Metaethics and Unconditional Mattering
We should oppose gratuitous suffering, no matter what's true
Here’s a moral datum: we should be unconditionally opposed to gratuitous suffering. (More broadly, we should be “pro tanto” opposed to all suffering—which leaves open that it may yet be outweighed in the event that there were sufficiently high stakes on the other side of the ledger. So we’re not talking about absolutist refusal to consider tradeoffs here. Rather, the claim is just that suffering must be counted as something that is “in itself” bad, as far as it goes.)
So our opposition to suffering should not be conditional on our own desires, on what our idealized selves would conclude at the end of inquiry, on the distribution of non-natural properties, or on the philosophical question of what metaethical view is correct. For each such candidate condition X, when we ask the question, “Is it still important to not torture, conditional on X?” we clearly ought to answer in the affirmative. Suffering matters, no matter what!
Capturing the Datum
It’s a tricky question what metaethical views are compatible with this datum. Non-cognitivists are often accused of making moral facts too contingent, as they might easily have endorsed or expressed pro-torture norms instead. Against this challenge, expressivists respond that they can endorse appropriately unconditional anti-torture norms. That is, they can express disapproval of [torture, even in that possible world where they approve of torture]. So their norms can take universal scope over possible worlds, even if it is contingent what norms they actually endorse and express.
Constructivists may face similar challenges. After all, it seems a contingent matter what your idealized self would conclude at the end of inquiry: you (or an ideally coherent Caligula) might turn out to be pro-torture, and it hardly seems fair to the torture victims to hold their moral status hostage to your attitudes in such a way. But perhaps constructivists could adopt the expressivist’s defense and insist that at least the norms they’ve contingently constructed are actually unconditional ones?
Moral realists, by contrast, have non-contingency secured: given that torture is bad in itself, it couldn’t have been otherwise. But critics still argue that realists make the badness perversely conditional on its possessing a further normative property, as though suffering by itself was not bad enough.
(I address this concern in my paper, ‘Why Care About Non-Natural Reasons?’ In short, the objection rests on a misunderstanding. Realists don’t hold that normative properties are what matter; their role is instead to mark which natural properties matter, objectively distinguishing them from the others that don’t. But it is just the identified natural properties—suffering and such—that do the mattering; and they matter fundamentally, rather than as a means to securing “properties that have the higher-order normative property of mattering”, or anything like that. Accordingly, what moral agents ought to care about is suffering and such, not “properties that have the higher-order normative property of mattering.”)
Can we stipulate “not mattering”?
Suppose that suffering didn’t matter. Would it matter then?
Something has gone wrong here. Obviously as a matter of logic, the answer must be ‘no’. But this is in some tension with our moral datum that we should be unconditionally opposed to suffering. (Should we be opposed to suffering, even conditional on its being the case that we shouldn’t be opposed to suffering?)
I’m tempted by the thought that our first loyalty should be to the actually-correct moral norms, rather than to the abstract idea of moral correctness (morality “de re” rather than “de dicto”, as philosophers say). Indeed, this is just what our datum was meant to capture: we should care more about suffering and such than about morality, such that our concern about the former is in no way conditional on claims about the latter. Furthermore, this is arguably just what morality itself requires. (It’s virtuous to care about others’ well-being, not to care about one’s own virtue.)
An important upshot: Even when entertaining some impossible scenario with contrary-to-fact (& necessarily false) moral stipulations, our evaluation of the scenario should be guided by the actual moral facts rather than the stipulated ones. Actual morality is greedy in this way, following us into hypothetical scenarios and stipulations, even when uninvited!
(This may not be true of all moral truths—there’s room for moral uncertainty, after all. Here I’m just talking about the core of obvious moral truths—that suffering is bad, and such—that we all ought to know.)
Some Realist Wagers
I’ve long been sympathetic to the argument that one “might as well” be a normative realist, as it can’t very well be a normative fault to believe in normative facts. (Either you’re right, or it doesn’t matter that you’re wrong.) Moreover, thinking recently about ways that metaethical views might affect our first-order normative beliefs, I’m drawn to the view that one should simply assume realism in forming one’s first-order views—at least if I’m right that the main effects of this are to push one towards greater impartiality and a more expansive moral circle. It could be a serious moral error for one’s moral vision to be too restricted, whereas it’s harder to see such grave moral risks to being too self-sacrificing or having an otherwise expansive moral view.
But one must be careful here. As Joe Carlsmith argues, we shouldn’t accept wagers that yield horrendous outcomes in the event that nihilism is true, even if (by stipulation) the horror “wouldn’t matter” then. (Carlsmith puts a subjectivist spin on this: “you may notice that you don’t want to be burned alive, even if nihilism is true, and it doesn’t matter that you’re being burned alive.” But it isn’t clear that a subjectivist account of mattering fares any better here. You may notice that you don’t want to be burned alive, even if it turns out that you DO want to be burned alive. Screw that guy’s preferences, right!? Even if he is me…)
So, again, I think it’s an important part of moral realism that certain core moral commitments should be unconditional. Roughly, we should assume that realism is true for determining what matters, and we should then proceed to take those things to matter unconditionally—so far as our practical commitments are concerned—even on the supposition that moral realism is false, or that somehow nothing “matters” at all.
We should be open to considering scenarios in which nihilism turns out to be true, but what we should be concerned about is avoiding, in that scenario, outcomes that would be evaluated negatively by the lights of the actually-objectively-correct moral view (no matter that we just stipulated nihilism to be true of the scenario). The nihilist world matters, but it never gets to provide the standard for evaluation, even of itself.
That’s just part of what we’re committing ourselves to, when we commit ourselves to caring unconditionally about that which actually matters (rather than about the mere concept of “mattering”).