Savage's framework grounds expected utility in preference axioms that are silent on whether relevant...