> Idk, it could just be a claim that the argument can be expressed in a logical system, combined with the normative axiom that rational agents should only trust arguments that can be validly expressed in such a formal system.

The term "valid" has both formal and informal senses. I agree some of its formal senses are non-normative, but I think its informal uses have a strong normative element. The original comment to which I was responding wasn't very clear about in what sense it was intending the term, but when people aren't clear, it is not unfair to assume they mean a term informally rather than in some specific formal sense.

> Regardless this seems to be missing the forest for the trees. Nobody has put forth any arguments for either position that are emperical in nature. Is there really a practical difference between impossible and possible but nobody knows how?

What is the proper role of empiricism in human knowledge, including moral knowledge? You seem to be making some assumptions here which you are not making explicit.

If one accepts moral sense theory – that moral intuitions are a form of empirical knowledge – then we can have direct empirical knowledge of the truth or falsehood of moral propositions.

If one accepts pragmatism – that beliefs can be justified by the pragmatic benefits of believing them, and that pragmatically justified beliefs can constitute genuine cases of knowledge – then empirical evidence can be used to establish the pragmatic benefits of a moral belief, and hence (indirectly) to establish its truth.

One may believe that some axioms are self-evidently true, and that one's knowledge of the truth of those axioms is innate rather than empirical – in which case, if one accepts that for non-moral axioms, how about accepting it for moral axioms as well?

The practical/empirical argument would essentially be to extrapolate out what the world will become like if we spread X or Y normative belief/practice.

You can argue that one will lead to a preferable world than the other.

Of course, we may have different preferences, so we may not agree even if we agree on that extrapolation. But often basic human preferences like not wanting to be the victim of unprovoked public violence do align.

Consider this (meta-)moral principle: For some moral proposition P, if universal human obedience to P would lead to human extinction, then P is false.

Example of application: Consider "It is always immoral to have children". Obviously, if humanity universally obeyed that principle, human extinction would inevitably and quickly follow. Therefore, that principle is false. From which obviously follows: "It is moral to have children (at least sometimes)".

Now, I would propose we accept that meta-moral principle as an axiom. To me, and to many others, it seems self-evidently and intuitively true. Who seriously doubts or denies this axiom? Well, most of the people who doubt or deny it, do so not due to its specific content, but due to a philosophical objection to accepting any moral axioms at all. The number of people who are happy to accept moral axioms but would refuse to accept this particular one appears to be rather small. It seems as good a candidate to be an axiom as any other. So why not accept it as an axiom?