Jake Chandler (Leuven) gives a talk at the MCMP Colloquium (6 February, 2013) titled "Reasons to Believe and Reasons to not." Abstract: The provision of a precise, formal treatment of the relation of evidential relevance–i.e. of providing a reason to hold or to withhold a belief–has arguably constituted the principal selling point of Bayesian modeling in contemporary epistemology and philosophy of science. By the same token, the lack of an analogous proposal in so-called AGM belief revision theory, a powerful and elegant qualitative alternative to the Bayesian framework, is likely to have significantly contributed to its relatively marginal status in the philosophical mainstream. In the present talk, I sketch out a corrective to this deficiency, offering a suggestion, within the context of belief revision theory, concerning the relation between beliefs about evidential relevance and commitments to certain policies of belief change. Aside from shedding light on the status of various important evidential ‘transmission’ principles, this proposal also constitutes a promising basis for the elaboration of a logic of so-called epistemic defeaters.