Internal Realism

Internal realism (e.g. Ellis) ‘> epistemic values are relative’> obvious response is to say, but some are better than others, as shown by their success rate ‘> but this is probabilistic’> and (not obvious) there’s no known, good way to evaluate probabilistic inference procedures (without a subjective prior, which is of course relative) ’> epistemic values ARE relative.

This is right! But it’s too ambitious for a paper. Maybe a book one day. Or a PhD topic.