How can moral epistemology taxonomies organize robust objectives?

Many alignment proposals attempt to capture the True Name of human value as a target to aim optimizers at. However, any attempt of implementing such proposals would inevitably carry over assumptions of moral epistemology and moral ontology: how is knowledge about morality gained and how is it shaped.

Moral ontology considerations:

• (moral) universalism: There's a universally-applicable normative framework out there. [[deontic-arrays]] generally buy into this. [[memetic-colonies]] partially buy into this, because although the seed memeplex is fixed, it is different from the forecasted ones.
• (moral) relativism: To each their own, normative frameworks are grounded in cultures or individuals. [[memetic-colonies]] might extend into this by cultivating multiple seed memeplexes. [[deontic-arrays]] might move towards this stance through hierarchical charters spanning a global-to-local spectrum.

Moral epistemology considerations:

• (moral) rationalism: Moral truth can be attained by rational deliberation or reflection alone. The two True Name-like proposals currently on this website don't align with this stance that much.
• (moral) empiricism: Moral truth is attained through observation. [[memetic-colonies]] are the closest to this stance, as they involves observing memetic simulations. [[deontic-arrays]] feel somewhat in the middle, as the counterfaction cross-validation scheme has a large statistical component, though the way the charter is obtained is not specified.