This must be the case if writing didn’t come along til a couple of centuries after the Buddha.
Such accomodations are obvious in the history of English orthography. Maybe Indic scripts don’t have quite the level of perfection we credit them with!
1 Like
Perhaps you are right, but Pali was a liturgical language and possibly never used as a daily spoken language. (Almost certainly not at the places where it was first written down.). So it seems, combined with the premium placed on memorization, that the sounds of Pali might have been well preserved.
3 Likes
Yes. The pronunciation of words is not what the Pali tradition must have cared about much (or at all). Their focus was evidently on getting the contents of the canon preserved for posterity, and they didnt mind if the words looked like gandhari, or pali, or sanskrit or BHS (they were merely orthographic variants - in my understanding). As long as the texts were preserved and they could make sense of what they meant, they weren’t particularly bothered about the writing being non-phonetic. But that calls into question which variant of the EBTs is in the original language (or close to the original language as it was spoken by the Buddha)? It must have been Sanskrit as it is the phonetic form that the other attested forms can evolve from. The reverse (pali or prakrit evolving into sanskrit phonetically) isnt linguistically feasible. But the prakritic form is the first written form of the canon.
That would also explain why Pali (as we find it in the canon) is a liturgical register unique to Buddhism, and not attested a general spoken language of the wider society.
2 Likes
Well that’s interesting, because Pali supposedly retained more forms than Sauraseni on which classical Sanskrit is based.
Sanskritization of Buddhist texts doesn’t appear to have begun until after 0 CE. They were still translating Prakrit to Chinese in the 5th century. By the 8th century, everything was in BHS and a couple centuries later Buddhist texts appeared that were fully Sanskritized in Tibetan translations. I would think Sanskrit forms in Pali sources would be indications of this later period of translation from Prakrit to Sanskrit. I.e., the Prakrit forms are the older readings.
1 Like
Yeh, I remember reading a long time ago that it’s not treated as a prakrit of Sanskrit, because it’s closer to vedic in a number of important ways. I don’t know whether it’s merely liturgical. I’ve read several places that there’s a fair amount of consensus that it is 1) educated and 2) likely the language Buddha spoke. Whether from the middle country or maghada seems at dispute.
And also, there’s something from Buddhaghosa about some monks asking Buddha if they could translate his words into sanskrit verse and he told them no. You teach it in the master’s words.
Anyway, busy. Nice chatting!
1 Like
Thanks. I have been harbouring similar thinking for a while now. From my rather bad attempts at production, I am beginning to understand Pali is a euphonic language, it places high emphasis on words that are easy to pronounce and “sounds nice”, so even things like the the spacing between long and short vowels are important. Again, very similar to Japanese which also has long and short sounds and and the relative occurrence and alternation between them are important.
That’s why we have no less than 4 causative affixes. It doesn’t mean they are interchangeable (even though Kaccāyana implies they are). In reality, one affix will be the dominant form for any given word, because it results in the nicest sounding word (that doesn’t collide with other words in the language).
I also get the feeling that Pali is a dynamic language rather than having a fixed vocabulary. A good speaker can construct “new words” on the fly with the right combination of roots, prefixes and affixes, and a good listener will be able to deduce the meaning, and some of the words used in the Pali canon may have been constructed this way. Which makes it a good language to convey Buddha’s teaching, since difficult philosophical concepts (which I doubt would have been in common usage had Pali been a language of the common people) can be constructed specifically to convey an aspect of the teaching. Over time these words became part of the “technical vocabulary” used to document the teachings.
I had been thinking Kaccāyana was an attempt to “reverse engineer” existing Pali vocabulary by deriving the rules used to construct them, but I think these rules are specifically there to enable new words to be constructed. So we have a language flexible enough to be “dynamically self-modifying” to be able to be optimised for conveyance of a specific philosophy. How cool is that?
1 Like
Yes Pali is a dynamic /generative language, as were its predecessors Sanskrit, Proto-Indo-Aryan, Proto-Indo-European, and most other early Indo-European languages.
Sanskrit (classical and Vedic) are even more grammatically analysable/generative in that respect than Pali (which you would see if you look at Pāṇini’s 4000 rules documenting Classical and Vedic Sanskrit).
So it is in Pali’s genes to be analyzable, but Canonical Pali had already started losing this ability even in the earliest attested suttas as it was starting to become merely a frozen literatary register with a large number of forms that couldnt be grammatically analyzed or explained. We dont see the production of non-Buddhist texts in Pali right from its earliest stage, it has remained exclusively a Buddhist literary register, rather than the natural spoken language of an ethnicity.
I do not believe that Kaccāyana’s rules are likely to be 100% valid for all of canonical pali i.e. there are still likely to be a large number of fossilized terms that dont fit the general rules for word generation that Kaccāyana lays down. So I think we have to take Kaccāyana’s with a dose of healthy skepticism about its ability to explain all/most attested words.
Not only was Kaccāyana reverse engineering Pali from attested source texts, but Pāṇini was doing the same a 1000 years before him. Kaccāyana does not claim that he was reverse engineering a spoken language of an ethnicity, but on the contrary Pāṇini and all his contemporaries and followers make that claim (that the grammar was a reverse engineered model of the spoken language of their time, and of the Vedic that preceded them).
2 Likes
Not really, there were several early Buddhist Nikāya (EBT) schools having full canons and commentaries in Sanskrit by the start of the 1st millenium CE. For example, the Sarvāstivādins apart from having a core canon, had a huge Abhidharma commentarial corpus (even bigger than the Theravāda tipitaka as a whole) in Sanskrit - called the Abhidharma Mahāvibhāṣa Śāstra (which today survives only in Chinese translation) - this was towards the beginning of the Common era. Aśvaghoṣa (the author of the Buddhacarita, a biographical novel on the life-history of the Buddha in perfect Pāṇinian sanskrit poetry), as well as Saundarananda (Biography of Nanda, the Buddha’s cousin) & Śāriputra-Prakaraṇa (the acts of Śāriputra) also lived sometime in the 1st century CE. He wrote in Sanskrit invariably and evidently his pre-existing Sarvāstivāda canonical sources were also Sanskrit.
Also you may know better about the Chinese and their āgama sources, but a source like pāli is probably out of question for the canonical texts that the Chinese translators had for their source. As far as I know, the Pali canon doesnt call its nikāyas the āgamas, and the Chinese tripiṭaka’s contents show sanskritic and gandhari underlying textual traditions, rather than any other kind of prakrit or pali sources. I consider Gandhari as the chronologically original prakrit, and all other prakrits (including pali) are later Prakrits.
However you are probably right in considering that the earliest written canon to have been in a Prakrit like Pali or Gandhari, rather than Sanskrit - and I believe that was because they were copying the orthographic conventions of Aśoka who would have found it difficult to write phonetically on rock edicts (for it required showing too many glyphs to represent Spoken Sanskrit phonetically accurate) - so the simplified orthography of the Aśokan edicts served as the model on which the language of the pali canon was standardized and simplified. It helped the Pali tradition to rapidly copy canonical texts (and carry them around all across the subcontinent) without having to focus too much on phonetic or phonemic exactness.
Since this is our first interaction here, I take this opportunity to thank you for your Chinese canonical translations.
Kaccāyana claims that his work is comprehensive with respect to jinavacanayuttaṁ hi (conforming to the usage of the words of the Buddha). His work is also the most succinct of all the grammars (with just 675 suttas) so probably has the least degree of reverse engineering compared to later works.
It’s true he doesn’t explain everything, and the book seems more focused on production rules rather than grammatical explanations (he doesn’t properly explain pronouns for example and doesn’t even give a full listing). I guess he considers his readers to be fluent in Pali and therefore do not need the obvious explained.
I imagine there would have been a lot of “massaging” of the Pali canon in the years after the Buddha’s death to convert his teachings into a form suitable for chanting, and perhaps after a few hundred years the text would have been somewhat frozen, but we have evidence the language continued to evolve in the latter commentaries and texts so the language wasn’t completely “fossilised” but perhaps was no longer organically adapting.
Of course, for all we know Pali may have never existed as a real language but an invention of Sri Lankan monks based on their Sanskritisation of Prakrit based languages plus a dose of Sinhalese.
With due respect, that is not what it means (as I understand it). It means that only those words that are found to conform to canonical usage are dealt-with in his grammar i.e. it doesn’t account for non-canonical word forms. i.e. it restricts its scope to the canon. It is not a statement on the comprehensiveness, rather a statement of limitation. It isn’t saying it can account for all canonical usage, but it is saying it should not be expected to account for non-canonical usage.
Possibly, but Kaccāyana also asserts:
Sabbavacanāna’mattho akkharehe’va saññāyate.
Akkharavipattiyaṁ hi atthassa du’nnayatā hoti.
Tasmā akkharakosallaṁ bahū’pakāraṁ suttantesu.
In other words, he is interested in “correct” productions. He may be deliberately omitting what he considers to be “corrupt” usage. He mentions “jinavacanayuttaṁ” quite a few times in the book, and always in the context of “the following suttas are referencing …” indicating his intention is to be comprehensive.
I dont understand what you mean by ‘correct productions’ - are you saying that he intends his grammar to lead to accurate word-forms - and not create improper or defective words?
I haven’t studied the Kaccāyana grammar, but here too I think what he is saying is something different (as it appears to me without any prior knowledge or study, so that’s the disclaimer):
Sabbavacanāna’mattho akkharehe’va saññāyate = The meaning of all utterances is understood with the (correct) use of syllables/alphabets.
Akkharavipattiyaṁ hi atthassa du’nnayatā hoti = Defects in syllables/letters/alphabets causes the meaning to be led astray (misinterpreted)
Tasmā akkharakosallaṁ bahū’pakāraṁ suttantesu = Therefore a knowledge of correctness of letters (as employed within words) is very helpful (in a study) of the suttas.
I don’t see where he’s saying anything about correct productions. He is simply making general statements about the importance of being letter-perfect, in order to not miscomprehend the meaning of the suttas.
yes, that’s how I am interpreting it. The structure of the text seems to be geared towards listing correct usage of the rules, as well as examples of incorrect or inapplicable usage. And always the reason is with reference to “jinavacanayuttaṁ”. I don’t think he is necessarily restricting the “scope” to the canon, but that he is using the canon as the ultimate reference for correct usage.