Extracting a Semantic Lexicon of French Adjectives from a Large Lexicographic Dictionary

First Joint Conference on Lexical and Computational Semantics (*SEM), pages 161–169, Montr´eal, Canada, June 7-8, 2012. c⃝2012 Association for Computational Linguistics Extracting a Semantic Lexicon of French Adjectives from a Large Lexicographic Dictionary Selja Sepp¨al¨a and Alexis Nasr Laboratoire d’Informatique Fondamentale Aix Marseille Universit´e 163, avenue de Luminy F-13288 Marseille Cedex 9 alexis.nasr@lif.univ-mrs.fr selja.seppala@lif.univ-mrs.fr Lucie Barque Lexiques Dictionnaires Informatique Universit´e Paris 13 99, avenue Jean-Baptiste Cl´ement F-93430 Villetaneuse lucie.barque@univ-paris13.fr Abstract We present a rule-based method to automati- cally create a large-coverage semantic lexicon of French adjectives by extracting paradig- matic relations from lexicographic definitions. Formalized adjectival resources are, indeed, scarce for French and they mostly focus on morphological and syntactic information. Our objective is, therefore, to contribute enriching the available set of resources by taking advan- tage of reliable lexicographic data and formal- izing it with the well-established lexical func- tions formalism. The resulting semantic lexi- con of French adjectives can be used in NLP tasks such as word sense disambiguation or machine translation. After presenting related work, we describe the extraction method and the formalization procedure of the data. Our method is then quantitatively and qualitatively evaluated. We discuss the results of the evalu- ation and conclude on some perspectives. 1 Introduction Formalized semantic resources are highly valuable in areas such as NLP, linguistic analysis or lan- guage acquisition. However, creating such resources from scratch is time-consuming and generally yields limited-size lexicons. Existing lexicographic dictio- naries do have a large coverage and present a reli- able content. They lack nevertheless the sufficient formalization. In this paper, we present a rule-based method to automatically create a large-coverage se- mantic lexicon of French adjectives by extracting paradigmatic relations from lexicographic defini- tions using lexico-syntactic patterns. Formalized ad- jectival resources are, indeed, scarce for French and they mostly focus on morphological and syntactic information. Our goal is, therefore, to contribute en- riching the available set of resources by taking ad- vantage of reliable lexicographic data and formal- izing it with the well-established lexical functions formalism of the Meaning-Text theory (Mel’ˇcuk, 1996). The resulting semantic lexicon of French adjectives can be used in NLP tasks such as word sense disambiguation or machine translation1. In section 2, we present related work. In section 3, we expose the method used to build the lexicon, i.e. the extraction method and the formalization procedure of the data, and outline the main results. Finally, in section 4, we present a quantitative evaluation of our method and a qualitative evaluation of our data, and discuss their results. We conclude on some perspec- tives for future work. 2 Related Work It is well established that there are different types of adjectives distinguished by properties, such as gradation and markedness, and by their seman- tic and syntactic behaviors (antonymy, selectional preferences) (Fellbaum et al., 1993; Raskin and Nirenburg, 1996). WordNet, for example, distin- guishes different types of adjectives according to their semantic and syntactic behaviors: descriptive, reference-modifying, color and relational adjectives (Fellbaum et al., 1993). However, it mainly accounts for the …

Content truncated to 500 words