Use of Features for Accentuation of ghañanta Words

Sanskrit is an accented language. The accent is very prominent in Vedic Sanskrit but classical Sanskrit does not make use of it. Accent is a disambiguation device which can help to understand the correct sense of a word. Words in Sanskrit tend to fall in certain grammatical categories like ghañantas, vocatives etc. Words in a category tend to have an accent governed by either a general or a special rule of accentuation. Pāṇini’s special (apavāda) rules of accentuation for a specific category of words can be studied along with the general (utsarga) rules of accentuation for the same class. The resulting body of rules states all the conditions necessary for accentuation of words from that class. The conditions reveal a set of features which can be used for determining the correct accent. If the features of accentuation for a class of words is understood, computers can be trained to accentuate unaccented words from that class. This paper discusses some features of ghañanta words and how the features can be used to restore accent on unaccented ghañanta words.