Abstract

Peter Walley argues that a vague credal state need not be representable by a set of probability functions that could represent precise credal states, because he believes that the members of the representor set need not be countably additive. I argue that the states he defends are in a way incoherent.