Highlights From "The Most Plausible Explanation Of Why God Allows Evil"

Many such criticisms

I recently wrote an article titled The Most Plausible Explanation Of Why God Allows Evil. In it, as particularly clever readers will be able to guess, I presented what I thought was the most plausible explanation of why God allows evil.

The short version is that God will put us in a diverse range of worlds—including this one—because a diverse range of worlds is conducive to improving the quality of our postmortem decisions, achieving various unknown goods, soul-building, connection-building, enhancing our life’s diversity, and making for good stories. I claimed it’s not implausible that God would put us in a diverse range of worlds not all of which resemble paradise, and that if this is right, it’s not implausible some would be like this one. I additionally claimed that even if there’s a low probability of that, because if this is right theism predicts our spending more time in worlds like this, anthropic considerations based on this will favor theism over naturalism.

At the time of writing this, there are 72 comments—the vast majority critical. Most of my readers are atheists, so this is not surprising. Seeing as so many people think my theodicy is nuts, I thought—given that I had so many commenters weighing in—I’d explain why you are all wrong.

Alex C says:

"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."

But why would god create souls that require suffering in order to undergo soul-building? In many real-world cases, suffering causes a degradation of character, not an improvement of it.

No offense, BB, but I find your theism-related posts to be baroque, hard to follow, and unpersuasive. But I don't have much background in philosophy, so maybe the fault lies with my ignorance.

I think that the soul-building element is one of the least persuasive elements of the theodicy. Still, I have a few main comments about Alex’s criticism. First, it might be that as free creatures, God can’t just make us with any old character. What we do is up to us, and if God made us constantly disposed to do the right thing, then that would eliminate the good of us doing the right thing even when it’s psychologically difficult. It would be like genetically engineering your child so that they never even consider doing the wrong thing. Perhaps then, given our freedom, experiences like those we have on Earth are needed for the best kind of long term character—one that freely chooses the good without compulsion.

Second, lots of kinds of virtue require us being in positions of hardship. We could not be brave without something to fear. We could not suffer for others without suffering. To exercise forgiveness, there must actually be something to forgive.

Third, it might be that forming desirable character traits freely—rather than having them directly implanted by God—is itself an intrinsic good. Coming to care about others freely—through seeing their suffering—might be more valuable than having God directly make you compassionate. Some virtues are best if formed freely through experiences rather than directly implanted.

Vikram V comments:

This is a repackaged version of generic explanations for evil. It shouldn’t convince anyone who wasn't already convinced.

I don’t think this is right, though of course it draws on explanations already in the problem of evil literature. Other theodicies have at least two general problems that mine avoids. The first is that even if they explain why there’s evil, they don’t explain why many of the evils function precisely as we’d expect them to on atheism. For instance, maybe some evils are good for soul-building, but what are the odds that the optimal set of experiences for soul-building would involve us having the precise set of faculties that evolution has delivered us? Theism makes little sense of why all the goods and evils in our world seem to follow from basic natural processes.

My theodicy can explain that. God places us in a bunch of different worlds. This one happens to be one that is governed either entirely or almost entirely by predictable natural laws. God doesn’t violate those. If he places us in tons of worlds, probably some will be like that.

Second, it’s hard to believe that this world is the best for many of the goods theists emphasize. Even if there are lots of ways this world is good for soul-building, what are the odds that this one is the best for soul-building? Wouldn’t being in the sorts of dramatic worlds that appear in fiction—e.g. Harry Potter—be better for exercising soul- and connection-building. My theodicy again answers this—a world like ours is not the best, but one of the best collection of worlds.

Third, other theodicies have a serious moral problem. If it’s really true that, say, the holocaust produced more good than harm because it caused soul-building, then it would seem we shouldn’t try to stop the holocaust. My theodicy avoids this worry by saying that the evils aren’t themselves desirable but instead are tragic side-effects of other things that deliver great goods.

NasalJack says:

Today is October 1st, 2024. What are the chances that in a timeline extending infinitely forward and backward that today would be that specific date? Infinitesimal. Instead imagine that we’re all stuck in a Groundhog Day loop but have our memories wiped every morning. In that scenario it is always October 1st 2024, which as a theory has infinitely better explanatory power for why that is today’s date. Therefore statistically it is infinitely more likely to be true.

That feels like the crux of a lot of these arguments. Something something infinite probabilities, therefore God.

Sarcastic quip: well maybe if you paid more attention to the arguments you would see that they are not, in fact, something something infinite probabilities therefore God but have premises and a conclusion that follows from the premise!

As for the substantive remark—it’s true that the theory that everyday is October 1, 2024 on loop makes it likelier that today would be October 1, 2024. But what the view makes up for in explanation, it loses in priors! The odds that every day would be October 1, 2024 = the odds that everyday would be some specific day times the odds that if everyday is some day, it would be October 1 2024.

Assume you think that if a day loops, it’s equally likely to be any of the days. Well, if the days don't loop but are experienced successively, the odds it would be October 1 2024 today would be 1 divided the number of days. This means that necessarily if the prior probability in days not looping is higher than the prior in days looping, it’s unlikely that today would loop. This is because:

Odds it would loop October 1, 2024 = odds it would loop/number of days (because its equally likely that it would loop any of the days).

Odds it would be October 1, 2024 if it doesn’t loop = odds it would not loop/number of days (because if it doesn’t loop it’s equally likely to be any of the days).

Ape in the coat says:

Anyone who wants to use the "what if suffering is metapysically necessary for goodness" argument needs to at the least engage with the fact that apparently even our world allows the existence of perfectly normal humans who simply do not suffer.

Citation from

Profile: The Far Out Initiative

Okay but that wasn’t my argument!

Dylan says:

Your argument for theism in "The Best Argument for God" is equally valid if God is evil-maxxing (i.e., "God in his perfection would create all possible people who he could give a bad life to").

Yep, but an evil God is implausible.

Shane says:

What does the atheist need to prove to satisfy you? It seems you've pretty much set the conditions so that no empirical evil could be evidence against God. Consider your analogy about watching someone play a big complicated strategy board game we don't know the all the rules to.

You: "I have prior reasoning that leads me to conclude that it's extremely likely that this guy is a truly God-Tier Magnus Stockfish player"

Me: "I mean, maybe, but it looks like this dude just hung a few pretty important pieces. This seems to call your theory into question"

You: "Not at all, we don't really know all the rules, and for all we know these supposed 'mistakes' could be brilliantly planned sacrifices"

Me: "It's true that we can't be certain one way or the other, but doesn't it seem a bit un-Stock-fishy that this guy is playing more or less exactly like we'd expect if he were just blindly and randomly moving pieces?"

You: "Not at all. In fact, in general, sacrifices in strategic games tend to lead to more beautiful play in the end, just like Paul Morphy games. Moreover, based on some mysterious heads/tails reasoning, if this guy is indeed a Perfect player, we should even expect to find ourselves in situations where we have this kind of dubious impression of his level of play."

As far as I could tell, your God's goodness is utterly immune to any observable evil

If we were all being grotesquely tortured all the time, then probably it would be reasonable to believe in some kind of evil God. There’s some large amount of evil at which point the thing that best explains it is an evil God—but a mere 80 years of bad stuff that’s 0% of total possible badness spread across 0% of our life isn’t that!

Is theism falsifiable? It’s true that in almost every world an agent could find themselves where they can consciously reflect, I think they should be a theist. But that’s not a fault. I think theism is mostly favored over atheism by predicting the existence of agents. So it’s no more objectionable to say almost every possible agent should be a theist than it is to say almost every possible agent should think they came to exist—though it’s possible that no agents would exist, after finding oneself in a world where they exist, one knows that agents would exist.

Finally, the fact that theism can’t be falsified by a lot of evil isn’t much of a problem. Neither can it be falsified by a lot of paperclips, cheese, and so on. Something need not not be able to be falsified by any old thing.

Both sides brigade says:

I 100% agree - I wrote a piece about this a while ago that uses exactly this metaphor:

The Other Half of the Problem of Evil

"Imagine you’re in some competition you don’t know much about, and your partner is a supposedly high-powered AI. Unfortunately, the AI might be malfunctioning in a way that makes it behave completely randomly. Near the end of the match, you start to think about what moves the AI has made. Some have been great - it made all the obvious moves that even you would have made, and a bunch of clever ones you would have missed. But quite a few of the AI’s moves didn’t get you any points at all, and they sometimes set your opponent up to take points for themselves. What would you think about all this?

Here, I can understand saying something like, “Well, the AI seemed to know what was going on with some of these moves, so maybe I just don’t get the rest of them.” Whether that sort of reasoning is actually justified is up for debate, but at least it makes sense; your knowledge of the game might just be too shaky for you to grasp the optimum strategy every single time. Personally, I’d still be inclined to think all the productive moves were just good luck - it seems much more likely that malfunctioning AI would occasionally do well than that a perfect one would occasionally do poorly - but as long as the AI sometimes shows a real aptitude for the game, it might be reasonable to throw up your hands and trust it.

Now imagine instead that, near the end of the match, you realize the AI hasn’t made a single move you can confidently say was anything more than random. Every decision it makes seems completely pointless or even intentionally destructive. How would that change your evaluation? Of course, no matter how many moves appear to go wrong, it’s still conceivable that the AI is working perfectly. Maybe your knowledge of the game is so deficient that you’ve failed to appreciate the strategy behind every move. Or maybe, by losing, the AI can get an advantage in a second round you don’t even know is coming. Who knows? All of those are metaphysically live options, but it’s hard to see how it could plausibly play a role in your actual evaluation of the AI’s status.

In real life, you would obviously decide the thing was malfunctioning, and you would be absolutely justified in doing so. Could you be wrong? Sure, it’s possible. But at the end of the day, you still need to work with the evidence you have, and none of that evidence even gets you started down the path of confidence. And it wouldn’t help for the designers to simply point out that the AI’s skill is so much greater than your own. After all, it’s not that you’re failing to trust the skill of the AI, once you’ve accepted it exists. It’s that you’re failing to trust any such skill exists in the first place!"

This is an interesting point but I don’t think it applies much to the theodicy. If God places us in a bunch of worlds—including ones where the laws don’t optimize for value and he doesn’t intervene—then probably some would be like this one. Rather than thinking of this world as like a bunch of independent moves, you should think of it as like some specific strategic aim achieved by a bunch of moves (e.g. capturing some specific territory in some game). But if you’re in the dark about how to play a game, what the rules are, and what the pieces are, and you see someone strategically take over some territory, then even if you don’t see a reason for taking over that territory, that doesn’t give you great evidence that the AI is malfunctioning. There might be unknown reasons for such a game.

In short, because the world has a few broad features that result in all its evils, it makes more sense to think of that as analogous to one move rather than a bunch of different moves. In addition, it simply isn’t true that there aren’t any evils that seem good overall—there are lots of theodicies that tell (very partial) stories of how great good comes from evil.

Metachirality says:

I don't think, say, being a factory farmed chicken is a particularly valuable experience on net. However, you can rectify this by saying although such experiences don't actually happen, valuable experiences may include experiences of worlds where it appears as if they happened. That is, if you see someone experiencing something very bad, they aren't actually experiencing it and it only just looks that way. However, this proposition is not exactly super attractive.

I agree that it’s not very plausible that being a factory-farmed chicken is valuable in the long term. This is why I think theists should think, rather than that particular horrors of our world (factory farming etc) are actually good, that instead they are tragic consequences of other desirable features of our world—in this case, a world with predictable laws and free agents where God doesn’t intervene.

Analogy: suppose your child is in a relationship with someone that you don’t think is good for them. In this case, you don’t permit it because you think the relationship is very good, but because some deeper thing (their freedom to enter relationships without your permission) is very good, and this produces bad relationships sometimes as an unfortunate side effect.

Supposing that God has a policy of not intervening makes best sense of the world’s evils, and so I don’t think it’s plausible that animals are simply not conscious during the times of their greatest agony. Theism poorly explains the world unless you posit that God has a near universal policy of not intervening in this world.

Okay, so now that I’ve responded to most of the significant comments (I didn’t respond to some clever ones because many of them had points that were also made by others) I trust that atheism has been DEBUNKED and you’ll all cease your fornication and begin going to Church!