many moral truths which we now take to be obvious were not considered true in the past
  2. Therefore there are many moral truths of which we are now ignorant
  3. Some of these may include justifications for the H evils that God ordained in the world
  4. Therefore we are not justified in inferring that there do not exist any goods which can justify the H evils in the world

This is an inductive argument, and it is hard to deny the force of it. Surely there are some people who will say that we have plateaued, and have discovered a very large portion of the moral truths that exist, and so our inference is justified. But I think these people are probably wrong: society continues to “discover” new moral truths with increasing rapidity. Even if I do not agree with all of these changes, most people who put forward this argument are probably people who will believe that recent changes in the moral zeitgeist concerning homosexuality and transgenderism are improvements. But those improvements have come about in the last 20 years or so. I think it’s exceedingly foolish to say that no further “improvements” or “discoveries” will come, or at least none significant enough to make us a bit more humble about our own moral knowledge.

To improve the strength of this argument, Howard-Snyder puts forward a supporting argument. Quoting directly from the SEP:

This argument begins by recalling that the data from which the strongest arguments from evil start are the profusion or seeming excess of evil in the world which, indeed, seems to be integrated into the fabric of nature and society. But for that very reason (their complexity and intricacy), any complex good whole of which these evils are a part would have to be exceedingly complex. Thus, he infers that it would not be surprising if it were beyond our ability to fathom.

These are some good starting points to consider a positive case for skeptical theism purely from secular epistemology and plain reason. However, the theist has more tools in their box than just this, and by considering some facts about the nature of God (and the particular nature of the Christian God), we can put forward an even stronger case.

Human Experience and Skeptical Theism

An interesting argument that I do not see discussed often is the argument from the theism of suffering people. It seems like the only people who really endorse the problem of evil, who really make the inference “there is no justification for this suffering”, are the people in the world who suffer the least. That is, the educated westerners. We live like kings, having all of our basic needs easily met and having the leisure time to pursue whatever we choose, including getting into arguments with strangers on the internet. We have the easiest lives anyone has ever had. And yet we are the only ones who really have any atheists among us.

On the other hand, the global poor, the suffering, the hurting, etc. are far more religious. They do not think that they can responsibly make the inference from suffering to the nonexistence of God. Presumably, at least some of them have considered the option and rejected it. Presumably, some of them are intelligent and clear-headed enough to be epistemically responsible in rejecting it.

Why this difference? I speculate that it is because us educated westerners have become so accustomed to the very mild suffering that most of us go through that any worse suffering has become unimaginable, and therefore we say that God would not allow it. But the people going through the suffering often don’t think that way. Maybe the educated westerner has underestimated man’s capacity to handle suffering due to our pampered position.

Some of the people who have experienced the worst horrors have remained theists, and indeed Christians, and have often said afterwards that they can see now what God was doing in them through the suffering and evil. The kinds of evil and suffering that they have experienced are not vastly qualitatively different from the kinds of evil the atheist is calling horrors and using to motivate their argument. But the people experiencing that evil are in a better place than the atheist is to talk about whether it’s unjustified, and many of them say that it’s justified. Plausibly then, all such evil is justified.

Does this argument work in reverse? Are those who don’t say it’s justified a problem for me? I don’t think so, I don’t think the observation is symmetrical. I think “I do see reasons, therefore reasons exist” is a lot stronger than “I don’t see reasons, therefore they don’t exist”. Seeing reasons is a very good reason for thinking that there are reasons, just like me seeing a chess tactic is a pretty good reason for thinking it exists: I’m wrong sometimes, but rarely. But when I don’t see a tactic, that is not a good reason to think that there is no tactic: I often miss tactics, much more often than I incorrectly see tactics. It is also my experience that the “I see reasons” often comes late in life, after a lot of reflection, and that puts it in a better epistemic position than “I do not see reasons”, which is often immediate and often supplanted by seeing reasons decades later. So by the fact that for any class of evil, some thoughtful, reflective people who have experienced that evil will eventually see reasons for it (or at least, will say that their lack of seeing reasons doesn’t justify an inference that there are no reasons), I think we can plausibly say that there possibly exist reasons for any evil in that class. This undermines the atheist’s judgement that there are no reasons for some particular horror.

Bare Theism and Skeptical Theism

If all we grant is that there exists the classical theist God, then I think we have some very good reasons to put forward a skeptical theism.

One such reason is the Parent Analogy, put forward famously by Wykstra. The analogy is quite simple: God under most conceptions of theism has a relationship to us similar to a parent’s relationship with their children. It’s no surprise to anyone who has ever taken a child to the dentist that sometimes, parents have to subject their children to suffering for reasons that the children don’t understand, but are nonetheless good reasons. Every child, if they were articulate enough, would be able to find some mysterious suffering that they are subjected to by their parents and say “I cannot see any sufficient justification for this suffering, it appears gratuitous to me”. Does the fact that the child experiences this suffering as mysterious, and can’t see any reason for it, mean they are justified in believing that there is no reason for it? We must say “no”. Sometimes children simply can’t understand the reasons that parents have, and they’re in no position to judge. If the relationship between man and God is relevantly similar, then this undermines the atheist’s judgement that the suffering that we see and experiences has no sufficient justification.

I think we can also, given theism, predict skeptical theism due to some goods that seem to arise from skeptical theism itself. The chief contenders here are trust and compassion. It seems to me that these attributes are impossible to develop in the absence of some mysterious and sufficiently great suffering.

Let’s think first about trust, specifically trust in God. It seems plausible that God wants us to develop trust in Him, at least under Christianity this is a central part of our relationship with God. So if God wants us to have this faith, how can He cause it to happen? It doesn’t seem like trust can exist if no H exists.

To see this, consider an example. If my wife says to me “I’d like our relationship to be one in which you can trust me”, can I safely respond “of course I trust you, I’ve got a camera and microphone hidden in your handbag!”? I don’t think I can, and not just because of the invasion of privacy, but also because such surveillance leaves no room for trust! I don’t have to trust, because I can surveil. Similarly, if I am able to see a corresponding good for each evil that exists, I never have to trust God to know what He’s doing. So in short: if God wants us to trust Him, horrors will exist. Plausibly God wants this, so plausibly God will allow horrors to exist. So the observation of the existence of horrors can’t be evidence against God.

Compassion is another attribute that God plausibly wants us to develop: people are simply better people when they’re compassionate. That’s part of what makes someone a good person. And compassion is developed by suffering. I think many of us know this intuitively too, having experienced this ourselves or seen it in others.

As John DePue says in his paper on skeptical theism and divine deception:

Another second-order justification for the appearance of gratuitous evil is that it produces some of the greatest acts of love and compassion. When we respond to seemingly pointless and unjust pain and suffering throughout the world, the exceptional character of the generosity and kindness that motivates these actions is fueled by the fact that the circumstances appear to be purposeless. Arguably, many of the highest degrees of human love would remain unrealized without the appearance of gratuitous evil that spurs us to extend these extraordinary acts of charity to others. Of course, positive skeptical theists are not claiming that these evil events themselves are justified on this basis, which would effectively disqualify their view as a version of skeptical theism. As skeptical theists, they insist that the justification for these evils remains beyond our understanding. What positive skeptical theists are claiming, however, is that if there is a God, we should expect the world to contain the appearance of gratuitous evils. This is the second-order justification for why God allows us to experience the world as appearing to have gratuitous evil.

The Bible and Skeptical Theism

The Christian has some further reasons to think that skeptical theism is plausible, following from some scriptural passages and various other facts that we know about God.

Consider first Job. The book of Job is probably the oldest book in the bible, and so constitutes the earliest Jewish (and therefore Christian) source. But the book of Job is all about mysterious suffering: Satan challenges God that job only serves Him because of God’s blessings on His life, so God removes these blessings and proves Job faithful. So right from the start, in the earliest source we have for our religion, we are told that deep mysterious suffering is not only possible but expected. So the Christian has a good reason to affirm skeptical theism, and the atheist has a lot of trouble motivating premise 3 of their evidential argument. In Bayesian terms, if Christianity predicts H, the observation of H can’t be evidence against Christianity.

The story is similar to that of Joseph. Sold into slavery, falsly accused and thrown into prison, he experienced great evil and suffering for which he could not see any justification. It was not at all apparent what greater good was coming from his suffering, if there even was one. If we are now justified in saying “there can be no reason for the evil I see”, so can Joseph. Joseph could have said that, if such an observation can be made at all. But if Joseph had said that, he would have been wrong. His suffering was used to preserve Israel, and later in his life, he would remark What you intended for evil, God intended for good”. Joseph would later see the good coming from it, and would praise God for bringing it about. Joseph thinks that the horror was not gratuitous. And that’s part of the story here: in another very early Christian text, the first book of the scriptures, we are told in no uncertain terms that great and mysterious suffering will befall us, but at the end we may see why. Christianity predicts H.

And Joseph serves as a type of Christ. In the Christian perspective, the death of Christ was the greatest evil to ever occur. The only perfect man, the only innocent man, the one who is Himself God and the rightful King and Ruler of the world, is betrayed by His friends, falsely accused, tortured, and publicly executed. For the Christian, there is nothing more evil than this. The Apostles, who did not understand, would see this as the clearest example of a horror to ever occur. But Christ rises and God is vindicated, as the greatest possible evil results in an incredible good, the salvation of God’s people from the power of sin. And that good was not possible without the blood of the Son. What I want to emphasize here is this: if God is able to use such a tremendous horror to bring about such a great good, then it stands to reason that there are great goods that can result from the evils we see around us. We do not know what they are, the Apostles didn’t know yet why Christ died, but if God can turn the cross into a good, God can turn anything into a good. So again, the Christian has grounds to reject the atheist’s move from “I see no reason for this evil” to “There is no reason for this evil”.

Another point to make here is the theology of transcendence. God is above and beyond us. All throughout the Old and New Testaments we hear things like His ways are not our ways, He is in heaven and does all He pleases, how unsearchable His judgements and unfathomable His paths. We are reminded again and again that He is God and we are not. This is what He says in Job too, somewhat sarcastically. We should not at all be surprised by mysterious events, God’s purposes are often mysterious. In fact, we would be surprised if we did see such reasons all the time. This point is regularly made by preachers and pastors, even outside the context of formal discussions on the problem of evil.

If one of God’s purposes in the world is the glory of the incarnate Son sacrificed to save the unworthy, then we again expect a world full of rampant sin and evil for which we can’t see any purpose. We expect a world that is entirely unworthy in every way. Because such a world magnifies the glory of Christ, as He steps into the world and dies for it. And this is not a justification for any individual evil like a theodicy would be, but is instead a principled prediction from the core truths of Christianity that the world will contain severe evil for which we can see no justification. The worse the world is, the more unrestrained evil in the world is, the greater the glory of the Son who stepped into it and died for it.

Objections to Skeptical Theism, and Responses

With any topic as important as this, philosophers have gone back and forth many times on various problems caused by accepting skeptical theism, and various ways of solving those problems. Here are some of the main ideas.

Ad Hoc

One important objection is that skeptical theism is ad hoc. To understand the force of this objection, consider the geocentric model of the solar system. As we all know, this model doesn’t quite fit the observations. So the proponents of the model added epicycles. These are ad hoc modifications of the model to make it fit the data, and if you add enough epicycles you can fit it to any data. You can always continue to modify a model in an ad hoc way to make it fit the evidence. So the evidence that should have refuted the geocentric model ended up being incorporated as epicycles, and now the model “predicts” that evidence.

If skeptical theism were an ad hoc modification do theism, then that would undermine our Bayesian argument. Yes, Christianity would predict H, but only because Christianity was modified in an ad hoc way to predict H. Therefore Christianity is made implausible. The good news is that we don’t seem to have this problem.

To avoid the charge of being ad hoc, what we need to do is demonstrate that the expectation of H under Christianity is entailed by the core of Christianity, rather than just being an extra bit that was tacked on to avoid the argument from evil. And I think we can pretty easily show that to be true. We talked about Job, and from Job we know that the expectation of H has been part of Christianity right from the start, as early as we can go. And Job is incredibly early, possibly as old as the patriarchs. There can’t be any ad hoc modification if there was never any modification in the first place, and Job significantly narrows the window during which such modification could have happened.

We can refute the accusation in another way: not only by showing that our faith was never modified in this aspect, but also by showing that it proceeds from a central part of the belief. And it does, central to Christianity is God’s transcendence. We have argued this above: that because of how vast the gap between God and humanity is, it is entirely expected that some (most!) of God’s purposes will remain mysterious to us. Alexander Pruss makes a similar argument on his blog: it would be surprising for the Christian if they were in a world where:

of this evil was such that its point (a) could be understood by us and (b) it would be on balance good for us to understand its point. In regard to (a), we can cite our cognitive limitations. In regard to (b), we can cite the fact that it is likely that some of the justifications for permissions of evil would involve soul-building, whereas it is very plausible that some soul-building would require techniques that are hidden from its beneficiaries.

If skeptical theism were ad hoc, then we wouldn’t have been able to give a bunch of reasons from epistemology, theism, and Christianity that motivate skeptical theism. But those reasons are all pretty plausible, so it does not seem like sceptical theism is ad hoc.

Normative Skepticism

Another objection to skeptical theism is that skeptical theism entails normative skepticism. The intuition here is easy to see: if we can’t make moral judgements like “this H should have been prevented”, then maybe we can’t make any moral judgements at all. Why be skeptical about some of our moral judgements, but not others? Any perhaps more dangerously, any reason we have for undermining the moral judgements we use in the problem of evil argument will end up undermining our normal, every-day moral judgements.

This is the argument of the philosopher Sharon Street, which we have talked about before. I still believe that the strongest solution here is an agent-relative ethics: it may be right for humans to try and do what they think is right, and it may also be right for God to frustrate them and counteract them. It seems quite plausible that God’s moral duties towards the world are different from a human’s moral duties towards the world, because of the vastly different kinds of things that humans and God are. This is what is argued in a recent paper by Philip Pegan:

In particular, moral common sense dictates that it would be seriously bad for a human being who knew she could easily prevent such an accident to fail to do so. Moral common sense obviously dictates nothing of the sort about God. So long as it is reasonable for the theist to believe that if God were to exist his relationship to the world would—or even just might—be relevantly different from that of any human being’s relationship to the world, reflection on this case need not give the theist any reason to doubt that it would be seriously bad for a human being who knew she could easily prevent this accident to fail to do so.


I think this is very plausible under Christianity. Most classical Christian accounts of ethics are something similar to virtue ethics, and in such a view the kind of thing that an object is determines what is right for that object to be or do. If God and humans are different kinds of things, it’s very easy for the Christian to say that they have different kinds of duties towards the world.

The atheist here may still respond “When I make ethical judgements, I am judging about what makes the world better or worse qua world, and acting based on that judgement. If skeptical theism were true, it would still undermine this judgement”. This is forceful because we are now looking at the value of a world qua world, whether or not a world will be better or worse based on the outcome of our actions. But I think this approach will fail as well, for a couple of reasons.

First, it’s not clear to me that moral judgements are about the world qua world. When I choose to save a drowning child, it’s not because I think that doing so will end up making the world better on average. I have no idea what kind of person that child will grow up to be, and even if I did know that they would grow up to be evil, I would probably still save them. I am not making a judgement about the moral value of the world under my decision, I am making a judgement about what kind of action that I qua human ought to take.

The uncertainty posed here is my second point: if we are attempting to judge which of our individual actions make a world better or worse, we are in a hopeless position. We have nowhere near enough information to make informed judgements of that kind. We have no idea what kinds of complex moral goods or evils might vindicate or condemn a particular world-choice. We have no idea what the long term consequences of our actions will be. If we have learned anything from the study of the discipline of economics, it is that well-intentioned and reasonable-sounding policies can have disastrous second-order effects.

So I think the normative skepticism objection does us no harm. We can appeal to a strong tradition of virtue ethics within Christianity to avoid Street’s argument, and avoid the arguments from the atheist that we need this kind of judgement to decide between good and bad worlds in our own actions.

Divine Deception

I will not spend too long on this point, since we are getting quite long as it is. Some atheists have argued that if skeptical theism is true, God is deceptive. The argument is obvious: if God makes a world under which it seems like X, but really not X, then God is deceptive.

I think that this paper from John DePoe presents a very strong counter-argument: God does not intend to cause people to believe false propositions and God does not provide sufficient evidence for someone to justifiably believe a false proposition. Since these conditions are not met, God is not deceptive. I do not think I can phrase it better than he did, so I will encourage you to read the paper instead.

Objections to CORNEA

Because of CORNEA’s strong restriction on what counts as evidence, some philosophers have suggested that if we endorse CORNEA, we have no ammunition to respond to the problem of global skepticism. That is, we have no ammunition to respond to the problem that we may all be brains in a vat, or deceived by an evil demon into thinking we exist in the physical universe, or something similar. The arguments here are quite technical and a bit beyond our scope for today, so I will simply refer you to this paper for a discussion on the problem and some suggestions on how to resolve it. I think that the principles required to solve the problem are quite plausible and can be endorsed by the theist without any danger.

There are also concerns that CORNEA is incompatible with another plausible and widely accepted epistemic principle: commonsense epistemology. Dougherty lays out the argument here. Similarly to the above, the arguments here are dense and technical and beyond the scope for today. Matheson responds here and here, and those responses satisfy me.

Conclusion

When faced with an evidential argument from evil, the Christian has very good reasons to reject the inference from “we cannot see a justification for some H” to “There is no justification from H”. These reasons are motivated by plausible principles in epistemology, from experience, from the content of theism, and from the specific content of Christianity. Therefore at the very least, the evidential problem of evil fails when deployed against Christianity.

Further Reading

One Reply to “Skeptical Theism: A Powerful Response to the Problem of Evil”

Leave a Reply

Your email address will not be published. Required fields are marked *