Creating a digital Sanskrit library

Prof. Scharf
File Photo

He calls himself a wandering ascetic, follows Ayurvedic principles in his daily life and is keen on conducting a basic course in Sanskrit.

Meet Prof Peter Scharf, a visiting professor at the Language Technologies Research Centre (LTRC) who has created a digital Sanskrit library by making it easy for researchers to have access to original Sanskrit manuscripts, texts, and lexical resources online.

How did you get exposed to Sanskrit in the US?

I was first introduced to Sanskrit at the age of fifteen when my brother returned from a teacher-training course in the Transcendental Meditation programme (TM) and taught my whole family to practice it. I had been practising for several years already by the time I was in college when I myself decided to become a TM teacher.

Did you formally study the language?

I began my formal study at a first-year Sanskrit class as a graduate student at Brown University at the age of 24. Then over the Summer, I got a tutor to help me study the first six chapters of the Bhagavad Gita. I memorised the first two chapters at the time because I’ve always had an interest in the oral as well as the written aspect. In the West, they teach Sanskrit only in the written form. To truly learn a language, why should you limit yourself to the eyes and the visual mechanism?

Tell us more about the Sanskrit Library.

In 2002, I founded the library. The first project we had was to take digital texts and digital dictionaries and integrate them with linguistic software. This would help people read the texts more easily.

If the sandhi (interword phonetic changes) in Sanskrit had been analysed, they could click on a word, and a morphological analyser would present the possible morphological analyses and corresponding stems, and you could click on the stem and look up the word in a digital dictionary.

Analysis is more of a challenge with normal Sanskrit text where the sandhi is not analysed. It is very difficult to locate the word boundaries. So, we collaborated with Gerard Huet, the computer scientist in Paris, who was creating a Sanskrit parser at his Sanskrit Heritage site.

Gerard creates a display page that shows a summary of all of the possible solutions. By clicking on checks to accept or reject a possible word analysis, one can narrow down the possible solutions and eventually come to an unique solution. We also started building a digital corpus where someone could save their analysis, and those results would be available later for more linguistic research.