Probability-like functionals and fuzzy logic (Q1378735): Difference between revisions

The author defines fuzzy logic on the set of functions \(I\to I\), i.e., fuzzy membership functions. He introduces a membership function \(\text{Val} (v,\pi)\) associated with a proof, meaning that, given information \(v\), the proof assures that the result holds at least to degree \(\text{Val} (v,\pi)\). The deduction operator of a theory \(S\) is \(D(v,\alpha)\) such that for every valuation \(v\), formula \(\alpha\), and proof \(\pi\) of \(\alpha\), \(D=\sup (\text{Val} (v,\pi))\). He also defines a fuzzy semantics as any fuzzy subset that does not contain the certain function. The theory so constructed, subject to a rule that implies inconsistency if the fuzzy membership degree is \(>1\), is shown to admit ultraproducts and to be logically compact. Applications are given to the description in this setting of the classes of super-additive measures, fuzzy necessity logic, upper and lower (Dempster) probabilities, and lower envelopes of probability measures.