The Most Plausible Explanation Of Why God Allows Evil

This might even make present evil a point in favor of theism

For a while, I’ve thought that evil is the best argument against theism—indeed, the only good argument against theism. While theists can sort of desperately flail in the direction of a solution, saying things that make it not enough to single-handedly raze their view, it’s still an enormous problem that renders theism much less plausible. But now I think the situation is much less bad than that; if the things I say here are correct, then evil—and its associated companion the divine hiddenness argument—aren’t very strong evidence against theism. More surprisingly, the fact that we’re presently experiencing evil of the sort present in our world may actually be a point in favor of theism—and not just for the cheap reasons commonly given by apologists that just smuggle in obviously unrelated theistic arguments (e.g. to have evil you need agents which needs fine-tuning, but that’s improbable if there’s no God).

Let’s see the three assumptions needed for the theodicy:

1. If there’s a God, he’d put us in a diverse and unpredictable array of worlds. So, for instance, maybe we’d spend some time in a Narnia-esque world, some time in a world where our experience just resembles a weird LSD-like trip (note, these are just possible examples), and some worlds that are extremely alien. The key point is that the worlds we’ll in won’t all resemble heaven—there will be a wide and varied mix of them, many of which seem quite flawed.

2. If God places us in a diverse and unpredictable array of worlds, probably some of them would be like this one.

3. If God would at some point place us in a world like this, then us being in a world like this isn’t evidence against the existence of God.

The basic idea is that God would be likely to place us in a bunch of different worlds over the course of our infinitely long cosmic tenure. But if he does that, probably some worlds will be like this one, so being in a world like this one isn’t evidence against theism. I think 2) is pretty straightforward and plausible—if God’s placing us in tons of different worlds, probably some would be like this one.

What do I mean by a world like ours? I think our world has two surprising features that entail its many evils. The first is non-axiarchic laws, meaning laws aren’t directly geared towards producing value. The law of gravity, for example, doesn’t only attract things when doing so is good. The second is that the laws don’t get suspended—God doesn’t meddle (except, perhaps, in very rare cases) but instead the laws remain consistent and uniform. Together those features entail a world like ours, governed by predictable laws, that produces lots of bad stuff. So then 2) claims that if we’re placed in a bunch of different worlds—potentially infinite different ones in our infinitely long life—maybe some of them would have non-axiarchic laws that aren’t suspended.

I think that 2) is a very modest premise (and 3 is definitely right, as I’ll argue later) so the argument mostly comes down to 1)—that God would put us in a diverse and unpredictable cocktail of worlds, many of which aren’t the kinds of worlds that would intuitively seem best.

Note that for the problem of evil to be very strong evidence against theism, it must be that 1) is almost certainly false. Even if 1) isn’t supremely obvious, as long as there’s some decent chance that it’s right, the problem of evil can’t have too much force, because for it to work, something must be false that might very well be true. For an argument to be extremely convincing, it isn’t enough for it not to be clear if the responses to it work—it must be very clear that they don’t work.

With that throat clearing out of the way, I think we have six reasons to think that 1) is at the very least not obviously wrong. The first two are, I think, very good arguments, and the rest are okay.

1. The afterlife contains a potentially infinite number of goods, many of which we’re totally in the dark about. Thus, we shouldn’t be confident that the optimal route towards producing those goods doesn’t involve a diverse array of lives. Analogy: suppose you only see the first two moves in a game that you don’t know all the rules to, and you think there might be thousands of pieces that you don’t know about. Are you in a position to be very confident that the optimal strategy in the game doesn’t involve a diverse array of moves? Obviously not! Yet how is that different from a person with our limited knowledge of afterlife goods confidently declaring that the best way of achieving them can’t involve being placed in a diverse array of worlds. If this is right, and it’s decently likely that God’s plan for us involves placing us in a massive and unpredictable matrix of worlds, it’s hard to be confident that none of them will be like this one.

2. The future decisions that we make are affected by what our lives have been like. What things you do depends on random stuff like what you had for breakfast. This isn’t just a contingent fact about our psychology, but an inevitable fact about anyone who has a psychology—having a mind of any valuable sort involves one’s actions being affected by their circumstances. But then that means that probably the sequence of experiences that will result in us taking the best actions in the long run are highly unpredictable—if we’re making significant decisions in heaven, the best heavenly decisions might be caused by a rather unpredictable mix of pre-heavenly lives. Analogy: suppose that you’re going to train for 50,000 years and then play 20,000 chess games. Probably the things that would make you play best in the 20,000 games are kind of random. But then God would orchestrate whichever situations in this life randomly make us make the best decisions in heaven. Even if you think of free will as just like a chancy process, chancy processes depend on their environment, and the circumstances that would lead a chancy process to maximally produce some outcome would be fairly random. (This response is defended here in a very clever paper by Kenny Boyce).

3. Probably a rich and diverse array of experiences is maximally conducive to soul building. One’s character is more enriched by tons of different experiences than just one kind—and probably some of those would involve lots of suffering.

4. A diverse array of experiences is probably also maximally conducive to connection-building—to strengthening our relationships with other people. Your relationships with other people will be stronger if you endure lots of different experiences with them, many of those involving hardships. Because those valuable connections last forever, they’re worth any finite amount of suffering.

5. Telling stories is a good thing—it’s a primary way that people bond. It’s not implausible that we’ll tell stories in heaven. But stories are more interesting if they involve a richer and more diverse array of experiences (I’ve defended this from objections here). So then if a diverse world makes us have more cool stories to tell, assuming we go on telling stories forever—thus producing infinite benefits—a diverse world is infinitely good.

6. It may be that a diverse life is an intrinsic good. If given the option, I’d choose to spend an hour as a fish, for instance, and such an action would seem valuable. Lives consisting of a diverse range of pleasures and experiences seem better than other lives that are more uniform. But then this means God would likely put us in a diverse range of worlds.

It may seem have to believe that these things could be worth the horrors of our world. But this world is just the blink of an eye—an infinitely small sliver of an infinitely large existence. Just as it’s worth undergoing a second of pain for years of benefit, this life is a tiny drop in a vast ocean.

These are just some of the reasons. Probably most of the reasons appealed to in most theodicies could be made as defenses of 1). So then the only premise in need of defense is 3)—that if God places us in a diverse array of worlds, many of them like this one, then being in a world with evil like this one isn’t good evidence against theism.

You might reject this on the following grounds—perhaps though theism makes it likelier that we’d be in a world like this one at some point, because we’d mostly be in worlds not like this one, it’s unlikely we’d be in a world like this one now. This might be analogous to the following case:

A fair coin is flipped. If it comes up heads, you wake up once in a green room. If it comes up tails, you wake up once in a green room. Then, you wake up again in a green room, with no memories of the first day. Then you wake up ten times in red rooms with no memories of the previous days. After waking up in a green room, what should be your credence in heads?

Note: this scenario is strictly analogous to the previous one. All the wakings are the same—the only difference is that in the first scenario the coin was flipped before the experiment and in this one it’s flipped after the first day. But surely that shouldn’t matter?

There are three main answers. One of them is 1/3—because there are twice as many green rooms if it comes up tails as if it comes up heads, you should think tails is twice as likely as heads. The number of red rooms is irrelevant. This answer is given by a view of anthropics called the self-indication assumption (which is obviously true) and means that 3) is correct—because theism means you’ll spend a lot of time in worlds like this one, it doesn’t matter that it’s a small share of worlds. In fact, this means that if 1 and 2 are right, being in a world like this one massively favors theism, because you’ll spend way more time in worlds like this if God places you in tons of diverse worlds, than you will under atheism where you probably only live once.

A second answer is 1/2—because both theories predict you’ll wake up in a red room, they both explain the data. This is the answer given by a view called compartmentalized conditionalization. This one is consistent with 3) as well, and means that 1) and 2) favor theism, because if you spend lots of time in a diverse array of worlds, it's likelier that one of them would be exactly like this one, than if you’re just in one world. Our world is very specific, so on theism, where you spend time in lots of worlds, it’s likelier that you’d, for instance, ever spend time in a world with our exact laws of physics.

A third answer is 6/7. The thought here would be that on heads you’re guaranteed to be in a red room, while the odds you’d be in a red room is only 1/6 on tails. Heads would thus be 6 times likelier than tails, and be 6/7 probability. This view would reject 3—if you spend a lot of time in a world like this one, but it’s a small share of your total existence, this view would say you being in a world like this is very surprising. What matters, on such a view, isn’t the total amount of time you spend in a world like this one but the percent of your existence spent in a world like this one.

Note that for this to be a convincing response to the theodicy, you have to be really confident that the third answer is correct. If you’re uncertain, then you’re uncertain that the objection works, and it can’t be massive evidence against theism. Thus, for 3) to fail, you must be really confident in the 6/7 answer.

But I don’t think you should be really confident in the 6/7 answer! In fact, I think we have overwhelming reason to reject it. And remember, if the 6/7 answer isn’t correct, then as long as 1) and 2) are correct, present evil majorly favors theism, because theism conjoined with 1) and 2) means we’ll spend more time in worlds like this one. If you think that 1/3 is probably the right answer, then even if there’s, say, a 1% chance of 1) and 2), you’ll get overwhelming evidence, from the fact that you’re in a world currently experiencing evil, that theism is right.

Why reject the 6/7 answer? Because the 6/7 answer assumes the self-sampling assumption—that you should treat the present moment as randomly drawn from the moments in your life (this is actually a more specific version of the self-sampling assumption, rather than the general version, so it doesn’t just assume SSA). But the self-sampling assumption is false, as I’ve argued in a bunch of other places.

I don’t want to retread too much old ground, but let me give one brief argument for the 1/3 answer. Imagine the following modification to the scenario:

You wake up once in a green room. A fair coin is flipped. If it comes up heads, that’s the end! If it comes up tails, you wake up again in a green room, with no memories of the first day. Then you wake up ten times in red rooms with no memories of the previous days. Suppose you wake up in a green room. Then you are told that it’s the first day currently. What odds should you give to the coin that will be flipped later coming up heads?

I’d answer: 1/2. Upon waking up in a green room, you should think that tails is twice as likely as heads. However, it being the first day currently is twice as likely if the coin comes up heads as if it comes up tails. After updating on it being the first day, you end up thinking that there’s a 50% chance that the coin will come up heads.

But people who give the 6/7 answer in the earlier problem would have to answer something very different. Upon waking up in a green room, they think there’s a 6/7 chance that the coin, when flipped, comes up heads. If the coin comes up heads then it must be the first day, while if the coin, when flipped, comes up tails, it could be the first or second day—each have 50% probability. So, when learning that it’s currently the first day, you get evidence that’s twice as strongly predicted on heads as on tails. This means that you should now think there’s a 12/13 chance that a fair coin that hasn’t been flipped yet will come up heads. This is super weird and gives us reason to reject such a view (if you found this point confusing, I expound on it in more detail here and here).

For this reason, when assessing theism vs atheism’s ability to accommodate evil, the relevant metric is which one means we’ll spend more total time in worlds with evil. But if we’re immortal and spend infinite time in worlds like this this one, containing laws that aren’t violated that aren’t directly responsive to value considerations, then theism predicts much more time spent in worlds like this one. Present evil, then, just might be an argument for theism.