You Can't Disprove Moral Realism By Just Assuming That It's False

Or: Where There is Only One Set of Footprints in the Sand, That Was When the Moral Facts Carried You

Pictured: Gilbert Harman and Nicholas Sturgeon

In this post, I’m going to talk about a very common objection to moral realism and why I think it fails. This objection has to do with the process of moral reasoning, and whether real, honest-to-goodness moral facts play any sort of role in it. The worry is that they don’t, which would be bad news for realists. In what follows, I’m going to draw heavily on the work of two philosophers, Gilbert Harman and Nicholas Sturgeon; Harman is the anti-realist who crafted the argument in the first place, and Sturgeon is the realist who (in my opinion) best explained why it doesn’t work. I’m not going to quote either directly, but this whole post should basically be seen as a friendly overview of one section in Sturgeon’s paper “Moral Explanations,” where he lays out his critique of Harman most directly. It’s one of the best papers ever written on the subject of moral realism, so if this post interests you at all - especially if you’re still unconvinced by the end - then you should go and read it in full.¹

I’ll start by laying my cards out on the table: I’m a moral realist, which I just take to mean, as Judith Jarvis Thomson put it, that there are some moral statements where we can find out if they’re true. We find out if these moral statements are true by considering all the relevant moral facts, which are just facts about what’s right and wrong, or good and bad, or what’s honest, gracious, cruel, patient, selfish, brave, and so on. Here are three moral facts I think are true, if you want some examples: Breaking a promise for selfish gain is bad; a good person wouldn’t torture someone just because they enjoy it; if you come across someone drowning in a lake, you should try to save them. These facts are as objective as any other type of fact, like the fact that penguins can’t fly or that riding a motorcycle without a helmet is dangerous. I’m also a moral naturalist, which means I think facts about how penguins get around and facts about what you should do when you come across someone drowning in a lake are ultimately the same sort of fact, and that we come to know all these facts through the same general process(es) of observation, experiment, and conceptual analysis. So on my view, the reasoning we use to come up with our moral beliefs is a lot like the reasoning we use to come up with our scientific beliefs - or at least it can be, if we try hard enough.

Moral naturalism is fairly popular among professional philosophers today, but many regular people still find it unconvincing or even obviously false. Some of those people are also moral realists - just another kind of moral realist, the kind who think moral facts are nothing like facts about the natural world. But plenty of other people are skeptical that moral facts exist at all. There are facts about what we believe is morally good and morally bad, they might say, just like there are facts about what we believe is beautiful, or boring, or funny. But these are just preferences, and there’s no real fact of the matter as to what’s objectively morally good or morally bad (or beautiful, or boring, or funny). To these people, moral realists are making a classic mistake; we’re taking something that only exists in our heads, projecting it outward onto the world, and then going out to “discover” it again.

One of the most common reasons people will give for taking this position is that moral facts, to them, seem unnecessary. I mean, think about yourself when you engage in moral deliberation: You make some observation, consider that observation in light of your own moral beliefs, and render your judgment. Case closed, it seems - so where, exactly, did the moral facts come into play? There are obviously objective facts about the situation you observed, and there are subjective attitudes possessed by you as an observer. But those two things are all you really need give a satisfactory account of the deliberation process, so even if moral facts do exist, it’s hard to see what sort of role they could even play. At best, they seem to just hover “above” our judgments; what we believe to be right or wrong might occasionally line up with the moral facts, if we’re lucky, but those moral facts are never the reason we believe what we do. Or, to put it another way: Even if nothing really was objectively good or bad, we’d all keep trucking along, making moral judgments as usual. So if moral facts are a bunch of ontological freeloaders - swooping in to take credit for our moral beliefs when they didn’t do any of the work needed to produce them - then what’s the point of even keeping them around at all?

If you’re reading this, there’s a good chance you find that sort of objection convincing. But instead of explaining why I think it’s a bad one, I’m going to ask you to consider another type of skeptic, who rejects not moral but physical facts. This skeptic says something like: “Well, of course there are facts about what we perceive to be round, or ten feet tall, or made of ice. But those are nothing more than facts about subjective perceptive states, with no objective truth behind them. It’s lucky for us that our perceptive states tend to roughly line up, but if you think there’s a table in front of you and I don’t, then no one is right or wrong - we’ve just got something different going on up in our heads, and it’s a mistake to think there’s some table ‘out there’ that could actually decide between us.” I’m sure most of us would think this position is insane, but be careful! This skeptic, if pressed, could make a very similar argument to the one a moral skeptic makes. All we really need to account for our judgments about the external world, after all, are the facts of our perceptive experiences and the subjective beliefs we have about them. I have an experience of a table, I think to myself, “There’s a table there” - what’s left to say? Even if there were no physical facts at all, our skeptic reasons, we would all still go on making the judgments we do about tables, or chairs, or any other objects. So why should we invoke real, honest-to-goodness physical facts lurking somewhere in the background?

Here, the response would be obvious: Our skeptic is just wrong to say that we would continuing making our normal judgments about tables or chairs in a world with no physical facts. This is because a world with no physical facts has no tables or chairs, or any physical objects at all. And because we rarely make judgments about physical objects that don’t exist, we’re perfectly within our rights to say that a world without those objects would also lack any judgments about them. Of course, this argument requires assuming that our own theory of the world is true. But crucially, so does the skeptic’s. He’s just assumed that judgments about physical objects have no relation to physical facts, and then he deduces from his assumption that those judgments would survive the elimination of those facts. But this is a trivial conclusion, right? We all agree one thing can survive the disappearance of another if there’s no dependence between them; what’s actually up for debate is whether there is such a dependence in the first place! We want to know whether I believe there’s a table in my kitchen because there really is one, and if that’s my goal, it doesn’t help me to know that, if there wasn’t a table, I would still think there was. I already know that I think there’s a table there. So if there isn’t one there in reality, then I’m mistaken, and something other than the table is causing my belief. But on the other hand, if I’m right about a table being there, then it’s obviously because of the table (get out of here, Gettier) and everything checks out. What matters is whether I have any actual reason to believe one way or another.

Getting back to the original objection, I’m going to say that the moral skeptic is making the exact same mistake. Their objection hinges on the assumption that moral judgments would stick around in the absence of moral facts, and therefore moral facts are unnecessary or unneeded. But the moral realist has no reason to accept this, because it’s just an assumption that moral realism is false! If moral realism is true, then moral facts play an obvious and direct role in our moral judgments; we judge that stealing from friends and family is wrong because it actually is wrong, just like we might judge there to be a table in front of us because it’s actually there. And if this is the case, then we have good reason to reject the idea that moral deliberation would continue uninterrupted in the absence of those facts. Of course, it’s always possible to assume some entity doesn’t exist, and from that assumption demonstrate that it can’t explain what it’s supposed to explain. But if, when you assume the entity does exist, it plays a perfectly functional role in an explanatory framework, then you can’t treat the mere possibility of catastrophic error as a blow against its explanatory power. Could it be that I make the moral judgments I do entirely in response to my subjective attitudes, totally unrelated to any actual objective fact? Yeah, it’s possible. It’s also possible that I judge myself to be sitting in this chair even though no such chair exists. But both of these possibilities are just specific instances of the general fact that every theory could be wrong. That possibility, by itself, gives me no reason to believe any particular theory actually is wrong; if it did, then every non-trivial theory would come already primed with self-refuting evidence.

The reason this point is obvious in the case of the external world skeptic and not obvious in the case of the moral skeptic is because we assign a different prior probability to the two theories involved, and that changes how we interpret the situation being posed. To see this, we have to look carefully at what the moral skeptic is really asking when they prompt us to imagine what things would be like if nothing was objectively wrong. Questions of the form “What would things be like if X?” can be interpreted in at least two different ways. One way would be to treat the question as an imaginative exercise, where we conjure up the closest possible world where X is the case and describe what that world is like. But we can also take a different approach, where we see the question as asking what the actual truth of some fact X would imply about the actual world we’re living in right now. Which interpretation seems reasonable depends on both the context of the question and exactly how it’s worded. Let’s say, for example, you ask me what things would be like if my wife was married to someone else. Unless I had reason to think otherwise, I’m going to assume you’re asking about another way my life could have gone, as opposed to asking me to consider what evidence there might be that my wife is actually, right now, in two different marriages. But if we’re detectives at the scene of a murder and I ask you, “What if a serial killer did this?” then obviously I don’t want you to start imagining how things would be different in some alternate universe where a serial killer was responsible; I’m asking you to consider what would have to be the case if, in fact, that’s what happened.

Similarly, when the moral skeptic asks us to consider what things would be like if nothing was objectively wrong, we can take the question in one of two ways: We can assume there are instances of objective moral wrongness in this world, and consider how things would be different if there weren’t; or we can assume there are not instances of objective moral wrongness in this world, and consider what else about this world has to follow from that. If you take the question in the second sense - the “assume this is the way things actually are and ask what follows” sense - then you should definitely conclude that moral facts play no role in moral deliberation (on the grounds that moral deliberation is, in fact, taking place in the absence of any objective wrongness). But what, exactly, does that conclusion tell me? Of course it’s true that, if there are no moral facts, then I’m making my moral judgments on the basis of purely subjective psychological states. (What else would even be left as an option?) But as I’ve said before, this just boils down to the fact that, if I’m wrong, I’m wrong. You’ll have to do better than that!

If, on the other hand, you take the question in the first sense - the “assume things are different from how they actually are and tell me about that world” sense - then the moral realist can give a perfectly realist-friendly answer. Remember that, for naturalists like me at least, objective wrongness supervenes on natural features like suffering, inequality, malice, and the opinion there are movies better than Tango and Cash (joking on that last bit, but only barely). So if you ask me to imagine a world without objective wrongness, my natural instinct will just be to imagine a world that doesn’t have those things. A world where nothing is objectively wrong would be a world without war, poverty, prejudice, and greed, because those things are objectively wrong and therefore, by stipulation, can’t exist in the world I’m imagining. This wrongness-free world would be very, very different from the world we live in today, and that actually strengthens the realist position, rather than harming it. After all, the moral skeptic is trying to get us to admit that moral facts don’t make any sort of difference. But if the realist is correct here, they make a huge difference, and th-

BRRRRR. Stop! The moral skeptic calls a foul. Obviously, this isn’t what he meant when he asked me to imagine what things would be like if nothing was objectively morally wrong. The point wasn’t to conjure up a world that just happened to have no instances of objective moral wrongness. Rather, it was to take the instances of objective moral wrongness that I recognize in this world and imagine what things would be like if we stripped that objective wrongness off. But asking the realist to imagine a world where horrendous cruelty or viciousness exist but are not objectively wrong is, again, just asking the realist to imagine that realism is false and the skeptic’s view is true. And what could anyone learn from that? Of course, I’m guilty of the same cheap move in the last paragraph when I said that the skeptic’s question was evidence for realism, because I was assuming the skeptic would accept a counterfactual informed by my theory. In reality, both counterfactuals are argumentatively inert, because the plausibility of one as compared to the other will be directly informed by the two theories they exist to test. If you believe moral facts exist, then you’ll naturally conceive of a world without objective wrongness as a world that just happens to lack bad things; if you believe there are no moral facts, then you’ll naturally conceive of a world without objective wrongness as… exactly like this one.

These dueling counterfactuals can be valuable in an illustrative sense, but they can’t actually tell us anything about the two theories in question, because each counterfactual will necessarily validate the theory that produced it. It would be like determining whether a certain medicine is effective by imagining what the world would be like if you gave it to everyone with the related illness; what that thought experiment would tell you is entirely a function of whatever preexisting beliefs you had about the treatment’s efficacy. So when the moral skeptic argues that we would still make the moral judgments we do, even in the absence of any objective moral facts, the realist’s response is simple: Yes, if moral realism is false, that’s true - but why should I think it’s false? Of course it must be that, if no moral facts exist, then our moral judgments have nothing to do with them. And if our moral judgments have nothing to do with moral facts, then our judgments wouldn’t change if those facts were added or removed. But it’s also true that, if electrons or amoebas or asteroids didn’t exist, then our judgments about electrons, amoebas, and asteroids would have nothing to do with them either. And if our electron/amoeba/asteroid-specific judgments have nothing to do with electrons, amoebas, and asteroids, then they wouldn’t change either if any of those things suddenly popped into or out of existence. But this is not a good reason to believe there actually are no electrons, amoebas, and asteroids. So why, exactly, is it a good reason to believe there are no moral facts?

The ultimate point here is pretty simple: No theory can survive the assumption that it’s false, and while you’re well within your rights to ask that a theory be tested in some possible world other than our own, but that possible world can’t be one in which the theory being tested isn’t true! Of course, this isn’t a license to adopt whatever insane theory you’d like, on the grounds that, assuming it’s true, then everything must work out. You still need reasons, from the outside, to explain why you think what you do. But if you do have those reasons - if your theory has plenty of explanatory power, parsimonious structure, inherent plausibility, and so on - then you’re well within your rights to assume, from the inside, that it’s true; the mere fact that, if you’re wrong, you’re wrong, shouldn’t move you. In my next post, I’ll tackle a positive case for moral realism and why I think we do have good reason to adopt the realist framework. But for now, I just thought I’d explain why you can’t rule it out from the start.

1. If you’re interested, it’s collected in Morality, Reason, and Truth: New Essays on the Foundations of Ethics, edited by Copp and Zimmerman. You can find a PDF online pretty easily, but feel free to comment and ask me for a copy!