The goal of this project is to provide a philosophically well-founded theory of rational belief based on "imprecise probabilities" (IP).
That is, instead of the orthodox Bayesian approach to modeling belief using probability functions, the IP approach uses a set of probability functions as the representation of belief.
This promises to be a more foundationally secure, more normatively compelling, more descriptively accurate and more representationally powerful theory of rational belief.
The project consists of three parts. The first analyses the basic concept of IP, and the methodology of formal epistemology more generally. The second explores learning and inference in an IP setting, and the third focuses on decision making with IP.