The Geometrization of Meaning

Abstract: One of the greatest problems in philosophy is that of meaning. The turning point in thinking on meaning was Tarski's definition of truth, and the rapid development of logical semantics and model theory was a consequence of this achievement. Perhaps less well-known among classical logicians and philosophers is that it is category theory that provides adequate mathematical tools to study the relationship between the syntax of formalized theories and their semantics. The aim of this article is to change this situation and make a preliminary philosophical analysis of the results obtained so far. They concern formalized algebraic theories with axioms in the form of equational laws, theories based on propositional logic and coherent Boolean logic, as well as decidable logic which is not necessarily Boolean. The syntactic-semantics relation for these theories takes the form of dualisms between the respective syntactic and semantic categories. These dualisms are given by the appropriate adjoint functors. In all the considered cases, the syntax--semantics dualism corresponds to the algebra-geometry dualism. We analyze the philosophical significance of these results. They allow us to look at the problem of meaning in a new light and formulate a criterion distinguishing formal theories from empirical theories. The disputes that were once fought over Tarski's definition of truth are transferred to a new context. The polemics as to whether Tarski actually succeeded in reducing the problem of meaning to purely syntactic terms has been superseded: between the syntactic and semantic categories there is not a relation of reducibility but rather that of interaction, and this relation is given by adjoint functors. We also touch upon other philosophical aspects of the categorical approach to the problem of meaning.