No label defined (Q4460673)

In this paper the authors offer, extending earlier joint work [``Reasoning about probabilities using fuzzy logic'', Neural Network World 10, 811--824 (2000)], a logical system, i.e. a logical calculus together with an adequate semantics, to reason explicitly about Dempster-Shafer belief functions, interpreting belief degrees as truth degrees of suitable belief sentences. The core idea behind the semantics is to treat belief degrees as probabilities of necessities. Thus the main semantical structures are suitable probabilistic Kripke models. The corresponding logical calculus derives from a combination of the (infinite-valued) Łukasiewicz and product logics, if necessary combined with truth degree constants for all rationals from the real unit interval, with a modal logic, particularly with S5 or KB4. Using these tools, the authors are able to mirror inside this logic naïve arguments which refer to Dempster's rule in belief theory.