The Need for a Novel Approach to Design Derivation Lexicon for Semitic Languages

Résumé En

Morphology knowledge is relevant in language learning, information retrieval and natural language processing. Derivation lexicons are organized and comprehensive collections of the morphological variants of a language's vocabulary. These lexicons can be developed either through analysis-based synthesis of large text corpora or through synthesis of surface forms from roots, stems, lemmas and morphological rules. Much of the research attempted in developing derivation lexicon for Indo-European languages, which are concatenative, focus on analysis-based synthesis, as they do have well-developed preprocessing tools and organized text corpora. However, the methods for these languages are not appropriate for non-concatenative languages such as Semitic languages. Moreover, most of the Semitic languages, except Arabic and Hebrew, do not have well-developed text corpora and language processing tools. Hence, a novel approach that can cater for the root-pattern and rich morphology of these languages is necessary. This paper is therefore a comprehensive survey of the literature, an analysis motivating an innovative and generic morphological synthesis approach with illustrated architecture. It is part of a larger project tailored for designing an innovative, generic, approach to derivation lexicon development for Semitic languages.