In this paper, we present a modeling of a syntactic lexicon for Arabic verbs. The structure of the lexicon is based on the recently introduced ISO standard called the Lexical Markup Framework. This standard enables us to describe the lexical information in a versatile way using general guidelines and make possible to share the resources developed in compliance with it. We discuss the syntactic information associated to verbs and the model we propose to structure and represent the entries within the lexicon. To study the usability of the lexicon in a real application, we designed a rule-based system that translates a LMF syntactic resource into Type Description Language compliant resource. The rules are mapping information from LMF entries and types to TDL types. The generated lexicon is used as input for a previously written HPSG grammar for Arabic built within the Language Knowledge Builder platform. Finally, we discuss improvements in parsing results and possible perspectives of this work.

L’interopérabilité entre les lexiques des grammaires d’unification passe par l’adoption d’une représentation normalisée de ces ressources. Dans ce papier, nous proposons l’utilisation de LMF pour établir la standardisation des ressources lexicales en HPSG. Nous présentons LMF d’une manière sommaire et nous détaillons son utilisation pour coder les entrées lexicales d’un lexique HPSG.