Abstract

This article provides an intuitive semantic account of a new logic for comparisons (CL), in which atomic statements are assigned both a classical truth-value and a “how much” value or extension in the range [0, 1]. The truth-value of each comparison is determined by the extensions of its component sentences; the truth-value of each atomic depends on whether its extension matches a separate standard for its predicate; everything else is computed classically. CL is less radical than Casari’s comparative logics, in that it does not allow for the formation of comparative statements out of truth-functional molecules. It is argued that CL provides a better analysis of predicate vagueness than classical logic, fuzzy logic or supervaluation theory.