Adding Information to Multiword Terms in Wiktionary

Abstract

We describe ongoing work dealing with the potential “auto-enrichment” of “Multiword terms” (MWTs) that are included in the English edition of Wiktionary. The idea is to use and combine information contained in the lexical components of the MWTs and to propagate this extracted and filtered information into the lexical description of the MWTs, as those are typically equipped with less lexical information as it is the case for their lexical components. We started our work with the generation of pronunciation information for such MWTs, on the base of the pronunciation information available for their components. We present in this paper first achievements but also issues we encountered. Addressing those issues lead us to consider additional resources for supporting our approach, like DBnary and WikiPron. This step was ultimately leading to suggestions of adaptations for those additional resources, which, in the case of DBnary, are already implemented. We are currently extending our approach to a morphosyntactic and semantic enrichment of the English MWTs in Wiktionary.