Language technology tools and the Phrasal Lexicon
María Fernández-Parra · ExLing Conferences · 2012
This paper explores the possibility of resorting to language technology to automatically extract lists of phrasal items instead of terms. Such items come in multiple and varied forms and constitute a very large part of what we say and write. Two types of tools—concordancing software (the Lexicon feature in Déjà Vu) and automatic terminology extraction software (MultiTerm Extract and PhraseFinder from SDL Trados, PlusExtract from Wordfast, Araya, and ExtPhr32)—were tested on a purpose-built corpus, and their performance was compared to the manual extraction of items. The results of the experiments suggest that many of these tools can yield high recall. Furthermore, because phrases are highly variable compared to terms, any software designed specifically for phrasal extraction should take such variability into account.