Some Questions Regarding Moral Nonnaturalism

Can't speak to (ii) so I'll just focus on (i). Consider, for example, relativism, which I'm taking to be speaker subjectivism--the view that moral terms have an indexical component (indexicals are words like 'I', 'here', 'now' etc.), so that sentences with moral language are only true relative to a given speaker. Not only does this view come with a metaphysical thesis (that there are no deep "objective" moral rules that oblige everyone, no intrinsically normative entities, at least in the sense realists would use these terms), but it also comes with a semantic thesis telling us what moral terms actually mean (that moral terms are like indexicals). Different metaethical views might agree on the metaphysical thesis but disagree with the semantic thesis; expressivism, for example, or MacFarlane's assessment-sensitive semantics. Now, the semantic thesis for speaker subjectivism has a clear problem. When Bob asserts "Euthanasia is wrong" and Alice asserts "Euthanasia is not wrong, Bob and Alice take themselves to be disagreeing about something--Alice can coherently make her claim to contradict Bob. Even if both Bob and Alice are Uber Enlightened Moral Relativists who believe that there are no neutral grounds on which one can convince the other of their claims, they are still ostensibly making assertions which are incompatible. If Bob says "Euthanasia is wrong" and Alice says "euthanasia isn't wrong," it wouldn't make sense for Bob to reply "Yeah, you're right" unless he's retracting his first statement.

Compare this to the uncontroversial cases of speaker subjectivism, such as the word 'I'. Bob might say 'I am tired' and Alice might say 'I am not tired,' and they will not take themselves to be disagreeing; their claims are not contradictory, and Bob might well respond to Alice's claim with "Yeah, you're right," without having to retract his earlier claim. There's a more purely semantic way of proving this point. Notice how if a sentence S has no indexicals, and someone utters S, you can report their statement by saying "That guy said that S." E.g. if you utter "grass is green," I can go tell someone "Leon said that grass is green." This doesn't work for indexicals. If you utter "I am tired," I can't tell someone "Leon said that I am tired," that would be false. Rather, I must evaluate the indexical according to the context in which it was uttered; I must say "Leon said that he is tired." But note how if you utter "Euthanasia is wrong," I can go ahead and tell people "Leon said that euthanasia is wrong," just like in the non-indexical case. If speaker subjectivism were true, we'd expect "wrong" to work like "I", which it apparently doesn't.

Now, for (i). Couldn't people just talk differently so that moral terms do work like indexicals? E.g. could we have a term "good-according-to-X" so that "Y is good-according-to-X" just means X approves of Y, and could we further make it so that "Y is good" means "Y is good-according-to-me" by default? If that were so, the above problems wouldn't arise. If you said "Euthanasia is wrong," I would have to report your utterance by saying "Leon said euthanasia is wrong-according-to-him." Could the relativist not claim that relativism really is true, people just don't talk this way because the superficial similarities between normative and descriptive language trick them into thinking different agents are disagreeing about something when they have different moral views? If I want to solve ethics, why should I care about mere linguistic conventions?

It would indeed be silly to act as though we could derive Deep Moral Truths just from conventions about how we use words. But this is why, earlier, I talked about the metaphysical and semantic thesis of moral relativism. While the moral relativist may have her conclusions about Deep Moral Truths unaffected, she must admit that her semantic thesis is false, in that they give an inaccurate picture of what moral terms actually mean. Perhaps there is a possible alternate language, call it English2, in which 'wrong' works differently. But the English2 word 'wrong' is a different word than the English word 'wrong'; metaethicists in part want to know what our moral terms mean, and by talking about the "moral" terms of another language, you're changing the subject matter. Just as you can't make substantive discoveries about pies by investigating the referents of the Spanish word 'pie' ('pie' means 'foot' in Spanish), facts about the English2 word 'wrong' don't necessarily tell us anything about what moral terms mean. The relativist, therefore, must revise her semantic thesis.

So while I agree with your apparent position that the ultimate normative principles don't answer to conventions of the English language, semantic facts can substantively inform the subject matter that metaethicists are concerned with.

The discussion above will way out a bit differently depending on what position is being criticized on semantic grounds, e.g. it would be different were I to replace speaker subjectivism with expressivism or naturalism. Let me know if you want me to elaborate on that.