Wiktionnaire’s Wikicode GLAWIfied: a Workable French Machine-Readable Dictionary

Wiktionnaire’s Wikicode GLAWIfied: a Workable French Machine-Readable Dictionary Nabil Hathout, Franck Sajous CLLE-ERSS, CNRS & Universit´e de Toulouse 2 {nabil.hathout, franck.sajous}@univ-tlse2.fr Abstract GLAWI is a free, large-scale and versatile Machine-Readable Dictionary (MRD) that has been extracted from the French language edition of Wiktionary, called Wiktionnaire. In (Sajous and Hathout, 2015), we introduced GLAWI, gave the rationale behind the creation of this lexicographic resource and described the extraction process, focusing on the conversion and standardization of the heterogeneous data provided by this collaborative dictionary. In the current article, we describe the content of GLAWI and illustrate how it is structured. We also suggest various applications, ranging from linguistic studies, NLP applications to psycholinguistic experimentation. They all can take advantage of the diversity of the lexical knowledge available in GLAWI. Besides this diversity and extensive lexical coverage, GLAWI is also remarkable because it is the only free lexical resource of contemporary French that contains definitions. This unique material opens way to the renewal of MRD-based methods, notably the automated extraction and acquisition of semantic relations. Keywords: French, Machine-Readable Dictionary, Free Lexical Resource, Wiktionary, Wiktionnaire 1. Introduction GLAWI1 is a large Machine-Readable Dictionary (MRD) extracted from Wiktionnaire, the French language edition of Wiktionary, and converted into a workable XML format. In a previous work, Sajous et al. (2010) introduced Wik- tionaryX, an electronic lexicon including lemmas, semantic relations and translations. Hathout et al. (2014b) described how GL `AFF, a large inflectional and phonological lexicon, has been extracted from the same source. The assessment of GL `AFF’s lexical coverage and the quality of its phone- mic transcriptions has shown that Wiktionnaire is a valu- able starting point to build lexical resources of good qual- ity. Sajous and Hathout (2015) introduced GLAWI, a dic- tionary built from an updated version of Wiktionnaire that merges the information stored in WiktionaryX and GL `AFF into a single resource. New information, such as etymology and morphological relations, has also been added. Sajous and Hathout (2015) focused on the parsing process and the standardization of Wiktionnaire’s heterogeneous data, a prerequisite to produce a workable MRD. In the current ar- ticle, we illustrate the richness of GLAWI’s lexical knowl- edge, leaving apart the extraction process. We also contem- plate different uses that can be made of this resource, either in academic research, or in concrete NLP applications. 2. Resource description The general structure of GLAWI’s entries is illustrated in Figure 1. GLAWI’s macro- and micro-structure are very close to the ones of Wiktionnaire: the basic unit is a word written form (hereafter, grapheme), associated with a given page/URL. When several parts of speech (POS) or homo- graphs correspond to the same grapheme, the article con- tains one separate POS section for each one of them. Each POS section includes definitions (glosses and examples), and several optional subsections described hereafter. Ta- ble 1 gives the number of lemmas and inflected forms by POS. 1GLAWI is freely available at http://redac. univ-tlse2.fr/lexicons/glawi_en.html POS Lemmas Inflected forms noun 179,340 272,170 proper noun 57,371 8,019 …

Content truncated to 500 words