The “Evil god” Objection

Dr. Craig this is a simple question in regards to your debate with Stephen Law.

Suppose someone hypothetically argued for an Evil God that exists. Could one use the "Problem of Good" as an objection, just as Non-Theists use the "Problem of Evil" against theism?

Would all the arguments such as Plantinga's Free Will Defense be flipped around, and actually work against the problem of good?

So far, it truly does appear that Evil is a privation of good, and the arguments used to counter the "Problem of Good" against an Evil God do not work very well as a refutation.


United States

Cornell, I’m grateful for your question because I think it’s very easy to misunderstand Stephen Law’s “evil god” objection as a result of conflating distinct questions.

First, let’s begin with the cosmological and teleological arguments. If successful, these give us a Creator and Designer of the universe. Notice, however, that they do not tell us much or anything about the moral character of the Creator/Designer. In my popular talks, I sometimes put this point by saying that the Creator/Designer might be an absolute stinker, for all we know!

That’s why, by the way, the widespread objection to Intelligent Design based on the cruelties of Nature is worthless. As I point out in my debate with Francisco Ayala, one might as well argue that a medieval torture rack does not need an intelligent designer because anyone who would make such a thing couldn’t be a very nice person.

Precisely because the cosmological and teleological arguments say little or nothing about the moral character of the Creator/Designer, they are immune to the atheist’s most important argument, the problem of evil and suffering. They are therefore powerful components of a cumulative case for theism. They cannot be ignored.

But that’s just what Stephen Law did in the debate. His response to these arguments, as you saw, is simply to say that even if successful, these arguments do not prove the existence of God, since in order to infer that the Creator/Designer is God, one has to prove that He is good. But for all we know from these arguments, the Creator/Designer could be evil. This is NOT, however, the “evil god” objection. Law is merely noting the incompleteness of the theist’s case so far: we’ve got a Creator/Designer, but we’ve as yet no reason to think Him good and therefore God.

In the debate, Law made the remarkable claim that the cosmological and teleological arguments are not even part of a cumulative case for theism! This is clearly wrong. The probability of God’s existence given the evidence for a Creator/Designer of the universe is obviously higher than without it. To borrow Tim McGrew’s illustration, suppose you’re expecting an afternoon visit from a friend in the military. That afternoon your wife tells you, “There’s a man coming up the walk.” Do you shrug this off with the comment, “Oh, well, it could be anybody!” She then says, “He’s wearing a uniform!” Should you respond, “Well, maybe it’s a policeman” and continue to go about your affairs? Of course not! The probability that your friend has arrived, though not certain by any means, is definitely higher given your wife’s testimony than it would have been without it. It is thus part of a cumulative case for the conclusion that your friend has arrived, and it would be folly to ignore it. Similarly, the probability that God exists is much higher given the evidence for a Creator/Designer than it is in the absence of such evidence.

So what argument does the natural theologian give for thinking that the Creator/Designer is good? Here Law mistakenly seems to think that the theist arrives at the conclusion that the Creator/Designer is good by an inductive survey of the world’s events. Seeing all the goods in the world, the theist supposedly infers that the Creator/Designer is (perfectly) good. That assumption is simply incorrect. As Michael Bergmann and Jeff Brower point out in their response to Law, “no traditional theists we know of have ever argued for God’s perfect goodness . . . by simply inferring it from the existence of some good in the world.”i They conclude that Law hasn’t “done anything to touch, much less undermine, traditional belief in the existence of a being which is at once all-powerful and all-good. If Law wants to mount a real attack on traditional theism, he will need at the very least to engage some of the actual support that has been given . . . for belief in God’s goodness, explaining why it fails, rather than completely ignoring it.”ii

What many natural theologians, including myself, do to justify belief in the perfect goodness of the Creator/Designer proved by the cosmological and teleological arguments is to offer various moral arguments for God. In so doing, one needn’t appeal to the good in the world at all; one can instead point to instances of objective moral evil. So, in our debate, I argued:

1. If God does not exist, objective moral values do not exist.
2. Evil exists.
3. Therefore, objective moral values exist. (Some things are evil!)
4. Therefore, God exists.

Law takes almost no cognizance in his published work of such an argument for God as the foundation of objective moral values and duties. All I could find is the brief comment: “it remains possible that a cogent moral argument along the above lines might yet be constructed. I suspect that . . . this is the most promising line of attack [for theists to take].”iii I concur.

It’s worth noting that Law agrees with premiss (2) because he is a moral realist. So in order to resist the force of this argument, he must deny (1). But on this score, he has very little to offer by way of explanation of objective moral values and duties in an atheistic universe. Indeed, after presenting the old Euthyphro dilemma, he admits, “None of this is to deny that there is a puzzle about the objectivity of morality—about how it is possible for things to be morally right or wrong independently of how we, or even God, might judge them to be.”iv But he has no solution to this puzzle to offer. Then he notes the theistic solution: “suppose that ‘God’ refers, not to the creator of this yardstick, but to the yardstick itself . . . then to admit that there is an absolute standard of right and wrong is just to admit that God exists. . . .”v That’s absolutely right! So what’s his objection to the theistic solution? He says, “this is a very thin understanding of what ‘God’ means.”vi This objection is based on a confusion between semantics and ontology. The theist isn’t offering a definition of what the word “God” means. The theist is claiming that God, in all His fullness, is the paradigm of moral value. God is the yardstick of moral value. By contrast Law more or less admits that the atheist has no explanation of the existence of the objective moral values and duties that we both apprehend.

So far the “evil god” objection has yet to appear on the scene. We have simply been discussing what grounds the theist might offer for thinking that God exists, i.e., that there is a perfectly good Creator/Designer of the universe.

It is at this juncture that Law raises the problem of evil. As we agreed in the debate, this problem can be stated in non-moral terms by substituting “suffering” for “evil.” The objection is that the suffering in the world provides, in Law’s words, “overwhelming evidence “ that God does not exist. For an all-powerful, all-good being, it is alleged, would not permit the suffering we observe in the world. Therefore, such a being probably does not exist.

Suppose that the theist responds, as I do, by saying that, for all we know, God may well have morally sufficient reasons for permitting the suffering in the world. We all know cases in which we permit suffering because we have morally sufficient reasons for doing so. What Law would have to prove is that it’s improbable that God has morally sufficient reasons for permitting the suffering in the world. But how could he possibly prove that? God’s justifying reasons might never appear in our lifetime or locale or even in this life. Suppose, for example, that God’s purpose for human life is not happiness in this life but the knowledge of God, which is an incommensurable good. It may be the case, for all we know, that only in a world suffused with natural and moral evil would the maximum number of people freely come to know God and find eternal life. Law would have to show there is a feasible world available to God in which there is a comparable knowledge of God and His salvation but with less suffering. That’s pure speculation.

It is at this point that the “evil god” objection finally comes to the fore. Law’s response to the above is to say that if such a response is tenable, then someone who believes in an evil god could also justifiably say that the goods in the world do not constitute refutation of the existence of such a deity because the evil god could similarly have reasons for permitting all the goods in the world, which Law just takes to be absurd.

A couple of comments: the “evil god” hypothesis is not suggesting that God could be evil. For, by definition, God is a being which is worthy of worship, and so no being which is evil could be God. That’s why Peter Millican, who independently formulated a similar argument, refers to the evil supreme being, not as “God,” but as “anti-God.”vii That is less misleading than Law’s terminology. One can refer to this being as “god” only by using the lower case “g,” as I have done. The idea is that there is a Creator/Designer of the universe who is evil. You can see immediately why this argument, which properly belongs to concerns of theodicy, gets conflated with arguments for God’s goodness. Notice, too, that Law is not giving reasons to think that an evil god exists. On the contrary, it is essential to his argument that such a supposition is absurd.

The claim of the argument is that given the existence of an evil god, it is highly improbable that the goods in the world would exist (Pr (goodsevil god << 0.5)). By the same token, given the existence of God, it is highly improbable that the suffering in the world would exist (Pr (sufferingGod << 0.5)). So just as the goods in the world constitute overwhelming evidence against the existence of an evil god, the suffering in the world constitutes overwhelming evidence against the existence of God.

I suspect that Law thinks that theists will try to deny the symmetry between these two cases. But that would be a mistake. The two situations strike me as symmetrical—I would just say that in neither case would we be justified in thinking that the probability is low. Just as a good Creator/Designer could have good reasons for permitting the suffering in the world, so an evil Creator/Designer could have malicious reasons for allowing the goods in the world, precisely for the reasons Law explains. My initial response, then, still holds: we’re just not in a position to make these kinds of probability judgements with any sort of confidence.

In our debate Law seemed flat-footed in the face of this response. He takes it as just obvious that an evil god would not permit the goods we see in the world—look at the rainbows, look at the children, etc.! But this is no better than the atheist who takes it to be just obvious that the suffering in the world would not be permitted by God—look at the tsunamis, look at the Holocaust, etc. This sort of response is basically an appeal to emotions and fails to grapple with the fact that a Creator/Designer of the world could well have sufficient reasons for permitting what he does.

I was gratified that other theists—like Steve Wykstra, Dan Howard-Snyder, and Mike Rea—who have specialized in the problem of evil share my assessment. Wykstra, for example, wrote:

any being (good or evil) big enough to make the heavens and the earth gives a high conditional probability that we'd regularly be unable to discern that being's ultimate purposes for many events around us. So our actual . . . inability to do so isn't strong evidence that those purposes (or that being) isn't there. . . . Just as the inscrutable evil in the world doesn't give much evidence that there's no totally good creator, so the inscrutable good in the world doesn't give much evidence that there's no totally evil Creator.viii

The point is that once you posit the existence of an evil Creator/Designer of the cosmos, all bets are off.

One final note: I talked earlier about reasons to think that the Creator/Designer of the universe is good. Suppose we concede for the sake of argument that an evil Creator/Designer exists. Since this being is evil, that implies that he fails to discharge his moral obligations. But where do those come from? How can this evil god have duties to perform which he is violating? Who forbids him to do the wrong things that he does? Immediately, we see that such an evil being cannot be supreme: there must be a being who is even higher than this evil god and is the source of the moral obligations which he chooses to flout, a being which is absolute goodness Himself. In other words, if Law’s evil god exists, then God exists.


i Michael Bergmann and Jeffrey E. Brower, “The God of Eth and the God of Earth, “ Think, (Winter 2007), pp. 36-7.

ii Ibid., p. 38.

iii Stephen Law, “The evil-god challenge,” Religious Studies 46 (2010): 365.

iv Stephen Law, Humanism: A Very Short Introduction (Oxford: Oxford University Press, 2011), p. 76.

v Ibid.

vi Ibid.

vii Peter Millican, “The Devil’s Advocate,” Cogito 3 (1989): 193-207.

viii Stephen Wykstra, personal communication, Sept. 8, 2011.