Construction of an HPSG Grammar for the Arabic Relative Sentences

Proceedings of Recent Advances in Natural Language Processing, pages 574–579, Hissar, Bulgaria, 12-14 September 2011. Construction of an HPSG Grammar for the Arabic Relative Sentences Ines Zalila Faculty of Economics and Manage- ment of Sfax ines.zalila@yahoo.fr Kais Haddar Sciences Faculty of Sfax kais.haddar@fss.rnu.tn Abstract The paper proposes a treatment of relative sen- tences within the framework of Head-driven Phrase Structure Grammar (HPSG). Relative sentences are considered as a rather delicate linguistic phenomenon and not explored enough by Arabic researchers. In an attempt to deal with this phenomenon, we propose in this paper a study about different forms of relative clauses and the interaction of relatives with other linguistic phenomena such as ellipsis and coordination. In addition, in this paper we shed light on the recursion in Arabic relative sentences which makes this phenomenon more delicate in its treatment. This study will be used for the construction of an HPSG gram- mar that can process relative sentences. The HPSG formalism is based on two fundamental components: features and AVM (Attribute- Value-Matrix). In fact, an adaptation of HPSG for the Arabic language is made here in order to integrate features and rules of the Arabic language. The established HPSG grammar is specified in TDL (Type Description Lan- guage). This specification is used by the LKB platform (Linguistic Knowledge Building) in order to generate the parser. 1 Introduction Relative phenomenon has a great importance in all natural languages and in all corpus kinds. That‟s way researchers in linguistics or in com- puter sciences pay great attention to this pheno- menon (i.e., (Belkacemi, 1998), (Elleuch., 2004) and (Garcia,2006)). Indeed, a phase of parsing of this phenomenon is fundamental for several types of Natural Language Processing (NLP) applications such as grammatical correction and machine translation. Nevertheless, the researches concerning the parsing of relatives, object of this work, have not reached an advanced stage yet. This is due, on the one hand, to the complexity of this phenomenon and, on the other hand, to the interaction with simple and complex linguistics phenomena. Thus, one of the objectives of this work is to study the various forms of the Arabic relative sentences. This study is based on old grammati- cal theories (Abdelwahed, 2004), (Belkacemi, 1998) and (Dahdah, 1992), and on discussions with linguists. From the study carried out, we also want to identify all possible syntactic repre- sentations of the Arabic relative sentences. The choice of the HPSG is justified by the fact that this formalism has shown great efficiency in sev- eral languages such as German. The elaborated HPSG grammar is specified in TDL (Type Description Language). Based on the elaborated TDL specification, an Arabic parser is generated using the LKB linguistic platform. The generated parser can process complex sentences containing relatives. The originality of this work consists, on the one hand, in the identification of a relative sentences typology, and on the other hand, in the proposition of a HPSG extension detailing under-categorization. This extension is specified in TDL (Type Description Language) (Krieger and Schäfer, …

Content truncated to 500 words