Resumen de Les problèmes linguistiques de la traduction automatique des prédicats nominaux entre l'arabe et le français

Béchir Ouerhani

• This paper explores the idea that translating nominal predicate constructions means correlating equivalents of predicative structures consecutively in L1 and L2. This demands detailed and exhaustive linguistic descriptions of the elements in question. We will try to show that formulating dictionaries of equivalents - starting from a model suggested by Gaston Gross - must necessarily respect the specificities of both languages. Concerning Arabic, for instance, it's necessary to state information about the compatibility and aptness between a given nominal predicate and the support verbs it selects, as well as the impact of morphological properties on reformulation by paraphrases. This last point refers to the Arabic features, namely the derivation system that establishes a link between varied forms springing from one given root.

  
  
  Plan de l'article
  
  Introduction
  1. Le passage de L1 à L2 : les cas de figure possibles
  2. Description uniformisée en vue de la traduction automatique et spécificités linguistiques
  2.1. La description linguistique uniformisée
  2.2. Spécificités de l'arabe
  a) La forme nominalisée des Vsup.
  b) Les dérivés de la même racine et les relations paraphrastiques possibles
  c) La construction en Npred=N0
  3. Échantillons de description de classes de Npred. en arabe