The Granularity of Meaning in Proof-Theoretic Semantics

摘要

The paper compares two conceptions of meaning both for logic and for natural language: 1. Model-theoretic semantics (MTS), basing meaning on reference and truth conditions (in arbitrary models). 2. Proof-theoretic semantics (PTS), basing meaning on canonical derivations in meaning-conferring natural-deduction proof-systems. It is shown that PTS induces a much finer granularity on meanings, in particular distinguishing the meanings of logically equivalent sentences. A certain coarsening by means of equating meanings based on identical grounds for assertion is proposed, useful in certain contexts.