The alleged limitations of fuzzy control

Charles Elkan (1994) claimed that, despite the popularity of using fuzzy logic, it could only be used in controllers and would never be successfully used in expert systems. This claim forms the basis of Elkan's 'paradox' and relies on a distinction between fuzzy controllers and expert systems. Yet it is unclear whether this distinction even exists and if Elkan is justified in using it to support his claim. An examination of this distinction is undertaken where it is shown to be faulty and Elkan's 'paradox' is rejected. The correctness of using fuzzy logic to model different kinds of vagueness and uncertainty is then examined in an effort to determine the kinds of phenomena fuzzy logic can be successfully used to deal with.