The Phasic Research Of The Arabic Pronouns Under The Centripetal Theory Model

Centering theory(Grosz,Joshi and weinstein,1986,1995;Walker etal,1998)is a relatively new discourse theory.As a niche theory of discourse coherence and conspicuous study,The proposition to the Centering theory has important theoretical significance to promote the development of linguistics and is widely used in the field of anaphora resolution.Pronoun anaphora is the most common phenomenon in natural language,Its also very common in Arabic discourse."Repetition" and the ambiguity of the meaning of the anaphora are important features of the anaphora.The repetition of the conceptual entity in the discourse is an important embodiment of discourse coherence and an important means of discourse coherence research.And the Centering theory is considered to be a typical theory of the interpretation of the coherence of the entity(Poesio,et.al.2004).Therefore,it is very important to study the process of pronoun anaphora resolution and to understand discourse coherence and discourse comprehension.In order to verify the relevant rules of the theory and the cross-linguistic applicability of the constraints.This paper attempts explore the application of the centripetal theory in the process of resolving the anaphora in the Arabic language from the perspective of computational discourse and define the core concept of the Centering theory,it also required parametric analysis to the concept in the specific digestion process,combined with the characteristics of the Arabic language to determine the centripetal theory of Arabic parameters.The paper proposes an algorithm based on a small annotated corpus,which is combined with the centripetal theory of Arabic pronoun anaphora resolution,and deduces the anaphora resolution algorithm for Arabic pronouns.The first chapter introduces the influence of rules,central constraints and central transition types in the basic mode of the Centering theory on the use of pronouns,and analyzes the distribution of centrality and centripetal model in Arabic discourse.The second chapter mainly make a statistical analysis to the use of the pronoun anaphora and the difference between the different forms of pronoun anaphora in Arabic discourse,As well as the discourse function and constraints of pronoun anaphora.In the third chapter,the solution to the digestion of pronoun anaphora in natural language processing is proposed,and the theoretical basis-syntactic,semantic and center are explored.And use this theory to deduce the algorithm of different forms of pronoun anaphora in the Arabic discourse.Centering theory was originally proposed as a textual linguistic theory,and has a cross-linguistic,so some of the core concepts of this theory are not clearly defined,which makes the different of the conclusions for the later researchers because of their own understanding to the basic concepts,and can not be horizontal comparison.However,the scope of the anaphora discussed in this paper is limited to the pronoun anaphora.The method of this paper is based on parameterized research method of Poesioetal(2004)'s CT theory.The undefined core concept mentioned above is used as a parameter in the theory,and the influence of the setting of pronoun parameters on the anaphora resolution of Arabic pronouns is explored by this method.