Possibility as Similarity: the Semantics of Fuzzy Logic
This work addresses foundational issues in fuzzy logic for researchers in logic and AI, but it appears incremental as it extends existing modal logic concepts.
The paper tackles the conceptual differences between probabilistic and possibilistic approaches in fuzzy logic by proposing a semantic model based on similarity relations between possible worlds, enabling the definition and interpretation of key fuzzy logic constructs like possibility distributions and generalized modus ponens.
This paper addresses fundamental issues on the nature of the concepts and structures of fuzzy logic, focusing, in particular, on the conceptual and functional differences that exist between probabilistic and possibilistic approaches. A semantic model provides the basic framework to define possibilistic structures and concepts by means of a function that quantifies proximity, closeness, or resemblance between pairs of possible worlds. The resulting model is a natural extension, based on multiple conceivability relations, of the modal logic concepts of necessity and possibility. By contrast, chance-oriented probabilistic concepts and structures rely on measures of set extension that quantify the proportion of possible worlds where a proposition is true. Resemblance between possible worlds is quantified by a generalized similarity relation: a function that assigns a number between O and 1 to every pair of possible worlds. Using this similarity relation, which is a form of numerical complement of a classic metric or distance, it is possible to define and interpret the major constructs and methods of fuzzy logic: conditional and unconditioned possibility and necessity distributions and the generalized modus ponens of Zadeh.