Sanskrit versus Pali and Nagarjuna

There is a Wikipedia article that might offer some contextualization. It is on so-called "Buddhist Hybrid Sanskrit" (BHS). From the article:

In many places where BHS differs from Sanskrit it is closer to, or identical with, Pāli. Most extant BHS works were originally written in BHS, rather than being reworkings or translations of already existing works in Pāli or other languages. However, earlier works, mostly from the Mahāsāṃghika school, use a form of "mixed Sanskrit" in which the original Prakrit has been incompletely Sanskritised, with the phonetic forms being changed to the Sanskrit versions, but the grammar of Prakrit being retained. For instance, Prakrit bhikkhussa, the possessive singular of bhikkhu (monk, cognate with Sanskrit bhikṣu) is converted not to bhikṣoḥ as in Sanskrit but mechanically changed to bhikṣusya.

In the paragraph immediately following this one, the article speaks of a Yale professor named Franklin Edgerton. If you follow the cited links on that Wikipedia entry, you'll see a great deal of them citing his groundbreaking paper The Prakrit Underlying Buddhistic Hybrid Sanskrit. Luckily for us, this paper is available for-free online to anyone with a Google account. The paper is here:

https://www.jstor.org/stable/608058

You can use your Google account to view the article for free and skip the Wikipedia summation of it. Basically, the gist is that the earliest Mahāyāna sūtras and several pre-Mahāyāna texts (such as the Mahāvastu associated with the Mahāsāṃghikas) were originally preserved in an unknown "Pāli-like" Prākrit and then were later "Sanskritized" and then were later eventually translated into a standardized "classical" Sanskrit. Of course, not all extant Mahāyāna sūtras are in a pure classical Sanskrit. The Lotus Sūtra for instance is in classical Sanskrit for the prose sections and in BHS for the verse sections.

With regards to reading Venerable Nāgārjuna, I've never heard of the text of Mūlamadhyamakakārikā (MMK) described as particularly Prākritic. AFAIK, its present-day redaction is in a classical Sanskrit. As such, you can read it as Sanskrit, but you'd have to be ready for two things:

1. difficulties related to the verse, and
2. Buddhism-specific terminology that ultimately has its root in the unknown Prākrit underlying Buddhist Sanskrit.

Take for instance the two terms "saṃvṛti" and "lokavyavahāra."

"Saṃvṛti" comes from a root meaning "to conceal." It is used in Buddhist Sanskrit to refer to a "convention." Its direct cognate is the Pāli "sammuti." You can see it used in Pāli in the Vajirāsutta. The Buddhist Sanskrit analyses the Prākrit term thus:

sammuti --> saṃvṛti

Why?

nibbuti --> nirvṛti

They understood the vowel as a phonological change derived ultimately from "vṛti." The Prākrit tradition (namely "Pāli" here) understands "sammuti" as a form of "sammati," meaning "consensus, agreement."

Why is this relevant? When Venerable Nāgārjuna discusses "saṃvṛti" or "the conventional," he is not referring to "saṃvṛti" in the traditional Sanskrit sense. He is using a Buddhist Hybrid Sanskrit sense, where "saṃvṛti" is understood to mean "lokavyavahāra" or "the speech of the world."

Lokavyavahāra is another example. In normal Sanskrit, this would refer to the daily life, perhaps even the daily processes, of the people in the world. In Buddhist Sanskrit, this has the specific sense of "the speech of the people of the world," which is once again closer to Pāli (Pāli: "lokavohāra") than Vedic or Classical Sanskrit.

As a final note of evidence, just look at how Buddhists use the terms "dharma" and "saṃskāra" compared to non-Buddhist Sanskrit usage. The difference is striking. Only Buddhists use "dharma" in such a way, which reflects the usages of the term from the original Prākrits that the buddhavacana was first preserved in.

PS: I just remembered this article hosted for free by the University of Toronto:

https://tspace.library.utoronto.ca/bitstream/1807/68342/1/Levman_Bryan_G_201406_PhD_thesis.pdf

It is Dr. Bryan Levman's thesis entitled Linguistic Ambiguities, the Transmissional Process, and the Earliest Recoverable Language of Buddhism. It discusses many of these linguistic ambiguities and anomalies, namely when one linguistic tradition of Buddhism (Pāli) reads words in such-a-such a way with such-and-such a meaning and another (substantiated in Sanskrit or Chinese) preserves them differently. To the best of my knowledge, this paper particularly does not deal with the ambiguities that the Chinese texts pose, but Chinese texts (which often come from Gāndhārī) have many of these same incongruities. Even things like the proper reading of "pratyekabuddha" was not h*m*gonous in early Buddhism, also appearing in forms like "緣覺" (*pratyayabuddha lit. "caused-buddha").