BENGALURU: Rishi Atul Rajpopat (27), PhD scholar from St John’s College, Cambridge, just solved a 2,500-year-old grammar problem in the ancient Sanskrit texts written by 5th-century BC scholar and philologist Panini, that had perplexed traditional and modern researchers all these centuries. A major implication of his breakthrough is that now the algorithm that runs Panini’s grammar can potentially teach this grammar to computers.
Mumbai-born Rajpopat’s PhD thesis, published on Wednesday in the journal Apollo — University of Cambridge Repository and titled ‘In Panini We Trust: Discovering the Algorithm for Rule Conflict Resolution in the Ashtadhyadyi’, has decoded a two-and-a-half-millennia-old algorithm which for the first time has made it possible to accurately use Panini’s ‘language machine’ in his famous Ashtadhyayi.
Rajpopat said Panini’s Ashtadhyayi has around 4,000 rules (sutras) that describe the structure of Sanskrit. “He has given one metarule (a rule that governs the application of other rules) — ‘1.4.2 vipratisedhe param karyam’ — to help us decide which rule should be applied in the event of ‘rule conflict’. But the problem in the last 2,500 years has been that scholars misinterpreted this metarule and added more metarules with grammatically incorrect results,” he explained. Often, two or more of Panini’s rules are simultaneously applicable at the same step, leaving scholars to agonise over which one to choose.
“Because of the misunderstanding and misinterpretation by traditional scholars, we were getting grammatically incorrect forms. My dissertation is on the understanding that ‘param’ (in ‘1.4.2 vipratisedhe param karyam’) meant, between rules applicable to the left and right sides of a word, that the user of the grammar should choose the rule applicable to the right side and not the serial order thereafter. I spent a lot of time with a lot of derivatives before I came to the conclusion, because Panini’s algorithms produce grammatically correct words and sentences without errors. His grammar is self-sufficient. I did not add any metarule to understand the master (Panini),” said the researcher.
Rajpopat said once he decoded Panini and his “language machine”, it became easy to feed in the base and suffix words and watch them turn into grammatically correct words and sentences through a step-by-step process.
“Computer scientists working on natural language processing gave up on rule-based approaches over 50 years ago… So teaching computers how to combine the speaker’s intention with Panini’s rule-based grammar to produce human speech would be a major milestone in the history of human interaction with machines, as well as in India’s intellectual history,” he said about the potential of teaching Sanskrit grammar to computers.
“Panini had an extraordinary mind and he built a machine unrivalled in human history. He didn’t expect us to add new ideas to his rules. The more we fiddle with Panini’s grammar, the more it eludes us,” he said.
Speaking about his journey as a research scholar, Rajpopat said he was a student of mathematics.
“I was always into patterns, problem solving, coding, numbers etc. I did my graduation in economics but for postgraduation, I wanted to take up a subject I loved. During one of my visits to a Navaratri pandal, I was informed about a retired professor in Mumbai, who taught Panini’s grammar. I approached her and joined her classes,” he narrated. Rajpopat did his Masters at Oxford, and met his supervisor at Cambridge, Professor of Sanskrit, Vincenzo Vergiani, who motivated and mentored him during his research.
Rajpopat’s research may now mean that Panini’s grammar can be taught to computers for the first time.
“I learnt that asking the right questions is a skill. One has to be obsessed about one’s work. The role of a teacher cannot be undervalued,” he said.
“What the Cambridge scholar (Rishi Rajpopat) seems to have done is to discover that a new interpretation of the word ‘Param’ (in Ashtadhyayi) can result in more natural derivations and without having to add many rules for exceptions that arise with the earlier interpretations… His new interpretation is that the word ‘Param’ in the metarule is applicable in the part that is on the “right side” of the word, not the rule that comes later in the Ashtadhyayi.
This seems to reduce the number of exceptions and hence be easier to explain and understand. Note that this observation has to be validated by scholars of Sanskrit after a close reading of the thesis and its applicability to the whole corpus of Sanskrit. Also, note that the structure of Ashtadhyayi is more or less the same except when it comes to selecting the next rule when there is contention in the contexts discussed earlier. However, the contribution is significant as this issue has occupied the great grammarians for the past two-three millennia.”– Prof K Gopinath, Senior Professor, Computer Science & Artificial Intelligence Department, Plaksha University, Mohali (formerly with Indian Institute of Science, Bengaluru)
