LINGUIST List 21.3947

Wed Oct 06 2010

Diss: Syntax/Comp Ling: Chakrabarti: 'A Study of the Complex ...'

Editor for this issue: Mfon Udoinyang <>

        1.    Debasri Chakrabarti, A Study of the Complex Predicates of Hindi

Message 1: A Study of the Complex Predicates of Hindi
Date: 06-Oct-2010
From: Debasri Chakrabarti <>
Subject: A Study of the Complex Predicates of Hindi
E-mail this message to a friend

Institution: Indian Institute of Technology Bombay Program: Ph.D Dissertation Status: Completed Degree Date: 2008

Author: Debasri Chakrabarti

Dissertation Title: A Study of the Complex Predicates of Hindi

Linguistic Field(s): Computational Linguistics                             Syntax
Subject Language(s): Hindi (hin) Language Family(ies): Indo-Aryan
Dissertation Director:
Vaijayanthi M. Sarma Milind S Malshe Pushpak Bhattacharyya
Dissertation Abstract:

The thesis aims to provide a principled way to identify Complex Predicates(CP) of Hindi and also discusses the issue of their storage in a lexicaldatabase. Two types of Complex Predicates namely, Noun+Verb (conjunct verb)and V+V sequences are examined. There are various multi-word verbconstructions in Hindi and not all of them are instances of CPs, hence, itis necessary to have a systematic way to identify true CPs. To achieve thisgoal different syntactic tests are applied to isolate a true CP from amongsimilar-looking multi-word constructions. After identifying CPs the nexttask was to determine how they are formed. To accomplish this we analyzedthe semantic and syntactic properties of the second member of the CP (i.e.,the verb in N-V sequences and V2 in V-V sequences). We conclude that theN-V sequences are formed in the lexicon. In case of the V-V sequences wehave found that there are two types of CPs. Type 1 is a product of syntaxwhile Type 2 is formed lexically. We call the lexical CPs Compound Verbs.

This thesis also addresses the issue of storage of these constructions in adatabase called Verb Knowledge Base (VKB). As these items are not listed instandard dictionaries, we provide a programme for the automatic extractionof Compound verbs from various corpora.

The work focuses on the following:A. isolating N+V combinations that are conjunct verbs from among other N-VsequencesB. isolating V+V combinations that are CPs from other V-V sequencesC. determining whether the CPs isolated above are formed lexically orsyntacticallyD. studying semantic properties of V2s (Supra-lexical properties of CPs)E. storing of lexical CPs in a lexical databaseF. providing an algorithm for automated extraction of CPs from a corpus

Page Updated: 06-Oct-2010