Has anyone done the analog of Cox's Theorem for belief function theories or possibility theory? That is, has anyone postulated a different set of qualitative requirements and then shown that these requirements lead inevitably to a system isomorphic to one of these non-Bayesian theories?
