Logic of change, change of logic

Understanding human behaviour, and indeed human beings more generally, requires an understanding of their attitudes: their beliefs, desires and preferences of course, but also a plethora of other attitudes or attitudinal factors, such as their presuppositions, their assumptions, their intentions, their attention, their awareness and their emotional states. But people’s attitudes change over time: to take the least controversial example, their beliefs change on learning new facts. It is not sufficient to understand the role human attitudes play at any particular moment; an understanding of how different attitudes change is also required. Formal models—be they from logic or mathematics—have historically played an important role in the modelling and study of some of these attitudes. Notable examples include the Bayesian model of (partial) beliefs as probability functions (de Finetti 1937; Ramsey 1931), the modal logician’s representation of belief and knowledge in terms of accessibility relations or partitions on sets of possible worlds (Hintikka 1962; Aumann 1976), typical models of preferences as order relations over sets of