On Spohn's rule for revision of beliefs (Q2640215)

This paper is about Spohn's theory of epistemic beliefs and provides a set of axioms mathematically equivalent to the ones proposed by Spohn himself. The motivation behind Spohn's theory is that probability theory should be (according to Spohn) inadequate for revising beliefs when new information is obtained. The author quotes only two examples of inadequateness: (1) if \(P(A)=1\), then \(P(A| B)=1\) for all B such that P(A\(| B)\) is well-defined (but \(P(A| B)\) is well-defined for any B such that \(B\neq \emptyset:\) when the cardinality of the family of events is greater than that of the set of natural numbers, there exist necessarily events B with \(P(B)=0\) and it is not difficult to construct standard counterexamples with \(P(A)=1\) while \(P(A| B)\neq 1)\); (2) if \(P(A)\geq 1-\epsilon\) and \(P(B)\geq 1-\epsilon,\) then it is not necessary that \(P(A\;and\;B)\geq 1-\epsilon\) (but it is easily seen that \(P(A\;and\;B)=P(A)+P(B)-P(A\;or\;B)\geq 1-\epsilon +1-\epsilon -1=1-2\epsilon)\).