Publishing Partner: Cambridge University Press CUP Extra Publisher Login
amazon logo
More Info

New from Oxford University Press!


Style, Mediation, and Change

Edited by Janus Mortensen, Nikolas Coupland, and Jacob Thogersen

Style, Mediation, and Change "Offers a coherent view of style as a unifying concept for the sociolinguistics of talking media."

New from Cambridge University Press!


Intonation and Prosodic Structure

By Caroline Féry

Intonation and Prosodic Structure "provides a state-of-the-art survey of intonation and prosodic structure."

Academic Paper

Title: Computational Identification and Analysis of Complicated Sanskrit
Author: Subhash Chandra
Email: click here TO access email
Institution: Centre for Development of Advanced Computing
Linguistic Field: Computational Linguistics; Morphology; Syntax; Translation
Subject Language: Sanskrit
Abstract: This paper presents a model for computational identification and analysis of/L/complicated Sanskrit noun phrases [(nominal morphology or Sanskrit subanta-padas)/L/(NPs)] in Sanskrit text. The simple ones or those forms which are strictly rule/L/governed and fall in to patterns are not very difficult to analyze. However, there are/L/several complicated and ambiguous forms which pose a challenge for analyzers. The/L/purpose of this paper is to put forth a strategy and algorithm which can enable any/L/Sanskrit parser to recognize and analyze these complicated NPs. Identification/L/includes separating the NPs from Verb Phrases [(tinanta) (VPs)] by a strategy of/L/isolating verbs and in-declinables. Analysis includes splitting the NPs into its subconstituents - base [{(praatipadik) (any meaningful form of a word, which is neither a/L/root nor a suffix) (PDK)}], case-number markers [(karaka-vacana-vibhakti) (KVV)]./L/Sanskrit is a heavily inflected language and depends on serial inflections on nouns/L/and verbs for communication of meaning. A fully inflected unit is called pada/L/(useable word) which are NPs or VPs. Therefore identifying and analyzing these/L/inflections are critical to any further processing of Sanskrit./L//L/According to Paanini, there are 21 nominal inflectional suffixes (seven/L/vibhaktis and three numbers 7 X 3 = 21) which are attached to the PDK according to/L/the category, gender, number, and end-character of the base. Some forms of Sanskrit/L/NPs can be very complicated for computational identification and analysis for the/L/examples. For examples: ramaah, bhavati, gacCati, etc. can be both a nominal as well/L/as verbal construction. The pronominal forms pose another challenge, as in most of/L/them; the inflected forms can not easily be related to their bases morphologically. We/L/may have to posit ad-hoc rules and processing to handle them. For example - ‘aham’/L/(first per sing), ‘tvam’ (second per sing), ‘sah’ (third per sing pronoun), ‘amu’ etc./L/are NP formed from respectively the base ‘asmad’, ‘yusmad’, ‘tad’, etc. by inflecting/L/for nominative singular and ‘adas’ by inflecting for nominative dual./L//L/The system first does punctuation, avyayas and verbs (non-NPs)/L/identification for NPs identification in Sanskrit text. After identification of these/L/words, system recognizes all remaining words as NPs and sends for analysis process./L/System does identification of Avyaya (AV) and VPs with the help of AV and VP/L/database. We have stored around 524 AV forms, commonly used in modern Sanskrit/L/languages and about 500 commonly used verb roots and their forms for verb/L/recognition. So we have around 90,000 verb forms stored in UTF-8 Unicode/L/devanagari scripts. Thus the NPs in Sanskrit text are identified by a process of/L/exclusion. After the verbs and avyayas are identified by their lexical pattern matching/L/search, the remaining words in the text are labeled NPs./L//L/The system also has some basic requirements for use- 1. JAVA installed to support/L/the Java Web Server. 2. Apache Tomcat 4.0 installed web Server. 3. Baraha software/L/for UTF-8 Unicode Devanagari input or any other. If the user’s machine does not have all of these then they can not use this system./L//L/The present work is an attempt to process Sanskrit NP inflections by way of/L/Paanini’s rule system, appropriate database and example-base. The system developed is an online system run on Apache Tomcat platform using Java servlet, MSSQL server/L/2005 as back end and JBDC for connectivity. The goal is to simplify Sanskrit text for/L/self reading, understanding, and also for any Machine (Aided) Translation (MAT)/L/from Sanskrit to other languages.
Type: Individual Paper
Status: Completed
Venue: Allahabad, India
Publication Info: ICCS, Allahabad, Proceeding
Add a new paper
Return to Academic Papers main page
Return to Directory of Linguists main page