Type-based HPSG analysis of Arabic verbal declension

Abstract

Arabic language is very rich in grammatical declen sion of nouns and verbs. In classical Arabic, there are Ive classes of verb lexemes and these are declined in four ways. Modeling the morphological effect of such a rich declension system is a challenging problem and is essential for processing of Arabic language. In this paper, we use Head-Driven Phase Structure Grammar GIPSG) to analyze the declension system of Arabic verbs based on lexical types. We design lexical type hierarchy of Arabic verbs and show the mapping of lexical type to declension type from this type hierarchy. We also design construction rules to capture the morphological and syntactic effect of declension types dynamically. Finally, we have implemented the lexical type hierarchy and construction rules in the TRALE platform to validate the proposed formalism. This work effectively extends the capabilities of existing HPSG framework for supporting declension of verbs in Arabic.