The study of the exact relationships between belief revision and belief update, on one side, and belief revision and nonmonotonic reasoning, on the other, has raised considerable interest, but still, the picture is far from being complete. In this paper, we add some new details to this line of research by making a foundational contribution to the discussions on the very nature of belief change operations, and by introducing universal inference operations as a proper counterpart in nonmonotonic reasoning to iterated belief change. Belief change is investigated within an abstract framework of epistemic states and (qualitative or quantitative) conditionals here. We show how belief revision and belief update can be realized by one and the same generic change operator as simultaneous and successive change operations. We propose general postulates for revision and update that also apply to iterated change. The distinction between background knowledge and evidential information turns out to be a crucial feature in our framework, in order to analyse belief change in more depth.