A Joint Characterization of Belief Revision Rules

This paper characterizes different belief revision rules in a unified framework: Bayesian revision upon learning some event, Jeffrey revision upon learning new probabilities of some events, Adams revision upon learning some new conditional probabilities, and `dual-Jeffrey' revision upon learning an entire new conditional probability function. Though seemingly different, these revision rules follow from the same two principles: responsiveness, which requires that revised beliefs be consistent with the learning experience, and conservativeness, which requires that those beliefs of the agent on which the learning experience is `silent' (in a technical sense) do not change. So, the four revision rules apply the same revision policy, yet to different kinds of learning experience.