The Evidential Problem of Evil

| | Comments (22)

This is the the twenty-eighth post in my Theories of Knowledge and Reality series. Follow the link for more on the series and for links to other entries as they appear. The last post looked at responses to the logical problem of evil. This post moves on to the evidential problem of evil.

Most philosophers have conceded that the logical problem of evil does not disprove theism. That does not stop them from offering a different version of the problem of evil against the existence of God. William Rowe is probably the most prominent philosopher defending the problem of evil today. He calls this version of the argument the evidential problem of evil.

The evidential problem of evil begins with the existence of evil, but it takes a different strategy. It does not seek to disprove the existence of God by finding a contradiction between the existence of evil and the existence of God. It simply seeks to show that the existence of God is unlikely given the kinds of evil, the amount of evil, how long evil has gone on, and so on. The facts about evil are that it is a lot more than just human free choices. Even if human free will explains some evil, some evil seems not to be explained so easily that way. Even if there are good reasons for God to allow some evil, is all evil explained so easily? What about cases where a little less evil would have accomplished the same purposes? Do we need to have all the kinds of evil? Did it need to go on this long? Does it need to be as widespread as it is? Do the individual cases of evil need to be as bad as they are?

The evidential problem of evil puts forward the lump sum of the facts about evil as its starting evidence. It does not conclude that this proves there is no God, as the logical problem did. It simply sees this lump sum of all the facts about evil as very strong evidence against the existence of God. Very strong evidence is still not a proof. Of course, if it is very strong evidence, and that evidence cannot be countered with alternative explanations, then perhaps no one should believe in God. We may be able to come up with explanations for a good deal of evil in the universe, but that does not mean we have explained all of it. If there seems to be no good explanation for a good deal of the evil that we can observe, is that not a good sign that some of it really has no good explanation? If that is true, then there probably is no reason why a good being would allow it. Therefore, there probably is no such being.

Responses to the Evidential Problem of Evil

It might be easiest to see the theistic responses to the evidential problem of evil if we separate the problem into several different questions.

A. What about kinds of evil that aren’t the result of the free actions of human beings?

B. God might allow free human beings to make evil choices, but why wouldn’t God immediately restore things to the way they were so that the world would not continue to contain evil?

C. Even if God would not restore things immediately, wouldn’t a good God prevent innocent beings from experiencing the consequences of that evil?

D. Even if God would not prevent the consequences or restore things immediately, why does it have to go on this long, and why does it have to be as bad as it is and as widespread as it is? Wouldn’t a good God want to minimize the effects of evil?

E. Even if these questions can all be answered somewhat satisfactorily, we do not have a complete explanation of evil. Isn’t that still evidence against God, if we can’t come up with a complete explanation of the amount of evil, the kinds of evil, the distribution of evil, the depths of evil, the length of time evil occurs, and so on? Explaining some evil helps resist seeing that evil as evidence against God, since it explains why God would allow that evil. But how can that remove all the evidence that the total sum of evil presents against God?

Theists have spent thousands of years looking at such questions, and these posts will not really do justice to everything that has been said, but I do hope to capture some of the more important things that recent philosophers have drawn on in responding to the latest versions of this argument in the form they have taken from philosophers such as William Rowe. In the next few posts, I will consider these questions, starting with questions A and B in the next post.


Huh. I don't know the relevant literature here that well, but I would have expected the dialectic to go a little differently here; rather than backing off from a "logical" argument to an "evidential" consideration, why not just alter the valid argument to get a clearer case of gratuitous suffering? Here's an argument schema:

(1) A morally perfect, omniscient, omnipotent being would not permit X to occur.
(2) X occurred.
So, (3) There is no MPOO being.

Fill in whatever gratuitous case of evil for X you like. Such-and-such natural disaster, freak accident, whatever. If there's any such true X, the argument goes through.

I had the same thought when I was teaching this earlier this semester, but that's not how any of the literature does it, including William Rowe, who is most associated with this. I believe David Lewis treats it the same way. All of the theistic responses to contemporary versions of this argument treat the problem exactly this way. So I've always taught it this way, since that's how the literature presents it. I haven't really thought through why until this semester, but here's what I've come up with.

You can't give anything but an inductive argument for premise (1), and then you're stuck in the same place, with something that you've merely got some good evidence for rather than a deductive proof. Since you might be able to come up with an explanation for X, for all you know, it seems that the best support for (1) is at best probabilistic. There isn't really any epistemic difference between giving a probabilistic, inductive argument for a premise of a deductive argument and giving a probabilistic argument to begin with. So you might as well present it more honestly as an inductive argument to begin with. I imagine that's what Rowe was thinking when he first put it this way.

I still wonder how philosophers who are skeptical about god deal with the nature of evil. Is something evil simply because most people agree to call it evil? How would one go about objectively defining evil within a non-theistic system of thought?

Or, put another way, if the argument from evil (against god) is successful, could it not also be self-defeating? And if it could be shown so, then does this not make it a fatally flawed argument?

I guess one could respond by saying that the argument is only about whether certain sorts of god claims are internally consistent, but I still think evil is just as much of a problem for the skeptic as it is for the christian, albeit in a different sort of way.

Paul, I the argument is exactly about whether theism is internally consistent. You don't have to believe in evil to present the argument against theism.

Now what you're suggesting is a common theistic argument. It's in fact the moral argument, which I've already treated earlier in the series.

ALright, but back to my question about defining evil then---if the argument is about whether theism of some sort is consistent, the skeptic MUST rely on the theist's conception of what evil is. There simply is no univocal definition of evil. So as long as the theist can arrive at an internal reconciliation of the existence of evil, however she defines it, and is able to do so in a manner which avoids some sort of formal fallacy, isn't the skeptic left without a response?

Let me back up for a minute.

I guess my problem really is that I am tied up thinking about actual discussions I've been involved in. I figure philosophy is only good to the extent that it can present you with arguments you can actually use within a discussion :)

In discussing the problem of evil with some skeptics I undertook to give some sort of short account of how christians have used scriptural accounts to talk about evil. In my case I used the book of Habakkuk since it was fresh upon my mind. Habakkuk complains to God about the evil paguing Israel. This implies that Habakkuk holds God to some extent to be responsible for the situation.

God's reply to Habakkuk is somewhat vague, but is certainly of the vein of "I'm going to do something about it." This seems to be a good nutshell of how christians deal with this problem: we aren't really sure how to explain the origin and continuance of Israel, we feel bothered that God, for now, isn't doing anything about it, but we receive the promise through the scriptures that God is doing something about it and will act with justice.

This, I think, needs to be paired with the question back to the skeptic: and how do YOU account for evil? I say that because all the skeptics I know DO believe in evil of some sort, but I have yet to see some sort of coherent account of what evil is froma non-theistic perspective. And, if one is a contemporary non-theist, the question of ultimate justice seems to be an overwhelming one. How could one possibly imagine justice to be done for all the deeds which have taken place, for those things done in secret, for those done by people "above the law", etc.

All of that is much broader that what you are talking about now, but it's just what has been on my mind. I say if someone wants to ask me how I deal with evil then I can ask them the same question.

So as long as the theist can arrive at an internal reconciliation of the existence of evil, however she defines it, and is able to do so in a manner which avoids some sort of formal fallacy, isn't the skeptic left without a response?

If the theist can offer an explanation for every little bitty kind of evil, then yes. But that's exactly what people like Rowe think can't be done. The question then would become whether the lack of such explanations constitutes sufficient evidence not to believe in God.

Your second comment, again, goes back to the moral argument. It's not the same kind of having to explain evil as this argument. It's basically just a reassertion of the issues in the moral argument but coming from the point of view of things being bad rather than the point of view of things being good. It's not a new issue in terms of the issues I've been covering in this series, though.

Something that I find puzzling about the POE be it logical or evidential.

The way the problem is structure would roughly be that:
1. There is an omniscient, omnipotent and benevolent God.
2. An omniscient, omnipotent and benevolent God would do away with evil.
3. Evil exists (and (3a) the consequence that God hasn't done anything to remove it or its effects)
4. An omniscient, omnipotent and benevolent God doesn't exist.

My point concerns (3). A God that could be compelled to correct the consequences of freely chosen evil and natural evil is by definition not omnipotent. I say compelled because as soon as something that might have evil consequences is happening, God has no choice but to act.
So when (3) and its consequence (3a) is proposed as a true statement, hasn't the atheist actually violated the assumption in (1) already?


I don't think that's true, because what would constrain God is not God's power but God's goodness. As I said in response to the logical problem of evil, a merely omnipotent being has a lot more options than an omnipotent, perfectly good being. But it's not because the latter being has less power and is thus not omnipotent. It's because a perfectly good being would not do certain things that a merely omnipotent being could do. I don't see how that point is a problem in an atheistic assumption. It's actually the most important point the theist would want to make to respond to the problem.

I read your post on the logical POE and agree about God's goodness part.

Setting that aside, I'm looking at the structure of the atheist's argument which involves (1) as an assumption which includes the omnipotent characteristic.

Now if you look at (3a), wouldn't that assumption of omnipotence (which I take as the power to act and also not act as well because God's freewill/sovereignity is based on his omnipotence) be violated because he is *required* to act to *rescue* his benevolent characteristic. I agree that it is out of God's goodness that he acts but to say that he must act before any evil occurs is to render him powerless to refuse.

Or is it that God's sovereignity is not located in his omnipotence?


So are you suggesting that an omnipotent being can violate its own moral nature? That just seems to make it impossible to combine omnipotence and perfect goodness in the same being. I'm still having trouble seeing how this is supposed to help the theist.

At this point, I haven't even looked beyond the structure of the POE - I'm focusing on it only.

About the question of omnipotence (as being able to do absolutely anything even the impossible) - Like you elaborated, a display of raw omnipotence of this kind is unaffected by contradictions. So by defintion, God's omnipotence should be limited to the possible.

However, in the POE, the atheist assumes omnipotence (of the possible variety) and further demands that in (3a) to violate this same omnipotence because he is also benevolent. In fact, the atheist has already done away with his initial assumption to obtain the supposed contradiction w.r.t benevolence. Or in effect, what I am saying is that the only way to obtain the contradiction w.r.t to benevolence is to implicitly falsify the atheist's initial assumption of omnipotence.

So the argument never takes off whether it is logical or evidential POE.

Seems to me that there is something in these characteristics taken together.

But is it possible to combine omnipotence and perfect goodness in the same being? Good question. The way I see it, it is possible only because it is God who decides when and where to act with omniscience and in a way to display perfect goodness. Demanding action on demand is to deny God's omnipotence in the first place.


I'm still trying to figure out exactly what your objection is. Is it the following?

You're taking the argument to say the following sort of thing. "God is perfectly good, and thus there are certain things God can't do." Does it get around your objection is you instead frame the argument in this way? "God is perfectly good, and thus there are certain things God wouldn't do."

I think once you put it that way, it doesn't interfere with omnipotence to say what the evidential problem of evil says. It just says that a perfectly good being will act according to that perfectly good nature, not that such a being couldn't act against that nature. But if you put it this way, you still get the conclusion that God would prevent evil to whatever extent is possible consistent with God's nature. If somehow you could establish that the amount of evil, kinds of evil, etc. that we experience in this life would be unlikely if such a being existed, then the argument can get off the ground, and it's then a question of whether the kinds of evil and so on are really evidence against that sort of being.

Jeremy, I think your substituting wouldn't for can't is the right move. A perfectly good, omnipotent being, like a merely omnipotent one, has the power to do anything that’s logically possible, but, unlike a merely omnipotent being, it will choose to do only good. Moral goodness lies in doing good even though other options are available. So moral perfection is not incompatible with omnipotence, but actually requires the power to do bad things (so that one can freely choose not to do them, not merely be prevented from doing them). Saying in this context that a perfectly good, omnipotent being ‘is required’ (as RI puts it) to prevent evil can only be interpreted as a loose way of saying that it will choose to prevent evil. This does not invalidate his omnipotence. Just like a good person always doing good things does not detract from her free will.

Jeremy, I’m not entirely confident of the free will defence you presented against the problem of evil. You sound as though your are not either, but you haven’t spelt out why.

The free will defence seems to assume that God could prevent evil only by forcing humans to act against their will or by turning them in some sort of automata, but I think there may be other ways. God could have contemplated all logically possible universes and seen how events would unfold in each of them before creating anything. I would have thought that it is logically possible for a universe to exist where nature is perfectly benign, and humans, or other intelligent species, have free will but always choose to do good. By choosing to create that universe rather than ours, God would ensure that there would be no evil.

I think that I should just get the POE out of the way first because as far as I can see, its a distraction and focus solely on the omnipotence characteristic.

You've illustrated the need for logical omnipotence and I agree with that.

Now going further with logical omnipotence, is it a logically valid argument to construct an argument that:
1) Assumes logical omnipotence
2) For proof of logical omnipotence, *requires* an omnipotent being to do something
3) Sees evidence that something *required* to be done is not done and concludes that this being is not logically omnipotent.

All I'm saying is that the definition of logical omnipotence (or omnipotence in general) means that you cannot *require* an omnipotent being to do anything.

Now, if you introduce Goodness as another characteristic that this being possesses, an atheist might wish to hold logical omnipotence and goodness in a contradiction but in order to do so, he *requires* this omnipotent being to do something but that is to invalidate the assumption of omnipotence in the first place.

The more I think about it, I think I might be muddling around the omnipotence characteristic but I sure appreciate the time you're spending on this comment.


Jacinto, I'm simply not a libertarian about free will. My philosophical reasons for that are here. I also believe it on biblical grounds, since I think compatibilism is assumed by all the biblical authors who write anything at all relevant to the issue. So I don't think a response relying on a libertarian conception of freedom is going to work if it requires asserting that view. Fortunately, I don't think it requires asserting that view, as I explained in a comment above.

As for your reasons, I don't actually agree. In fact, I think what you've described is impossible. It's certainly true that God could create the world in such a way that people might end up doing only good. But what's not possible is that God could create a world in a way that guarantees that the people in it both have libertarian free will and do only good. So you can speak of God looking at all the worlds and choosing one, but what does it mean to choose one and actualize it? The only thing it could be is creating a world in a way that guarantees what people would do.

That sort of action would indeed remove libertarian freedom, which is why God wouldn't do it if libertarians are correct about what freedom is. It only makes sense for God to actualize a world in all its history in a way that doesn't violate freedom if compatibilism is true (and I believe it is, so I can talk that way, but libertarians can't).

The one way people try to reconcile them that could do it is with what's called middle knowledge. I don't know how familiar you are with the literature on this issue, but middle knowledge is God's knowledge of what free beings would do in various situations. If God has that knowledge (as I think he does), then he can predict what people will do in any situation he allows them to be in, and then you could get God knowing which world would develop given certain initial conditions, because God could anticipate every free being's actions. The only problem is that if these beings have libertarian freedom, what makes it true that they will do certain things in certain contexts? The only things I can think of that would explain how middle knowledge is possible are things that aren't true if we have libertarian freedom. So I don't think that view ultimately works. The foreknowledge issue is going to come up later on in this series, though, so perhaps we should save some of the questions on that for then.

RI, see Jacinto's comment above in response to you. I think I agree entirely. I'm not sure it's a different point from what I'm saying, but it's coming at it from a different direction. At this point I don't have anything else to say besides what I've already said.

Jeremy, I think we agree on the issue of omnipotence and goodness, I was just emphasising different aspects.

In the free will issue I was indeed assuming God's perfect knowledge of everything past and future even before creation. So I take this sort of omniscience is possible with compatibilist freedom but not libertarian freedom. Should we then not talk of the libertarian freedom defence against the problem of evil, rather than simply the free wil defence?

I'm just using the term used in the literature. It's Plantinga's argument, and the philosophical community uses the name he gave it.

I don’t really know much about this, but I would have thought that God would choose the laws of causality and what type of freedom he would give us. So we could restate the problem of evil like this: a perfectly good, omniscient and omnipotent god would consider only worlds with compatibilist freedom, so that he could screen out all bad worlds in advance and create one without evil. So a defence based on free will would have to argue either that libertarian freedom is the only logically possible type of freedom or that libertarian freedom is so much better than any alternative that it compensates for the possibility of evil that it necessarily entails.

Yes, but libertarians think a world where people have libertarian free will is far superior to a world where what compatibilists call free will is all people have (because they don't count it as genuine freedom).

Leave a comment


    The Parablemen are: , , and .



Books I'm Reading

Fiction I've Finished Recently

Non-Fiction I've Finished Recently

Books I've Been Referring To

I've Been Listening To

Games I've Been Playing

Other Stuff


    thinking blogger
    thinking blogger

    Dr. Seuss Pro

    Search or read the Bible

    Example: John 1 or love one another (ESV)

  • Link Policy
Powered by Movable Type 5.04