Phraseology in specialized resources

Abstract

In English, the international language of communication (Tono in Lexicography 1(1):1–5, 2014), complex nominals (CNs) are frequently used to convey specialized concepts (Sager et al. in English special languages. Principles and practice in science and technology. Brandstetter Verlag, Wiesbaden, 1980; Nakov in Natural Language Engineering 19(03):291–330, 2013). These phraseological units have a nominal head that is modifed by another element (e.g., hydropower production). Problems can arise in relation to their identifcation, their bracketing or internal structure disambiguation, their meaning access, and their translation or production in another language. Although they are not marginal phenomena in specialized language, they are rarely included in specialized resources. Even when they are included, their treatment is not systematic (Cabezas-García and Faber in Computational and corpus-based phraseology. Springer, Cham, pp 145–159, 2017a). This article describes the representation of CNs in EcoLexicon (http://www.ecolexicon.ugr.es), a terminological knowledge base, whose new phraseological module will include verb collocations (e.g., a volcano spews lava) as well as CNs. For that purpose, we used a wind power corpus in English and Spanish for term extraction, semantic analysis, establishment of interlinguistic correspondences, and defnition crafting. We propose diferent access points to information (Kwary in International Journal of Lexicography 25(1):30–49, 2012), such as the CNs formed from a given term, a bilingual view in English and Spanish, or the syntactic–semantic combinations in CNs. The structure of the CN module is based on the semantics of these phraseological units, which facilitates the specifcation of mapping rules as well as knowledge acquisition (Faber in A cognitive linguistics view of terminology and specialized language. De Gruyter Mouton, Berlin, 2012).

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 50 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.