There’s been a lot of media buzz lately about an “Indian PhD student at Cambridge University” solving a “2,500-year-old Sanskrit puzzle”. Here are a few examples of recent press coverage—The BBC, The Hindu, Indian Express. If you’ve read one you’ve read them all because they are basically all copy-pastas of each other. Unfortunately the coverage is pretty scant on details, so I thought I’d provide a quick primer for those interested in understanding what the whole thing is about. Note that I am omitting irrelevant micro-details in the interest of overall clarity.
Panini’s Ashtadhyayi can be thought of as a kind of generative grammar that can generate only and all grammatically correct utterances of Sanskrit as spoken during the time it was written. More specifically, it has an inventory of base / root forms, along with an inventory of affixes (prefixes, suffixes etc.) and a set of transformation rules for applying one or more of these affixes to the root to form valid words. There are also a set of metarules that provide guidance or constraints on how / where these rules should be applied.
This (and a bit more) is put together in the form of close to 4000 sootras, which can be thought of as very terse / compact phrases / sentences. Think of the compact English spelling mnemonic “An I before E, except after C, or when sounding like A, as in neighbor or weigh” as an example of the Paninian style of conveying rules and exceptions, with the first “sootra” establishing the default rule, and two “sootras” following it specifying departures from the default behavior.
Although a magnum opus, the Ashtadhyayi isn’t amenable to being read cover to cover like a textbook. So alongside the text has evolved a very rich culture of discourse about the Ashtadhyayi, with critical analyses, illustrative examples, and commentaries filled with arguments and counter arguments about the interpretation of the text.
Since it’s very terse, the precise meaning of some sootras of the Ashtadhyayi can be unclear or might require reading between the lines. Some of these gaps can be filled using common sense, but many others need more detailed investigation.
As an analogy, imagine Python code presented without any indentation. In some cases the indents can be guessed based on common sense, such as after a line ending in “:“, while in other cases you might apply reasoning like “a statement following a return statement won’t be at the same level of indent because then it would never be executed”. However, that will still leave complex cases that might need to be figured out by looking at valid input-output pairs and back-calculating.
In the case of some sootras, following their surface meaning might result in forms which are obviously ungrammatical (based on actual attested usage), so grammatical traditions have in many cases fine-tuned their meanings or developed heuristics (and sub-heuristics) to be borne in mind when applying them.
This brings us to the issue at hand—Dr Rishi Rajpopat’s thesis on one such metarule of the Ashtadhyayi. This rule (#1.4.2 विप्रतिषेधे परं कार्यम्) has been interpreted over the millennia to mean that whenever you reach a branch in the derivational process where multiple rules can be applied (in other words, there is a rule conflict), you should apply the rule that comes later in the Ashtadhyayi. However, applying this blindly generates many obviously incorrect forms, so traditional analyses have also come up with several (50+) heuristics to overrule this metarule.
Dr Rishi Rajpopat’s thesis argues that the interpretation that ties be broken in favor of the numerically later sootra of the Ashtadhyayi is fundamentally an incorrect interpretation of Rule 1.4.2, necessitating the creation of a litany of exceptions to justify grammatical forms that would otherwise be deemed ungrammatical. He provides a different interpretation of what he thinks Panini meant by “rule conflict”, and claims that the “later” (पर) in the sootra signifies the later half (RHS) of the operand of the rule, rather than a numerically later rule. For instance, if at a point in the derivation, the operand was “A B” and there were two candidate rules, one that modified A and one that modified B, the tie should be broken in the favor of the rule that modifies the RHS of the operand, aka B, rather than the rule that comes numerically later in the text.
His claim is that his interpretation gives rise to more correct derivations out of the box, and relies on far fewer heuristics and exceptions compared to the current tradition. Since Rule 1.4.2 is a very widely applicable metarule that shows up in many derivations, Rajpopat’s interpretation, if it bears out, would be a noteworthy change in the way certain aspects of derivation are thought about and would considerably simplify the mental model for derivations that rely on Rule 1.4.2
A brief comment on the media coverage on this issue. Most coverage seems to be referring to all of this as a 2500 year-old puzzle / a mystery of Sanskrit grammar / the key that finally unlocks Sanskrit grammar etc. This is a bit sensationalist—1.4.2 wasn’t a mystery or the missing piece of a puzzle. It was interpreted a certain way so far and it’s not been settled yet if Rajpopat’s interpretation is better overall, with fewer false positives and false negatives. If his claim survives the critical analysis it’ll be subjected to in the weeks to come, I hope he gets due credit for it and that modern Sanskrit scholarship adopts it. Also, 1.4.2 is far from the only ambiguous / open-to-interpretation sootra of the Ashtadhyayi, so ‘solving’ it doesn’t really ‘unlock’ all of Sanskrit grammar in the way the media seems to be representing.
Edit (21 Dec 2022): Neelesh Bodas (the creator and maintainer of ashtadhyayi.com and a teacher and student of Paninian grammar) has released his critique. You can access the full document here and a video summary here.