Towards Better Understanding of Anaphora

TOWARDS BETTER UNDERSTANDING O15' ANAPHORA Barbara Dunin-K@pH cz Institute of Informatics, Warsaw University P.O. Box 1210 00-901 Warszawa, Poland ABSTRACT This paper presents a syntactical method of interpreting pronouns in Polish, Using the surface structure of the sentence as well as grammatical and inflexional inlormation accessible during syntactic analysis, an area of reference is marked out for each personal and possessive pronoun. This area consists of a few internal areas inside the current sentence and an external area, i.e. the part of the text preceding it. In order to determine that area of reference several syntactic sentence-level restrictions on anaphora interpretation are formulated. Next, when looking at the area of pronoun's reference, all NPs which number- -gender agree with the pronoun can be selected and this way the set of surface referents ol each pronoun can be created. It can be used as data for further semantic analysis. I INTRODUCTION Reference is one of the central concepts of any linguistic theory. In recent research into anaphora the term "reference" has been used in three different senses ( Szwedek, 1981): (a) as a relation between the name and the thing named (Hall Partee, 1978) (b) as an association between noun phrases and mental entities in the language user's (Nash-%~ebber, 1978) (c) as an association between the occurrence of phrases in the text (Reinhart, 1981) However the reference is understood, irl order to interpret correctly anaphora on the semantic level ((a) and (b)), first a stage (C) is necessary. in this paper I have taken the point of vie~ presented under (c). i shall discuss the problem o~ onaphora in Polish ser Atences. rvly altentioF, is focused on personal ond possessive pronouns expticitely occurring in the text and moreover on zero pronouns, i.e. ellipsis of NP in the subject position, specific for Slavonic languages. My purpose in the description of regularities of the reference in the Polish language. I shall express them by defining the area of pronoun's references, i.e. those regions of the text where its antecedents should De found, q hese surface referents will be selected from among NPs occurring in the sentence. The research on anaphora made for English has led to the formulation of some structural rules using such relations as command, c-command and precede-and-command (Reinhart, 3.981). I have been searching for analogous rules for Polish. But two essential differences have to be considered: (i) grammatical and morphological properties of Polish and English; (ii) different grammatical traditions. For English the rules concernig the coreference of entities were forrrulated on the basis of generative-transformational grammar. For Polish the first precise description of Polish syntax was formulated only recently by Szpakowicz, who based his work on the framework created by Saloni (Saloni, 1976; Saloni and Swidzinski, 1981). It is a kind of in,mediate-constituent grammar; the grammatical categories (case, ~ender, etc) are applied not only to single words, but also to compound phrases. In my present ~vork I have limited my attention to the subset of Polish described by Szpakowicz (Szpako~Jvicz, …

Content truncated to 500 words