Zero Pronouns and Their Resolution in Sanskrit Texts

Abstract

This paper attempts to formulate an algorithm for resolving Sanskrit null/zero pronouns, specifically null subjects. For this purpose we have conducted an extensive study of two Sanskrit texts, namely, Panchatantra and Hitopadesha – the neeti texts of Sanskrit literature – focusing on null arguments of finite verbs. Use of zero or null pronouns in the language is a massive phenomenon. Null pronouns are empty slots of obligatory arguments of a verb in a finite clause. Sanskrit verbs may take one, two, three or four arguments, depending on their subcategorization. We have observed in the corpus that subject, direct object, indirect object and possessive are the syntactic categories which could be dropped. We have formulated an algorithm using heuristic rules, depending mostly on agreement features of the verb and its subject. We have used the POS annotated data of the above mentioned texts for this study. The whole enterprise involves linguistic aṇalysis of zero pronouns and then determining antecedents of zero subject pronouns after detecting them automatically in the input text.