When L.J. Savage used three categories to classify theories of probability he put Jeffreys in the category labeled .necessarist..In Jeffreys's refinement of his earlier position, reasonable degree of belief is a primitive concept and initial probabilities are ultimately a matter of convention rather than a correct description of ignorance.Because personal judgment is not fundamental in Jeffreys's scheme, I disagree with Zellner (1982) when he describes Jeffreys's philosophy as "individualistic".