We’ve been talking a lot about meta-ethics in other threads lately. I had planned a while back to eventually start a thread about my own meta-ethical system, but since my last thread on my principles of commensurablism basically turned entirely into a discussion just about the implications on that, I thought I would instead start a thread about the meta-ethical possibilities generally, and why I fall where I do among them. I hope and expect that this will lead eventually into a discussion about philosophy of language more generally, as that’s where my thoughts had to go to figure my way out of the apparent trap that the usual span of meta-ethical options left me in.

Wikipedia has a great breakdown of common meta-ethical positions and how they relate to each other. The rest of this OP is assuming you’re familiar with that, so give it a glance if you’re not

Cognitivism seems like the default position most people start from: moral assertions are the kinds of things that can be true or false. I certainly started out and ended up there myself.

The most common kind of cognitivism seems to be a form of subjectivism, divine command theory (what is good is what god commands). But for those who reject that view, like me, the most obvious alternative seems to be a form of moral realism: there are some facts about reality that make some things moral and other things not.

The most straightforward candidates for such facts, especially for someone like me who had rejected the supernatural generally at the same time as rejecting divine command theory, are natural facts. Most Modern-era normative ethical theories like utilitarianism seem to just assume this, and that seems natural enough (pun intended).

But there are a lot of good arguments like those put forward by the father of meta-ethics, GE Moore, against the possibility of natural facts constituting moral facts; notably the Open Question Argument, which is much akin to Hume’s is-ought problem or the fact-value distinction. Moore thought that this entailed that moral facts must be non-natural; and yet, he was still otherwise a realist.

For those like me whose physicalist ontology cannot admit of some strange non-natural but nevertheless real moral facts, non-naturalism is a non-option. (This is also called the Argument from Queerness: wtf is a non-natural moral fact like?). But together with the defeat of ethical naturalism, that rules out all of moral realism. That puts us back in the subjectivist camp, but without a God to defer to, we’re left with a more ordinary subjectivism, a kind of cultural relativism: what is good is what your culture says is good.

Yet that seems vulnerable to the same Euthyphro-esque attacks as Divine Command Theory: is what culture says is good good just because they say it is? (if so that seems rather arbitrary and unjustified) or do they say it's good because it actually is? (if so then it's not them saying it that makes it good and we're back to where we started: what actually makes it good?). Subjectivism also seems to undermine a lot of the point of cognitivism in the first place: if everybody's differing moral claims are all true relative to themselves, then it doesn't seem like any of them are actually true at all, they're just different opinions, none more right or wrong than the others.

With both realism and subjectivism ruled out, the only cognitivist option left is error theory: everyone is trying to make truth-apt claims when they make their moral assertions, but all of them are categorically false. That seems just patently absurd, though, that everyone categorically are trying but failing to do something with language that just cannot possibly be successfully done. So then it seems much more reasonable that instead, people aren't trying to do that with language at all; they're trying to do something different altogether.

That takes us finally out of cognitivism. The most straightforward kind of non-cognitivism is emotivism, a kind of expressivism which says that rather than making assertions of the kind that can be true or false, moral claims are just expressions of emotions, like "boo this" and "yay that". Like error theory, this ends up effectively a kind of moral nihilism: no moral claims are ever actually right, though unlike error theory, they're not all wrong either, because they're all not even wrong. That, I think, runs counter to most people's impressions of what they're trying to do with moral language, as evidenced by the fact that people actually have moral disagreements where they think that one person is right and the other is wrong.

There are other non-cognitivist possibilities. Some hold moral discourse to be a kind of shared fiction, a projection of our feelings as though they were real facets of the world. That, to me, seems too much like error theory, mixed with some of the flaws of subjectivism: nothing is actually moral, but we agree to pretend like certain things we agree on are actually moral.

There's one other non-cognitivist theory, universal prescriptivism, which holds that moral claims are actually more like imperatives, commands that we're giving each other, but specifically commands that are universalizable: commands that we're committed to issuing and adhering to in all similar circumstances. That one, I think, gets very close to a solution, because in being non-cognitivist it gets around all the problems with those views detailed above, but also escapes the nihilistic implications of other forms of non-cognitivism, and tries to salvage not the robust moral realism we discussed above, but a more minimal moral "realism", a mere moral universalism.

...

I'm going to hold off on discussing where I went from there for now, and let everyone discuss whether all of these conventional options are so far insufficient, and we're in need of something new and different, which is where I found myself before trying to come up with such a new and different option.

It looks to me like we need something that’s compatible with ontological naturalism or physicalism, that honors the is-ought divide, and yet still allows for moral claims to be genuinely truth-apt and not mere subjective opinions — and none of the conventional options above satisfy those conditions.