This article explores some fundamental issues of definition-based lexical semantic research through a critique of the Natural Semantic Metalanguage (NSM) theory of semantic and grammatical description (Wierzbicka 1996, Semantics. Primes and Universals, Oxford University Press, Oxford). NSM is criticized for attaching excessive importance to explanatory definition, for its adoption of the reductive requirement that a definiens be simpler than a definiendum, and for its use of ‘canonical contexts’ to disambiguate meaning. The principle of substitutability, according to which a definition of (...) a term is accepted if it can be substituted for the term itself, is also critically examined, and the theory’s use of syntactic phenomena as evidence for polysemy is shown to be inconsistent. Finally, suggestions that NSM may be a valid analytical method for only a subpart of the lexicon are rejected. (shrink)

Explicatures are not cancellable. Theoretical considerations.

In this article I challenge the claim that nuclearity is a central principle in the organization of texts. I propose the Framework for the Relational Analysis of Texts which accounts for the paratactic and hypotactic realization of coherence relations. Within this framework, the taxis of coherence relations is co-textually dictated. I consider the writer choices in the distribution of discoursal salience and the intertextual and intercultural variation of these choices. It is suggested that divergence between approaches that perceive text as (...) exhibiting both hypotactic and paratactic organization and those that see nuclearity as a basic characteristic of text structure arises from differences in the linguistic corpora examined during the construction of respective theoretical frameworks. (shrink)