The Best Argument Against a Perfect God Isn’t an Argument for Naturalism
A neglected argument against omni-theism
Note: When I talk about Bayesian arguments, as I do in this post, I often write as if I accepted the uniqueness thesis about epistemic probability—that for any given body of evidence, there is exactly one uniquely correct probability function that assigns a single uniquely correct epistemic probability to every proposition, conditional on that evidence. It’s often convenient to write as if I believed that, especially when I’m discussing the work of someone who does, so I write that way here. But I don’t actually accept the uniqueness thesis.
Let’s call the view that there’s an eternal, nonphysical, omniscient, omnipotent, omnibenevolent God “omni-theism.” Paul Draper offers a very powerful argument against that view in, of all places, the Stanford Encyclopedia of Philosophy.
Draper asks us to consider a rival view: “aesthetic deism.” Aesthetic deism posits a God just like the God of omni-theism but with different motivations:
An omni-theistic God would be morally perfect and so strongly motivated by considerations of the well-being of sentient creatures. An aesthetic deistic God, on the other hand, would prioritize aesthetic goods over moral ones. While such a being would want a beautiful universe, perhaps the best metaphor here is not that of a cosmic artist, but instead that of a cosmic playwright: an author of nature who wants above all to write an interesting story.
Because of its similarity to omni-theism, aesthetic deism should have a prior probability no lower than that of omni-theism. And as Draper points out, aesthetic deism accounts for the relevant data far better than omni-theism does. An aesthetic-deist God would be just as likely as an omni-theist God to create human consciousness and fine-tune the fundamental physical constants to make the universe life-permitting. (You need consciousness and a life-permitting universe for a good story.) And the kinds of suffering we observe would be much less surprising given an aesthetic-deist God than given an omni-theist God. (You can’t have a good story without conflict, and conflict leads to suffering.) So after considering all the relevant evidence, aesthetic deism seems much more probable than omni-theism. Which means omni-theism is probably false.
As Draper puts it:
(1) Aesthetic deism is at least as probable intrinsically as omni-theism.
(2) The total evidence excluding “the data of good and evil” does not favor omni-theism over aesthetic deism.
(3) Given the total evidence excluding the data of good and evil, the data of good and evil strongly favor aesthetic deism over omni-theism.
It follows from (1), (2), and (3) that
(4) Aesthetic deism is many times more probable than omni-theism.
It follows from (4) that
(5) Omni-theism is very probably false.
It follows from (5) that
(6) Atheism (understood here as the denial of omni-theism) is very probably true.
I find this argument totally compelling. The premise asserting that two very similar views have the same prior probability seems more modest and easier to defend than, say, the claim that naturalism has a much higher prior probability than omni-theism. And precisely because aesthetic deism concedes so much to omni-theism, it can handle the familiar evidence for omni-theism, like fine-tuning, without breaking a sweat.
You might object that something about the motivational structure posited by omni-theism makes it somehow “simpler” than aesthetic deism and thereby gives it a higher prior probability. Personally I don’t see any difference in simplicity between a “morally perfect” God and an “aesthetically perfect” God—in any event, not enough of a difference to give omni-theism’s prior probability a big enough boost to overcome aesthetic deism’s vastly better fit with the evidence.
There do seem to be some arguments that are available to omni-theists but not aesthetic deists. For example, there are modal-logic party tricks like Plantinga’s ontological argument. And there’s the familiar claim that we’ve solved metaethics and the only acceptable view is some version of divine command theory. But those arguments don’t move me at all. They definitely don’t seem strong enough to overcome aesthetic deism’s advantage in accommodating evidence of various kinds of suffering.
Suppose you’re skeptical of Draper’s claim that aesthetic deism fits the evidence much better than omni-theism does. Then you could turn his argument into a more modest one.
At the very least, aesthetic deism fits the evidence no worse than omni-theism does. In that case, the posterior probability of aesthetic deism is no less than the posterior probability of omni-theism. But it’s not certain that either omni-theism or aesthetic deism is true. In other words, the posterior probability of the disjunction of omni-theism and aesthetic deism is less than one. If you grant all that, it follows that the posterior probability of omni-theism is less than one-half—omni-theism is probably false.
A related idea is Stephen Law’s evil-God challenge. Law argues that we have just as much reason to believe in a perfectly evil God as we have to believe in a perfectly good God. Omni-theism faces the problem of evil, and evil-God theism faces a parellel problem of good. But the evil-God theist, like the omni-theist, can rely on variants of familiar theodicies to save her hypothesis from the evidence. (For example, surely it would be more satisfying to an evil God if people did evil freely. But free will, alas, leaves open the possibility that people will choose to do good instead.)
As someone who’s sympathetic to both Draper’s argument and Law’s claim of parity between omni-theism and evil-God theism, I find the following argument fairly compelling:
(1) The prior probability of omni-theism is no greater than the prior probability of aesthetic deism.
(2) The prior probability of omni-theism is no greater than the prior probability of evil-God theism.
(3) The total evidence does not favor omni-theism over aesthetic deism.
(4) The total evidence does not favor omni-theism over evil-God theism.
From which it follows:
(5) Omni-theism is no more probable than aesthetic deism.
And:
(6) Omni-theism is no more probable than evil-God theism.
But since the probability of polytheistic views and nontheistic views like naturalism can’t be zero, we have to grant that:
(7) The probability of the disjunction of omni-theism, aesthetic deism, and evil-God theism is less than one.
Hence:
(8) The probability of omni-theism is less than one-third.
These kinds of arguments are often ignored. Most atheists who engage with philosophy of religion are probably naturalists. They don’t want an argument for an alternative conception of God. They want a defense of naturalism. But Draper’s argument from aesthetic deism is as compelling as any argument I’ve ever encountered for the conclusion that omni-theism is false. When we ignore arguments like Draper’s, we let omni-theism off the hook much too easily.