Last week, I put out a short piece on miracles and the problem of evil that ended up getting a lot more attention than I’d expected. One of the many thoughtful replies it received was from Noah McKay, a recent Catholic convert who’s been doing very interesting work on novel arguments for the faith. In it, he made the following point:

God is an infinitely wise and intelligent being who *almost certainly* acts on considerations to which we’re not privy. And it’s not that surprising, in light of this, that much of what God allows isn’t justified by considerations we know about.

This is, of course, an extremely common theistic response to the problem of evil — in fact, it’s arguably the default response in the modern world today. In many apologetics spaces, it’s taken for granted that pointing to God’s infinite wisdom is enough, by itself, to at least seriously weaken the problem of evil, if not undercut it entirely. So in this piece, I want to take a bit of time to explain why I think that approach is just totally misguided, and why the most sophisticated forms of the argument have essentially nothing to do with God’s wisdom at all.

Right away, it’s important to be clear about what the “problem” in the problem of evil really is, and how it relates (if at all) to God’s reasons for acting. Of course it’s true that, if God exists, he’s going to consider all sorts of things that we don’t understand when deciding what to do.1 But that’s also the case when it comes to anyone with even a modest level of field-specific expertise. My accountant also makes decisions on the basis of things I haven’t got a clue about, and yet there’s no “problem of evil” for accountants — or, at least not for this reason.

Why? Because even if I don’t grasp the considerations that lead my accountant to act the way he does, I can still make perfect sense of the goods he brings about. So long as I find myself with manageable debt and no outstanding warrants for tax evasion, I’ll be satisfied with how things are going. And that’s true even if I’m totally in the dark as to why, exactly, he handles my money the way he does. If the human experience of horrendous suffering was anything like this — with a steady supply of identifiable goods accruing to us, even as all the little details go over our heads — then the problem of evil would be at best a morbid curiosity.

But unfortunately for us, that’s just not the way that horrendous suffering works. Instead, we almost never come to comprehend the goods that (supposedly) accrue to us as a result of life-destroying misery. There are, of course, occasionally times where some gruesome and destructive event happens to work out for the best. But those rare examples are easily chalked up to coincidence. What we don’t see, on the other hand, is anything like the clear, consistent relationship between human suffering and human good that we’d expect to easily grasp if every single instance of evil was guaranteed to leave us better off than before.2

And so, at some point, theists have to confront the real problem: Not that God sometimes acts on the basis of considerations we can’t understand, but that the goods God secures through those acts are so consistently obscure to us, the people being acted upon. What theists are stuck defending, in other words, isn’t just the banal claim that God knows quite a bit more than we do about what it takes to secure valuable goods. What needs a defense is the much less obvious claim that human beings should never expect to even comprehend those goods in the first place.

So how does God’s infinite wisdom help secure that conclusion?3 Theists might want to fall back on a general epistemic principle. Here’s one option:

The likelihood that one agent will comprehend the goods being secured by another agent’s actions is inversely proportional to the difference between the wisdom of the two agents (with the acting agent’s wisdom being greater than the wisdom of the agent being acted upon).

But a premise like this is just obviously false. If I’m standing on the sidewalk one day, and Terence Tao pulls me from the path of a runaway bus, his mathematical genius will have nothing to do with whether or not I’m able to make sense of his actions. In fact, God himself could be the one to pull me to safety, and I’d still have no more trouble figuring out why than if a child had done it instead. I just need to see the bus careen past, and it’ll all make sense regardless. So even massive gaps in wisdom, by themselves, aren’t enough to show that goods obtained by one agent will likely be incomprehensible to another.

With that said, I get that this very general principle isn’t really what the theist is arguing. Instead, it’s probably something more like:

The likelihood that one agent will comprehend the goods being secured by another agent’s actions is inversely proportional to the difference between the relevant wisdom of the two agents (with the acting agent’s relevant wisdom being greater than the relevant wisdom of the agent being acted upon).

This screens out cases where an agent’s extensive wisdom has nothing to do with the situation at hand, which is certainly an improvement. But as we just saw with my accountant, this premise still doesn’t work. In general, it’s trivially easy to imagine scenarios where two agents differ wildly in even their relevant wisdom, and yet the duller of the two agents still comes to understand whatever goods are being secured by the wiser one’s actions. All that really matters is whether the duller agent reaches some sufficient level of understanding themselves — and what’s needed to reach that level of understanding has nothing to do with any additional understanding that the acting agent may or may not have.

To put it another way: For any action you could ever imagine, there’s going to be some set of facts that, if known, are sufficient for comprehending whatever goods the action will secure; if we have access to those facts, then we’ll comprehend them, and if we don’t, then we won’t. So where, exactly, does the acting agent’s knowledge even come into the picture? Whether I can grasp a certain set of facts is totally unrelated to whether some other agent has grasped even more of them. The acting agent might only know one single fact that I don’t, and my ability to reach an appropriate level of comprehension would be just the same as if they knew a trillion more instead.

Once you recognize this, it’s easy enough to see that the problem of evil is all about our wisdom, not God’s. It’s true, of course, that having limitless knowledge might put you in touch with enough facts to make horrendous suffering extra-super-duper comprehensible — but a lack of extra-super-duper comprehension isn’t what makes anyone an atheist. What makes so many people atheists is that we seem to have essentially zero understanding whatsoever of the goods our own horrendous suffering supposedly produces. And what makes that surprising is our sense that, if those goods really did exist, human beings would be likely to notice them.

So if theists want to give a convincing response to the problem of evil, then just pointing out that God is a whole lot wiser than us won’t do it. What theists need to target instead is the assumption that human beings would have a reasonable shot at recognizing the outweighing goods our horrendous suffering is bringing us. And that means defending a very immodest premise like:

The relevant facts sufficient for understanding the goods secured by horrendous suffering are almost always unknown to human beings.

But why would anyone believe this? Here’s where I think theists will try to bring God’s infinite wisdom back into the picture: Let’s say, generously, that human beings know about 1% of all the facts in any particular domain, while an infinitely wise God knows them all. Assuming that all morally relevant facts are roughly equally distributed throughout the known and unknown portions of the domain, that means there’ll be a 99% chance that any facts about the goods that justify suffering will happen to come from a portion of the “fact space” that we simply don’t have access to. And so, in any particular case, God’s decision to allow for horrendous suffering is overwhelmingly likely to depend in some way on facts that we shouldn’t expect to know. Simple enough, right?

The problem with this line of reasoning is that domains of human knowledge don’t really work this way. Like, at all. For one thing, there are plenty of subjects where it’s safe to assume that even limited human beings have everything pretty much figured out. If I were to take the most comprehensive account of Walter Mondale’s 1984 presidential campaign that we have, and compare it with God’s own take on things, I don’t think we’d see too much of a difference. There’s just only so much to know! But even in domains where important unanswered questions do exist, the epistemic situation isn’t automatically as hopeless as theists are suggesting.

This is because, in reality, human beings don’t form theories by reaching blindly into a big hat labelled “facts about X” and pulling out a few bits of knowledge at random. Instead, we come to understand particular domains through a process of theorizing that tends, over time, to draw us towards the most essential facts in that domain. It might be true that, for example, our best scientists only know 1% of all the biological facts there will ever be — but that 1% we do know obviously includes a disproportionate amount of what really matters when it comes to understanding biology.

Moreover, at a certain point in the process of theorizing, our picture of the world can start to give a reliable, self-reflective assessment of its own completeness. A great example of this would be dark matter in astronomy — our best minds know enough to notice when there’s something they don’t know, and to begin tracing out the shape of that gap in their knowledge. Whereas other theories, while still equally incomplete, have disparate parts that all point towards each other in a way that suggests no major element is missing. So any vague, generalized appeal to “how little we know” about some particular subject isn’t nearly enough to demonstrate that its essential aspects are beyond our comprehension. That kind of hopelessness requires a concrete, domain-specific justification.

If we actually want to address the problem of evil, then, what we really need to tackle is one very specific question: Whether our grasp on the nature of human well-being is such that we should expect to almost always miss out on recognizing the extremely valuable goods that are consistently accruing to us as a result of horrendous human suffering. Every consideration beyond that — including the exact nature and extent of God’s wisdom — is just background noise. And in my mind, no theist has ever given any plausible explanation for why we ought to answer in the affirmative.

I have two reasons for saying this. First, if there’s anything at all that we know about human good, is that things like famines, earthquakes, wars, plagues, and so on are not generally conducive to our welfare. (A shocking result, I know.) And this isn’t just some gut feeling — we actually have many remarkably detailed and reliable methods for evaluating the degree to which particular human beings are flourishing, and our consistent finding is that those measures universally trend downwards in the aftermath of these sorts of events. In many of the most extreme cases, the effects of these tragedies can last for decades or even centuries, almost always without any compensating increase in welfare popping up anywhere else.

In fact, I’d go so far as to say that, if the amount of evidence for the net-negative impact of horrendous evil on human flourishing was replicated to roughly the same degree in any other field, it would be seen as totally dispositive and beyond any legitimate dispute. And that’s before we reflect on the fact that the theory of human good we’re relying on is itself incredibly robust and self-consistent, with zero signs of any unknown “dark goodness” we should be expecting to miss. It would be one thing if experts on human welfare were constantly scratching their heads, trying and failing to explain why some human lives go so much better than others. Then the theist might have grounds to identify these elusive outweighing goods with whatever variable the experts were struggling to find.

But instead, we see the opposite: A remarkably stable convergence across a huge number of distinct fields — psychology, biology, sociology, economics, you name it — when it comes to the basic picture of what makes a human life go well. And so far, that picture doesn’t seem to include starving to death, being crushed in a building collapse, or succumbing to AIDS as a toddler. But it also doesn’t include any conspicuous gaps where the source of our flourishing remains stubbornly unexplained. So there just isn’t the sort of theoretical empty space you’d need to hide away all the goods we’re meant to be gaining from the evil God permits.

Once you recognize all this, it becomes undeniable that we’re nothing at all like the confused and irrational creatures that theists love to compare us to. We’re nothing like sick dogs, who snarl at the veterinarians trying to treat them; most dogs, as far as I’m aware, aren’t working off a comprehensive empirical theory of canine well-being when they decide whether or not to bark. We’re also nothing like infants who wail when a parent has to bathe them; any infant who’s rational enough to wrestle with the justificatory status of bath time to begin with would probably figure out why it was needed pretty quickly. And similarly with every other trite analogy you might come across.

All this is to say: If these poor creatures understood their situation with the same relative clarity that we understand ours, then they very likely would come to grasp whatever goods are secured by their suffering. And that’s all the more reason to be surprised that we don’t come to grasp those same goods in our case. We can, if we want, just assume without reason that we’re catastrophically in error, and that an unimaginable amount of human good really does goes by unnoticed every day. But it seems much better, in my mind, to accept the incredibly productive theory of human good that we’ve developed instead, and then trust what it tells us: That these instances of horrendous evil are not, in fact, accompanied by any goods that outweigh them.

Of course, there are still ways for theists to get around all this. The most obvious option would be to say that these outweighing goods will accrue only in the far distant future. Or perhaps they diffuse to the whole of humanity, through some unimaginably complex causal chain. If this really is the case, then God’s infinite wisdom would be a good way to explain why he’s able to make such elaborate trade-offs succeed, and why we would be mostly unable to recognize them. But before any theists get too excited, I’ll have to explain the very basic reason for why this sort of thinking fails.

So put aside, for a moment, all the ethical issues that come along with inflicting extreme, uncompensated harm on one person just to benefit an unrelated party. Ignore the fundamental absurdity of God, the all-powerful creator of reality itself, reduced to tinkering with history like it’s a giant Rube Goldberg machine. And accept, against all reason, the idea that some logical constraint can make certain goods completely inaccessible in the absence of tremendous human suffering, while at the same time accepting that this constraint is arbitrarily absent in the heavenly state. That’s all fine. Even then, the pivot to diffuse or far future goods can only ever restate the problem of evil, as opposed to actually solving it.

Here’s what I mean: Recall that, if you want to defeat the problem of evil, what you’ll need to show is that we should expect to almost always overlook whatever facts are required to identify the outweighing goods our horrendous suffering produces. But on the assumption that we can reliably identify immediate, obvious human goods — which this pivot implicitly accepts — that would be equivalent to explaining why the goods in question are almost always obscure, diffuse, or far off in the future. But if you’re paying attention, you might notice that our explanation for why the goods in question are almost always obscure, diffuse, or far off in the future can’t simply be that they’re…. well, almost always obscure, diffuse, or far off in the future.

An analogy might help here. Suppose you have a doctor who puts you through an endless number of painful and invasive procedures, all of which leave you feeling much worse off than before. Understandably fed up, you storm into his office and accuse him of being a quack. Your main piece of evidence is the firm conviction that, if he really was a legitimate doctor, you would expect to recognize at least some of the benefits these procedures are supposedly bringing. He responds by reminding you that he’s so unbelievably talented, he can benefit you in ways that you can’t even comprehend. Would an answer like that do anything at all to make you feel better?

Of course not! The whole problem is that he’s (supposedly) benefiting you in ways you can’t even comprehend. So while his exceptional medical skill, if it does exist, might help explain how it’s possible for him to give incomprehensible treatments, it does nothing to explain why those are the only sorts of treatments he ever ends up giving, or why he’s unable to communicate their benefits effectively. In exactly the same way, merely explaining that God can bring incomprehensible goods out of horrendous suffering does nothing to explain why those are the only sorts of goods he ever ends up bringing about, or why he’s unable to make them obvious.

This point is so easy to overlook because we’re not used to thinking about what life would be like if this wasn’t the case. But it’s trivially easy to come up with all the examples you’d ever want. You could imagine a world where lightning bolts regularly strike serial killers dead, localized floods put out massive urban fires, and corrupt dictators die of cancer at fifty times the average rate. In a world like that, people would regularly grasp the outweighing goods produced by various evils — and in turn, it would be significantly more reasonable for them to accept the existence of unknown goods in the cases where things weren’t so obvious.

Alternatively, we could imagine a world more like our own, but with a God who directly intervenes to make the outweighing goods comprehensible. He could extend a profound divine comfort to anyone whose suffering was part of God’s plan, in a way that brought them to peace with their role in human history. He could give us a little burst of identifiable spiritual light every time we benefited from another person’s misery in some roundabout way.4 Or, if he wanted to, he could just obtain all the same goods in ways that are intuitively comprehensible all by themselves — exactly the sort of thing a being with infinite wisdom and power should be able to do!

In either sort of world, and in countless others, instances of profound misery and pain would pose no serious challenge to theism. That our actual experience with suffering is nothing at all like that just is the problem of evil. So any response that attempts to answer this problem by presenting the very incomprehensibility that motivates it in the first place is almost perversely inadequate. What explains why God never brings about comprehensible goods from our suffering can’t just be the fact that he does it some other way instead. How enlightening!

So where does that leave us? Well, the theist does have one move left to make here: Introducing the possibility of post-mortem reconciliation. And while you might not be surprised to hear that I think this defense is also very, very bad, it doesn’t ultimately matter for this discussion in particular — if post-mortem reconciliation really does solve the problem, then it’d solve things just as well with a God who was as dumb as a box of rocks. So as it regards his supposedly infinite wisdom, retreating to heavenly considerations is barely more than an admission of defeat. And so here I’ll end my case.

In this piece, I’ve tried to make a few basic points: First, that the problem of evil is about our own inability to grasp the goods that our suffering produces, rather than God’s ability to understand how best to bring about those goods; secondly, that we have extremely good reason to believe we would comprehend those goods, if they did exist; and finally, that shifting towards a long-term perspective based on unimaginably complex causal chains can never address the problem. My hope is that, even if these observations can by no means settle the debate, they can at least help us move past one particularly unproductive response.

All of this should only take you about six minutes to read, maybe even one or two less if you’re skimming things. Six minutes isn’t very long. I would imagine that, day to day, six-minute chunks of your life come and go without much notice. Six minutes cleaning up your kitchen counters after dinner. Six minutes zoning out in the shower. Six minutes scrolling on your phone, a dozen times a day. It’s the sort of time you don’t feel bad about wasting, a collection of moments that doesn’t feel big enough to hold anything important inside.

But in the six minutes you’ll spend reading through this, something important will happen: Children will die. About eight from pneumonia, another five from diarrhea, ten within an hour of their birth. Two or three will have their bodies crushed beyond repair in a car crash. Five will succumb to the hemorrhaging and seizures brought on by malaria. One will die of cancer. All in all, by the time you finish reading, this earth will be burdened by the weight of fifty or sixty new corpses, each one tiny enough for a mother or father to hold in their arms while they weep.

I think a lot about this now that I have a child of my own. She’s going to a party with my wife this afternoon, so I spent the morning watching her. We colored in her dinosaur coloring book and danced to Weird Al music videos. On the way to the park, I imagined for one single second what it would be like for her to die, and my eyes were filled with tears before I could even realize what I was thinking. And then I remembered: She does die, every day, just somewhere else. She dies in the Central African Republic. She dies in Gaza. She dies in the hospital right down the street. And every day, I watch it happen through someone else’s eyes, even if I don’t know it. When we get to the park, I tickle her feet when she swings high enough to reach me.

There’s been a lot of talk about miracles lately: Heavenly discs, reappearing eyeballs, a man with gashes in his hands that smell nice. I’ll be honest: I try to investigate the ones I can, spending a few minutes here and there checking if an Italian word has been translated correctly, Googling certain cloud formations to see if they really reflect light like some say they do. But there’s just so much to search through, so many loose ends to chase, that separating truth from fiction can feel like an impossible task. Most of the time, I’m tempted to just say: Okay, fine. Sometimes it really looks like people levitate. Sometimes tens of thousands of pilgrims see the sun turn dim and spin in circles. Sometimes a guy shows up with a leg that, by all accounts, just wasn’t there before. Maybe those really are miracles, if that’s the word you want to use.

But then you know what else is a miracle? The deaths of those children. Not a good miracle, of course. But a miracle nonetheless: Something utterly alien, unfathomable, beyond any hope of rational accounting. Every single second on this earth, there’s a child who is dying, screaming with pain and fear in the arms of their helpless mother. If God does exist, then that utterly pointless defeat of everything good and holy is infinitely more inexplicable than any clever spin on the precise details of Padre Pio’s scars. So yes, I can swallow my pride and admit that sometimes I struggle to offer the sorts of accounts for a miracle that can impress a zealous convert. Let’s see if that same convert can offer his own account for a still-warm child’s corpse that would impress the mother clutching it.

It’s important to remember that there is no God of Naturalism. Carl Sagan’s spirit isn’t looking down on us, promising to always ensure that the world will appear as a neat and orderly comprehensible thing. As a naturalist, I can only dare to claim that reality is systematic, and regular, and free of capricious breaks with fundamental laws — not that it will always appear to be. Nothing in my sense of the world so much as hints that I should be able to know what exactly took place on some Portuguese pasture more than a century ago. In fact, given naturalism, I should think it’s absurd that I ever could. And so I’m content, as they say, to merely dwell in the mystery.

But there is a God of Theism. There is a being, supposedly, whose very essence is suffused with a love so pure and overwhelming that no human being could ever possibly imagine it. And that being watches thirteen thousand children die every day while he plans out his next round of inscrutable stigmatas and flashing lights in the sky. He cherishes us, and cares for us, and desires with the purest heart that we come into the knowledge of the truth. And yet the deaths continue — five million lives a year, each one as beautiful and precious and innocent as my own little girl, all while demanding submission to blog posts about blurry apparitions three thousand miles and sixty years away. I’ll take on the thankless task of explaining away an endless number of Fatimas before I sign up to defend that.

We can imagine a just world, if we want. A good world. The sort of world we’d expect God to bring about for his beloved. And in that world, the miserable death of one single child from diarrhea would be an unprecedented, unfathomable, world-stopping suspension of the natural order. It would make the evening news. It would be the subject of every preacher’s sermon the next Sunday. A hundred years later, we’d be debating the details on Substack, with theists scrambling to explain how it could possibly be so. Instead, it’s happened five times since you’ve started reading, and no one bats an eye.

There might be more complicated arguments I could draw on to defend the point I’m making, written out with all sorts of impressive symbols and equations and fancy philosophical argot, all aiming to overcome our habituated sense that the suffering and death of the most vulnerable among us isn’t really too big of a deal. But in my mind, it’s that gruesome and disfiguring familiarity itself that testifies most strongly. At least naturalists are surprised when people’s legs grow back.

So God, if you’re listening: I don’t need any more miracles. I don’t need any more priests whose bleeding sores smell like perfume, or who have the power to fly up and perch on top of churches. Not while the children are dying. What I need is a truce: Just twenty-four hours where the wind never carries the sobs and screams of another human being like me. Do it on Christmas, or Easter, if you really want. But please, just do something. Give up on making a few hosts bleed every couple decades and instead stanch all at once the bleeding of this whole wretched world and the people trapped inside it. Not later, for some sliver of humanity, while the rest of us burn forever. Do it now, before more children die.

And if you don’t, then at the very, very least: Don’t expect a light show in the sky to make up for it.

This is the second half of a two-part piece on moral causation. You can read the first half of the piece here. In it, I give a brief overview of my preferred account of causation, and why we should accept that higher-order properties can have legitimate causal power that isn’t shared by the base-level physical states that instantiate them.

Up until this point, the illustrations I’ve been relying on have involved properties of the same general “kind” competing for causal power: The property of having a certain temperature versus the property of being in some specific microphysical state, the property of being red versus the property of being scarlet, and so on. Before I turn to the concept of moral causation directly, I first want to show that this account also handles situations where the base-level properties and higher-level properties are fundamentally distinct, or avoid any obvious reduction from one to the other.

Take, for example, the question of mental causation. Physicalists believe that all mental states are ultimately physical states, in the sense that a creature’s mental life is nothing “over and above” the neurobiological workings of their brain. And yet, at the same time, (most) physicalists will also want to say that our emotions, desires, beliefs, and so on still play a meaningful role in causing our actions. This is an issue for many accounts of causation that take a “base-level only” approach, since it seems as though an entirely non-mental description of the physical states in a creature’s brain at any one time should, in theory, be able to explain whatever behavior results with no mention of any additional mental properties that those physical states also realize.

The question, then, is how mental properties could have their own distinct causal power, as opposed to just “freeloading” off the work being done by the physical properties of their substrates. There are many different solutions to this problem that various philosophers have offered, but what matters here is that the account of causation I’ve been defending has a straightforward, natural way of resolving the apparent tension. It would simply say that certain base-level physical brain states are causally responsible for the base-level physical states that result from them, and that whatever higher-order mental properties those brain states realize are, at the same time, responsible for the higher-order properties of the action itself.

Consider, as an example, someone who desires something on a high shelf and, for that reason, raises their arm to reach for it. According to the causal analysis I’ve laid out, the particular physical brain states that realize this person’s desire for the object are causally responsible for whatever particular physical states in the body cause their arm to move in precisely the way it did; those physical brain states, in other words, causally explain the specific manner in which the person’s arm moved upward. However, if we ask what explains why the person reached upwards at all, the proper explanation would be their higher-order desire to reach the shelf, and not whatever physical states happened to realize that desire in this one specific instance.

This is because, given any remotely plausible account of human neurobiology, the general desire to raise your arm upwards will be multi-realizable — that is to say, the desire to raise your arm upwards can be realized by a wide range of particular physical brain states, including those that differ slightly from whatever states did actually happen to realize the desire in any specific instance. And this means that those particular physical brain states will fail the so-called counterfactual test we covered earlier: They can’t be the cause of the person moving their arm because, in the closest possible world in which those particular states don’t obtain, a very slightly different set of states would still realize the desire and the arm would still reach upwards.

Meanwhile, what does pass the counterfactual test is the desire for the item itself, since that’s what the “arm-raising process” is actually sensitive to. Therefore, the accurate causal relationship would look something like this:

Just like it did previously with the freezing water, this account of causation easily secures all the commonsense conclusions we’d hope for when it comes to mental causation: We avoid both strong emergence and over-determination, while at the same time finding a meaningful causal role for both physical brain states and the mental properties they realize. This is, in my mind, just more reason to believe that this account is the correct one, and that we can rely on it to determine causal relationships even in situations where the properties involved are meaningfully distinct, or share non-reductive supervenience relationships. Therefore, with all this in mind, we should now consider what this account would say about the role that moral properties can play in causal explanations.

One obvious context in which moral properties seem to play a causal role would be in the analysis of an agent’s character and its influence on their actions. For example, we can imagine someone who operates a successful local business, and decides to offer all his employees a substantial bonus after they complete a major sale. In this case, it seems (almost tautologically) clear that his actual offer of the bonus was caused by his desire to offer it. However, we can also push further and ask why it is that he desired to offer the bonus to start with, which is another way of asking what caused him to form that desire. And here, his generous character seems like the appropriate answer, which gives this general causal chain:

At every point, distinct base-level physical states have causal relationships on the microphysical level: One particular brain state brings about another, which brings about another. But crucially, these base-level states could all differ slightly while still realizing the same general psychological properties and intentional acts. So if it’s that final intention act that we want to explain, the counterfactual constraint on causation should lead us to say that the general existence of the desire is what explains the offer — and the desire is, in turn, explained by the owner’s generous character, since had the owner not been generous, he would not have desired to offer the bonus. This is a simple example of a moral property that seems to exist as a straightforward distal cause for a real-world event.

We can also consider situations in which moral properties not only play a causal role in explaining our actions, but also in explaining how it is that we evaluate the actions of others. Consider someone who walks by a stray dog and kicks it, purely out of spite. In this case, the analysis above would plausibly assign a causal role to that person’s cruel character. But it can also show that another person’s negative judgment of that act was caused by the cruelty of the act itself. The causal chain involved here is slightly more complex, but should still be comprehensible given what we’ve seen so far:

Here, we start with the man’s kicking the dog, which causes me (through contact with my visual faculties) to form the belief that the man kicked the dog.1 We also see that the man’s specific act of kicking the dog realizes a more general property of cruelty, since wounding a vulnerable creature for no reason is paradigmatically cruel. In parallel, we see that my belief about the man kicking the dog guarantees, in a non-causal sense, that I believe the man’s behavior has what you might call “cruelty-making features” — that is to say, it has the sorts of features sufficient for making some action a cruel one.2 And it’s this feature of my belief that then causes me to recognize that the man did indeed behave cruelly.

But how do all these pieces hang together in terms of their causal relationships? What’s important to note is that, according to the counterfactual test, the base-level fact that the man kicked the dog doesn’t properly explain why I believe that his behavior had cruelty-making features. This is because, in the closest possible world in which the man doesn’t kick the dog out of spite, he still behaves cruelly in some slightly different way — and I, in response, still come to believe that his behavior had cruelty-making features. On the other hand, if the man’s behavior hadn’t actually been cruel, then (assuming I’m perceiving everything properly) I wouldn’t have formed a belief that attributed any cruelty-making features to his behavior.

For this reason, the cruelty of the man’s behavior, rather than any specific detail of the act, is the proper candidate for the ultimate cause of my judgment that the man’s behavior was cruel — a straightforward example of clear moral causation. In my experience, many people are seriously skeptical of this conclusion at first. However, the general structure it relies on is relatively uncontroversial in almost all other non-moral contexts. For example, we can imagine seeing the stray dog who was kicked at some later time and recognizing that it has a broken leg. In this case, the causal chain that leads us to our judgment that the dog is injured would look something like this:

Here, the dog’s leg being broken realizes an injured state, which counterfactually grounds the injury-making features present in the content of my observation; in response to my recognition of these injury-making features, I judge that the dog is injured. Had the dog’s paw or hip been broken instead, then the content of my original observation would have differed, but some injury-making feature in that observation would still have been present, and therefore I still would have come to believe that the dog was injured. For this reason, it would be appropriate to identify the dog’s general property of being injured as the cause of my judgment, rather than the specific avenue by which that injured status was realized.

If this account seems reasonable enough to you in a non-moral context, then in the absence of any special pleading, you should consider it just as reasonable in an otherwise identical moral context. And once we do accept that both actions and beliefs can be caused by moral properties, it becomes clear that morality’s practical role in the world is significantly greater than what most people would naturally assume. Human actions and beliefs are, obviously, extremely important when it comes to understanding politics, economics, culture, religion, art, and essentially any other domain of human life. So if at least some of these actions and beliefs are themselves causally grounded in moral facts, then moral facts necessarily figure into our best ultimate explanations of a wide range of events.

We could justifiably say, for example: That a certain regime collapsed because it was oppressive, or that a certain activist succeeded because she was courageous, or that a certain social practice was eventually abolished because it was seriously wrong. These sorts of explanations, while obviously incomplete by themselves, aren’t just emotive restatements of more fundamental non-normative accounts. Rather, they’re legitimate causal accounts of their own with real, independent explanatory power. There are, of course, situations in which moral explanations are not productive, and where non-moral explanations are sufficient. But the assumption that this is always the case, and that no moral facts could ever be proper causal grounds for events in the world, is simply unjustified.

A huge number of possible objections could be raised in response to the analysis I’ve laid out here. However, I want to close out this piece by focusing on three of the most common in particular. The first, which I think can be dismissed fairly quickly, is that the account is somehow unacceptably circular. The idea seems to be that, before you can use some property in a causal explanation, you must first demonstrate that the property exists through some “objective” means, and then confirm that the pre-theoretic conception of it you’re relying on is accurate. But this is just a mistaken understanding of how empirical theorizing works, in at least two distinct ways.

First, there’s no general prohibition on starting out your theorizing by relying on folk concepts. When early modern physicians began to rigorously study disease, for example, they weren’t required to erase the everyday notion of a disease from their minds and then rediscover it from scratch. Instead, they began with basic, commonsense assumptions about what disease was and proceeded to investigate further. At times, those investigations forced revisions to their folk concepts, in the same way that moral theorizing has also at times forced similar revisions in our moral concepts. But the idea that we should have a default suspicion towards widely accepted folk concepts until their accuracy can be “proven” in a way that makes no reference to those concepts themselves is just unreasonable.

Similarly, the idea that we should first “prove” that some property exists before we use it to construct causal explanations for observations ignores the fact that, in most cases, it’s precisely the causal role those properties play that gives us a reason to accept their existence. The reason physicists first came to believe that electrons exist, for example, is that their existence best explained the results of various experimental observations; it would make no sense to expect that those physicists first prove the existence of electrons in some other way before being allowed to use them in explanations. In much the same way, the explanatory power of moral properties is itself very good reason to believe those properties exist, and moral realists aren’t obligated to show their existence through other means before then seeing if they “pull their weight,” so to speak, explanatorily.

The second objection, which I think is more interesting, would be something like this: Rather than accepting the existence of truly moral properties like cruelty or righteousness, we should instead simply accept properties like “being conducive to judgments of cruelty” or “being the sort of thing that makes observers believe that the act was righteous,” and so on. In other words, the idea seems to be that we can avoid ontological commitment to moral properties by instead adopting other non-moral properties that are framed in terms of the actions or judgments they provoke. These properties would also trivially pass the counterfactual test, since if some act didn’t have the property of compelling a particular judgment in observers, then by definition, that particular judgment wouldn’t be made.

Because these sorts of properties seem to earn the same causal power that moral properties do, without requiring additional ontological commitments, it’s easy to see why many would consider them to be preferable. The issue, however, is that these sorts of properties may actually be moral properties. Let me briefly explain what I mean with an analogy. When I look outside and form the belief that a tree is in my front yard, it’s tautologically true that my belief can be explained by the presence of some “tree-in-front-yard-belief-causing property” in what I saw. But at the same time, when I ask what in the world that property actually picks out, it shouldn’t be controversial to say that it picks out there being a tree in my front yard.

In other words, for any person whose visual faculties are functioning reasonably reliably, the yard’s property of “having the sort of thing that provokes a belief in the presence of a tree” just is the property of actually having a tree be present. I don’t think this is a particularly radical claim, and I think moral realists should say something similar in this case. That is, for any person with reasonably well-functioning moral faculties, the property of compelling a judgment of cruelty just is the property of being cruel, since what compels a judgment of cruelty in people with reasonably well-functioning moral faculties just is the presence of cruelty in the act itself. So on the assumption that the person in question does have those well-functioning moral faculties, this objection doesn’t avoid a commitment to moral properties, but instead merely relabels them.3

Now, it is worth asking what this analysis would say in situations where someone forms an inaccurate moral judgment, or only forms the correct judgment through some unreliable inference. Consider, for example, someone whose racial biases lead them to assume that an acquaintance is dishonest, and that their unjustified assumption happens to be correct. In this case, had the acquaintance not been dishonest, the bigot still would have believed that they were, and therefore this analysis would return the correct conclusion that the acquaintance’s dishonesty did not cause the bigot to form his belief. Instead, it would have been caused by another “dishonesty-judgment-compelling property,” which in this case is the bigot’s prejudice. And this is independent reason to think that moral properties can’t be cleanly replaced by non-moral properties framed entirely in terms of what responses are being provoked without altering the basic causal structure.

Finally, I want to very briefly address the objection that the moral properties I’ve made reference to — cruelty, generosity, and so on — don’t qualify as truly moral properties at all, because they can be given entirely reductive definitions in terms of non-moral properties. Personally, as a non-reductive naturalist, I don’t believe there are any plausible non-normative accounts of at least some moral properties. And this means I can simply deny the objection. However, it’s also important to point out that, even if entirely descriptive reductions for these concepts do exist, this is completely compatible with those properties still being moral properties. (Just as many chemical properties are completely reducible to more basic physical facts, but are still chemical properties.) Nothing about moral naturalism requires or even implies that no reductive account of moral properties exists.4

I want to finish this piece by being very clear about what it’s not meant to show. It doesn’t show that moral properties are always causally relevant in explanations of human action or belief, or that moral explanations always trump non-moral ones. It also doesn’t show that moral properties have any necessary normative “grip” on human beings, or that the presence of any particular moral property automatically gives us a reason to behave or act in any particular way. Instead, all I aim to show is that, given an independently persuasive account of the nature of causation, we have good reason to see moral properties as being causally efficacious aspects of reality, alongside all sorts of other natural properties that play uncontroversial causal roles.

Whether or not this one fact is, by itself, is enough to establish moral realism is a complex debate for another time. What it does establish, however, is that morality has at least some objective component that extends beyond our own stances, and that this objective component can play a meaningful and valuable role in the best possible understanding of our lives and of the world.

In this piece, I’m going to argue that moral properties are causally relevant, in the sense that the best explanation for why an event occurred will sometimes require a reference to moral properties. If this is true, then it would be an extremely good reason to believe that moral properties exist. The argument I’ll be giving is drawn from the work of Nicholas Sturgeon, Frank Jackson, and John Gibbons, and Peter Yablo, and also borrows from the illustrations Judith Jarvis Thomson has used in her analysis of Sturgeon’s arguments in particular.

I want to start off by making a few general points about the way causation works. Consider this scenario: One winter day, when the air temperature is 18°F, a bottle of water left outside freezes and turns into ice. We could represent the causal chain here as follows, with the arrow standing in for the “X caused Y” relation:

Immediately, however, we notice an issue. One extremely plausible constraint on a causal explanation is that it has to pass what you might call the counterfactual test: If X is the cause of Y, then if X hadn’t taken place, Y wouldn’t have taken place either.1 This constraint helps to capture the obvious and uncontroversial connection between causation and responsibility; to say that Y was caused by X, but also that Y would have taken place even without X, seems like a basic failure to grasp what it means for one thing to cause another.

The issue, then, is that the explanation given above clearly fails this test. In the nearest possible world in which the temperature outside wasn’t 18°F, it would be 19°F, or 17°F, or some other similar temperature instead, and in that case the water would still have frozen just the same. (Or at least similarly enough to still make it true that the water froze.)2 For this reason, we might feel pressure to say that, strictly speaking, the temperature outside being 18°F wasn’t the true cause of the water freezing. Instead, it might be more reasonable to say that the water was caused to freeze by the temperature outside being below freezing:

Unlike last time, this new causal analysis does pass the counterfactual test: In the nearest possible world in which the temperature outside wasn’t below freezing, the water (obviously) wouldn’t have frozen. But this brings up a new issue, which is that the temperature outside being below freezing and the temperature outside being 18°F aren’t completely separate states of affairs. After all, the temperature is only below freezing in the first place because that temperature is, specifically, 18°F. In light of this, there are three rival ways of framing the causal relationship. The first would be:

In this framing, the temperature outside being 18°F is what causes the water to freeze, while also (non-causally) instantiating the supervenient property of being below freezing.3 This latter property, however, simply “floats above” the specific 18°F temperature responsible for all the causal work. While it might be tempting to frame things this way, the lower-level property’s failure to successfully pass the counterfactual test is still an issue. In response, another option might be:

In this situation, both the lower-level property and the higher-level property are causally responsible for the water freezing, while the former is also responsible, in a non-causal sense, for the latter. This framing allows for the water’s freezing to be explained in a way that passes the counterfactual test, but it also requires that we accept extensive overdetermination in almost all everyday instances of causation. Moreover, this framing also continues to grant causal weight to the first-order property of being 18°F, which is an issue for reasons we’ve already discussed. This leaves us with one more option to consider:

In this framing, the lower-level property of being 18°F non-causally grounds the higher-order property of being below freezing, which is then causally responsible for the water freezing. While the temperature’s lower-level property of being 18°F is obviously relevant, in some general sense, to the water freezing, its relevance is just that it happens to realize the actually causally efficacious property, which then goes on to do all the causal work. This would be the framing that properly accounts for causation in the counterfactual terms we expect it to.

But how could this be possible, given that the temperature’s lower-level property of being 18°F is enough, by itself, to provide a robust physical explanation for the water’s freezing as it actually took place? You could, after all, explain the entire physical history of this specific bottle of water and its transition from liquid to solid without ever invoking the higher-level property of being a below-freezing temperature. So how could it be that an explanatorily unnecessary higher-order property is, in fact, the proper basis for the causal relationship being described?

The answer, I think, comes from looking more closely at the relation’s right side, which we haven’t considered in detail so far. You might notice that the property of being frozen is itself a higher-order property, just like the property of being a below freezing temperature. (Similarly, the property of being 18°F isn’t a truly “base-level” property either, given that it supervenes on a huge number of more fundamental physical states.) When we make this explicit by splitting both sides into their foundational and supervenient aspects, we might land on something like this:

This framing returns us to the more comfortable base-to-base causal relation we naively expect, and this time the counterfactual test is easily passed. The problem, however, is that the water being frozen now has no causal explanation at all. The base-level physical state B didn’t cause the water to freeze (because supervenience isn’t a causal relation) and the property of being frozen has no other causal link to any earlier state, basic or not. This is arguably an even worse situation than what we had before; whatever theory of causation you want to choose, it ought to say that properties have causal histories. For this reason, it would be best to consider another framing instead:

Here, base-level physical states are responsible for grounding their supervenient higher-level properties, but both the higher-level and lower-level states only ever enter into “intralevel” causal relationships. That is to say, a specific arrangement of particles in the outside air causally produces another specific arrangement of particles in the water, while a more general property supervening on the former arrangement causally produces another more general property supervening on the latter arrangement. This framing can give a reliable causal history to all the relevant properties, whether lower-level or higher-level, while at the same time avoiding the uncomfortable situation wherein higher-level properties seem to causally produce lower-level ones.

The only question left to ask, then, is just whether this framing is philosophically defensible. We can see that it most easily accommodates our commonsense views on the causal relationships in question. But do we have any independent reason to believe that it’s actually true? I think we do, and the reason has to do with causal sensitivity. Stephen Yablo uses the example of a chicken who can be trained, with some effort, to peck at red squares, but can’t be trained to peck at squares with a more specific shade. If you present the chicken with a scarlet square, it’ll peck at it, but what caused this pecking is the fact that the square was red, not the fact that the square was scarlet. The chicken is causally insensitive to the specifically scarlet tone of the square, and this is true even if the fact that the square was red supervenes on the fact that the square was scarlet.4

The basic principle is this: In order for X to be the cause of Y, it has to be the case that the process by which Y comes about is sensitive to X, such that a change in X would have made a difference in the process even with all other related properties or features remaining fixed. In the case of the chicken, the redness of the square is the cause of its pecking because whether or not the square is red is what makes a difference to its pecking. In this particular instance, the square that was pecked is red, and scarlet, and colored, and the redness, scarlet-ness, and colored-ness all supervene on the same base-level physical state. But only the redness caused the pecking, because the chicken’s pecking procedure was only sensitive to redness:

I think this account of causation gets things right. It’s important to note here that nothing I’ve said so far requires a belief in “emergence,” or any spooky downward causation from higher-level properties to whatever lower-level physical states realize them. After all, whatever level a fact about the square being red is on, that would very plausibly be the same level that facts about the chicken pecking are on too. So in this model, physical states cause other physical states, chemical states cause other chemical states, psychological states cause other psychological states, and so on. No causal power is moving up or down the chain on either side of the relation.

Now, very quickly, let me explain (in light of all this) why assigning causal power exclusively to the base-level physical states is especially hard to defend. The argument is simple: If base-level physical states carry all the causal power, then either those physical states share their causal power with whatever higher-level properties they realize, or else they don’t. If they don’t, then we’ll have to accept that no properties ever cause things to happen, and that no properties are ever brought about by causes. We’ll also have to accept that the causal sensitivity requirement is false. Both are high prices to pay.

But if the base-level physical states do share their causal power with the higher-level properties they realize, then they share that power with all of them. And this means that all of the higher-order properties realized by that base-level substrate will, together, act as the cause of the resulting physical state (and possibly all of its properties as well). The problem is that, as we’ve just seen, this clearly isn’t how it works. The property of being colored, or of being scarlet, didn’t cause the chicken to peck at the red square. The property of being red did. Any account of causation that can’t make that sort of distinction is a bad one.

To make a broader point: When we identify one particular property as the proper cause of some event, we aren’t (as some might argue) just changing how it is that we describe things, or how it is that we model them in our minds. This is because an objective and meaningful extensional difference exists between claims about the causal role of properties and the causal role of whatever base-level physical states ground them. Asserting that the redness of the square is what caused the chicken to peck it isn’t the same thing as asserting that the square’s scarlet-ness, or colored-ness, is what caused the chicken to peck it. Those are different claims, not just different ways of speaking. A world in which the chicken was caused to peck by something else would have to be a different world.

With all this in mind, we can return to the original example of the bottle of water freezing to see that this analysis works things out right. Of course, freezing isn’t a conscious process in the way pecking is, so the process by which the water comes to freeze isn’t causally insensitive to specific base-level physical states (or specific temperature properties) in the same exact way. Still, when it comes to the process of freezing in general, that process is relevantly insensitive; just as minor changes in the exact reflective profile of the square make no difference as to whether the chicken pecks it, minor changes to the exact temperature of the outside air make no difference as to whether or not the water freezes.5

For this reason, we can affirm two things without contradiction: That the specific base-level properties of the frozen water were caused by the corresponding base-level properties of the surrounding air, and that the water’s higher-order property of being frozen was caused by the surrounding air’s high-order property of having a below-freezing temperature. Crucially, these aren’t merely two different ways of speaking, but rather two distinct (although obviously related) causal relationships involving two distinct properties, each with distinct counterfactual implications.

This basic account, while obviously incomplete, is enough on its own to generally secure everything we could reasonably expect from a theory of causation. It aligns with both the counterfactual and sensitivity constraints, avoids emergence or any instances of downward causation, and also gives an independent backing to the commonsense idea that higher-level properties are sometimes causally relevant. For this reason, I consider it the most plausible account being currently offered, and I think we ought to rely on it when trying to answer the question of whether moral properties in particular can be causally relevant.

This is the end of Part 1 — in Part 2, I’ll apply this account of causation to the question of moral properties directly, and give my argument for their being causally relevant to the production of both other moral properties as well as many non-moral properties.

A few days ago, I put out a piece on moral relativism, and why I think it ends up falling victim to many of the same classic objections that people generally raise against emotivism. I’m happy with how it turned out overall, but there’s one big section that I really wanted to include even though I couldn’t ultimately find the space for it. See, near the end of the piece, I make a brief offhand reference to the Frege-Geach problem, and just mention in passing that I think relativism has its own issues with explaining how moral claims are supposed to work when they’re embedded in conditionals. I had originally planned to give a much more detailed account of what, exactly, I think those issues are, but it seemed like a bad idea to basically double the original piece’s word count — so now, dear reader, you get a follow-up piece with some highly sought-after unreleased material! How lucky.

First off, let me clarify what I mean when I talk about “moral conditionals.” When I use that term, I don’t just mean any moral judgment that depends on some sort of context, like “If your friend is planning on robbing a bank, then you should try to talk them out of it.” Relativism doesn’t have any serious problem making sense of claims like that, or at least the problem isn’t any worse than it would be for any other sort of claim. Rather, I’m talking about conditionals where the antecedent and consequent are both themselves moral assertions, like in the classic example that people often use to illustrate the Frege-Geach problem: “If it’s wrong to tell lies, then it’s wrong to get your little brother to tell lies.”

The issue for relativism, in my mind, is how we should understand where exactly the relevant moral framework “comes into the picture” — that is to say, whether the entire conditional should be taken relative to the framework, or each element individually, or some split down the middle. I’ll call these three basic approaches the external view, the internal view, and the partial view. In my mind, the internal view is pretty much dead in the water, so that would be a good place to start.

On this view, the conditional “If it’s wrong to tell lies, then it’s wrong to get your little brother to tell lies” breaks down into something like: If, relative to a moral framework M, it’s wrong to tell lies, then, relative to that same moral framework M, it’s also wrong to get your little brother to tell lies. But the problem with this view is that the resulting assertion just simply isn’t true, or at least it doesn’t have to be. After all, it’s not as if there’s any cosmic guarantee that a moral framework containing a prohibition against lying will also, as a matter of necessity, contain any additional prohibition against getting your little brother to lie.

It would certainly be weird, of course, to run into someone who believed that lying was wrong, while at the same time also being fine with getting your little brother to lie. But then again, people have exceedingly weird moral commitments all the time, right? At the end of the day, whether or not the elements of a given person’s moral framework will come together properly — whatever that even means given relativism! — is an empirical question, not a logical one. So while there are plenty of moral conditionals that you can reliably expect to come out true, none of them would actually be capable of expressing any legitimate entailment, which seems like a serious problem for the internal view.

So what about the external view instead? This would see the entire conditional as existing inside a moral framework, producing something like: Relative to a moral framework M, [if it’s wrong to tell lies, then it’s wrong to get your little brother to tell lies]. I would imagine that most relativists, if asked, would probably say this is how they personally understand moral conditionals — as an expression of what their own moral framework says about the implications of one commitment or another. And personally, I think it’s the most reasonable interpretation out there too. But there’s still a pretty major issue with it that I’m not sure relativists can really get around.

The issue is pretty simple: It just doesn’t seem to me that a normal person’s moral framework actually does give content to conditionals like these, especially when it comes to conditionals that involve moral assertions they don’t actually affirm. For example, I think it’s true that, if it’s morally wrong for a man to date another man, then it’s also morally wrong for a man to marry another man. That just seems true on a logical level. But of course I don’t think it is actually wrong for a man to date another man, and I don’t think my own moral commitments are what explain why the conditional holds. To me, what explains it instead is just my basic belief that moral judgments have to have consistent entailments — but that’s precisely what I’m worried that relativism can’t explain!

One possible solution to this dilemma might be for relativists to step back a level or two and identify some meta-principle that structures the first-order content of their framework, like “If X is wrong, then anything relevantly similar to X is also wrong.” But while that certainly closes the gap, it doesn’t really seem like a moral principle, so much as a logical one — and once you start introducing those sorts of structural second-order constraints, the “relativism” in question starts to look more and more like a degenerate constructivism.

Furthermore, what are we supposed to do with people who, for whatever reason, don’t have a framework governed by this kind of consistency principle? Would a relativist be forced to accept that “If it’s wrong to torture babies, then it’s wrong to torture toddlers too” could ever actually be false? That certainly seems crazy! But arguing that moral frameworks simply must be committed to consistency is, as I said, a dangerous move for relativists to make. And here’s the thing: Even if there is some relativist-friendly way to establish the constraint, the conditional statements it would produce still don’t seem like they’re about that constraint at all. Like, do I think “If segregating schools is good, then segregating restaurants is good” is true? I mean, I guess. But it’s obviously not as though, when I affirm that conditional, what I’m communicating is a personal commitment to keeping judgments about racial apartheid consistent!

Let me finish things off with the third option, the partial view. This would be the view that the moral framework only applies to the consequent of the conditional, giving something like: If it’s wrong to tell lies, then, relative to moral framework M, it’s wrong to get your little brother to tell lies. Personally, I can see why a view like this might appeal to some people, since it captures the commonsense notion that moral conditionals express what we think follows from a universal claim — but the problem, of course, is that there are no such universal claims about moral reality given relativism, so any analysis of moral conditionals with a non-relative aspect to it can’t possibly be true. And once you do try to ground the truth of the antecedent properly, you’ll need to give it its own relative framework, which just collapses things back into the untenable internal view.

I want to be clear that I’m not saying these problems are totally insurmountable, or that they count as some knockdown proof that relativism is false. I’m sure it would be more than possible for a committed relativist to find a semi-plausible way of dissolving the tension, or at least work up the courage you’d need to bite whatever bullets need biting. Still, I think these sorts of concerns are, if nothing else, yet another good example of all the work it takes to make relativism match up cleanly with the full range, depth, and complexity of everyday moral thought. And so long as the realist position gets all these things for free once you pay the ontological cost upfront, I think that’s a serious point in its favor, and one that doesn’t get acknowledged nearly enough.

In the online philosophy space, relativism seems to be significantly more popular than emotivism among the anti-realist crowd. And for a while, that made sense to me, given that emotivism (supposedly) faces all sorts of problems that relativism can avoid. However, after going back to reread Judith Jarvis Thomson’s section in Moral Relativism and Moral Objectivity, I started to ask myself: is there really that much daylight between the two positions in the first place? They aren’t the same, of course, and the differences that do exist can matter quite a bit. But in terms of which theory it would be better to actually accept, I’m beginning to think that the supposed benefits of relativism pretty much disappear once you make some very plausible assumptions about how exactly it would have to work.

In what follows, I’ll be working with the fairly simple account of moral relativism that Gilbert Harman gives in his section of Moral Relativism and Moral Objectivity:

For the purposes of assigning truth conditions, a judgment of the form It would be morally wrong of P to D, has to be understood as elliptical for a judgment of the form In relation to moral framework M, it would be morally wrong of P to D.

Let’s start with a simple question: Just what, exactly, is a “moral framework” even supposed to be? For most other intrinsically relative statements, it’s easy enough to lay out the specific content of whatever standard makes them true. When I say that I was driving at sixty miles per hour, for example, basic physics tells us that the truth of my assertion depends on a reference frame, in the same way that the truth of my moral assertions (supposedly) depends on a moral framework. But in the case of my car, I can say more than that — I can say that the reference frame that makes my assertion true is, specifically, a coordinate system with its origin point at such-and-such a place. So what would be an equally informative account of the moral framework that, under relativism, makes my moral assertions true?

In everyday life, we generally think of a person’s moral framework as a collection of all the moral principles they rely on to navigate life, like “demeaning others is wrong” or “you ought to help the less fortunate.” And so, at first glance, we might be tempted to think that the moral frameworks we’re talking about here are also sets of moral propositions, at least once they’re systematized in a certain way. But what many people — including many relativists — don’t seem to realize is that relativism’s core commitments explicitly rule this option out.

Here’s the problem: According to relativism, moral propositions are essentially… well, relative. And that means they’re ill-formed and vacuous apart from whatever framework makes them true or false, in the same way it would be meaningless to assert that something is “intrinsically stationary” apart from any reference frame. But if that really is how moral propositions get their truth values, then how could we ever make sense of the propositions that comprise a moral framework? Either each proposition would require its own moral framework, which of course would produce an infinite regress, or else the one moral framework you’ve got would be nothing but a collection of pseudo-propositions that are literally incoherent.

As an example, let’s say that Frank’s assertions about the rightness of X are made relative to a moral framework F, and that F contains the proposition “stealing is wrong.” If that proposition has no truth-making standard of its own, then it fails to mean anything at all, and so does his original assertion about the rightness of X. (This is because “X is right, relative to [ill-formed proposition]” is itself an ill-formed proposition.) But if, on the other hand, F’s anti-stealing proposition does come with another proposition that fixes its truth value, then that truth-making proposition would also need to be justified, in which case you inevitably end up with a framework full of infinitely complex propositions like: “Stealing is wrong, relative to a framework that says ‘stealing is wrong, relative to a framework that says “stealing is wrong, relative to…”’” and on and on forever.

This isn’t just an issue for unqualified moral claims of the form “X is right” or “X is wrong,” either. The same dynamic applies to any normative assertion, including ones like “It would be good to X” or “X is valuable.” Because relativism frames all normative content as being tied to a particular framework, there’s no collection of irreducibly normative propositions (or, more accurately, pseudo-propositions) out there that could ever pull double-duty as their own truth-making grounds. But if moral frameworks can’t contain these sorts of pseudo-propositions, where else is there to look? If we want to keep the contents of the framework truth-apt, then it looks like we’ll have to consider purely descriptive propositions instead.

The issue here, however, is two-fold: First off, there’s the well-worn criticism that making moral statements relative to descriptive facts about attitudes robs them of any actual normative heft; “X is wrong, relative to facts about what this particular community disapproves of” seems more like an anthropological observation than an actual moral assertion. But the bigger issue is that, if relativists really do insist on seeing these expressions as legitimately moral ones, then they’ll have to go on record as saying that at least some purely descriptive facts can have substantive normative entailments all by themselves — and if you’re comfortable saying that, then why not just be a moral realist? If the natural fact that some critical mass of people in a community think X is wrong can produce actual, honest-to-goodness moral wrongness, then why not think other facts could too?1

In my mind, it would be better for relativists to just avoid the whole mess and rule out any propositions at all, descriptive or normative, as candidates for the content of a moral framework. And that leaves them with a choice between two different kinds of non-proposition expressions: Commands or attitudes. But if we consider which of the two seems fundamentally prior in a normative account, it strikes me as obvious that attitudes are more basic. After all, if I ask you why you’ve shouted “Don’t murder!” at me, it would be natural to refer to your anti-murder attitude; meanwhile, if I ask you why you have an anti-murder attitude, it would be totally unhelpful to reply by pointing out that you assert the command “Don’t murder!” So if we hope to have at least some sort of explicability to our moral judgments, attitudes seem like a much more fitting truth-maker than arbitrary commands.

All this is to say: If you’re a relativist, you should think the content of the moral framework by which a person’s moral judgments get their truth values is a set of non-propositional attitudes. And for what it’s worth, this isn’t even a particularly controversial conclusion — in fact, my guess is that most sophisticated relativists don’t see our moral frameworks as being themselves truth-apt, and would prefer attitudes over commands when surveying all the non-propositional alternatives. (Plenty of anti-realists online don’t seem to be aware of this, but I digress.) So it’s not as if, by pointing this out, I’ve done anything to disprove relativism. But with that said, I do think problems start showing up very quickly once you ask exactly how it is that these non-propositional attitudes ground our moral judgments.

The first issue, which I’ll just flag briefly before moving on, is that it’s not clear to me what standard is supposed to distinguish a specifically moral framework from any other set of attitudes, given that we all experience feelings of both moral and non-moral disapproval from time to time. For example, I have a powerful negative attitude towards country music, but enjoying country music doesn’t seem wrong, relative to that attitude — it just seems like something I don’t personally enjoy. Whereas it could be that I have an equally strong negative attitude towards some minor act of immorality, like cutting me off in traffic, that does end up grounding the assertion that cutting me off in traffic is wrong. So what, exactly, explains why one of these attitudes grounds wrongness and not the other?

This leads to a bigger question, which has to do with the incredibly broad range of real-life moral discourse. When we engage in moral deliberation or express our moral views, we don’t just limit ourselves to assertions like “X is wrong” or “Y is bad.” We also judge that things are brave, or dishonest, or unkind, and we lay out complex relations like “Erick’s greed was worse than William’s carelessness” or “Contempt is a correct response to shamelessness.” So how are we supposed to understand the way these so-called “thick” moral assertions end up with their truth values, if the frameworks involved contain only assortments of attitudes?

The way I see it, there are two general options available. The first would be to say that moral frameworks contain multiple distinct sorts of attitudes, each differing phenomenologically based on the concept they relate to or pick out. On this view, there’s a certain “felt sense” of greed-specific disapproval, another “felt sense” of justice-specific approval, and so on for every other basic moral concept. So when someone says that Sue’s actions were greedy, that claim would be made true by a framework containing some specific BOO_greediness attitude towards whatever it is that she did — where, to be clear, BOO_greediness would be a unique kind of attitude that grounds the truth of the claim “Sue was Greedy,” rather than just expressing generalized disapproval towards her greediness.2

The problem with this approach is that I just cannot for the life of me understand what these distinct “felt senses” could even possibly be. Whenever I reflect on my own experience of moral disapproval or approval, I don’t get the impression that any of those feelings come in different flavors at all. What I disapprove of varies, of course, but the disapproval itself sure feels like the same attitude in all cases. I guess you could define these attitudes entirely extensionally, if you really wanted to — something like “the BOO_greediness attitude is the attitude that makes ‘X was greedy’ true” — but then you’d be stuck with a perfectly circular, uninformative account that obviously doesn’t capture what’s actually going on.

So unless I’m missing something here, and most other people really do have a way to make incredibly fine-grained distinctions between their various disapproving attitudes, then trying to ground the full range of moral judgments in a one-to-one matchup between concepts and attitudes seems like a nonstarter. The alternative, then, would be to say that moral frameworks contain only the basic pro- and con- attitudes we all recognize, but just applied to particular specifications that “draw out” the intended moral concepts. I’ll symbolize this as BOO(X), where X stands in for some specification.

For example, we can imagine that Sue is selling her house, and she pulls off some sleazy maneuver to squeeze a couple thousand more dollars out of the buyer. In a case like this, it might be that the judgment “Sue is greedy” is made true relative to the speaker’s BOO(Pursuing more than what you’re fairly owed) attitude, with that specification intended to “draw out” the moral concept of greediness. But the issue here, which you might have noticed already, is that the description given is compatible with “drawing out” multiple different moral concepts; pursuing more than you’re fairly owed is, after all, not only greedy but also unfair and unkind. So no matter how carefully we try to describe the situation in a way that isolates the proper moral concept, there will always be situations where the same pro-attitude is forced to ground multiple moral judgments that are obviously distinct.3

Stepping back, we see an interesting sort of symmetry. When it comes to the first account, our disapproving attitudes just aren’t fine-grained enough to ground the full range of our varied moral concepts; meanwhile, in the second account, it’s our ability to make sufficiently fine-grained distinctions in our descriptions that falls short. Personally, I’ll have to agree with Judith Jarvis Thomson when she says, “It seems to me that the prospect of supplying either one of the required accounts is most kindly describable as dim.” The structure and function of moral discourse, it seems, is just fundamentally more complex than our reactive attitudes alone.

Now, if you’ve been following along closely (or if you just read the first paragraph) then you might have noticed something: These are all big problems for emotivism too! In fact, this basic objection was originally meant to challenge non-cognitivist theories, rather than relativist ones. But given the very plausible assumption that moral frameworks are comprised of pro- and con-attitudes, the same basic issue reasserts itself. So with this in mind, my question is simple: What, exactly, would be the point of adopting relativism? After all, if making relativism work requires solving all of emotivism’s problems, plus having answers to all the extra questions that pop up when you introduce an actual property of wrongness, then why not just leave that latter bit off and cut your losses?

As far as I can tell, there are two main concerns that tend to scare people off from emotivism: The Frege-Geach problem, first off, and then the more general sense that seeing moral assertions as being non-truth-apt expressions requires denying the self-understanding of everyday speakers. (Presumably, most people who make moral judgments take themselves to be doing more than just expressing feelings that are neither true nor false.) I won’t take any time to tackle the first concern in any detail, except to say that I think modern expressivists have actually produced some pretty solid responses, and that relativism itself has its own trouble making traditional moral logic work. But what about the idea that expressivism requires some illegitimate presumptuousness about the intentions of normal people?

My response, which you might be able to predict, is that relativism has this exact same problem! Recall that relativists see moral assertions as intrinsically relative to a particular moral framework, which means that claims like “X is wrong” will always carry some elliptical standard waiting in the wings. But it seems patently obvious to me that the vast majority of everyday speakers don’t take their moral assertions to be made true solely by their own attitudes, or by anyone else’s. That doesn’t mean, to be clear, that everyday speakers are all moral realists instead — it just means that the average person has a folk conception of moral properties that doesn’t fit with the specific picture relativism paints.4

In fact, if you gave the average person an interpretive choice between seeing their moral judgments as true, but only relative to some non-truth-apt set of attitudes, or just seeing their judgments as expressing those attitudes directly, it’s not at all obvious to me which one would feel “closer” to the folk conception that underlies general moral discourse. Relativism probably wins out on structural grounds, but I wouldn’t be surprised if someone thought the sort of wrongness it provides was unsatisfying enough to make a direct reliance on attitudes more appealing. So if the problem with emotivism is supposed to be that it doesn’t match up with how everyday speakers understand their own judgments, then in my mind, relativists are throwing stones from glass houses.

None of this means that I think emotivism is particularly plausible, of course; the issues I just went over are enough for me to basically rule it out, at least as long as competing realist theories are available. But if I did end up rejecting realism, and had to go looking for an anti-realist theory to fill the gap, I just don’t see any good reason to side with relativism if I’ll be stuck with emotivism’s baggage either way. So while I can’t imagine many relativists will be coming to me for any advice on this one, I would at least encourage them to consider if the extra pain of a truth-apt semantics is actually worth it. Personally, I think anti-realists have enough to deal with as it is!

If you spend any time at all in the debate between moral realists and anti-realists online, you’ll almost certainly come across two very common claims:

First-Order Independence: Normal, everyday moral discourse doesn’t require or imply any particular metaethical commitments, and whether moral judgments are stance-dependent or stance-independent would make no major difference when it comes to how your average person navigates important ethical issues.

Evolutionary Origins: We have good reason to doubt whatever realist intuitions we might have, because we can explain those intuitions purely in terms of the evolutionary advantage that comes from widespread acceptance of objective, stance-independent categorical moral obligations.

Now, personally, I think both of these claims fall somewhere between misguided and just plain false. (See here and here if you want my thoughts on each in more detail.) But today, I won’t take any time to criticize either individually. Instead, I just want to make the obvious point that they can’t both be true at once! If First-Order Independence turns out to be correct, then Evolutionary Origins can’t be, and vice versa. Anti-realists, it seems, will have to pick a side.

Just seeing the two claims juxtaposed together should be enough to notice where the tension lies, but let’s quickly do some work to make the internal contradiction clear. Start with First-Order Independence, which says that our everyday moral lives are totally isolated from whatever second-order views we hold on the ontological status of the moral properties themselves. If this is true, then any moral concepts that play a role in everyday moral reasoning have to exist firmly on the first-order side of the divide, since otherwise they’d be vulnerable to metaethical meddling. And this means that, assuming First-Order Independence, whether or not we have any realist intuitions (or, for that matter, anti-realist ones) could have no possible impact on how we develop and apply those moral concepts.

So what’s the problem? Well, think of it this way: In order for natural selection to favor some trait, that trait has to be fitness-enhancing, and that means our realist intuitions would have only been preserved had they offered ancient humans some noticeable benefit in the past.1 But crucially, the only way for realist intuitions to offer that benefit would be by impacting those ancient humans’ first-order moral practices! After all, it’s not as if my ancestors’ evolutionary fitness ever hinged on them having purely theoretical beliefs about the ontology of moral properties; an interest in philosophy isn’t exactly known for increasing reproductive success in the present day, I’m afraid, and I doubt things were different in the paleolithic.

What did benefit those ancient humans, on the other hand, was the chance to live in a cooperative social groups where reciprocity, fairness, honesty, and so on were entrenched and widespread communal values — and what made those conditions possible was the intuitive sense that selfishness, inequality, deception, and so on aren’t just subjectively undesirable but categorically forbidden. The earliest human societies (and perhaps even many non-human ones) were able to grow and thrive so effectively in part because their members saw pro-social behavior as carrying an intrinsically motivating and authoritative prescriptivity that transcended any purely hypothetical “If you want X, do Y” framing.

If you’re anything like me, you’ll think this conception of morality is just obviously an implicitly realist one, given that it directly invokes all sorts of basic normative features that metaethical relativism explicitly denies. But First-Order Independence is committed to seeing all this talk of categoricity and authoritative prescriptivity as being entirely contained inside first-order discourse, with no connection at all to the exact same concepts as they exist in a metaethical context. I’ll leave it up to the reader to decide just how plausible that sort of division is, but all that matters here is that whatever does get left on the metaethical side, natural selection can’t be responsible for it.

We can also run things in the other direction to see the same issue in reverse: If Evolutionary Origins is true, and natural selection is responsible for giving us our second-order realist intuitions, then it just has to be the case that intuitions like those somehow make a difference in human behavior and the basic structure of human society. But the claim that second-order realist intuitions are capable of having that impact is just a straightforward rejection of First-Order Independence. If second-order beliefs about stance-independent, categorical reasons for action do play a fundamental role in how we understand our pro-social moral concepts, then it would make perfect sense for natural selection to preserve them. But then First-Order Independence would have to be false.

So either way you look at it, you’re stuck. If realist intuitions have no bearing on how our everyday moral lives play out, then there’d be no reason for evolutionary pressures to promote them. Meanwhile, if it is advantageous for a community to accept and promote widespread realist intuitions, then it simply can’t be the case that those intuitions have no impact on first-order moral reasoning, since that’s precisely where the advantage is supposed to come from. Or, to put it even more simply: First-Order Independence needs our second-order intuitions to be causally inert, while Evolutionary Origins needs those exact same intuitions to be causally efficacious. And if you’ll forgive me for slipping briefly into AI-speak, that’s not just a tension — it’s an outright contradiction.

Now, lest I fall prey to the Goomba fallacy, let me be clear: I get that moral anti-realists aren’t actually required to support either of these claims, much less both of them at once. It’s perfectly possible to believe that our second-order realist intuitions play a major role in first-order moral discourse and that evolution had nothing to do with those intuitions developing, all while still remaining an anti-realist. But that’d be a pretty weird position, honestly, and I would imagine most anti-realists find it extremely unappealing. So the better question to ask is which of the two claims, First-Order Independence or Evolutionary Origins, should end up on the chopping block.

I don’t think too many anti-realists will be coming to me for advice on this one, but for the record, I think it’s pretty easy to make either option work. If you can’t give up First-Order Independence, then you should just go the extra mile and deny that we have any second-order realist intuitions at all — in which case there’d be nothing for natural selection to explain anyway. And considering that First-Order Independence already sees literally every relevant moral concept as metaethically neutral, there isn’t much for those second-order intuitions to even do in the first place. So at least from an internal perspective, giving them up doesn’t seem like too much of a bullet to bite. (Although, from an external perspective, I would be remiss if I didn’t quickly point out that it’s, you know, totally absurd.)2

If, on the other hand, you’d prefer to keep Evolutionary Origins while ditching First-Order Independence, then the most natural move would be to embrace error theory. That way, you can affirm that our pro-social moral commitments rely on implicit realist intuitions — which provides those intuitions with the evolutionary oomph they’d need to be selected for — and then just deny that those commitments are, strictly speaking, correct. In other words, you’d be arguing that natural selection has given human beings the false (but highly advantageous) belief that anti-social behavior is categorically, stance-independently wrong. And hey, if you’ve got the stomach for such a radical move right at the start, then I think the whole thing fits together pretty well.

There is, however, one method for reconciling this whole mess that I do think has a serious issue, and I want to close things out by taking a look at it. As I just said, Evolutionary Origins fits perfectly well with error theory, which denies that moral properties actually exist at all. But as far as I can tell, most people I see defending Evolutionary Origins are just plain ol’ relativists, which means they do believe that some acts really are morally right or wrong (even if that rightness and wrongness will ultimately be relative to some particular standard). So then it’s worth asking whether these three views — affirming Evolutionary Origins, denying First-Order Independence, and traditional moral relativism — can fit together as well as you might hope.

In my mind, there’s reason to be skeptical! Here’s my worry: If Evolutionary Origins is true and First-Order Independence is false, then at least a sizeable portion of our most fundamental moral commitments have an ineliminable stance-independent categoricity baked in. But if you endorse relativism, then you’ll have to deny that any such categoricity actually exists — and that means you’ll be forced to accept that everyday first-order moral discourse is, at least in one sense, systematically distorted by ingrained intuitions that inaccurately represent the normative state of the world. Yet, at the same time, relativism says that these exact intuitions are the truth-makers for our moral claims. And that’s at least a little weird, right?

It might help to make things a little more concrete here. Let’s say you’re going out to eat with your friend Phil, and while walking to the bathroom, he overhears the restaurant’s manager say something hateful or politically offensive. So when Phil gets back to your table, he suggests getting up and leaving without paying. (And let’s also stipulate that there are no cameras or cops around, so there’s no risk of being caught if you do.) Now, hopefully, you’d speak up and say no — not because you personally care about the jerk who runs the restaurant, but because you have a more basic moral conviction that taking something without paying for it would be theft, and that stealing from someone when you don’t really have to is just plain wrong, no matter how much you’d enjoy sticking it to them.

That categorical, self-restraining commitment to pro-social behavior is a perfect example of the sort of moral intuition that natural selection would likely explain, at least assuming Evolutionary Origins. But someone who endorses relativism and denies stance-independence will also have to admit that, in reality, there are no such categorical restrictions on human behavior, and that your sense of stealing as carrying some intrinsically prescriptive wrongness is merely an evolutionarily advantageous illusion. If you didn’t have that illusion — that is, if you could only consider hypothetical, desire-based normative constraints — then it seems very likely to me that you’d have no good reason to reject Phil’s proposal.

But reject Phil’s proposal you do, and according to relativism, your (evolutionarily informed) rejection is what determines the wrongness of what was suggested. So here we have a situation where the wrongness of stealing depends on your stance towards it, which in turn depends on a widespread cognitive distortion with roots in entirely non-truth-tracking natural processes. Isn’t this a problem? Of course, if relativism of any kind is the correct moral theory, then we shouldn’t expect our moral judgments to be backed up by any sort of essential, intrinsic Moral Truth — but you’d at least hope they aren’t backed up by actual error, right?

This is an especially difficult issue to dodge because, in many other contexts, the fact that some moral view can be traced directly to a non-truth-tracking natural process is often given as a good first-order reason to abandon it! For example, it’s fairly common for people to frame homophobia as a consequence of an overactive disgust response, or nepotism as a relic of evolved kin preference, with the clear implication in both cases being that these etiologies undercut the reasonableness of the judgments in question. But if the evolutionary origins of our “undesirable” moral convictions are enough to rob them of independent weight, why not think the same way about our general intuitions regarding categoricity and overriding prescriptivity? Without the realist’s ability to call on objective moral truths as a comparative standard, it’s hard to see how even the relativist’s most fundamental moral commitments would survive this sort of debunking.

A relativist might respond to all this with something like:

Sure, natural selection has given us a sense of universal, categorical wrongness that doesn’t reflect reality. But it’s also given us the sense that external objects have various colors that aren’t truly present in those objects, and no one thinks that undermines our judgments about color. All we ever have to work with are the representational schemes that evolution has given us, and expecting us to step outside those subjective senses in order to deliberate “correctly” is just a form of crypto-realism. Even if some normative judgment carries a misleading or illusory felt sense of categoricity, the judgment’s content is still a perfectly good truth-maker, since there’s no “right way” to establish normative content (in the same way there’s no “right way” to see red).

There’s something to this response, I’ll admit, but the problem is that judgments about color can easily survive acknowledging these evolutionary origins in a way that moral judgments can’t. After all, I can affirm that my subjective experience of a sunset’s color will misrepresent the sky as having primary properties that it doesn’t really have while still enjoying its beauty, since the experience of beauty doesn’t hinge on beautiful properties “really being there.” But for someone who accepts Evolutionary Origins, the picture is different, since the whole point is that our first-order normative judgments do depend on there being actual categorical restraints on human behavior. (Otherwise, we wouldn’t have needed to develop the corresponding intuitions in the first place.)

In other words, “I know the sunset itself isn’t intrinsically orange and red, but it’s still beautiful” is a perfectly coherent thing to say — whereas acknowledging that theft isn’t actually categorically impermissible will directly undercut the internal, first-order justification required for taking a universal anti-stealing stance. And that’s because the felt sense of categorical impermissibility that accompanies the anti-stealing stance isn’t something extraneous that can be easily discarded while leaving the moral judgment intact. Instead, that categoricity is itself the normative grounds for the judgment, and once you deny it, there are no conceptual resources left over for the relativist to rely on.

There are, of course, a million ways a relativist might try to close that gap, and I think many of those responses are at least plausible within a broader anti-realist framework. But with that said, I do think this tension between relativism and an evolutionary account of our second-order moral intuitions is obvious enough that it deserves much more attention than I see it getting now — especially since the solutions to it will end up emphasizing all sorts of otherwise unremarked-upon conflicts between distinct anti-realist frameworks that are too often carelessly grouped together.

This is the final entry in a three-part series on the fine-tuning argument. In the first entry, I gave my reasons for thinking that both theism and naturalism struggle to explain finely-tuned physical constants in a satisfying way. In the second entry, I argued that our most plausible account of divine psychology gives us good reason to believe these finely-tuned constants would be even less likely under theism than under naturalism.

Having given my fine-tuning argument for naturalism, I want to end this piece by addressing a common response and why it strikes me as demonstrably misguided. The response goes something like this:

Yes, a universe finely-tuned for sparse amounts of biological life in particular might be seriously implausible on theism, given the issues you’ve mentioned; however, it can’t possibly be as implausible as the original one-in-10¹²⁰ chance of the same universe coming to exist on naturalism! In general, so long as the auxiliary hypothesis theists are adopting to neutralize hiddenness concerns is significantly more probable than that one-in-10¹²⁰ chance — which almost all possible candidates will be — then the conjunction of traditional theism and that auxiliary hypothesis will still be more plausible than the story naturalism tells. Therefore, all the theist needs to do in order to salvage their fine-tuning argument is identify any possible explanation for the creation of biological life that has a non-infinitesimal chance of being true.

Call this the “but not that implausible!” objection. Put simply, the assertion is that essentially any modified version of theism will accommodate the fine-tuning data better than naturalism can, since any modification a theist could imagine will still have a higher probability than what our best current physics assigns to those life-permitting constants appearing randomly. Many theists are fond of pressing this objection as a way of driving home just how (supposedly) hopeless naturalists are in the face of the fine-tuning argument. However, while the logic of the response might be intuitively appealing at first, it quickly falls apart under scrutiny.

To understand the error involved, it’s worth taking a moment to remember where, exactly, the one-in-10¹²⁰ figure comes from. This figure is meant to represent an estimate of the chance that every physical constant present in our standard model would happen to land within the ranges required for biological life to develop, on the assumption of a uniform probability distribution in each particular case. The value of this figure, then, obviously depends entirely on the precise details of our standard model, which in turn depend on our best current physical theories.

For this reason, the most accurate way to understand the one-in-10¹²⁰ figure is as a representation of the probability that our universe’s physical constants would support biological life, given the assumption that our best current physical theories are perfectly correct. It’s crucial to note that, by stipulation, these best current theories haven’t yet been modified to address the embarrassing reality that our universe is fine-tuned. Rather, the odds given for fine-tuning are being theorized prior to (or at least apart from) the recognition that fine-tuning has, in fact, taken place.1

Given that this one-in-10¹²⁰ figure is the product of our best current physics, the proper point of comparison would be with whatever probability is given our best current theology — that is, we should ask what the likelihood of observing finely-tuned constants would be, given the assumption that our best current theories of God’s divine psychology are perfectly correct. To preserve the proper symmetry, we should also stipulate that these theological accounts are equally unmodified in response to embarrassing data. Therefore, we should consider the most reasonable overall account of God’s divine psychology before the observation of divine hiddenness is factored in.

But what is this best current theory? Unlike with physics, there is no widespread consensus among theists, much less atheists, when it comes to which account of God’s divine psychology would be most reasonable. However, all atheists are well within their rights to argue, as I have, that any remotely plausible account of this divine psychology will include an overriding and easily accommodated desire to avoid making his existence maximally obscure. If atheists are correct about this, then the probability that God would desire a world with embodied biological life in particular would be literally zero. It would simply be incompatible with a basic feature of his divine psychology.

Theists might balk at such a bold assertion, but there’s nothing illegitimate about assuming that, given a certain conception of God’s divine nature and psychology, some acts would be impossible. Most theists would, for example, hold a view of God’s nature such that it would be impossible for him to lie, or allow gratuitous evils. If atheists have reason to believe that God’s desire for relationship with his creation is sufficiently strong — or that his perfectly good nature requires him to make that relationship maximally accessible — then a similar sort of restriction on the intentional concealing of his existence would be equally valid.

Of course, this isn’t to say that any atheist should actually accept this probability overall, because it could always be the case that their preferred account of God’s divine psychology is somehow mistaken. However, by the same token, theists will have to acknowledge that the probabilities given by our current standard model might also be mistaken in a roughly similar way. What this shows is that we have to make sure we keep two very different questions separate: First, what our best current unmodified theories in physics and theology tell us when we take them at face value; and second, how likely it is that those predictions are actually correct.

With this distinction in mind, it should be very easy to see where the “but not that implausible!” objection goes wrong. If we wanted to do things properly, we would compare the probability for fine-tuning given by our unmodified standard model with the probability given by whatever unmodified account of divine psychology we find most compelling; alternatively, we could compare the probabilities of the two theories after both have been modified to align with our observations. What we can’t do, however, is take the one-in-10¹²⁰ probability given by our unmodified physical theories and compare it with the probability given by a modified theism.

Unfortunately for theism, when we do take care to keep these things straight, it’s very possible (at least from the naturalist’s perspective) that naturalism will come out on top regardless of which comparison we choose to make. If our best current theology — that is, the most plausible overall account of God’s divine psychology and his relationship with creation, plus a few other related concerns — rules out the possibility that he would actively obscure his own existence, then any chance of observing a universe finely-tuned for plausibly naturalistic biological life will be greater on naturalism, no matter how small that chance might be.

If we compare modified theories, on the other hand, then what we need to ask is whether auxiliary hypotheses aimed at solving the problem of divine hiddenness are more likely to be true than auxiliary hypotheses aimed at solving the problem of fine-tuning. Here too, naturalism has the clear advantage, since its options for solving the problem include both explaining the fine-tuning data as well as just rejecting it. Regardless of how confident any particular naturalist (or theist) is in the project of modern physics, it would be unreasonable to dismiss at least some chance that observed fine-tuning is a relic of our as-yet incomplete physics that will be eliminated by some future, deeper model. Perhaps this is a one-in-twenty chance, perhaps it’s one-in-fifty, but whatever the chance might be, it’s clearly non-negligible.

Combining this possibility with companion theories that do seek to explain the fine-tuning data taken at face value — like, for example, a multiverse model — it becomes very easy for the naturalist to argue that the overall probability of fine-tuning being resolvable given naturalism is greater than the overall probability of divine hiddenness being similarly resolvable inside a plausible theistic worldview. If this is the case, then the existence of a universe with constants like ours will be more probable overall on naturalism than on theism. However, even if the theist doesn’t believe the conjunction of possible naturalistic solutions to fine-tuning is more probable, that probability is still the bar they need to clear, and not simply the original one-in-10¹²⁰ figure.

In light of this, the entire fine-tuning debate could be restated (with only a small amount of editorializing) as a question about what is more likely: That relatively recent developments in an unambiguously incomplete physical theory have given us a pseudo-problem that will be resolved at some point in the future; or whether a similarly effective solution to the perennial problem of divine hiddenness that has not yet been discovered after thousands of years of theorizing is still waiting in the wings.

I would choose the former option, obviously, even given a fairly pessimistic view of its chances. At the same time, however, I also recognize that theists may well disagree, depending on how seriously they take the issue of divine hiddenness in the first place. What ultimately matters is just that both sides are clear on what, exactly, their relative epistemic burdens are. As comforting as it might be for a theist to believe that “winning” only requires overcoming the naturalist’s one-in-10¹²⁰ chance, the odds are, in reality, significantly less favorable. So long as naturalists hold a sufficiently low opinion of the prospects for explaining away divine hiddenness — a perspective that I see as not only plausible, but seriously compelling — then they are well within their rights to think that fine-tuning is not a meaningful challenge to naturalism overall.

As an atheist, one of the most common questions I get is what it would take for me to become a Christian. How about a near-death experience? Some ingenious new argument from this-to-that harmony? Or perhaps a freshly discovered letter of Paul’s where he declares himself a male feminist? To be honest, apart from the angel Gabriel descending on the White House lawn and casting Donald Trump straight into hell, I don’t really have a good answer. But what I do have, if I may say so myself, is a really good reason for not having that good answer — and in this piece, I’ll try my best to explain what that reason is.

Let’s start by talking about 2 Peter for a bit. (If you don’t care at all about Biblical criticism, then feel free to skip down maybe twenty or so paragraphs until you see the next divider.)

Now, if you don’t read your Bible too often, then you might not have even known there was a 2 Peter — or, for that matter, a first one. Both of the Petrine epistles are solid B-tier entries in the New Testament canon, and apart from a few famous verses that play an outsized role in some long-running theological disputes, they usually end up taking a back seat to real heavy hitters like Romans or Galatians. That’s a bit of a bummer in my eyes, since 2 Peter in particular still has plenty of really good stuff for the thoughtful reader to enjoy. Plus, at only 1500 words or so, it’s the perfect size to sit down and read through in one sitting.

There’s just one problem: The actual historical Peter pretty obviously didn’t write it. Instead, we have extremely good reason to think it was actually written at least thirty years after Peter’s martyrdom, by someone deliberately attempting to pass it off as an authentic work. Both Christians and non-Christians are often shocked by this claim, and treat it with skepticism — but among experts in the relevant fields, it’s something close to an uncontroversial consensus.

There are, of course, plenty of evangelicals out there with a prior commitment to biblical inerrancy, and it’s natural that they’d disagree. A few have even managed to get degrees and write semi-convincing papers desperately working to explain the facts away. But among those scholars who are actually able to honestly follow the evidence where it leads, the conclusion is pretty straightforward: The text is a late first-century or early second-century forgery, produced by an author with no direct connection to Peter himself.

But why is this the view held by the vast majority of scholars? Well, first of all, it’s important to recognize that the sort of forgery we’re talking about was undeniably widespread in early Christian period. There are dozens of examples of apocryphal gospels and pseudepigraphal epistles from the first two hundred and fifty years of the church, and Paul himself warns against accepting false letters from imposters claiming to be him. (Although, ironically, that message is from 2 Thessalonians, a letter that some scholars believe to also be a forgery!) For example, around 200 or so, Tertullian also mentions a text that was forged by an overzealous presbyter:

But if certain Acts of Paul, which are falsely so named, claim the example of Thecla for allowing women to teach and to baptize, let men know that in Asia the presbyter who compiled that document, thinking to add of his own to Paul's
reputation, was found out, and though he professed he had done it for love of Paul, was deposed from his position.

What this shows is that, during the second century, early Christians had a habit of writing false histories and letters from important Christian figures of the past, as a way of advocating for novel doctrinal positions or addressing disputes that were dividing the church. (It also shows that other Christians at the time did not think this practice was acceptable, but we’ll get to that later.) So even before we look at any specific evidence for 2 Peter in particular being forged, we should accept that the prior probability is already pretty high.

Meanwhile, once we do start looking at the specific evidence in question, it’s hard to even know where to begin. For one thing, the early Christians themselves seem to have had no knowledge of 2 Peter for the first one hundred and fifty years after it was supposedly written. The first unambiguous mention we get of the epistle is from Origen in the middle of the third century, and even there he isn’t sure if it’s actually authentic or not! It takes another hundred years (at least) before we have record of the epistle being widely accepted by the church at large, something that makes absolutely no sense whatsoever if you think the text was actually from Peter.

It’s also worth noting that, in the book of Acts, Peter is explicitly identified as an illiterate (or at least uneducated) peasant, as pretty much all Galilean fishermen would have been at the time. It’s difficult, then, to explain why 2 Peter is written in a dense, ornate, and sophisticated style of rhetorical Greek — a language Peter wouldn’t have even spoken before meeting Jesus, much less grasped with enough familiarity to produce the text we have today even after years of study.

But let’s say you don’t find any of these external arguments to be convincing at all. That’s fine! You could realize 2 Peter was a forgery just by looking at the contents of the letter alone, and the multiple anachronisms that all point directly at a late first-century or early second-century origin. One particularly obvious example of this would be 2 Peter 3:14-16:

Therefore, beloved, since you look for these things, be diligent to be found spotless and blameless by Him, at peace, and regard the patience of our Lord as salvation; just as also our beloved brother Paul, according to the wisdom given him, wrote to you, as also in all his letters, speaking in them of these things, in which there are some things that are hard to understand, which the untaught and unstable distort, as they do also the rest of the Scriptures, to their own destruction.

Two things immediately jump out here as suspicious. First, the author mentions “all [of Paul’s] letters,” which is an odd way to describe a corpus that had yet to be collected, standardized, or made available to the (supposed) first-century audience being addressed — but one that makes perfect sense for an author writing several decades later with access to Paul’s collected works.

Even more importantly, though, the author of 2 Peter casually refers to the letters of Paul as scripture. This is a very, very big deal! Apart from this one passage, we don’t see anyone else speak this way about the Pauline epistles for at least another hundred years, and the notion that Christians in 65 AD were already seeing Paul’s words as equivalent to the Old Testament goes against pretty much everything we know about how these traditions developed. In all the undisputed New Testament books, the term “scripture” always refers to books like Genesis and Deuteronomy, never to contemporaneous writings by the apostles of Jesus or anyone else — just one more instance where the author of 2 Peter unknowingly indicates a clearly anachronistic frame of mind.

Finally, there’s also just the content of the letter itself and what the author seems to be concerned with. In 2 Peter 3 3-9, we read this:

Know this first of all, that in the last days mockers will come with their mocking, following after their own lusts, and saying, “Where is the promise of His coming? For ever since the fathers fell asleep, all things continue just as they were from the beginning of creation.” For when they maintain this, it escapes their notice that by the word of God the heavens existed long ago and the earth was formed out of water and by water, through which the world at that time was destroyed by being flooded with water. But by His word the present heavens and earth are being reserved for fire, kept for the day of judgment and destruction of ungodly people.

But do not let this one fact escape your notice, beloved, that with the Lord one day is like a thousand years, and a thousand years like one day. The Lord is not slow about His promise, as some count slowness, but is patient toward you, not willing for any to perish, but for all to come to repentance.

In other words, the author is confronting an anxiety among his audience over the delay of Christ’s return; non-Christians have started to mock believers by saying that their first generation is already dead, and yet Jesus still hasn’t come back for them. The text gestures at this sort of mockery being a future occurrence “in the last days,” but it’s pretty obvious from what follows that the guidance being given is meant for Christians living at the time it was written.

Now ask yourself: Is this the sort of concern that you would naturally address in 65 AD, thirty years or so after Jesus’ death? Not really! Around the time Peter is supposedly writing this epistle, most first-generation Christians would have still been alive and kicking, and the sorts of things that early Christians looked to as harbingers of the end times — war, oppression, instability, and so on — were all becoming more widespread, not less. If a Christian in 65 AD was mocked by an unbeliever in the way 2 Peter describes, he would be baffled; however you might describe the situation a year before the Jewish-Roman war broke out, it sure as hell wasn’t “things continuing just as they were!”

We can see this even more clearly by looking at an excerpt from the undisputed Pauline epistle of 1 Thessalonians, specifically 1 Thessalonians 4:13-17:

But we do not want you to be uninformed, brothers and sisters, about those who have died, so that you may not grieve as others do who have no hope. For since we believe that Jesus died and rose again, even so, through Jesus, God will bring with him those who have died. For this we declare to you by the word of the Lord, that we who are alive, who are left until the coming of the Lord, will by no means precede those who have died. For the Lord himself, with a cry of command, with the archangel’s call and with the sound of God’s trumpet, will descend from heaven, and the dead in Christ will rise first. Then we who are alive, who are left, will be caught up in the clouds together with them to meet the Lord in the air, and so we will be with the Lord forever.

Here, Paul addresses the same basic issue of Christians dying before the return of Christ — but this time, unlike in 2 Peter, it doesn’t even cross his mind to explain why that return hasn’t happened yet. Instead, Paul tackles a totally different fear the early Christians had, which is that those who died before the second coming would somehow miss out on the coming resurrection. This is the fear that would make sense in the context of the time, and the fact that Paul takes it for granted that his audience still has a faith in the second coming is just more evidence that the concerns we see addressed in 2 Peter are hugely anachronistic.

Put all the evidence together — including plenty of things I haven’t even covered here —and what you have is a text where every single piece of data you can point to just screams an early second-century (or late first-century at best) origin. Now, do conservatives have their strategies for addressing all these points? Of course they do. But the story you’d have to tell for traditional authorship to make good sense is just so convoluted, and so implausible on its face, that I really don’t see how it comes together without an obvious precommitment to biblical inerrancy.

And one more thing: I say “forgery” here because, as part of their efforts to avoid admitting the obvious, inerrantist scholars will often make completely ridiculous claims about literary practices in the ancient world. Most absurdly, they’ll often argue that first-century Greeks and Jews had no issues with anonymous authors attributing their work to prominent religious figures of the past, and that these false attributions weren’t considered deceptive or immoral. But this is, to be very clear, not true! It’s trivially easy to show that early church figures cared a whole lot as to whether or not a particular piece of writing was actually from a real-life apostle, and the idea that no one would have seen 2 Peter as fraudulent had they known its true origin is just flat-out cope.

In reality, applying any term other than “forgery” obscures 2 Peter’s actual nature as an intentionally deceptive text, meant to be understood as having been written by someone who didn’t actually write it. Now, did the actual author pass 2 Peter off this way for some nefarious purpose? Probably not! It’s much more likely that they recognized a profound anxiety among Christians in their day and felt like a reassuring letter from their spiritual hero would be the best way to soothe it. We see people doing the same thing today all the time. But even the best intentions don’t change the basic fact that Peter didn’t really write it.

So 2 Peter is a forgery. Who cares? What does this have to do with the question of what it would take for me to be a Christian?

Well, I want you to do a little thought experiment with me. Imagine, for a second, that you don’t know anything whatsoever about Christianity, or Christian history, or anything else like that. Now imagine that someone comes to you and tells you the whole Christian story, from the Old Testament up to the New, and then gives you a Bible to read. You see Jesus tell his disciples that the Holy Spirit will guide them into all truth. You see Paul identify all scripture as theopneustos — literally “god-breathed” — and refer to the church as the “pillar and foundation of truth.” You see Jeremiah proclaim that, in the age of the messiah, all believers will live with the law of God in their hearts and none will need to be instructed. And on, and on, and on.

Now, let me ask you: After all that, would the notion of 2 Peter being an outright forgery even occur to you? Is there anything in the Old or New Testaments, or in the early Christian tradition, that gives any grounds whatsoever to think that the scriptures themselves could be shot through with deliberate falsehoods and lies? Of course not! That’s precisely why Christians like Eusebius were so concerned about 2 Peter in the first place: Because the idea of a forgery being contained in sacred scripture was obviously unacceptable. It’s just that, unfortunately, those early Christians got it wrong, and failed in their own stated goal.

What’s crucial here is just how trivially easy it would have been for God to avoid this outcome if he’d wanted to. He could have inspired the actual Peter to write the exact same letter in 65 AD — sans all the anachronisms — and saved anyone the need to forge it later. Or he could have legitimately inspired the later forger and had him express a similar message without the fraudulent identification, and then made sure it was received by anyone who would benefit from it. And if God didn’t feel like it was something that needed to be shared around at all, he could have guided the church to identify the forgery as false, as he promised he would do, rather than permitting it to unknowingly enter the sacred canon.

But God didn’t do any of that. Instead, he allowed the church to fall into universal error for no clear purpose, despite this obviously undercutting the probability of Christianity being true. I know some theists will try to deny this point, but come on! Is it really controversial to say that the presence of deliberate, unrecognized falsehood in a religion’s scriptures makes that religion’s claims less likely? Is it really some unfair, ultra-skeptical standard to expect that the church God himself established on earth would be able to discern and exclude the sorts of texts that they uniformly saw as unfit for acceptance?

No one is saying that 2 Peter’s presence in the canon disproves Christianity all by itself, of course — but arguing that it has no impact whatsoever on the likelihood of Christianity being true is even sillier. This makes 2 Peter’s acceptance a perfect example of what I like to call an anti-miracle: Something that, if taken seriously as an act of God, seems purposely designed to make unbelief rational. If the God of Christianity does exist, and has anything even resembling the relationship with his church that scripture explicitly asserts him to have, then he allowed that very church to irrevocably fail at an essential task he promised to guide them through. If that doesn’t lower the probability of Christianity, in the same way an apparition or healing might raise it, then it’s hard to see how anything could.

Once you go looking for anti-miracles like these, it’s not too long before you start finding them everywhere. For one thing, 2 Peter isn’t even the only likely forgery to wind up in the Bible! Critical scholars also make convincing arguments for the presence of similar anachronisms in the pastoral epistles — 1 and 2 Timothy and Titus — and even 1 Peter, Ephesians, James, and others are doubted by many as well. There are also major historical errors throughout the gospels and Acts that point towards a history of composition that contradicts church tradition, and of course huge portions of the entire Old Testament contradict the near-universal consensus of archeologists (not to mention our near-universal moral consensus against genocide, slavery, the abuse of women, and so on).

This is, let’s admit, exactly what you’d expect to see if God did not exist. And that means that, if God does exist, then he’s going far beyond just “leaving us room for doubt” or “encouraging sincere seeking” — instead, he’s actively working to make sure those sincere seekers have good reasons to give up on their search! So please, Christians, don’t disingenuously reduce these complaints to nothing but “Hey, it looks like there are some errors in the Bible.” The problem is so much bigger than that: It’s that the errors we’re talking about serve no purpose whatsoever beyond clearly maximizing the rational case for rejecting what God supposedly wants us to accept.

Any way you slice it, that’s a theological problem way before it’s a historical one. If the Christian God really does exist, and really does desire for me to know him, then his decision to seed history with so many utterly pointless opportunities for rationally compelling doubt would be totally inexplicable even on his own terms. So this isn’t me, a mere man, talking back to God and asking why he isn’t doing things the way that I’d prefer — it’s me, a mere man, asking why God isn’t doing things the way that he said he prefers. We’ve got Jesus on record, after all, telling his followers that the spirit will lead them into all truth. Is it really so weird that someone like me would take him at his word?

This, more than anything else, is why I’m not a Christian: Not because there’s any one single thing I can point to that immediately rules it out, but because there are so many little reasons for disbelief that have piled up over history. And it’s not just that all those little reasons ultimately add up to a broader cumulative case against the faith (although they do) — it’s that the existence of all those little reasons for doubt is itself a devastating issue, over and above whatever lesser issues combine to create it. In other words, while the 2 Peter’s presence in the canon might make Christianity only a little less likely on its own, the idea that God would allow the likelihood of Christianity to be diminished in such an embarrassing, internally inconsistent way drags down its probability to a significantly greater degree.

For the record, this is also why I’ve always been dissatisfied with the term “divine hiddenness” being used to describe our epistemic relationship to God’s existence. In reality, if the Christian God does exist, then he isn’t exactly shy about revealing himself in history, or in the text of his sacred scripture, or even (so the story goes) through indisputably miraculous intervention. So to the degree that he still ends up “hidden,” it’s only because he’s chosen to counteract that explicit revelation by engineering equal and opposite reasons for doubt. And that’s not being “hidden,” so much as just being schizophrenic.

Had God so desired, he could have easily remained “hidden” in the classic sense of the word by just avoiding decisive evidence in either direction; he could have created a world with neither horrendous evils nor awesome miracles, pushing us towards theism with nothing but the overall goodness of the world or some inner sensus divinitatis tuned to the proper strength. But instead, the Christian God has landed us in the supposed epistemic sweet spot by providing a jumbled mess of evidence that just ends up canceling out. To excuse the clumsy metaphor, he isn’t like a nervous first date, dropping little hints to avoid coming on too strong; he’s more like a guy who proposes the moment you sit down, and then spends the rest of the night rolling his eyes at all your jokes and telling you your breath stinks.

Inside that epistemic context, I just can’t find a story that makes it all work. And that means there isn’t any one piece of bombshell evidence, or any one irrefutable argument, that I’m sitting here waiting on before I’ll believe. Instead, what I need is just some way to understand how the distribution of evidence is even remotely compatible with God’s existence in the first place; I need to understand, in other words, why a God who desires for me to believe would go out of his way, over and over again, to invent new reasons why I shouldn’t.

Until that fundamental mystery is solved, asking me to tally up a bunch of first-order pros and cons I can’t even explain just feels like missing the point. So who knows? Maybe tomorrow, I’ll have some miraculous experience that really truly feels divine. But if that does happen, do you know what my first thought will be? “Hey, thanks for the show, Father, but you didn’t need to warp the very fabric of space and time to help me consider that your revelation is true — next time, just let Eusebius know that 2 Peter was a fraud!”

This is the middle entry in a three-part series on the fine-tuning argument. For context, you can read the first entry, in which I give reasons for thinking finely-tuned constants are equally unlikely given either theism or naturalism.

Having seen that both theism and naturalism struggle to give an easy explanation for our finely-tuned constants, we might be tempted to declare a draw; maybe the entire fine-tuning argument is just dialectically inert, and provides no real benefit to either side of the debate. Although this conclusion would be good news for the naturalist by itself, I think an opportunity exists to go even further — atheists can push a modest fine-tuning argument against theism and for naturalism.

How might this argument work? Let’s start by considering two possible worlds, both of which (as we’ve seen) are live options for how God might create:

Celestial Kingdom: The physical laws of the universe are such that, across the whole of spacetime, nothing physical exists beyond occasional diffuse clouds of hydrogen gas. However, at the same time, every cubic inch of that physically empty space is entirely saturated by unfathomably many disembodied spiritual beings who exist in eternal spiritual societies alongside various spiritual goods.

Terrestrial Kingdom: The physical laws of the universe are such that trillions of galaxies eventually form, each containing hundreds of billions of individual stars and planets. Over time, small pockets of biological life gradually develop on a vanishingly small proportion of these planets, while the rest of existence remains entirely and eternally lifeless. The creatures that do exist in this world are born, grow old, and die in physical bodies composed of whatever physical matter is present in their surrounding environment.

In both of these worlds, atheism would be a reasonable position for at least some well-informed, rational agents to take. (Although exactly how reasonable it would be depends on plenty of other factors beyond just the natural laws.) Still, all other things being equal, it’s likely that atheism would be significantly more plausible in the Terrestrial Kingdom, for at least three reasons.

• Most obviously, it would be impossible for creatures in the Celestial Kingdom to doubt that disembodied minds exist, which immediately rules out all forms of atheism that involve a commitment to physicalism or naturalism. Atheists in the Celestial Kingdom could still rely on other non-naturalistic, axiarchist theories instead, but the range of available options for rejecting theism would be significantly more limited.

• The dense, uniform distribution of conscious minds in the Celestial Kingdom would also make it much easier for theists to argue that reality was ordered intentionally towards the production of conscious agents. On the widespread assumption that producing conscious agents is something God would want to do, this would be very strong evidence for theism, and many countervailing examples of apparent cosmic indifference that are available to atheists in the Terrestrial Kingdom would be completely off the table.

• Finally, atheists in the Celestial Kingdom would have no easy way to provide an illuminating explanation for their own existence in the first place. In the Terrestrial Kingdom, atheists can point to evolution through natural selection as a plausible candidate for an unguided mechanism by which consciousness could develop from simple, non-conscious precursors. However, absent some parallel process of “mental evolution” in the Celestial Kingdom, the best its atheists could do is rely on the costly claim that complex, widely distributed conscious minds are just a brute feature of reality.

What this shows is that certain sets of natural laws are capable of making God’s existence appear more or less obvious to conscious creatures. Laws that allow for huge numbers of psychologically complex but metaphysically basic minds to be evenly distributed across all of creation increase the likelihood of theism, while laws that only allow for small numbers of physically contingent minds to develop over time make theism less likely — which, in turn, makes atheism more rational for creatures in that world.

In fact, if we were to imagine the sort of laws that would make atheism maximally rational, they would likely resemble those that govern existence in the Terrestrial Kingdom: Laws that result in the sporadic development of plausibly naturalistic biological organisms. These laws would allow for reasonable skepticism towards the possibility of both disembodied minds and life-directed creation, while also providing satisfying non-theistic explanations for the existence of consciousness. Moreover, the existence of biological organisms would also ensure many unique opportunities for pointless suffering (disease, aging, injury, and so on) that would give additional support to other non-cosmological arguments for atheism.1

The twist here, of course, is that the Terrestrial Kingdom is just the actual world we find ourselves in today — which is to say, if God did create this universe, then he chose to create in such a way as to make atheism a maximally rational position for his creation, at least up until the discovery of the finely-tuned constants. This is a serious issue for all defenders of the fine-tuning argument, and especially for Christians, Muslims, and other members of historical monotheist traditions who hold that knowledge of God’s existence is a tremendously valuable good.

To understand the dilemma these theists have to deal with, it’s worth considering another popular objection to the fine-tuning argument that I don’t think succeeds. This objection, which is often referred to as the objection from divine psychology, argues that we can’t know enough about God’s motivations to predict whether he would even create at all, much less whether he would create the sort of creatures who need finely-tuned constants to survive. Therefore, the atheist can argue, the fact that we see finely-tuned constants in our own universe can be no evidence for the existence of God, since the existence of God doesn’t make those constants any more likely.

A common response to this objection would be to argue that God’s psychological properties aren’t entirely obscure, and that we can confidently determine at least some of them purely on the basis of his tri-omni nature. Most importantly, theists will argue, we can conclude that God’s perfect goodness would guarantee a strong desire to bring about conscious minds, given that conscious minds are capable of realizing incredible amounts of otherwise unobtainable goods.

For reasons I’ve already given, I think the leap from the value of conscious minds to the value of specifically biological life is completely illegitimate. However, even as an atheist, I’m inclined to agree with theists on the basic assertion that an all-powerful and perfectly good being would desire to create other conscious beings as well; it does seem to me that an individual being’s perfect goodness includes a desire to maximize overall goodness, and that creating conscious beings would be at least one obvious and uncontroversial way to do so. This conclusion, however, is hardly good news for theists.

Why is it not good news? The issue is that, once we accept God’s desire to bring about as much goodness as possible, we’re committed to acknowledging another desire as an obvious consequence: A desire to reveal his existence to his creation. On every plausible version of theism, conscious beings entering into relationship with God would be an extremely valuable good, and such a relationship requires those creatures to first be aware that God exists. This means the primary reason we have for believing that God would act to create conscious creatures in the first place is also a strong reason to believe that God would create them in such a way as to maximize the obviousness of his existence.

Yet, given what we’ve just seen, it’s clear that bringing about biological organisms through plausibly naturalistic physical processes — the only processes that would require finely-tuned physical constants to succeed — would accomplish precisely the opposite. This leaves advocates of the fine-tuning argument in a very perilous position. For their argument to succeed, it has to be the case that our insight into divine psychology extends up to the point where we can reliably assume that God desires to create, but not so far as to reliably assume that God desires to make his existence maximally clear.

I see no principled reason to think our insight exhausts itself at such a convenient stopping point. In fact, the idea that God’s goodness would compel him to create at all seems more contestable than the idea that, once he did create, he would then necessarily provide the ultimate good to that creation. There are, of course, many theistic strategies for explaining why God’s existence would be hidden. However, the vast majority of these strategies are clearly ad hoc, and rely on a wide range of theistic assumptions that no naturalist considering the fine-tuning argument can be expected to accept. (For those who disagree, I’ll have more to say about this in Part 3.)

For this reason, any naturalist who is unconvinced by theistic solutions to divine hiddenness should feel comfortable assuming that God would have a strong bias against creating conscious creatures through the sorts of physical processes that require a world with finely-tuned constants. This is a very impactful conclusion, since it fundamentally alters the probability calculus of the fine-tuning argument overall. Prior to this conclusion, we had good reason to believe that the existence of finely-tuned constants would be roughly equally likely on either naturalism or theism. However, learning that God has an additional bias against creating finely-tuned worlds tips the scales past neutrality and back into naturalism’s favor.

We can see this impact more clearly by returning to the hotel analogy from the previous piece:

Imagine, for a moment, an unfathomably massive hotel with trillions upon trillions of equally comfortable rooms. Under normal circumstances, guests will have their room assigned by the manager, who has no reason to put them in any one room over another. However, there’s also one single room in the entire hotel that has a door with a broken lock, and it would be possible for someone to enter this one single room without being assigned there by the manager. Now, imagine that you wake up with amnesia in that exact unlocked room. What would be more likely: That you were assigned to it by the manager, or that you entered on your own?

Assuming you have only one chance to choose a door on your own, it’s easy to see that the probability of randomly choosing the room with the broken lock is equal to the probability of being randomly assigned to that same room by the manager. Things change, however, if we stipulate that the manager strongly desires to make it clear that he was the one who chose your room. In that case, he would actively avoid placing you in the one room that made it possible to doubt his involvement, and you would therefore have very good reason to conclude that you had entered the room yourself. Crucially, this is true even if the odds of you randomly picking the room in question were still astronomically low when considered in isolation.

The same is true when it comes to the question of divine fine-tuning. Since God would have a strong desire to make his existence known to his creation, he would have a positive reason to avoid creating conscious minds in ways that maximize the possibility for rational doubt. Yet, as we’ve seen, physical constants that allow for the sporadic development of plausibly naturalistic biological creatures would be precisely those doubt-maximizing features. Therefore, we have good reason to believe that existing in a universe like ours is even less likely given theism than it would be under the uniform probability distribution naturalism predicts.

That’s all for now — in Part 3 of this piece, I’ll address a tempting theistic response to this argument, and explain why it’s demonstrably misguided. Stay (finely) tuned!

In Neil Sinhababu’s wonderful paper “Divine Fine-Tuning Vs. Electrons In Love,” he identifies these two premises as the basic elements of a fine-tuning argument:

1. For intelligent life to exist (as is the case), the fundamental physical constants must have values within very narrow life-permitting ranges.

2. If fine-tuning is true, it’s more probable that God set the values of the constants than that they took those values without God’s intervention.

In this piece, I want to explain why I think both of these premises are false, or at least would be false if theism were true. My argument has nothing to do with any alternative, non-theistic explanation that can solve the fine-tuning problem on its own. I’m happy to grant, for sake of argument, that naturalists like me don’t have an easy explanation for why the constants are the way they are.

Instead, in this piece, I want to argue for three more modest claims: First, that no one has an easy explanation for why the constants are the way they are; second, that the problem is even worse for theism in at least one important respect; and third, that a common strategy theists use to avoid admitting so is demonstrably misguided.

Let’s start by recalling the bad news for naturalists: The best physical theories we currently have really do seem to suggest that our existence is wildly improbable. I won’t waste time going over all of the technical details, but suffice to say, almost any combination of physical constants other than the ones we actually see would have quickly resulted in a universe totally inhospitable to biological life. Despite this, we do exist, and that means it would be nice to have an explanation for why it is that our constants have landed so perfectly inside these unimaginably narrow (and unimaginably lucky) ranges.

Unfortunately, it’s not obvious what sort of explanation for this cosmological data there could even possibly be given naturalism. Whatever process was responsible for fixing the value of our universe’s constants, it was presumably uninterested in whether or not biological life would one day emerge. Yet chalking these values up to either chance or brute necessity is radically unsatisfying on its own. Of course, most naturalists (myself included) will argue that these aren’t the only options for explaining the data. Nevertheless, the fact that we don’t yet have a fully developed and empirically testable explanation on hand is still a serious embarrassment.

This is what you might call the first half of the fine-tuning argument, and I would imagine we’re all familiar enough with it by now. The bad news for theism, on the other hand, comes in the second half: It’s also not at all obvious that the existence of God can explain these finely-tuned constants any better than naturalism can. Of course, it’s easy to tell a nice little story about God being good and conscious creatures being valuable, and then leave it up to everyone else to work out all the details — but the devil, as they say, is in those details, and making the argument fit with theism’s other metaphysical commitments is harder than you’d think.

Most notably, on theism, the first premise of the fine-tuning argument is simply false: It just isn’t the case that the existence of intelligent life requires constants to be finely tuned towards any particular value. Finely-tuned constants are very likely necessary for biological life to develop, of course. Without stars, planets, or certain essential elements, there would be no opportunity for complex physical organisms to evolve over time. However, if theism is true, then complex physical organisms aren’t needed for intelligent minds to exist. In fact, if theism is true, then nothing physical is is needed for intelligent minds to exist.

This might sound like a radical claim to some theists, but it’s really nothing more than a straightforward implication of the idea that God himself is a non-physical, disembodied mind. If one such disembodied mind exists, then it should obviously be within any omnipotent God’s power to create as many similarly disembodied minds as he’d like. In light of this, it’s easy to see that any finely-tuned constants are already dispensable, since God could always bring about an unlimited number of arbitrarily complex conscious minds without establishing any physical reality at all.

Moreover, if these disembodied minds are metaphysically possible, then whatever physical correlations exist between mental states and physical states in our world must be necessarily contingent; since minds don’t have to interact with anything, what they do interact with is a matter of God’s free choice. For us, God happened to decide that our minds would interact with brains — and that, in turn, required tuning the constants in such a way as to allow for brains to arise. However, God could just as easily have established psychophysical laws that allow for the same interactions to take place between minds and gallbladders, or toenails, or piles of rocks, or diffuse clouds of helium gas.

For this reason, no theist can coherently claim that the values of the constants we observe are in any sense required for intelligent life to emerge. Since the laws that control psychophysical interactions can themselves be “finely tuned” to allow for consciousness to arise from whatever the physical constants end up producing, no values for the physical constants, by themselves, are any more or less amenable to the emergence of conscious minds.

This is a significant problem for theists, since it undercuts their explanation for the values of the constants that we see. If only biological creatures could develop conscious minds, then there would be a clear and non-arbitrary reason for God to ensure that our constants fell within the narrow ranges that permit biological life to emerge. If conscious minds are equally compatible with any physical constants, however, then God’s desire to bring about conscious minds gives us no reason, by itself, to see any particular set of values as more or less likely. Therefore, without some further argument for the value of biological life in particular, the theist has no better answer for our specific cosmological data than the naturalist.

In response to this point, theists will often argue that embodied biological life is uniquely valuable, and that God has additional reasons for preferring physically embodied conscious minds. This position is hard to defend, however, for at least three reasons. The most obvious issue is that God himself is a disembodied mind that necessarily contains every possible perfection. We should, therefore, be very skeptical of the claim that physical embodiment could be required for obtaining any important goods, given that theists are already committed to the position that every possible good can be (and has been) maximally realized by a non-physical mind.

Secondly, as the author of the psychophysical laws, God has the power to ensure that any valuable mental state will be available to his creation. Consider a few of the examples Sinhababu gives in his original paper on this subject:

Protons and electrons could yearn to be together, feeling delight at the presence of the other as their opposite charges drew them closer. When they formed a hydrogen atom, they could fall in love. Whenever two electrons were a prime number of centimeters apart, they could have the mental states involved in heartfelt communication about their histories. Every subsequent time they were a whole number of meters apart, they could fondly remember each other. The remaining strong nuclear force between protons, despite being too weak to bring them into one atomic nucleus, could at its moments of greatest intensity realize a tantalizing but forbidden attraction. When any six particles formed a regular hexagon, they could share awe at the grandeur of the universe.

To reject this reasoning, theists would need to argue that certain morally valuable conscious experiences are necessarily impossible to experience apart from a very particular sort of embodied, biological existence. This is difficult to defend, given God’s omnipotence and (as we just discussed) his own ability to realize all value in his own disembodied state.

The final issue is that, even if some essential good did truly require the existence of physical bodies, God isn’t required to bring those bodies about through lawlike physical processes. Rather, he could simply sustain them miraculously. Although opinions on this question differ, I would imagine most theists who accept a post-mortem state don’t believe that state will be governed by laws or regularities that could be neutrally described in non-theistic terms. If this sort of world is possible after death, then it would also be possible for God to create a similarly miraculous world for his creation from the start. Therefore, we have no reason to believe that finely-tuned constants are necessary even if physical bodies are required to realize certain essential goods (which is already difficult to defend).

In summary: It’s very likely that, given God’s perfect goodness, he would desire to create morally valuable conscious minds. On this, theists are correct. Their error, however, is the assumption that creating morally valuable conscious minds would require finely-tuned constants that allow for embodied biological life to develop. Once we reject this assumption, the finely-tuned constants we do observe become just as inexplicable on theism as they are on naturalism.

You might start to notice an ironic symmetry here. For naturalists, the question to answer is why we exist at all — and yet, given that we do exist, our existence as biological creatures is essentially guaranteed. Meanwhile, theists find themselves stuck in the opposite position. To them, existence itself is essentially guaranteed, while the unanswered question is why we exist as biological creatures specifically. What’s crucial to note is that the probability of the overall conjunction will be the same for both sides; the only difference will be which half is considered extremely likely and which half is considered extremely unlikely.

A simple analogy might help to drive this point home. Imagine, for a moment, an unfathomably massive hotel with trillions upon trillions of equally comfortable rooms. Under normal circumstances, guests will have their room assigned by the manager, who has no reason to put them in any one room over another. However, there’s also one single room in the entire hotel that has a door with a broken lock, and it would be possible for someone to enter this one single room without being assigned there by the manager. Now, imagine that you wake up with amnesia in that exact unlocked room. What would be more likely: That you were assigned to it by the manager, or that you entered on your own?

In my mind, the most obvious conclusion to draw would be that you had checked each door individually, and eventually opened whichever one had a broken lock. (Arguably, this is comparable to a naturalistic multiverse explanation.) However, stipulating that you only have one chance to check a door, then it should be clear that the probability of you randomly choosing the door with the broken lock and the probability of you being randomly assigned to the door with the broken lock are the same. All that changes is whether the extremely low chance in either case comes from your choice or the manager’s.

Similarly, when it comes to our existence in a universe finely tuned for biological life, all that differs between naturalism and theism is why, exactly, that existence is so unlikely. For the naturalist, what’s hard to explain is how we “got a room,” so to speak, at all; for the theist, what’s hard to explain is why we got this room and not any of the others that God could have chosen. These two situations might feel very different, since we’d exist no matter what in one case but be almost certain to not exist in the other. Still, when we consider only the probability of each specific outcome obtaining, both theism and naturalism fail to explain our incredible luck — and this means the fine-tuning argument simply cannot be the sort of evidence for theism that its proponents present it as being.

That’s all for now — in Part 2 of this piece, I’ll give my reasons for thinking that theism is actually in a worse spot than naturalism when it comes to explaining the cosmological data. Stay (finely) tuned!

Some virtue ethics discourse has broken out over the last few days, and I’ve been surprised to see a few people I wouldn’t expect throwing their hats into the ring. (check out this piece by Joseph Rahi in particular.) This is wonderful news, since I’ve thought for a long time now that virtue ethics should be seen as the default, obvious moral theory and that very few objections to it really end up landing. It’s good to know the right view is more popular than I thought!

Unfortunately, however, I don’t find writing about virtue ethics very interesting myself. Part of the problem is that, as I said, so many of the arguments against it are just fundamentally confused, and there are only so many ways to repeat over and over that critics are totally missing the point. But it’s also that even relatively theoretical takes on virtue ethics end up intersecting with all sorts of first-order questions about the good life and what it means to live righteously, and that’s not always the kind of thing that fits well in a quick little Substack post. Personally, I save my deep moral wisdom for a place where it really belongs, like Twitter.

Still, I thought it might be fun to celebrate Virtue Ethics Week anyway by looking at one objection that I do think we ought to take seriously, which is the question of why we’d even care about the virtues themselves, apart from the role they play in bringing about good consequences. In my experience, many consequentialists just find the entire concept baffling. “Sure,” they’ll say, “bravery, compassion, self-restraint, and so on are valuable whenever they’ll ensure that things turn out best. But in situations where being cowardly, callous, or impulsive will lead to a better outcome, what possible reason could you have for preferring a worse future just because it was brought about ‘virtuously?’ Aren’t virtues only good in the first place because they actually make things better?”

This objection had a hold on me for a long time, up until I actually sat down and read the work of virtue ethicists themselves — at which point, it wasn’t too long before consequentialism is what started sounding ridiculous. In this piece, I won’t take a shot at offering any overarching defense of virtue ethics itself, or giving a serious critique of consequentialist alternatives. Instead, my goal is just to show where that shift in my mindset came from, and to convince consequentialists that approaching things from the perspective of virtue at least makes sense.

To start, let’s steal an example straight from Philippa Foot’s incredible defense of virtue ethics, Natural Goodness: The oak tree and its roots. (Judith Jarvis Thomson also uses oak trees as an example in her equally incredible book Normativity, for some reason, so I guess it must be one of the more philosophically insightful trees out there.) If you live anywhere in North America, Europe, or Asia, you’ve likely seen these beautiful behemoths at some point in your life. But if you haven’t, just know that they’re notorious for their dense, expansive root systems that can often stretch hundreds of feet from the base of the trunk and deep into the soil.

For any particular oak tree out there, we can ask if it has good roots — and this, in turn, requires figuring out what would make an oak tree’s roots good or bad in the first place. I’m certainly no botanist, so I can’t hazard much more than a guess at what exactly the answer would be. But whatever that answer is, we can be sure it’ll have something to do with the positive role those roots play in the health and flourishing of oak trees as a species. At the very least, good roots will need to fully penetrate the soil, absorb water and nutrients, keep the tree stable, and otherwise help the oak tree thrive.

So far, so good. But let’s complicate things a little by imagining a pair of oak trees growing on the site of an old chemical factory. The first has robust roots that can draw plenty of water from deep underground, while the second has stunted roots that barely reach below the surface. However, because the site they’re planted on has been soaked with industrial waste, the aquifer below is now highly toxic; as a result, the oak tree with deep roots ends up being poisoned, while the one with shallow roots can hang on by surviving exclusively on fresh rainwater.

Now ask yourself: Does the oak tree that survived have good roots? Intuitively, it seems like the answer is still no. Having withered roots turned out to be good for the tree, of course, but that doesn’t mean the roots themselves were good. Instead, it just means that sometimes it’s good for a tree to have bad roots, much like it’s good to have a bad knee when a war breaks out and they start looking for people to draft. Meanwhile, when you judge the roots as roots, they’re still bad ones, even if their badness just happened to stave off something worse.

But hold on a second — didn’t we just say that good roots are the roots that help trees flourish? Last time I checked, nothing messes with your flourishing quite like being poisoned. So if the withered roots helped to keep the tree from dying, then why isn’t that enough to call them good? Why reserve that designation for the deep roots, when their “goodness” is what ends up killing the tree?

Here, I think the answer is pretty simple: Once we make the goodness of the roots depend entirely on how things actually turn out for the tree, then we’re not really talking about the goodness of the roots anymore. Instead, we’re just talking about the goodness of the outcome, and then imputing that goodness back to the roots themselves. But everyone already knows that it’s good when trees aren’t poisoned, right? It’s not as if, when we ask whether a tree has good roots, we’re just looking for an evaluative ranking of all the different states that it’s good for trees to be in.

Instead, what we’re asking is whether (as I said before) the roots are good as roots — and that means our answer shouldn’t depend on any contingent circumstances entirely external to the roots themselves. Rather, it should be determined by the actual features of the roots and how they measure up to a general root-specific standard. This sort of goodness isn’t the only thing that matters when it comes to evaluating the oak tree’s situation overall, of course. But it is the proper thing to consider when you’re asked to evaluate whether the roots themselves are good.

In our everyday lives, we constantly make this distinction without even thinking; we say things like “It’s good that we were matched with such a bad team,” or “It was bad to have such a good movie playing when I was supposed to be studying.” In general, it’s just a basic aspect of our commonsense normative framework that judgments about “a good X” aren’t taken to be judgments about whether the X in question will bring about some generally good state, but rather as questions about whether the X is a good one of those — that is, whether it has the features that an X ought to have.

I hope nothing I’ve said up until now has been particularly controversial, at least when it comes to oak trees. But here’s the twist: If you’ve been nodding along so far, then congratulations! You have everything you need to understand why virtue ethics makes sense.

In my mind, the central insight of virtue ethics is this: Just like an arborist might be concerned with the question of what constitutes good roots for an oak tree, an ethicist is properly concerned with the question of what constitutes good action for a human being. As agents in the world, we don’t just choose which futures to actualize from a distance, like we’re sitting back and pressing the OUTCOME 1 or OUTCOME 2 button in the abstract. Instead, we bring those outcomes about ourselves, by acting — and it’s precisely this acting that ethics as a project exists to guide.

But in order to evaluate the goodness of human action in this way, we’ll first have to step back and ask of ourselves the same questions that we asked about the oak tree and its roots: What, exactly, does it mean for a human to flourish, and how do our actions contribute to that flourishing? This is where the virtues come in. To a virtue ethicist, a central aspect of human flourishing is cultivating and exercising these qualities of compassion, bravery, wisdom, generosity, and so on — virtue is as much a part of proper human living as a dense canopy and sturdy trunk are for the oak tree.

In other words, the virtue ethicist sees virtues as central to the well-lived human life and constitutive of proper human action, regardless of whether the outcomes they bring about are necessarily best when considered apart from those actions themselves. Of course, prudence, benevolence, rationality, and vigilance are all essential aspects of a well-lived human life, which means proper consideration for the outcomes our actions bring about is something that every virtue ethicist can and should expect from a moral agent. But to see those outcomes as the end-all, be-all of moral reasoning is to fundamentally misunderstand the question of what makes an action good as an action, and a human person good as a human person.

Let’s take, for example, the classic question of whether it would be right to arrest and torture an innocent man if you could be absolutely sure it would reduce crime rates overall. (Let’s also assume this crime-free future is legitimately better than a future in which the man hadn’t been tortured, even accounting for his individual suffering.) So long as you stipulate enough to close off all the easy escape routes, consequentialists are going to be stuck saying yes, you ought to. And if you ask why torturing the innocent man would be the right thing to do, all they’ll have to gesture at is the goodness of the outcome torturing produces.

But just like the goodness of the oak tree’s roots can’t just be the goodness of the oak tree’s survival in disguise, virtue ethicists will say that the goodness of the act of torture can’t be given by the goodness of whatever that torture secures. When we ask whether torturing an innocent man is a good way to act, we aren’t trying to figure out which of the two possible outcomes would be better overall — the scenario already stipulates that directly! Rather, what we’re asking is whether the torture of the innocent is good as an act, and that question can only ultimately be answered by examining the intrinsic features of the act and how they relate to the goodness of the human being who performs them.

(Compare this, again, to how we consider what constitutes good roots for an oak tree: First, we figure out what it means for the specific life-form of an oak tree to flourish, and then we consider what role the roots play in the oak tree reaching that state — not individually, in any particular circumstance, but in terms of the very life-form the question is oriented towards in the first place.)

From this perspective, the consequences of the action won’t be totally irrelevant, since, as I said before, a good human being ought to have concern for the harms that will come if the innocent man isn’t tortured. But at the same time, that won’t be their only concern, since a good human being also cares about fairness, justice, and plenty of other equally virtuous ethical commitments. The decision to torture an innocent man might come from legitimate concern for the community at large — but it also expresses reckless, unjust disregard for the dignity of the innocent person being tortured, and it’s exactly this sort of conflict that first-order ethical reasoning exists to handle.

The question, then, isn’t so much whether the good consequences of torture will “outweigh” the injustice and cruelty, but rather whether one virtue (the virtue of benevolence) will outweigh all the other virtues that the torturing would violate. In this way, virtue ethics can naturally accommodate the intuitive importance of outcomes, while at the same time giving us a coherent explanation for why it is that other considerations seem to matter too. Aiming at good results is a major part of what it means for an action to be good as an action, after all — but so is aiming at fairness, justice, and a respect for human dignity.

Now, to the consequentialist, I realize this might all sound like one big exercise in begging the question. “Sure,” they might say, “I get the point you’re making about evaluating actions as actions — but why think there’s anything that makes actions good as actions beyond bringing about good consequences? Tree roots aside, you haven’t given any argument for why these virtues would matter intrinsically. Why can’t I just say that actions are the sort of thing that ought to bring about good consequences, and that’s it?”

To this, I respond: Fair enough! Like I said at the start, nothing here is meant to be a knock-down argument for virtue ethics being indisputably correct. Even if you accept that the goodness of actions and the goodness of consequences aren’t identical, there’s still plenty of work to be done to identify what else the former’s goodness might consist in. And that work will require, among a bunch of other things, answering complex questions about what sort of creature a human being is and providing a fleshed-out theory of human action and how it intersects with intention, desire, motivation, and so on.

Virtue ethicists like Michael Thompson, Rosalind Hursthouse, Martha Nusbaum, John McDowell, and Philippa Foot have all done great work to develop their own answers to these questions, and I would encourage anyone who’s unconvinced to take a look at what they’ve written. It probably won’t appeal to consequentialists at first glance, but it’s all still worth taking seriously; my sense is that most of the critics who chide virtue ethics for not having answers to these questions haven’t seriously gone looking for them.

But either way, my goal here is just to offer a reasonable justification for virtues as our central ethical concern, and to show that it’s grounded in the same general commonsense normative framework that we’re already happy to accept in other domains. The bottom line, to me, is this: “Why should we evaluate the goodness of an X apart from the actual results it brings about?” is the sort of question we can already answer when it comes to tree roots (and books and baseball teams and lawyers); it’s really not so crazy to think that the same kind of answer might also apply to human beings, their actions, and the lives they choose to live.

Earlier this morning, a man in Shreveport, Louisiana systematically hunted down and murdered eight children, ranging in ages from one to fourteen. After a short chase, he was eventually shot dead by police. Some, and maybe all, of those who were killed were his own children. Two more victims, likely the mothers of these dead children, are currently fighting for their lives in intensive care.

When tragedies like these occur, atheists have an easy answer as to why. There might be specific factors we point to, depending on the situation: Mental illness, gender inequality, political strife, whatever. But in a broader sense, the atheistic explanation for horrendous evil is simpler than that. It’s just that human beings, like many other animals, often decide to hurt and kill each other, and sometimes there’s no one around to stop it.

Theists, on the other hand, have a much harder time explaining what happened in Shreveport this morning. Of course, there are plenty of theodicies out there, and each one has its own unique approach to the problem. But no matter how they get there, the majority of them will all end up saying that no instances of evil are truly gratuitous — that is to say, no evil is ever ultimately purposeless, or pointless, or without some countervailing good that makes our suffering worth it in the end.

Theists often frame this conclusion as a comforting one, if not for the people who are suffering themselves, then at least for those who witness their suffering and wonder how a just God could allow it. And for some, I’m sure this belief in God’s all-encompassing providence really does bring a measure of peace that I have no urge to sneer at. As for me, though, the idea that every evil fulfills some purpose has always felt like a dismal, and even nightmarish, conclusion. I can say with absolute certainty that I would much, much rather some evils be pointless.

To some, that preference might sound absurd. But the basic chain of reasoning behind it is fairly simple. To say that no gratuitous evils exist is to say that every actual instance of evil obtains some outweighing good that justifies it. And this, in turn, is to say that no actual instance of evil could ever have been done away with by God, without something of greater value being lost. But if this is the way that reality is, then all these evils — all the murders, all the rapes, all the family annihilations and genocides and hurricanes and famines and everything else — must be ineliminably central to the goodness of this world. If no evils are ever truly pointless, then these atrocities aren’t just occurring alongside the joys and wonders of our lives; rather, they must be deeply, essentially intertwined with them.

Further, assuming that God would desire to maximize the goodness he provides to his creation, it’s hard to see how we can avoid the Leibnizian conclusion that this is, in some meaningful sense, “the best of all possible worlds” — or, at the very least, that this world isn’t dramatically far off from the best that it could be. And if this is true, then we have to accept the grim conclusion that there was no better way for these goods to be purchased, even with the power of God. If these theodicies are right, then the maximal goodness of our world itself depends on what was done in Shreveport this morning — not contingently, as a tragic matter of happenstance, but on a foundational level that constrains God himself.

This, more than anything, is what terrifies me: These theodicies tell us that the creator of all things, in his perfect love and wisdom, looked out at all the possible worlds he might bring about and realized that, by some sinister brute necessity, the best world, the most perfect world, the world that brings the greatest joy and flourishing to all, still has men in it who kill their children one by one. They tell us that creation has an essential ceiling of goodness, one set far below any world without an Auschwitz, or a Black Death, or a bright Louisiana morning where eight children’s bodies slowly grow cold. They tell us that an immutable bedrock of evil is woven into the fabric of reality itself. What, I ask, could ever be more hideous?

As an atheist, I have no expectation that this world will one day be made perfect, and that the evils we wrestle with at night will ever be done away with in my life or in the lifetime of any conscious creature. But what I don’t have to accept is the dreadful conclusion that what I see in this world represents a global maximum of goodness that could never be escaped. When men slaughter their families, or set off brutal wars, or otherwise unleash endless amounts of suffering into the lives of decent people, I can look that random, pointless evil in the eye and say that it didn’t need to happen. And that is, often, the only comfort I can find; to trade it for a purpose almost feels like a betrayal.

There’s a pretty common view in philosophy of mind that I’ll call the “experience-first” perspective. It goes like this: For any conscious creature, the knowledge that you’re experiencing will necessarily be prior to any knowledge you can have about the objects of your experience; before you can know there really are tables, trees, rocks, and so on, it’s just a basic conceptual truth that you first need to know that you’re having table-like, tree-like, and rock-like experiences, from which you then extrapolate or infer the actual existence of what those experiences represent.

This Cartesian approach is (in my experience) something close to the default, and it’s obviously a big part of why so many people find dualist and idealist theories of consciousness appealing. Unfortunately, I think it’s also totally wrong, and in this piece, I’ll be relying on the work of a very astute philosopher who I just dropped off at daycare.

To start, let me brag a little bit about my daughter. She’s coming up on two years old now, and she’s right in the middle of what developmental psychologists call the “language explosion.” Just a few months ago, this little slacker had maybe three or four words that she’d managed to grasp; now, it feels like she’s learning three or four new words a day, just by hearing them in passing.

One of the most wonderful things about this period is that, as a parent, you start to hear your kid reporting back all the little things they’ve figured out about the world. Thanks to the language explosion, I can finally know that she knows quite a bit. She knows there are cars, and books, and dogs, and shoes. She knows the difference between apples and tomatoes. She knows where I put her toys at night so she can find them in the morning, and she even knows which parent is more likely to give her jelly beans when asked. (It’s me, for the record.)

But even if I can be sure my daughter knows these things about the world, getting a handle on what she knows about herself is quite a bit harder. When she points at her copy of The Very Hungry Caterpillar and says “cabbapuwuh buk,” that’s plenty good evidence that she knows what she’s seeing. But what could she ever say or do to demonstrate the knowledge that she’s experiencing? To some people, it’s silly to even ask this question — of course she knows that she’s experiencing, they’ll say, or else she’d have no way to know about whatever it is that her experience presents her. Unfortunately, things aren’t that simple!

To explain why, let me pull out a slightly clumsy metaphor. Imagine your friend Jack wakes up one morning, looks out his window, and sees his neighbor walking down the street. However, it later turns out that Jack was never actually making visual contact with the neighbor herself; instead, a massive television screen had been secretly pressed up to his window the night before, and what he saw in the morning was the live feed of a camera that was mounted behind it. In a case like this, can Jack know that his neighbor is walking down the street without first knowing that he was looking at a television screen?

It seems obvious to me that he can! Assuming the camera’s feed is reliable, what Jack has been presented with is an accurate picture of his neighbor’s movements, which is a perfectly fine source of knowledge. Whatever medium was responsible for that presentation, on the other hand, is completely unknown to Jack, even as it’s causally responsible for presenting him with all the visual data his knowledge is based in; all Jack knows about is what has been presented, not what’s doing the presenting. So even though Jack can only know about his neighbor’s movements by first looking at the screen, the knowledge that he’s looking at the screen isn’t needed for the knowledge that his neighbor is walking down the street.

Of course, even the most diehard Cartesian doesn’t think our perceptual data is literally presented on some internal screen. But the point I’m making here is just that it’s possible, at least in theory, to distinguish between knowledge of objects in the world, and knowledge of the means by which you’d come to know them. Based on the fact that Jack knows about his neighbor, I can know that Jack has been looking at the screen — but it isn’t until Jack reports features of the screen itself that I have any reason to think he’s aware of it.

In the same way, I can know that my daughter is experiencing, and that she knows what it is that she’s experiencing when she does so; if not, she wouldn’t be able to use her newfound language skills so consistently (and adorably). But whether that outward-facing knowledge of the world is accompanied by any additional inward sense of experience as experience seems, to me, like an open question. So it isn’t enough to just say my daughter must know she’s experiencing, as some necessary byproduct of her external awareness. What I need is evidence that she has beliefs about her experience, and not merely the objects that experience presents her.

But what could that evidence even be? Unless Sesame Street starts a new segment where Elmo studies philosophy of mind, it’ll probably be a while before my little girl learns to express any explicit conceptual distinction between the experience of an object and the object of experience. Still, a more primitive non-conceptual distinction might make itself known in other ways, most obviously through the appearance of “seeming” language.

See, when children first learn to speak, their development tends to follow a very specific path. They start by using words to pick out specific physical objects, like ‘milk’ or ‘car.’ Then, around the time of the language explosion, they’ll begin to pick out specific features of objects with words like ‘blue’ or ‘tall.’ And finally, a few years after that, they’ll develop the ability to report what features an object seems or appears to have. That last step is what matters here, since it requires being able to acknowledge that our perception of an object can fail to represent an object accurately — which is to say, it requires grasping the possibility of a perceptive error, where the way something looks isn’t the way that something is.

Grasping this concept, in turn, requires the ability to understand that a seeming or appearance can be analyzed apart from the external object it represents. It’s this recognition, I think, that marks off the point at which we can coherently say someone knows they’re experiencing: When they can consciously consider their experience as an experience, rather than just considering the world as experience presents it. And since my daughter isn’t at this point, at least as far as anyone can tell, I can’t see an obvious reason to believe that her knowledge of external objects is accompanied by the knowledge of her experience.

And before anyone gets the wrong idea, let me be clear that nothing I’m saying requires denying the obvious mechanical primacy of experience in the process of knowledge formation. Every sensible person is going to agree that experience is necessary for people to learn new things; I’m certainly not going to argue that my daughter somehow knows there are dogs before she experiences them, or that her knowledge of dogs has nothing to do with that experience. But, as with Jack and the television screen, those are just mechanical facts — facts about what has to happen before my daughter can know about the world, not facts about what she has to know before she can know about the world.

What the “experience-first” position requires is something much more than that. It requires that our epistemic starting point be an actual, firsthand knowledge of our experiences, with any further external world knowledge being extrapolated or inferred from those epistemically prior experiential states. And if I’m right, then that just isn’t how my daughter sees the world at all. In fact, it seems more likely to me that her internal sense of things is something close to the exact opposite.

All this is to say: At some point in the next few years, I imagine my daughter will start to wonder (in more toddler-friendly terms) just how it is that she can know what she does about the world, and why the beliefs she forms through perception sometimes end up being wrong. And once those sorts of thoughts start popping up, I think she’ll slowly, and even perhaps semi-unconsciously, start to grasp a profound distinction between (as Harman puts it) the objects of her experience and her experience of those objects. Before that point, however, I just don’t see any reason to believe that she knows she’s experiencing at all, much less that she knows it prior to knowing all the other things she knows about cars and shoes and dogs and books.

Now, there’s a very simple rebuttal available for everything I’ve said so far. It goes something like this: Sure, a toddler might not be able to grasp the conceptual (or even non-conceptual) distinction between external objects and their own internal experience. But that doesn’t mean she’s lacking a more primitive and immediate awareness of her own phenomenal states, even if doesn’t conceive of them in the terms laid out by modern philosophy of mind. After all, when I experience things, I have a direct awareness of that experience, and it’s where I end up getting my knowledge of the external world. So why should I doubt that she has this kind of immediate, foundational knowledge by acquaintance, given that I do?

To this, I’d say two things. First, I’m skeptical that non-propositional “knowledge by acquaintance” exists at all. In my mind, to know something is always to know that it is a certain way, with that certain way being expressed as (or at least being capable of expression as) a proposition. If my daughter knows her experience of a red stuffed animal, but not that she is experiencing it, then I’m just completely at a loss as to what exactly it is that she knows. It’s about as meaningful, to me, as saying you know your reading of a book, but not that you’re reading it.

More importantly, though, I’d also say this: Even if knowledge by acquaintance does exist, it definitely isn’t how my daughter figures out that she’s experiencing. In fact, I’m pretty sure it isn’t how you figure out that you’re experiencing either! To show you what I mean, let me run a little experiment.

Right now, for some reason, I’ve got an old polyester American flag the size of a playing card balled up on my desk, and I’m looking at it closely. If you’re a patriot like me, you’ll have a flag of your own to look at too, but otherwise just grab some other object from around the house and fix your eyes on that instead. As you look at whatever object you chose, you’re presumably going to know that you’re having an experience of looking at it. Now here’s a simple question to ask yourself: How, exactly, do you know that? What can you point to that explains your knowledge of this supposedly primary, epistemically undeniable experience going on in your head?

At first, you might be tempted to say you “just know it,” based on that ineffable acquaintance we talked about earlier — in which case, fair enough, but then I’d like you to at least list out what it is, exactly, that you’re knowing so directly. And when you try to do this, you’ll inevitably start talking about the actual content of whatever experience you’re having. In my case, I might say things like: I know I’m having the experience of seeing a little American flag because I see the red and white of the stripes; because I see the little white patches of the stars; because I see the darker regions where the scrunched-up fabric casts a shadow; and so on.

So far, nothing I’m saying should seem too crazy. But now it’s time for the second half of the experiment: Try to imagine, for a second, that you don’t know anything at all about whatever object it is that you’re looking at — not its size, or its shape, or its color or texture or location in space. And once you’ve put yourself inside a total epistemic blackout as best you can, I want you to return to the first question and take another shot at explaining how you know you’re having an experience of the object you picked. Go ahead, try! Can you do it? Is there any way to explain your knowledge of the experience you’re having? Is there even any way that you could have that knowledge to begin with?

It seems obvious to me that the answer is no! Under normal circumstances, your knowledge of the perceptive experience you’re having just is your knowledge of the features of the external object — and if you don’t know anything about how that external object actually is, then you couldn’t possibly know anything about what your experience of it is like. When I look at my little American flag, what I see first is something out there with red and white stripes; to the degree that I’m also aware of any experience of redness, and whiteness, and stripe-ness, it’s only as a sort of cognitive delivery mechanism for those facts about the flag.

In other words, I don’t first know about my experience of the flag, and then from there determine that whatever is being represented by that experience must really exist. Instead, I come to know that the flag is on my desk, and that it has certain features, and it’s this knowledge that tells me I must be experiencing. You might even go so far as to say that my experiences are, in some sense, just theoretical postulates for explaining the presence of my external world knowledge, and that any sense of direct contact is an illusion that comes when we fail to distinguish between what’s been presented and the means of presentation.

What this says to me is that the developmental process by which my daughter will come to a knowledge of her experience as experience isn’t really so different from how we as adults come to know we’re experiencing in any particular instance. In both cases, we first come to recognize external world knowledge, and then come to realize that experience must be what produced it. My daughter won’t hit this realization until later, when disparities between experience and reality start to make themselves known. But as an adult who already has these concepts at the ready, I can reflect on this process at any time and see the dynamic for myself.

With this framework in mind, it’s easy to pinpoint where the “experience-first” people go wrong. They start with the reasonable observation that, for any adult who can grasp the concepts involved, the knowledge that you’re experiencing is going to be more certain than the knowledge you develop from that experience. But then they erroneously assume that, because our knowledge of experience is epistemically certain, it must therefore be epistemically prior; since we know it more than we know anything else, they reason, it must be that we know it before we know anything else.

But as we can see, that logic just doesn’t hold up. Here’s another brief analogy to help explain why: Let’s say you order a shirt online and come home one day to see the box waiting at your door. Once you see the box, you can be absolutely certain that the order was delivered, whereas you can’t be absolutely certain that what’s in the box is really what you ordered. But that doesn’t mean that you must have known the box was delivered before you knew there was a shirt on your doorstep; you knew (without 100% certainty) that there was a shirt on your doorstep first, and from that knowledge you deduced (with 100% certainty) that your package had been delivered. So certainty and priority can easily come apart!

My argument is that perception works roughly the same way. When we look out into the world, we first gain our knowledge of the properties of external objects: Their size, their shape, their color, or whatever else. Then, after we develop this knowledge, we come to know that we must be experiencing (or else we wouldn’t know what we do about those objects.) This knowledge that we are experiencing is necessarily going to be more certain than whatever knowledge tipped us off to it, because sometimes experiences can be misleading. But just because the fact that we’re experiencing is knowable with a certainty that our knowledge of the outside world can’t match doesn’t mean the former must come prior to the latter.

In other words, we can know that experiential presentation has taken place with more certainty than we can know what’s been presented, while still knowing that presentation has taken place because we know what’s been presented. Accepting the epistemic priority of external world knowledge doesn’t require rejecting the idea that our knowledge of experience is epistemically privileged; it just requires recognizing that certainty and priority aren’t the same thing. And once we accept this, there’s just no good reason to assume that our knowledge of experience is actually epistemically prior in the first place — not for me, not for my daughter, and not for anyone.

Last week, the folks over at Secular Pro-Life posted a piece by Hugh Geraghty that poses a simple question: Should Consciousness Determine Who Deserves Human Rights? In it, Hugh answers with (surprise!) a resounding “no,” and he wraps up his case with a pretty harsh take on the basic concept of personhood:

History’s greatest atrocities often begin with the premise that we can distinguish between humans and “persons.” In fact, I believe that the only real purpose of separating the idea of a “human being” from a “person” is to justify the abuse, subjugation, and often slaughter of whole classes of human beings. And so, I tend to reject “personhood” as a worthwhile concept when determining human rights.

Now, regardless of your views on abortion, a sweeping statement like this should automatically set off alarm bells. I mean, come on — the only reason to ground a being’s moral status in something other than their species is to excuse oppression and genocide? What about, you know, explaining why we shouldn’t kill and eat E.T? There are countless examples of moral issues where the distinction between personhood and humanity helps to clarify things, even if it turns out that human beings are also automatically persons. That’s why other staunchly pro-life outfits like Catholic Answers still happily admit that the two categories can’t be, and shouldn’t be, collapsed into one.

Of course, Hugh isn’t the first person in the anti-abortion movement to draw this dubious parallel between non-reductive conceptions of personhood and historical atrocities. The idea that abortion advocates are doing today what slaveowners and Nazis did in the past is actually an extremely common anti-abortion talking point. Here’s noted “abolitionist” Jeff Durbin, for example, saying much the same thing while visiting the Colorado state legislature to explain his support for executing women who get abortions:

They used to say, “I know it’s technically a human, I know it looks like a person [but] it’s not a person, it’s a Black man. I know it’s technically a human, it looks like a person [but] it’s not a person, it’s a Jew.” Humanity is filled with this evil and this tyranny where we draw circles around ourselves and we dehumanize and butcher and tyrannize other human beings… We all agree with the injustice of a white man killing a black man unjustly. But the pro-choice position is essentially saying we don’t have a problem with that as long as it’s in the womb.

And here’s Kevin Roberts (of Heritage Foundation fame) making the same direct comparison in a speech from back in 2024:

Every slave auction, every lynching, every concentration camp, every abortion mill, every pogrom, every terrorist bombing from the Middle East to Kermit Gosnell, from Herod to Hitler to Hamas, has been justified on the same inhuman pretense that the victims aren’t really people. That human dignity is not universal, not inherent to our nature, but contingent on the convenience to the powerful.

Et cetera, et cetera. Now, look: Generally, I doubt that much productive discourse ever gets started with both sides arguing over which one sounds more like Hitler. I’d rather just talk about the actual legal and philosophical issues at the heart of the issue. But since anti-abortion pundits are constantly pushing this ridiculous (and frankly insulting) insinuation that abortion supporters share the ideological lineage of every genocidal maniac in history, I think I’m owed an opportunity to explain why their logic isn’t just ridiculous, but actively backwards — if there’s anyone who “sounds like Hitler” or “sounds like a slave owner,” it’s the so-called pro-lifers, not me.

Before going any further, we should pause to get a better grasp on what exactly the anti-personhood narrative is even supposed to be. It seems to go something like this: Throughout human history, powerful people have always wanted to exploit and abuse less powerful people; but they can’t justify that oppression in light of their victim’s obvious humanity; so they cook up a scheme to artificially restrict the circle of moral concern to “persons” only, with that concept being haphazardly defined to exclude whichever human beings they want to brutalize.

The implication, I guess, is that sticking with “human being” as our fundamental moral category would make it harder to justify all sorts of atrocities, and that pro-abortion advocates want to maintain the distinction between humans and persons just because we need an easy excuse to keep on killing babies.

There are, to put it delicately, some issues with this story — the most obvious one being that it has essentially zero connection whatsoever with the actual historical development of personhood as a legal and philosophical concept. If you take time to study pioneering theorists like Locke and Grotius, you’ll see very quickly that their motivation was, at least in part, to secure the moral status of human beings as human beings, and to prevent the powerful from arbitrarily abrogating it purely on the basis of social status or class.

Now, did these early modern thinkers have biases and blind spots that kept them from fully extending that status to women, non-Europeans, and other oppressed groups? Of course! But it isn’t as if those groups were seen as any more deserving of moral consideration before the concept of personhood developed; slaves and housewives weren’t exactly doing great on the civil rights front when all they had to work with was their basic species category. So this idea that the cultural shift from ‘human’ to ‘person’ as a basic moral category even correlated with, let alone contributed to, any major violation of human rights is just historically indefensible.

But let’s put all that aside and just ask a much more basic question: Whenever the most notorious oppressors of the past couple centuries did start looking for ways to shrink the circle of moral concern, how exactly do they do it? Let’s zoom in on Nazis and slave traders, the pro-lifer’s two favorite points of reference. How did Hitler justify his slaughter of obviously human Jews? How did John C. Calhoun defend the buying and selling of obviously human Africans? Neither of them just waltzed in one day and announced that those people weren’t persons just because. They had to find some grounding for their newly restricted circle. So what was it?

Well, here’s one thing it definitely wasn’t: The capacity for self-aware experience! At no point did the Nazis or the Confederates try to argue that their victims were literally insentient p-zombies — in fact, essentially no oppressor in all of human history has justified their atrocities that way. And how could they? Even the most vicious bigot would have to admit that Jews and Black people (and women and homosexuals and immigrants and everyone else) have the ability to feel pleasure and pain, and a desire for joy over suffering.

This point is so obvious that you might be wondering why I’d even take the time to make it. But it actually matters quite a bit if you want to understand where this anti-abortion rhetoric goes wrong. See, people like Hugh are trying to push a very specific narrative, one that says pro-abortion activists are just the latest in a long line of tyrants using a shifty conception of personhood to justify abusing human beings. But if you look at the actual standards for personhood that supporters of abortion generally rely on — consciousness, sentience, self-awareness, and so on — then you’ll quickly see that, both formally and substantively, the arguments we’re making have nothing whatsoever to do with the appeals of historical tyrants.

Meanwhile, if we look at what all these historical tyrants did rely on to cynically restrict the circle of moral concern, we’ll see they invariably turned to the same basic standard that Hugh relies on today: Biology. For the Nazis, only “Aryans” mattered, while in the Antebellum South, full moral status was reserved for the “white” or “Caucasian” citizen only — but in both cases, the oppressor’s perverse logic ultimately agreed with Hugh and his anti-abortion allies that a creature’s moral status depends not on their capacity for suffering or flourishing, but rather on some particular fact about that creature’s DNA.

And for the record, it’s not just the Nazis and southern slave owners who excused their atrocities by reference to some essential biological category. Ever since the end of the scientific revolution, biological difference has actually emerged as the predominant justification for genocidal violence in the modern era. In fact, even conflicts where no obvious biological distinctions exist at all regularly set off a desperate search for some difference to latch onto; in the lead-up to the Rwandan genocide, for example, Hutu radicals used earlier British race science to cast their (almost biologically identical) Tutsi victims as a degenerate sub-species. And of course Hugh’s own species criteria is currently trotted out to defend the abuse and exploitation of very plausibly non-human persons like elephants and great apes.

Now, to be very clear: I’m not calling Hugh a closet Nazi here, or comparing him in any meaningful way to any other genocidal monsters. What I’m saying instead is just that, by any reasonable measure, his belief that only species membership matters is obviously closer to the logic of fascists and white supremacists than the pro-abortion advocate’s belief in the moral centrality of conscious experience. All that ultimately distinguishes Hugh’s moral framework from the moral framework that excused the Holocaust is which biological category ends up being responsible for fixing the binary yes/no of moral status; for him, it’s species, whereas for the scientific racists of the Third Reich or the American South, the focus was slightly narrower.

Pro-abortion advocates, on the other hand, embrace an approach to moral status that explicitly denies the foundational assumption of both parties. Of course, this doesn’t mean we think both positions are equally wrong; I would take Hugh’s own standard over the alternative in a heartbeat, since the human beings excluded by the racialized pseudoscience of Nazis and slavetraders are also morally valuable persons who deserve to be defended from abuse and exploitation. It’s just that I think Hugh is wrong about why those people deserve that moral status. It’s not because they happen to be the same species. It’s because they can think and feel and hope and dream — and that undeniable capacity is infinitely more difficult for any wannabe tyrants to work around.

But don’t take my word for it here. Pick any historical atrocity that anti-abortion activists constantly bring up and ask yourself which standard, mine or Hugh’s, would have made that brutality easier to justify. Which sort of population would Hitler have preferred to have, one that conceptualizes basic moral worth in terms of biology or in terms of psychology? Obviously, it would be the former — finding some way to shift from species to race (as tyrants have done countless times over the last few hundred years) would be undeniably easier than convincing the public that Jews literally have no inner lives at all! So there’s just no plausible standard, whether practical or ideological, that honestly groups the pro-abortion position with these historical bigotries in any meaningful way at all.

My point here is very simple: If anyone tells you that a person’s moral worth can be securely grounded in their species membership, or that an emphasis on the inherent moral importance of any particular biological category can act as some essential bulwark against oppression and atrocity, then whoever you’re listening to has absolutely no idea what they’re talking about. None. The entire history of the modern world points in exactly the opposite direction, from the trans-Atlantic slave trade to Treblinka and on to Tyson Foods today. In fact, body for body, the belief that a creature’s biology matters more than their ability to suffer just might be the most destructive fantasy in all of human history.

And once again, I’m not saying this is even good reason, by itself, to dismiss the anti-abortion position. Maybe species membership really is the right category for grounding moral status, and it’s just an unfortunate fact that the correct standard for determining moral worth happens to brush up so closely against the logic of our greatest historical monsters. Could be! All I’m asking here is just that these anti-abortion pundits stop throwing stones from glass houses. If anyone here has a problem with thinking a little too much like Hitler, it’s the people who agree that facts about your DNA are central to the question of whether or not it’s okay to kill you.

Last week, Tim Seyrek published a really interesting piece on physicalist assumptions in the sciences and how researchers might benefit from considering their theoretical commitments more carefully — and even though I’m a diehard physicalist myself, I still thought there was a lot to appreciate in what he had to say. Honestly, you’d hope that everyone could agree on the importance of acknowledging presuppositions and avoiding dogmatic insistence on any one particular approach, regardless of whatever view they ultimately endorse. In fact, just in purely selfishly terms, I think it would be good for physicalism if scientists started tackling these philosophical issues directly, rather than just taking for granted that everyone around them will agree.

But at the same time, when it comes to the actual question of whether or not idealism is compatible with scientific investigation, I think there are reasons for physicalists to be very, very skeptical. Individual non-physicalists can do science just fine, of course, and no one is arguing that they deserve to be run out of their labs with pitchforks and torches. What actually matters is whether idealism as a philosophical system can give a plausible story to ground the work those scientists are doing — and in this piece, I’m going to argue that the answer depends on just how closely idealism can be made to resemble the physicalist model it’s meant to replace.

Let me start things off with a very unflattering comparison that I promise I’ll qualify later: When I was growing up, my family and I attended a few different conservative Christian churches where doctrinaire young Earth creationism was an inexplicably central commitment. (This was in the 90’s, at the height of the so-called “intelligent design” movement, so I guess it wasn’t too surprising.) As a result, I spent most of my Sunday mornings in the children’s classes, where we were constantly bombarded with talking points meant to counter basically every aspect of modern evolutionary biology, geology, astronomy, history, and so on.

During one particularly intense struggle session, I remember someone bringing up a very reasonable objection to creationism, which is that there’s no way we should be able to see light from stars that are a million light-years away if the whole universe is only about six thousand years old. Luckily, the decrepit old man who ran the classes had a perfectly good explanation: God had created the entire universe with the light already in transit, because of course he did. This explanation wasn’t very convincing to me even back when I was eleven years old, I’ll admit, and it hasn’t grown any more convincing with time. However, a quick Google search will show you that plenty of young Earth apologists are still running with it today.

So why bring this up? Well, imagine that someone asks you whether creationism is compatible with astronomy. It’s likely that your first instinct would be to say no, since there are just so many basic astronomical observations that fit extraordinarily poorly with a straightforward, literal reading of Genesis. But that might be jumping the gun a little bit, since a creationist can immediately reconcile the two by adopting this kind of “mature creation” model that says God spoke everything into existence as if it had all been around for the last 13.8 billion years. Once they do that, all astronomical data is now necessarily compatible with the creationist theory, since the creationist theory just is the mainstream scientific theory, with an extra set of caveats slapped on.

In other words, if you really want to be precise, then you’ll have to grudgingly admit that creationism is indeed compatible with astronomy — but only to the degree that it can be made to artificially resemble whatever non-creationist theory is actually doing the productive work. And while idealism is certainly nowhere near as goofy as young Earth creationism, I think the same general dynamic applies: You can absolutely do idealist science, but only by doing good ol’ physicalist science first, and then putting the whole thing in some pretty ad hoc quotation marks.

To see what I mean, just take a second to consider what your own mental states seem to be like. Just speaking for myself, they seem to be essentially private, in that no one else can observe them directly; inconsistent and ephemeral, in that they tend to come and go for unclear reasons; easily altered, in that I can dramatically shift their basic features with very little conscious effort; unconstrained, in that I can think whatever thoughts I want and imagine anything I can conceive; and experientially immediate, in that they disappear whenever my consciousness isn’t present. Of course, it could be that all these perceptions are illusory in some way, and that “actual consciousness” is completely different. Still, we have to start somewhere, and it seems undeniable that these are the sorts of features that define our pre-theoretic conception of the mental.

Meanwhile, when we turn to our pre-theoretic conception of physical objects, things work very differently. Physical objects “out there” are publicly accessible and open to observation, and they exist apart from any particular person’s experience. They’re also completely impervious to the influence of our own internal states: They don’t change to match our emotions, we can’t create new ones just by thinking the right thoughts or taking the right drugs, and they stick around even when we close our eyes or go to sleep. Physical objects are, in other words, the exact sort of thing that mental entities like thoughts, emotions, and sensations are not. They play by their own set of rules, and that basic distinction between the nature of our “outside” and “inside” worlds is one of the most fundamental dividing lines in our conceptual map of reality.

(Again, could that sort of distinction be an error? Of course! If idealism is true, it is an error. But if creationism is true, it would also be an error to think the starlight we see really started out a million light-years away — all that really matters here is just that, before we start making those sorts of revisions, our folk model of reality affords the mental and the physical a commonsense set of mutually exclusive features.)

The question now is this: Which of these two sets of features does the world need to have in order for rigorous, reliable scientific investigation to take place? Obviously, it would be the latter — if you want to do good science, you’re going to need the objects of your investigation to be publicly accessible, ontologically persistent, and free from the direct influence of the scientist’s own mental states. But precisely none of that is what we’d expect given the proposition that the (supposedly) outside world is entirely mental, since the only unambiguously mental entities we’re familiar with are the ones we encounter inside our heads, and none of those have the properties we would need to make science work. And yet, lo and behold, science does work! So what gives?

The only explanation the idealist has for this extremely lucky coincidence is that, for some reason, some number of fundamentally non-physical entities just so happen to have the exact sorts of features they would have if they were physical. But ultimately, this is just a nice way of saying that the scientific picture of reality only really comes together once the idealist’s theory is redrawn to resemble the physicalist’s as closely as possible. If you go into the world assuming that trees and stars and human brains will behave like literally any other mental entity you’ve ever encountered in your life, then botany and astronomy and neurobiology will be absolutely hopeless projects; the fact that you can rescue those projects by stepping into a physicalist frame of mind just demonstrates how essential those commitments are to the structure of science.

Another way to put this might be that physicalism predicts what idealism can merely accommodate. Given only our pre-theoretic conceptions of the physical and mental, the physicalist has a straightforward, obvious explanation for the sorts of regularities that are required for effective science to take place; the idealist, meanwhile, can only get to that same spot by first arbitrarily tweaking their theory to imbue (supposedly) external objects with all the scientifically essential features that physicalism grants them automatically. Do both theories eventually end up in the same place, all ready to do some science? Sure! But the science they’re ready to do is physicalist science, built from the ground up around the features of reality that physicalism naturally asserts.

Now, I’ll pause here to acknowledge that idealists obviously do have some internally consistent explanations of their own for the features of (supposedly) external objects. But the problem with these theories, which usually invoke a singular consciousness at the base of reality to ground the shared content of individual perspectives, is that the singular mind in question would need to have the sorts of properties we’ve never seen any actual mind unambiguously possess. None of us have any experience whatsoever with mental states that are as stable, persistent, and fine-grained as even the simplest physical object, nor can we even really grasp the idea of one consciousness sustaining an experience for others — in fact, the essential privacy of consciousness is a common talking point for anti-physicalist critics in pretty much every other context!

In other words, this sort of idealism is working backwards: Rather than starting with our actual experience of the mental and using what we observe to ground the features of (supposedly) external objects, it just surveys the features of those objects and then arbitrarily assigns them to a limitless cosmic mind — a mind that doesn’t resembles our original experience of the mental in any way. This is, ironically, the exact sin that so many idealists accuse physicalists of whenever we argue that physical matter can sometimes instantiate qualitative properties, but on a dramatically greater scale and with a dramatically lower chance of any future explanatory progress. Whatever stop-gap, “somehow it works” strategy that physicalists are using to temporarily ground consciousness, idealists are using to ground literally everything else, forever.

Idealists don’t have to take this kind of pseudo-physicalist approach, for the record. In fact, we have endless examples throughout history of those assumptions being directly rejected. Parapsychology, ayurvedic medicine, anthroposophy — if you really want to, it’s actually very easy to find people who are trying their best to investigate the world without first assuming that it exists as a persistent collection of physical objects apart from the mental states of human beings. It’s just that, unfortunately, all that “science” is universally junk! As points out in a replies to Tim’s piece, human beings have spent a good chunk of the last fifty thousand years trying to do science without a commitment to physicalism, and the results haven’t been particular convincing.

Of course, this isn’t to say that idealist scientists working today want to introduce this sort of quackery back into the research lab. In Tim’s piece, he actually gives some very interesting examples of what a more restrained, sober-minded idealist interpretation of neuroscience and psychology might look like — and honestly, I don’t have anything close to expertise in these areas, so I won’t hazard any specific critique. But I will say that even these more legitimately scientific approaches seem to take for granted the sorts of regularities that physicalism naturally entails, rather than actually explaining how an idealist could ground them from the opposite direction.

For example, Tim suggests that certain illnesses might ultimately be manifestations of a person’s unhealthy psychic activity. But what sort of psychic activity that we know anything about is persistent, regular, and law-like enough to stand in as a reasonable candidate for the essential nature of cancer, or psoriasis, or diabetes? And what could explain why it is that those psychological traits manifest with a complexity that seems to radically outstrip our sense of it from the inside? Put disease aside for a moment and just ask about the basic functioning of a human body. What sort of mental state is so fine-grained that it could possibly conjure up and sustain the appearance of all 1.3 billion neurons flashing in my frontal cortex, or the trillions of cells in my digestive tract? Is there even the possibility of an idealist science that can give us any answers beyond something completely isomorphic to the physical structures our non-idealist science has already discovered?

The bottom line is this: A physicalist can start with the physical as the physical and end up producing all sorts of explanatorily powerful theories that emerge directly from the straightforward commitments of their framework, whereas idealists are stuck taking those higher-level conclusions and just recapitulating them in non-physicalist terms. What they can’t do is start with the mental as the mental and then work their way back to trees and stars and human brains. There is no idealist atomic theory that can tell us which primitive mental states show up as carbon versus helium, or an idealist theory of gravity that can explain why the manifestations of mental states all seem to clump together — and if we’re being honest, it’s hard to even imagine how there could ever be, since I don’t see empirical access to fundamental psychic reality getting any easier in the future.

Until we have these sorts of answers, it makes sense for physicalists to feel a little like idealist scientists are just copying their work and slapping “The cosmic mind makes it look like…” before each observation — much like the creationists who ran my Sunday school had a creation model that said God made the world look like whatever all the secular astronomers determined it actually was like. Of course, like I said, this doesn’t mean idealism and creationism are equally problematic, or that any of the first-order empirical absurdities that plague creationism will also plague an appropriately sober-minded idealist approach. But still, it absolutely should set off alarm bells any time we see one theory working backwards from another.

Let me close by saying that, even with these concerns, the project Tim recommends is still far from worthless. It’s good to have people exploring new ways to approach these perennial questions, even if I’m personally skeptical that the method they’ve chosen is going to succeed. And as the process plays out, physicalist scientists will be pushed to consider, adapt, and defend their own approach in response, which can only be good for everyone. And who knows? Maybe once we do abandon dogmatic physicalism and make more space for idealists in the sciences, we really will stumble across a radical consciousness-first revolution that I’ve just been too cynical to see coming. But until that point, I hope even the most optimistic idealist can understand why physicalists like me have our doubts.

It’s not at all surprising that moral anti-realists would reject the existence of stance-independent reasons for action, since the position that only stance-dependent reasons for action exist is pretty much just the definition of anti-realism. However, some more militant anti-realists go even further and argue that stance-independent reasons not only don’t exist, but are literally inconceivable — that is, they report a total inability to even imagine or entertain the concept of a reason for action that doesn’t in some way depend on some agent’s goals or desires.

Now, I certainly don’t think the people who make these claims are lying when they say they can’t conceive of moral realism being true. But, at the same time, I have to admit that I struggle myself to imagine how that could possibly be the case, given some basic features of reasons-discourse that are close to undeniable. So in this piece, I want to take a shot at giving my best explanation for why stance-independent reasons at least seem like the sort of thing that should be easy to imagine, and perhaps offer a couple of “intuition pumps” (yuck, I hate that term) for anyone who might be on the fence.

To start, I just want to point out the basic fact that we often speak about and consider reasons without any essential reference to what exactly it was that “gave them to us.” Take these two informal syllogisms as an example:

Hungry: If you desire to eat a delicious slice of your favorite pizza, then you have a reason to order a pizza from Big Frank’s Pizza; you desire to eat a delicious slice of your favorite pizza; therefore, you have a reason to order a pizza from Big Frank’s Pizza.

Charity: If Big Frank’s Pizza is holding a fundraiser where half of what you pay will go to your local women’s shelter, then you have a reason to order a pizza from Big Frank’s Pizza; Big Frank’s Pizza is holding a fundraiser where half of what you pay will go to your local women’s shelter; therefore, you have a reason to order a pizza from Big Frank’s Pizza.

In either case, it’s trivially easy to just apply the law of detachment and see that both result in the same general conclusion, which is that you have a reason to order a pizza from Big Frank’s Pizza. And crucially, from this point on, it doesn’t really matter which syllogism got you there — so long as you know you have a reason to order the pizza, all the same relevant normative conclusions fall out regardless: You’ll know, for example, that you also have a reason to pick up the phone and dial, or that ordering the pizza is the right move in the absence of any countervailing reason to do otherwise. (The only thing that might differ is the relative strength of the reason itself, but we can account for that easily by just stipulating that your desire for cheap pizza is exactly as strong as the corresponding moral weight of helping to support a women’s shelter.)

Crucially, this isn’t how things work in other situations where clearly inconceivable options are on the table. You can imagine, for example, an impossible world in which both square and round circles exist. In that world, a syllogism about square circular objects and a syllogism about round circular objects could both end up giving you the conclusion that you’ve got a circular object, period — but if you try to consider what would follow from that conclusion, you come up empty, because it’s just not possible to understand what a circular object would be if that understanding needs to allow for both the round and square versions. In contrast, a reason to act by itself carries all the sorts of normative implications you’d want from a reason, regardless of whether we stipulate that reason as coming from a desire or from some objective moral fact.

What this suggests to me is that the essential conceptual content of a reason doesn’t come with any reference to subjectivity or objectivity “baked in,” so to speak. (And on this point, many anti-realists agree with me.) However, if that’s really the case, then conceiving of a stance-independent reason involves nothing more than conceiving of of a reason that you already accept, and then just stipulating a different non-essential origin or grounding for it. Why, exactly, would that be difficult? Presumably, the anti-realist accepts that someone in the Hungry case has a reason to order the pizza, and that such a reason carries all sorts of important normative entailments. Why not just fix the presence of that reason and the nature of its entailments in place and then imagine the whole thing being given by something else instead of a desire?

This is, ultimately, what I find so earnestly baffling about the idea that two reasons for ordering a pizza could differ radically in conceivability, solely because one of them is given by facts about desires and the other by facts about the world: To paraphrase the esteemed philosopher Pam Beesly, they’re the same reason! It’s the same fundamental X-counts-in-favor-of-Y relation, with the only difference being what we choose to fill the first variable. Of course, anti-realists are free to think that relations like this only ever do take desires as their first component — they’d have to, or else they wouldn’t be anti-realists — but to go the extra step and say that it’s literally inconceivable for a non-desire fact to play that antecedent role is just legitimately hard for me to fathom.

If you’ll excuse the clumsy metaphor, it’s a little like someone who says they can easily imagine a pair of pants that you buy at a store, but can’t wrap their head around a pair of pants that you order online and have shipped to your mailbox. It just doesn’t make sense! After all, there’s nothing about a pair of pants that’s essentially store-bought, and there’s nothing particularly weird about the idea that Amazon could deliver them. (Or, even if you do think that’s weird, there will at least be plenty of other seemingly reasonable people who confidently testify that Amazon does deliver them.) So why not just imagine the pair of pants you picked up from the store and say to yourself, “This, but from Amazon instead?”

My feelings about stance-independent reasons are very similar: I just don’t see why it would be hard to focus on the normative entailments of a desire for pizza, fix them in your mind, and then just imagine those same entailments being realized by another fact instead. You don’t even have to believe that the other fact in question could ever actually play that normative role, just like you don’t need to believe that any human being could ever actually wake up as a bug in order to understand The Metamorphosis. All you need to do is patch together two things you already accept — the normative entailments of desires on one hand, and some objective fact about the world on the other — inside the structure of a general counting-in-favor relation. If there really are people out there for whom this sort of move just doesn’t compute, then I guess I can’t fault them — but hopefully they won’t fault me for thinking that it seems like a move that shouldn’t be that hard to make!

Now, by this point, I’m sure a few anti-realists have their responses already typed out. You fool! they’ll say. You moron! Don’t you understand that we don’t believe reasons as you describe them even exist? When we say someone has a reason to act based on their desires, we aren’t trying to say that those desires have generated a new “thing” called a “reason” that can exist over and above whatever “generated” it; rather, we’re just using the term “reason” to pick out a particular sort of means-ends consistency that is necessarily desire-based, since it’s very obviously the case that only desires can fix an agent’s actual ends. So of course we’re going to think that stance-independent reasons for action are inconceivable — reasons for action, to the extent that we speak of them existing at all, just are facts about what would or wouldn’t help to accomplish an agent’s stance-dependent goals.

To this, I reply: Yes, trust me, I’m aware. (You guys aren’t exactly shy when it comes to making this point!) But what conception of reasons you actually have isn’t particularly relevant here — what matters is whether you could wrap your head around a different conception, one that also seems compatible with all the important first-order features of reasons that we’d expect an account to preserve. After all, I’m absolutely sure right now that Mark Twain was Samuel Clemens, and that means any alternative theory is going to strike me as necessary false. But if one of those alternate theories can vest all the essential Mark-Twain-defining features in someone else instead, then even if that theory will never be convincing, it should still be trivially easy for me to imagine the sort of situation it describes.

Moreover, it isn’t as if the alternate model I’m talking about here is some obscure or controversial way of thinking that was dreamed up just to test the limits of an anti-realist’s imaginative skills — it’s one that millions and millions of people successfully rely on every day to make the exact same sorts of decisions that anti-realists do! In fact, a central commitment of anti-realism (or at least the particular version of anti-realism favored by so many who say they can’t conceive of stance-independent moral reasons) is that first-order moral discourse is metaethically indeterminate, in the sense that it doesn’t require any specific second-order understanding of the terms involved. But isn’t that, by itself, a good reason to think that all the second-order frameworks functioning successfully behind the scenes are, at the very least, coherent?

Think about it this way: An anti-realist and I could easily have a straightforward, first-order discussion about abortion, and if we did, it’s likely that “reason talk” would play a major role — we might talk about the reasons women choose abortions, what reason a legislature might have to ban it, our reasons for supporting or opposing those bills, whatever. And even if we don’t ultimately end up agreeing, the conversation would at least qualify as some real, honest-to-goodness moral discourse. But if that anti-realist wants to turn around and say that a stance-independent reason is truly inconceivable, then they’d also have to say that I was talking (or at least thinking) nonsense the whole time, right? So it seems like a very remarkable coincidence that moral discourse still somehow works!

I mean, think of all the other inconceivable things out there that anti-realists are fond of comparing to stance-independent reasons, and try to imagine what it would be like having a conversation about them with a true believer. If I was stuck in a debate with someone who thought all bachelors were married, for example, I don’t think we’d get very far discussing what society should do about the rising number of single men. Or if I landed on the side of the road with a flat tire, and the only person available to help me put on a new one was convinced all circles were square, then I imagine it wouldn’t go well either. And if I asked why, the answer would be obvious: Because beliefs that are literally meaningless or impossible to grasp don’t tend to be productive!

But unless anti-realists want to argue that their realist interlocutors are just getting really, really lucky whenever they try to morally deliberate, it seems like a reasonable conclusion would be that something at least somewhat meaningful is actually going on inside their heads. And since those realists report that what’s going on inside their heads involves considerations of stance-independent reasons for action, then (barring some extreme introspective error I see no evidence for) the anti-realist ought to take them at their word. But once the anti-realist accepts that realists are able to consider and apply the concept of a stance-independent reason for action, then I just don’t see how that could possibly be compatible with the claim that such a reason is virtuously inconceivable.

I say “virtuously inconceivable” there because, for all I know, the anti-realists who say they can’t conceive of stance-independent reasons for action might just be reporting a brute psychological fact about themselves, rather than trying to make any judgment as to the actual coherence of the concept in some ideal sense — in which case, I guess all I can do is wish them the best in eventually figuring out how to conceive of it like the rest of us can! But if the idea is that their inability to conceive of stance-independent reasons for action indicates any legitimate issue with the concept itself, then the fact that so many realists are able to apply the concept so effectively should at the absolute very least put pressure on anti-realists to explain what they think is really going on.

These, then, are the two reasons I’d give for thinking that stance-independent reasons for action ought to be conceivable: First, because it seems as though the concept of a reason is separable from any particular notion of stance-dependence or independence, and it should just generally be possible to conceive of something having any of its non-essential aspects altered; and second, because we have empirical evidence that a fair number of self-aware and thoughtful people are successfully relying on the concept of a stance-independent reasons for action when engaging in moral discourse, which would be incredibly surprising if the concept was actually totally meaningless (as the claims of inconceivability seem to suggest).

Of course, neither of these two points can prove that stance-independent reasons for action actually do exist, nor should we expect them to make the concept automatically conceivable for anyone who’s currently struggling with it. But at the same time, I do think they’re enough to, as I said, establish conceivability as a default assumption that anti-realists should be expected to argue against. And who knows? Maybe there really is some argument out there that could snap me out of whatever delusion I’m stuck in that makes me think stance-independent reasons for action are coherent — but until I hear it, I hope everyone will forgive me for thinking the problem is probably on the other end.

Back in January, I published a piece on evolutionary challenges to moral realism and why I think they ultimately fail. In that piece, I called these sorts of arguments (which I’ll refer to as EDAs from now on) some of the most serious objections to realism out there — however, after recently stumbling across a brilliant paper by Louise Hanson from back in 2016, I think I’ve changed my mind. The paper, which you can read here, raises a fatal issue with the basic structure of an EDA that I find very compelling; in fact, I don’t really get how I could have missed it until now.

Everyone who’s interested at all in the topic of EDAs should definitely give the whole paper a thorough reading, but just to help introduce Hanson’s basic point to a broader audience here on Substack, I thought it would be fun to write a short overview of the critique she lays out. I’ll be quoting her at length in a few spots, but otherwise you should take all this as a rough paraphrase — if there’s anything that you think sounds a little off, I’m sure it’s just me glossing over some necessary details. Still, I hope what follows can at least give realists (and others on the fence) a solid starting point for pushing back on a particularly widespread and influential anti-realist objection.

As I wrote in my original piece, there are multiple different “versions” of the EDA out there, and each one is going to emphasize something slightly different. Some target the possibility of moral knowledge or the likelihood that our moral beliefs would be true, while others frame things more in terms of explanatory power and parsimony. Still, each version will almost certainly start out with two basic premises:

1. Natural selection is responsible for the moral judgments that we make.

2. Natural selection is not a truth-tracking process.

Depending on the specific aspect of moral realism the argument is seeking to target, both the next few premises and the ultimate conclusion might differ. But either way, the basic worry is pretty simple: If our moral judgments are, in fact, explained by a process that isn’t concerned with what’s actually true or false, then we have no good reason to think those judgments reflect any reliable mind-independent truth, nor do we have any reason to think those mind-independent truths actually played any role in their formation. And it’s easy to see why a conclusion like this would be bad news for the realist, since a moral theory isn’t particularly appealing if it can’t explain why we should have any confidence in the truth of our moral judgments or the explanatory importance of the moral facts that they supposedly relate to.

So how does Hanson’s rebuttal work? Ultimately, it all boils down to pointing out a simple but easily overlooked ambiguity in the claim that natural selection can explain our moral judgments. Her basic contention is that, understood one way, the claim is obviously true but mostly irrelevant; meanwhile, if it’s understood in a way that would actually undercut our confidence in realism, then it’s almost certainly false. Hanson refers to these two distinct interpretations, respectively, as the quantificational reading and the predicative reading.

The distinction between these two interpretations, and how that distinction ends up impacting EDAs, can be a little tricky to grasp in the abstract — but luckily, we’ve got Donald Trump here to help us make things a little more concrete. So let’s all imagine, for a brief (but painful) moment, that we’ve been assigned as journalists to cover an invite-only midterm GOP rally where he’s going to be the headline speaker. Our boss, desperate for yet another article psychoanalyzing MAGA voters, tasks us with getting an answer to one fundamental question: Why does everyone in the audience think Donald Trump is a great president?

In this case, it should be pretty easy to see that there are two distinct ways you could interpret what we’re being asked to figure out. We could understand the question, on one hand, as something like, “Why does this specific audience consist of people who believe Donald Trump is a great president, rather than consisting of people with other political views?” This is the quantificational reading of the question, and if that’s really all we want to know, then the explanation should be pretty obvious: The reason people in the audience think Donald Trump is a great president is because anyone who felt differently wouldn’t have been invited.

If, on the other hand, we’re interpreting the question in a predicative sense, then what we want is an explanation for why any particular person in the audience has the sort of belief that would get them invited in the first place — which is to say, we want to know why that guy specifically thinks Trump is a great president, and that guy, and that guy, and so on, for every individual in attendance. Here too, there are plenty of obvious answers, none of which would be particularly flattering for the people involved. (The phrase “ghoulish moron” might make an appearance once or twice.) All that matters here, though, is that these two sorts of explanations are very different, even if both can be seen as explaining the same basic state of affairs.

In short: A quantificational explanation for some judgment being widespread within a group will tell us why the group is made up of people who tend to share that judgment, whereas the predicative explanation will tell us why the people in the group do, in fact, tend to share it. The ambiguity is that, if I just say “The crowd has the beliefs they do because of Donald Trump’s campaign managers,” it’s not clear whether I’m just saying that Trump’s campaign managers selected the crowd from a preexisting population of supporters, or if I’m making the much stronger claim that his campaign managers are directly responsible for the fact that those people are Trump supporters in the first place. To resolve that ambiguity, I’d have to clarify whether my claim was meant in a quantificational or predicative sense.

So what does this have to do with EDAs? As Hanson points out, a claim like “Natural selection explains why we make the moral judgments that we do” can also be read in a predicative or quantificational sense, with the two different readings giving two very different results. Quantificationally, no one (or at least no one worth taking seriously) would deny that many of our fundamental moral intuitions are fitness-enhancing, and that natural selection has played a major role in ensuring that those sorts of intuitions were passed along. So if we want to know why we exist, and not some other creatures who make very different sorts of judgments, then natural selection is a perfectly good explanation, or at least a big part of one.

On the other hand, reading the EDA’s central claim in a predicative sense commits us to a significantly stronger position: That natural selection explains why every human being, in an individual sense, makes one moral judgment instead of another. But why would anyone think this stronger claim is true? At the very least, it’s not generally the case that natural selection can explain traits in a predicative sense, since evolutionary pressures only act on traits that are already present. Quoting Hanson directly here:

When we talk about a particular trait common to members of a given species being the result of natural selection, we are saying, roughly, that individuals who had that trait or tendencies towards it had more offspring than individuals who didn’t, resulting, ultimately, in a situation where individuals with that trait existed rather than individuals who lack this trait. When we say, for example, that the long-necked-ness of giraffes is the result of natural selection, we are saying, roughly, that the ancestors of giraffes were such that, those among them who had slightly longer necks than the others tended to produce more offspring, and that this led to a situation where a greater and greater proportion of the creatures that were born had long necks, which eventually led to the long-necked giraffes that exist now existing instead of some other creatures with shorter necks. This is a claim about the process that led to there being individuals who had the trait in question, not a claim about the process that led to these individuals having this trait.

What’s more is that, even if we do want to say that natural selection can explain our individual moral judgments predicatively, it could only ever do so through a broadly quantificational process; as Hanson says, it’s not as if there’s some other way natural selection could influence our individual moral judgments, except by slowly pruning off any competing dispositions in our ancestors. So regardless of whether we end up calling that sort of process a predicative explanation or not, a more robust predicative reading — one that frames evolutionary pressures as literally “changing our minds” from one judgment to another — is pretty clearly off the table.1

All this is to say, the best a defender of the EDA could plausibly claim is that natural selection has slowly filtered out human beings who happen to develop very different moral intuitions, until only the ones who share our moral judgments are left — not that natural selection is responsible, here and now, for the judgments being made by the ones who remain.

To see why this distinction is so devastating for the EDA, let’s imagine we’re back at that Trump rally one more time. As you sit there listening to him ramble, you start to develop the radical notion that — hear me out — people who think Trump is a great president might not have the most reliable political judgments. The question, then, is this: If we all try our best to explain why the crowd supports Trump, then what sort of explanation, predicative or quantificational, would give us good reason to think their judgments aren’t truth-tracking? Or, to put things another way: If all we can give is a quantificational explanation for why everyone in the crowd thinks Trump is a great president, is that enough by itself to show us why they shouldn’t be trusted?

Clearly, a robustly predicative explanation for their judgments absolutely could do the job. Let’s say, for example, that we do some interviews after the speech has ended and realize that 95% of the audience forms their political views solely by reading Trump’s posts on Truth Social. In that case, given the extremely controversial assumption that Trump’s posts aren’t always made with a rigorous devotion to fairness and accuracy, this sort of explanation would be a pretty solid reason to think the views of the crowd aren’t reliable: Their individual political judgments are explained by something that’s indifferent to actual political reality.

Things look very different, however, when it comes to the impact of a quantificational explanation. As we said before, the quantificational explanation for why the audience feels the way it does about Trump is just that the rally’s organizers only invited people they knew would be supportive. But this, by itself, obviously isn’t enough to show that the supporters who did get invited are wrong, since the organizers played no role in the supporters actually forming their individual judgments. Crucially, this is the case regardless of whatever moral or intellectual qualities the organizers have; even if they themselves operate with absolutely zero regard for the truth — which I assume would be the case for a Trump campaign manager, honestly — that’s no reason (by itself) to think the judgments that attendees previously formed would be similarly unreliable.

The same thing is true, Hanson argues, when it comes to the EDA: If the most we can say is that natural selection happened to preserve a species-wide disposition towards certain moral judgments once those judgments developed, then that by itself doesn’t tell us anything particularly meaningful about their overall reliability — even once we acknowledge, as we should, that natural selection isn’t a truth-tracking process. I’ll quote Hanson at length here:

Read quantificationally, [the EDA] doesn’t entail that individuals only have the moral beliefs they have because they are fitness-enhancing; it entails only that people who have these moral beliefs, whatever their reasons for believing them in fact are, have tended to live longer and have more offspring, so that over time, more and more people exist who have them… [A]ll this is entirely compatible with the picture that moral realists want to endorse — that moral facts are mind-independent features of reality that we have an ability to discern. Beyond the dispositions being heritable and the beliefs being fitness-enhancing, [the EDA’s] claim, in the quantificational sense, doesn’t say anything about what these dispositions that underlie our moral beliefs are. Many kinds of dispositions could play this role, and the [EDA] read quantificationally does nothing to rule out what the realist wants to say, which is that the disposition that plays this role is an ability to discern moral facts.

Or, to put it another way: Every moral realist believes that human beings end up with moral knowledge somehow, whether through intuition, or by empirical observation, or something else — but just because that knowledge ends up being advantageous in an evolutionary sense doesn’t mean the realist has to think those evolutionary pressures themselves are responsible for it. Ironically, this is equally true for the anti-realist who thinks our moral judgments reflect nothing but fact-free subjective stances; even if natural selection can explain why a propensity for those stances is widespread among human beings, it’s still the propensity natural selection preserved that properly explains those judgments, not whatever process was responsible for preserving them.

So many anti-realists miss this point, I think, because they’re implicitly assuming that the sort of moral knowledge realists are talking about doesn’t really exist. And if you grant that assumption, then the EDA will make perfect sense: Without the possibility that we ever actually learn moral truths, all natural selection would ever have to work with is a revolving collection of randomly arising whims — and since we all agree that natural selection isn’t truth-tracking, the chance of it somehow wringing out a bunch of true judgments from random dispositional fluctuations is basically zero.

However, this way of thinking very obviously begs the question against moral realism, which necessarily involves the belief that (at least some) widespread moral judgments really do come from making contact with objective, mind-independent moral facts. If you think that sort of picture is reasonable, then the truth of the moral judgments in question doesn’t need to come from natural selection — those judgments will already be justified by whatever independently truth-tracking ability the realist has in mind, and natural selection just explains why that ability became widespread over time.

Of course, this doesn’t mean that moral realists are totally off the hook when it comes to defending the possibility of moral knowledge in the first place; without a plausible account for how it is that we as individuals gain moral knowledge, and an explanation for why that process counts as legitimately truth-tracking, anti-realists are right to say that moral realism is dead in the water. But that sort of individual account will have nothing whatsoever to do with natural selection, and to the degree that it is plausible, the further fact that natural selection preserved the capacity being described won’t matter one way or another. To quote Hanson at length one more time:

Again, it is often objected that realists owe — and are unable to supply — a convincing positive story about how [moral knowledge is obtained]. But again, EDAs purport to do something different from that kind of objection: They claim to show that whatever view realists hold about how it is that we discern moral truths, evolutionary considerations show that we can’t be doing that thing. If I’m right, this argument fails to establish this. Even if moral knowledge is deeply implausible on the realist picture — as well it might be! — the claim that our moral beliefs were selected for doesn’t make it more implausible.

So, in closing: If you’re an anti-realist, and you’re confident that our moral judgments are best explained by a purely psychological account, then feel free to make your case! There are all sorts of legitimate reasons for doubting the realist picture, and it could always end up being the case that anti-realism does ultimately win out on the merits. But just be aware that, whatever the problems with realism might be, our evolutionary history very clearly isn’t one of them.

Imagine, for a moment, that you work as the hiring manager for a local department store. Your boss, Mr. Wilson, is universally beloved by your colleagues for his steady leadership, stellar communication skills, compassionate demeanor, and impeccable business sense. However, because he works out of the corporate headquarters, you’ve never actually met him before, and he communicates with you solely through email. One morning you open up your inbox and see this message from him waiting: Hey, we had fifteen new applications yesterday that I need you to look through. And if any of them are from Jews, make sure to toss them out right away.

Naturally, you’re horrified, and decide right away that this couldn’t possibly be a real message from Mr. Wilson — especially since there are only seven applications waiting for you, not fifteen. But right as you’re about to dismiss the message as some nasty, antisemitic prank, a coworker tells you that he also saw it, and also thought it was a fake, until he realized that all the applicants would need to work a Saturday shift. So, he says, Mr. Wilson was probably just talking about Orthodox Jews, who can’t work on the Sabbath. He also points out that it’s possible to withdraw an application after you send it, which means that even if there are only seven right now, there still could have been fifteen yesterday, just like Mr. Wilson said. “So even though I know it looks bad at first glance,” he says, “I think the message really is from him.”

Right away, you see a major problem. Your coworker’s explanation certainly makes the message seem much less egregious. But still, you know Mr. Wilson would be horrified by the thought of his employees mistaking him for an obnoxious antisemite. (Or, for that matter, mistaking him for a boss who can’t keep track of how many applications are coming in each day.) Why, then, would such a thoughtful leader communicate in a way that so easily invites those exact misunderstandings? Thinking things over, you come to an obvious conclusion: Even if your coworker’s explanation can help to make the content of the message seem like something that Mr. Wilson might actually express, it does nothing to make the message itself seem like something that Mr. Wilson would ever actually send. For this reason, you decide it’s still best to ignore it.

In this piece, I want to lay out a similar meta-argument against Christianity — that is, I want to show that even if many historically and morally problematic passages in the Old and New Testaments can be interpreted in ways that avoid committing Christians to the glaring factual or ethical errors endorsed by fundamentalists, the notion that an all-powerful, perfectly good God would communicate his revelation in ways that made such destructive misunderstandings possible is ultimately no more plausible than the truth of those fundamentalist claims themselves.

If my argument succeeds, then the common accusation that popular atheist critiques of Christianity “lack nuance,” or fail to engage with the most sophisticated forms of the tradition, will turn out to be misguided; actually existing fundamentalism, at least with regards to certain important theological and moral doctrines, will ultimately set a firm ceiling on just how plausible Christianity itself could ever be.

I’m sure many readers will think my opening analogy was unnecessarily outrageous. However, even a cursory reading of the Pauline epistles, which form a major portion of the New Testament, will turn up plenty of passages whose straightforward readings are every bit as morally abhorrent (and whose progressive interpretations are every bit as strained). Take, for example, this passage from 1 Timothy 2:

Let a woman learn in silence with full submission. I do not permit a woman to teach or to have authority over a man; she is to keep silent. For Adam was formed first, then Eve, and Adam was not deceived, but the woman was deceived and became a transgressor. Yet she will be saved through childbearing, provided they continue in faith and love and holiness, with self-control.

I think it would be difficult for someone to argue that the plain meaning of this text is anything other than a blanket prohibition on female church leadership. There are, of course, other interpretations that come to a different conclusion, usually by pointing to historical context and certain quirks of Greek grammar and vocabulary. Still, the fact that all major Christian traditions universally interpreted these verses to forbid female leadership for more than 1500 years is, by itself, compelling empirical evidence that “I do not permit a woman to teach or have authority over a man” is most easily understood to mean just what it says; if the author of 1 Timothy really did hope to communicate something else, then his audience missed that intention immediately.

Similarly, the New Testament is also shot through with serious historical errors. One prominent example is how the Gospel of Luke’s describes of the census of Quirinius, and the many ways in which it directly contradicts our best understanding of when and how that census actually took place. As was the case with Paul’s letters, it’s easy to find apologetic responses that seek to harmonize those contradictions. However, the majority consensus among secular historians is still that the author of Luke was either simply mistaken, or else decided to use the census as a literary device that was (again) immediately taken as historical fact by his readers.

These two passages are obviously very different, but they share two related features that are relevant here: First, their straightforward readings are seriously implausible, whether for moral or historical reasons; and secondly, those straightforward readings have resulted in serious negative consequences, either by leading believers to adopt morally atrocious views, or else by undermining the confidence of non-believers in the reliability of the text. In what follows, I’ll refer to passages with these two malign features as destructive passages. Other examples of destructive passages would be all similar instances of Paul’s homophobia and misogyny, the references to eternal hell in Revelation, historical tensions between the Pauline epistles and the Book of Acts, and of course the myriad historical absurdities and divinely sanctioned atrocities of the Old Testament.

When examining these destructive passages, I’m going to make the uncontroversial assumption that, if God really has has chosen to reveal himself through the Christian scriptures, then those scriptures are generally as God wants them to be. Note that this assumption doesn’t require the scriptures to be necessarily inerrant, literally true, or even perfectly reliable. Instead, it only requires that the content of the scriptures be fixed by what God considers best overall. If we accept this assumption, then for any destructive passage in the Christian scriptures, there are only two options:

1. The straightforward reading of the destructive passage is true.

2. God intended to communicate a non-destructive proposition by means of a passage whose straightforward reading is destructive.

The crux of my argument is that, as a general rule, the implausibility of (2) will be at least as great as (1) — which is to say, if the straightforward reading of some Biblical passage seems to contain some major factual or moral error, then the probability that God would communicate his revelation through that passage will generally be as low or lower than the probability that the straightforward reading is true.

To see why this is the case, recall that the God of Christianity is traditionally defined as both perfectly wise and perfectly powerful. This means that God should have both a faultless understanding of what would communicate his intended revelation in the maximally effective way, as well as an unconstrained ability to actually communicate his revelation in that maximally effective way; God can’t have false beliefs about how revelation will be received, and he can’t be unable to give that revelation in whatever way he wants. For this reason, it must be that God is at least capable of successfully communicating his revelation in a maximally effective way, if he intends to do so.

Further, the God of Christianity is also essentially perfectly good. This means, among other things, that God will always do whatever is ultimately best for bringing about his morally perfect will — central to which is the sanctification and edification of his saints (Ephesians 4; 1 Corinthians 6; Hebrews 12) and the spread of the Gospel to as many persons as possible (2 Peter 3; 1 Timothy 2; Mark 16). Because God’s revelation plays a major role in facilitating both of these aims, it therefore must be the case that God does intend to communicate that revelation in a maximally effective way, so long as “maximally effective” is taken to pick out whatever method of communication will maximize the moral excellence of his followers and the factual reliability of the text.

I take these conclusions to be relatively uncontroversial, since it seems very odd for any Christian to argue that God could ever fail to be maximally effective in any of his actions, especially when those actions relate to his pursuits based in his fundamental goodness. (Near the end, I’ll consider the cost for a Christian to give this assumption up.) All that needs to be settled, then, is just whether God’s decision to communicate his revelation through destructive passages like these has been maximally effective in the way we should expect. If this isn’t the case — which is to say, if we have extremely strong reason to believe these passages have ultimately failed to maximize the moral excellence of believers and the factual reliability of the text — then we will be equally justified in believing that the Christian scriptures are not a legitimate revelation from God, and therefore that Christianity is false.

However, once we start to consider this question, what becomes immediately obvious is that the plausibility of a passage being maximally effective is directly related to the plausibility of that passage’s straightforward reading. As an example, we can return to the analogy I started with. The message supposedly sent by Mr. Wilson contained two issues: First, it seemed to miscount how many applications were received, and much more importantly, it also seemed to endorse obvious antisemitic bigotry. If we then ask ourselves how likely it is that an extremely gifted communicator like Mr. Wilson would ever send such a message, we see that the plausibility of him actually being the author has a ceiling set by the severity of the straightforward reading’s error.

In other words, as it becomes less likely that Mr. Wilson would endorse the apparent content of the message, it becomes correspondingly less likely that Mr. Wilson would express himself in a way that appeared to endorse that content; however seriously the straightforward reading of the message conflicts with Mr. Wilson’s true character, the decision to send that message would conflict with his character to a similar degree. The same is true, I argue, in the case of the Christian scriptures.

Returning to Paul’s clear misogyny in 1 Timothy 2, we’re presented with the dilemma I mentioned earlier: Either God does desire that women be excluded from positions of spiritual authority, or else he decided to communicate some other (possibly temporary or church-specific) set of standards in a way that, given his perfect foreknowledge, he was absolutely sure would meaningfully contribute to centuries of morally atrocious discrimination towards women that he neither desired nor endorsed. Crucially, just as in the case of Mr. Wilson, the implausibility of the latter option depends entirely on the implausibility of the former option. That is to say, however unlikely you think it is that God would actually endorse the marginalization of women, you ought to think it’s proportionately unlikely that he would ever communicate as though he did.

The same is true for historical errors like those in the Gospel of Luke. If you think the claims being made about the census of Quirinius are only slightly implausible, such that they only reduce the reliability of the text to a small degree, then you have good grounds for thinking that God may have permitted that small decrease in reliability so as to obtain some outweighing good. However, if you think the claims being made are very obviously false, such that they seriously degrade the reliability of the text, then you ought to also believe that the odds of God obtaining some outweighing good by allowing that reliability to be so seriously degraded drop in roughly equal proportion.

What this means is that, in general, the plausibility a person assigns to fundamentalist interpretations of Christian scriptures — that is, interpretations that rest on a plain and literal reading of the text — should set an approximate ceiling on the plausibility they can justifiably assign to any interpretation, on the grounds that those alternative interpretations have a plausibility tethered to the severity of the error involved in the straightforward interpretation. The idea that God communicates in ways that lead his followers to believe he’s a brutal and capricious bigot shouldn’t be any more palatable than the idea that he really is a brutal and capricious bigot; in fact, to whatever degree we’re confident that he isn’t such a wicked God, it should be correspondingly absurd to imagine that he’d willingly portray himself that way for millennia.

As I said earlier, this observation undercuts any defense of Christianity that involves painting its critics as “lacking nuance” or failing to engage with alternative views on how best to approach scripture. Many non-Christians who condemn the Bible for its obvious moral and historical errors do admittedly target fundamentalist approaches over their progressive alternatives. However, the assumption that the progressive’s alternatives will be more plausible overall is mistaken; all the concerns that lead non-Christians to criticize fundamentalism’s claims about the nature of history and ethics should also lead them to criticize, with equal confidence, progressivism’s claims about how God has chosen to reveal himself.

All this is to say: If we have good reason to believe that fundamentalist Christianity is false, then we have roughly similar reason to believe Christianity of any sort is false.

I want to finish with a few clarifications and qualifications, and then offer what I see as the best (and worst) way for a Christian to respond. First, I don’t believe the general argument I’m making will necessarily apply to every fundamentalist belief, since many have developed apart from the straightforward reading of any particular biblical text. Nor do I think it will apply to beliefs that have developed on that basis, but only rarely or in isolated pockets, like Appalachian snake-handling cults. However, I do believe it applies to many of the most controversial and destructive beliefs that the Christian church has adopted historically, and that continue to exert powerful influence in the modern day.

Secondly, I don’t claim that the progressive interpretation of a destructive passage will always be roughly as improbable as the straightforward reading of the passage. This is because a straightforward reading of some biblical passages, like those that seem to teach eternal conscious torment for all non-believers, would require Christians to take on some of the most radically implausible positions imaginable; in that context, no alternative interpretation could ever be as absurd. However, even in those cases, the unimaginably horrendous nature of the straightforward reading still does dramatically lower the plausibility of that alternative interpretation. Further, I do believe the rough proportionality I describe does apply, again, to many historical and current-day beliefs that have been adopted by large numbers of Christians.

As for what I consider the most plausible Christian response, I would think the only way to avoid the basic conclusion I’ve presented is to argue that God’s involvement in the scriptures is radically limited, such that many of these passages are not what he desires them to be. This would be the equivalent of saying that Mr. Wilson’s emails are sometimes edited after the fact by a bigoted and careless assistant, and that Mr. Wilson has simply never intervened to prevent the catastrophic misunderstandings that he knows are guaranteed. In the same way, God would be aware of the fact that his revelation had been seriously corrupted in ways that fundamentally misrepresent his nature, but fails to address the error.

Taking this approach would certainly free Christians from having to account for why God would ever author such problematic texts. However, it would also come with the heavy cost of radically reducing reasonable confidence in any of scripture’s teachings, while also raising the seemingly identical question of why a perfectly powerful and morally righteous God would allow such horrendous error to persist in his name. At this point, it’s unclear to me whether just “pushing things back a step,” as it were, has the power to solve this basic problem. So long as God is both perfectly powerful and perfectly good, his responsibility for the content of scripture seems unavoidable, and it’s this responsibility that ultimately dooms any apologetic explanation.

Meanwhile, the least plausible — but unfortunately most common — response would be to simply assert that God’s consistent, willful misrepresentation of both his own moral character and basic historical reality has, in fact, obtained some outweighing good that justifies it. What exactly such an important outweighing good might be is often left unspecified, but the implication seems to be that God permits error as a way of spurring on collective moral development over time, or else to encourage believers to wrestle with the text in a spiritually fulfilling way.

It would take an entire separate piece (like this one) to explain all the reasons I have for finding this explanation to be unbelievably implausible. Suffice to say, I’ve yet to see any even remotely convincing attempt to actually weigh up the harms caused by these passages and identify any competing spiritual or social benefits obtained by them that could not have been obtained by God simply directly communicating the actual moral and historical truths that he desires his followers to eventually discover. In the absence of any such account, the notion that our optimal moral development required catastrophic moral and historical error, driven largely by God’s deliberate misrepresentation of his own fundamental character, is hardly more convincing than the fundamentalist’s notion that those misrepresentations are actually true.