This was a good post. 
Focusing on (2) for a moment, there are two significantly different approaches that I want to single out, and they have to do with the idea of “retrospect.”
Let’s simplify the idea:
2a. There is some moral sentence that is true even if everyone believes and acts otherwise.
On a “classical” view 2a is true in the sense that everyone who disagrees with the moral sentence would, in retrospect, see that they were wrong and that the sentence was in fact true. Elaborating, the idea is that, given enough time, effort, and fortune, each person could be moved to a different vantage point in which the retrospective claim obtains. That vantage point would itself both be and be seen to be superior to the previous vantage point. “Vantage point” is meant to be very broad, and could involve things like habits, ways of thinking, speculative insights, breadth of experience, logical capacity, affective patterns, etc. The classical view is in no way limited to forms of rationalism, and therefore it is possible (and probable) that the shift of vantage point would not necessarily occur simply by (for example) reading and writing text on a philosophy forum.
An opposed view, which I think is quite popular, denies all this. On the opposed view there is no possibility of a retrospective acknowledgement of the moral truth. For this view the moral truth is somehow per se unknowable by one or more individuals.
This goes hand in hand with the idea that “morality” is undefinable or nonsensical (which is tied to the Kantian inheritance). The paradox is that, for the second group, “moral truth” is translated as “something that cannot be reliably known.” It is understandable to deny the existence of some species of truth, but the uncontroversial error lies in translating the interlocutor’s statement about “truth” as if it were a statement about a per se unknowable proposition. As if the interlocutor thinks they know of a truth which large quantities of people could never come to know even in principle.