In my book (Cambridge University Press, 2017) I try to develop a generalized approach to probabilistic learning in the tradition of Richard Jeffrey’s ‘radical probabilism’. I show that the core principles of rational Bayesian learning, which are based on dynamic consistency and symmetry, apply to probabilistic learning in general, including so-called boundedly rational models of learning. In addition, I explore the sense in which dynamic consistency is a principle of rational belief change, and I argue that it is in fact the basic rationality principle of learning within a radical probabilist framework. The book also considers a number of questions that a radical probabilist should be able to answer, including the status of prior probabilities and the roles of disagreement and consensus in a broadly Bayesian epistemology.