Abstract
Understanding the semantic relations between terms in specialized texts is of critical importance in translation and terminology, and generally speaking in learning from texts. Our research highlights the advantages of formalizing them in order to build hierarchies and sets of horizontal conceptual relations (i.e. process-oriented relations) for knowledge acquisition. This paper discusses a method for extracting domain-specific semantic relations in specialized texts. Obviously, some texts are more appropriate than others in this regard. ‘Knowledge-rich’ texts such as encyclopaedia and textbooks are considered good materials because of the density and richness of thematic information. Considering them as such, we used the encyclopaedic articles of the Dictionnaire analytique de la distribution/Analytical Dictionary of Retailing. We retrieved over 3000 terms semantically related to all 350 headwords of the Dictionary, and grouped them into 28 classes of relations (paradigmatic, i.e. generic, specific, agent, goal, instrument, recipient, location, etc., and also syntagmatic, such as related verbs and adjectives). This paper discusses in particular the generic, agent and property relations and examines the linguistic markers that permit their retrieval.
Publisher
John Benjamins Publishing Company
Subject
Library and Information Sciences,Communication,Language and Linguistics
Cited by
3 articles.
订阅此论文施引文献
订阅此论文施引文献,注册后可以免费订阅5篇论文的施引文献,订阅后可以查看论文全部施引文献