No Free Will Theodicy is Even Remotely Plausible

Or: Why all the world's most terrible evils aren't properly justified by... all the world's most terrible evils

Pictured: a different kind of incoherent libertarian

Over the past several thousand years, theists have come up with a hefty set of theories that aim to reconcile God’s goodness with the huge amount of horrendous moral evil we encounter in the world. One of the most well-known strategies for doing so centers on the notion of libertarian free will; the argument, put simply, is that God has to allow human beings to choose evil, because a world without “true” free will is even more dismal than one overflowing with suffering and pain. We can call this general approach a free will theodicy. Nowadays, free will theodicies have largely fallen out of favor in the analytic philosophy world, and you’re more likely to see theodicies centered on soul-building or universal reconciliation getting academic attention. But because these ideas are still very influential at a popular level, especially in many religious communities where they’re used as a first line of defense against any and all doubts, I thought it would be valuable to lay out my position: Not only that free will theodicies ultimately fail, but that they are obviously false and should not be considered a reasonable response to the problem of evil.

Let’s start with a bang: The first reason libertarian free will can’t justify the evils we see in our world is that libertarian free will doesn’t exist, and goods that don’t exist can’t justify anything. Of course, people who defend the free will theodicy aren’t going to agree with me on this. But they aren’t going to agree with anything else here either, so I might as well say it: For my money, libertarianism is the single most implausible and poorly-defended “mainstream” position in analytic philosophy today. I’m no expert on the subject, but I’ve read all the heavy hitters libertarians are likely to recommend, and none of their work has struck me as much more than vague gesturing in defense of a single doctrinaire intuition. (This is also the case for libertarians in the political sense, but I digress). Going forward here, I’ll set this flat rejection aside and engage with the free will theodicy on its own terms - but still, it’s worth remembering the entire concept can really only ever get off the ground if you accept a view that (according to the latest PhilPapers survey) more than 80% of professional philosophers reject.

I’m also going to set aside another issue, which is that giving people the ability to choose evil doesn’t necessarily require giving them the ability to actually carry out evil. Almost all free will theodicies treat the two as interchangeable, even though it would be trivially easy for an omnipotent being to make note of our volitional intentions towards evil acts while also intervening in every case to make sure those evil acts we’ve chosen don’t succeed. Now, theists tend to argue this would have the unacceptable consequence of making all evils hypothetical, in the sense that humans would quickly learn it wasn’t actually possible to behave too badly. But the claim that this consequence would undercut the significance of our free will choices is just as implausible as the same claim made about the impact of human institutions like police or regulatory bodies. If true freedom requires the ability to actually succeed in harming others, then human beings would be freer in failed states and war zones than in modern liberal democracies. But this is absurd! No one believes that human societies ought to make evil easier, just so the option is tempting enough to make our refusal of it meaningful. That God is somehow bound by this same perverse logic on a cosmic scale is a careless assumption, one that most free will theodicies uncritically take for granted.

But like I said, ignore all that. Let’s assume libertarian free will really does exist in our world right now, and that it requires blanket non-intervention to have any moral value. We’ll call our world, stipulated to have these assumed features, W1. Then we can imagine W2, a world where infinitely many beings exist in a state of unshakeable moral perfection that leads them to act in ways that promote universal flourishing and the total absence of any and all suffering, but whose decisions are deterministically fixed by the sum total of their motivation set with no possibility for alternate choices. My argument here will be that it is absolutely insane to think W1 is preferable to W2 - that anyone in W2 would be wildly irrational to endorse the transformation of their world into W1, and that anyone in W1 would be wildly irrational to resist the transformation of their world into W2. If this is the case, then it follows that God would be equally irrational if he were to actualize W1 over W2 when creating. And since God (hopefully) isn’t irrational, it can’t be the case that he did actualize W1 over W2. Therefore, not only is the existence of libertarian free will not an adequate justification for evil, but the existence of libertarian free will would actually be a strong argument against the existence of God, if it were in fact the case that we had it.

To see this, we can imagine that a slice of W2 exists right now as an isolated Shangri-La hidden away in the Himalayas. The denizens of this kingdom, because of some neurological malfunction or spiritual quirk, have no libertarian free will to speak of and instead always do exactly what their deterministic “agential algorithm” recommends. But luckily, the motivational set this agential algorithm operates on is perfect in every case; everyone, from the peasants to the king himself, is entirely virtuous and seeks only the greatest possible good for themselves and others. Now imagine that an intrepid explorer from our world were to stumble upon Shangri-La and offer its people the chance to have their libertarian free will restored, at the cost of inevitably introducing tremendous evil and suffering. Would they be rational to accept the offer? The drawbacks would be obvious, of course: War, murder, rape, poverty, and all the other nightmares we’re saddled with in our own world. But a more interesting question would be how, exactly, the denizens of Shangri-La would benefit, and whether that benefit could possibly be great enough to outweigh the sum total of every possible evil.

My answer here is not only that the benefits of introducing libertarian free would fail to outweigh the costs - although I do think that’s obvious - but that there are no benefits to introducing libertarian free will at all. This is because, in practical terms, the only new power granted to the denizens of Shangri-La by libertarian free will would be the power to occasionally make a morally compromised choice that is not, all told, what they desire most. Remember that, prior to the introduction of libertarian free will, a denizen of Shangri-La always does what their deterministic “agential algorithm” identifies as most in line with their motivational set, and that motivational set is always oriented towards the greatest possible good. Therefore, if libertarian free will introduces alternate possibilities for action, those alternate possibilities are definitionally neither morally optimal nor maximally desired; if they were, they would just be the choices denizens of Shangri-La already make, and libertarian free will would be completely inert. So, to the degree that libertarian free will would result in any observable difference whatsoever in the world of Shangri-La, it could only ever do so by making things worse - not only in a moral sense, but also in the sense that it could only degrade the reliability of their decision-making process, severing the connection between what they all told most desire and what they actually choose to do.

This is, ultimately, the point I want to make: What free will theodicies frame as the overriding good of libertarian free will just is the evil it supposedly justifies. It isn't as though the denizens of Shangri-la receive, on one hand, an expanded set of alternate possibilities for action, while on the other hand receiving some set of newly accessible evils. If that were the case, then we could go along with the free will theodicy and reasonably ask whether the former was valuable enough to justify the latter (although finding any set of new actions valuable enough to outweigh every possible evil is still an absurdly tall order). But in this case, the two sets are the same! The previously inaccessible actions that libertarian free will makes available are entirely comprised of just those actions that bring about various evils; if you were to literally tally up the pros and cons of introducing libertarian free will by hand, the result would just be various forms of “Now people are able to do [X evil act]” entered into both columns. Clearly, this is incoherent, and it undermines even the bare possibility of a free will theodicy succeeding.

We can also imagine the opposite scenario, in which a wise man from Shangri-La ventures out to make contact with our world. If this wise man brought with him some arcane ritual or magic potion that could transform the denizens of our world into his - that is, make it so that we deterministically choose to act in accordance with a morally perfect set of desires - would we be right to take him up on his offer? I’ll say again that the answer is obviously yes. Consider Frank, a relatively decent denizen of W1 who desires to act morally but, like all of us, sometimes gives in to anger, selfishness, and cowardice. What would actually happen to Frank if he were to be transformed in this way? The only real answer is that Frank would become a better person. He would gain a more wholesome set of values and desires, and he would be guaranteed to act on them consistently. What he would lose, on the other hand, is nothing more than the ability to sometimes behave poorly in ways that he’d all told rather not. But how could this possibly be any kind of loss at all, if it’s at the same time precisely the transformation’s benefit? We’re stuck in the same incoherent situation once again, but this time in reverse.

Whenever I make this point, I tend to get one of two responses. The first is to fall back on the classic libertarian assertion that what matters here is not the specific set of alternate possibilities available, but just the general fact that some alternatives exist at all. This is what the libertarian thinks is necessary for moral responsibility in the first place; if the denizens of Shangri-La don’t have the power to do otherwise, so the story goes, then they aren’t even the sorts of beings who can be considered in moral terms at all. But without going too far afield into a general critique of libertarianism, all I can say is that I find this claim totally incomprehensible. It just can’t possibly be the case that whether we’re responsible for doing what we desire most hinges entirely on whether we could have also done things we desire less. This is especially true when you consider that, for every deterministic choice a denizen of Shangri-La makes, there’s a possible world in which that agent uses libertarian free will to make the exact same choice on the basis of the exact same motivational set. Arguing that only one of these qualitatively identical decisions would come imbued with moral weight, entirely on the basis of whether the agent could have acted in a way less grounded in the sum total of their deeply-held desires and values, is just absurd. (For a more detailed critique of the idea that moral responsibility requires alternate possibilities for action, as well as a great analysis of why the misconception develops in the first place, see this recent Substack post from Sebastian Montesinos.)

The second response, which I think is more interesting, goes something like this:

Well sure, you might be willing to give up your libertarian free will if it guaranteed that you’d always do the right thing. But there are plenty of other people who wouldn’t - most obviously, those who have no current interest in behaving morally, or those who believe that libertarian free will is immensely valuable - and it would be unjust for God to force them into righteousness against their will, even if they’re wrong. You wouldn’t sneak someone pills that made them nicer without them knowing, so why is God justified in bringing about an even more dramatic change in our motivational sets without our consent?

I have two things to say here. The first is that consent is clearly not so valuable that it justifies permitting endless unchecked evil, especially if you’re the kind of theist who believes in eternal conscious torment as a punishment for sins. I won’t rehash everything point I made in the second section of my previous post on Christian justifications for hell, but suffice to say, a loving parent should have no problem whatsoever violating or altering their child’s will so as to spare them (or others) from the worst possible outcomes. And even if you disagree with me on that point, this sort of objection has a much bigger problem: Consent doesn’t even enter the picture when we’re talking about what sorts of beings God will decide to create in the first place, because those people don’t exist to grant or withhold consent at the time the decision is made. So even if you believe that it would be unacceptable to take someone who currently exists and alter their motivational set against their will, our current desire for autonomy in this world can’t retroactively constrain God’s decisions while creating. After all, had God made me with perfectly good desires instead, I wouldn’t currently consent to being made with imperfect ones either!

To sum it up: Free will theodicies start off on a bad foot, seeing as they require taking for granted a seriously unpopular (and, in my mind, extremely implausible) conception of free will. So if you believe libertarianism is false, then you should obviously reject any free will theodicy that requires it. But even if the libertarian freedom these theodicies rely on did exist, it couldn’t possibly justify the evils we see in our world, because libertarian free will in this case just is those evils; it would be like taking a pill that does nothing beyond make you nauseous, and then arguing that the chance to take the pill justifies the nausea. So if you believe libertarianism is true, then you should be confident that a perfect God doesn’t exist, because no perfect God would allow for something that only makes the world worse. And that means, either way, that the free will theodicy doesn’t work.