KB-N (KunnskapsBank for Norsk økonomisk-administrativt domene): presentasjon av et korpusbasert terminologiprosjekt

Abstract

This paper reports early results of a 3-year project at the Department of Professional and Intercultural Communication, NHH, aiming to establish a knowledge-bank for economic-administrative domains. The underlying assumption is that domain-focal special knowledge is embedded in text produced typically by domain experts for documentary, argumentative, didactic or general communicative purposes. It is further assumed that the essential knowledge content is embedded in relatively language independent concepts and manifested through relatively language specific terminology (in casu English and Norwegian), and that such terminology is stratified with respect to domain specificity ranging from general shared terms down to a small set of domain-focal terms. KB-N represents the culmination of efforts to refine and integrate computational strategies and tools in NLP for corpus design and analysis, automatic and semi-automatic extraction, representation, and retrieval of terminology, dynamic thesaurus building, dynamic display of authentic collocational and phraseological evidence, etc. The paper attempts to demonstrate (via screen-shots) a working version of the integrated KB-N software suite for handling corpus-based concordancing, term extraction and selection, thesaurus building on-thefly etc., in the context of a discussion of general theoretical and methodological issues. A range of applications is envisaged for the knowledge-bank. Designed as a web-enabled resource it will be available for e.g. systematic terminology registration and look-up, textbook authoring, e-learning as well as machine translation.