The data of semantic description Summary

We have been concerned with developing a form of semantic description that is oriented to the generation of behavioral data, as well as of sentences. Thus, eliciting techniques and test sentences have been described which aim at isolating the semantic components that comprise the paradigmatic meanings of items in a language. A semantic component is seen as a schematic sentence or function, a form borrowed from symbolic logic. The items so defined by components are also in the form of schematic sentences or functions and include not only lexical items, but syntactic constructions and such elements of obligatory choice as verb tenses. Definitions are statements of equivalence between the defined and the defining sentences. The results of semantic eliciting techniques and tests can thus be integrated into a transformational semantic grammar which generates surface structures of a language from these underlying sentences or functions. A metatheory based on the intertranslatability of languages would include a stock of semantic components which could in turn be combined to define and generate many of the underlying functions in the description of a particular language. One such universal component discussed at greater length was be/have, expressing a state or state relation.

It was also argued that a semantic theory should contain ‘ rules for breaking rules ’ to account for code manipulation in such creative uses of language as irony and referent splitting and in the hearer process of interpreting, reinterpreting, and establishing or not establishing logical consistency in sentences.