Maybe my reaction there is that it's a symptom of a disease? But in the incomplete preferences case I think the preference for the lottery is (more) easily explained once we have the sets of utilities idea on the table (or for that matter ex ante pareto). But I may not be responding to your concern?

Oh I was kind of thinking the opposite. Most of the arguments I’ve seen against imprecise credences appeal to principles that you show can’t be right in full generality.