Proof-theoretic semantics for subsentential phrases (Q969428)

In their previous works the authors developed a proof-theoretic semantics for a fragment of English constituted by nouns, determiners, verbs, and the copula ``is a'', which is reviewed briefly in this paper as well. The main purpose of this paper is to present a method for deriving proof-theoretic meanings for sub-sentential phrases down to lexical units (words) by appealing to (a version of) Frege's context-principle (specified to fit type-logical grammar) which determines the decomposition of sentential meaning according to the function-argument structure. In contrast to Montague's model-theoretic approach, a type is interpreted proof-theoretically as a collection of derivations in the associated ``dedicated'' natural-deduction proof-system or of functions therein, which therefore do not refer to models, truth-values, or elements of a domain. The semantics thus introduced is analyzed with respect to its compositionality.