Why indeterminate probability is rational

This essay has two objectives: (1) To offer a general motivation for the use of sets of probabilities to represent states of credal probability judgment. Doubt as to which of rival answers to a given question is true is representable by the set of potential answers that have not ruled out as impossible. Doubt concerning probability is representable by the set of probability distributions over potential answers that are not ruled out as impermissible to use in delibration. (2) To defend a more specific thesis that the set of permissible probability distributions ought to be convex.