A thing that occurred to me in another discussion:

A big problem with labels (in the sense of identity politics) is their context. If we're working to discuss things in a formalized way, maybe to classify or quantify particular things in as part of a study, clear meanings that can encompass all possible data points are necessary, even for some qualitative work. Outside of that context, labels and spectra are mere social convenience and not everyone must have a place in regard to them.

I think a lot of argument comes from a confusion of ideas between those two contexts.