What is called Gandhari is some of the earliest Indic attested on record, in the inscriptions of Aśoka in the Gandhara region (now in northern Pakistan).
How is Gandhari itself clearly later than Pali? Based on what evidence?
It is known for a fact that Kharosthi script (used to write Gandhari) has far clearer links with the Aramaic script introduced to India by the Achaemenid empire that comprised Gandhara in the lifetime of the first generation of Buddhists, so the Yona and Kamboja janapadas referred to repeatedly in the Pali canon and other EBTS were Greek and Persian ruled states respectively in the north-west of the subcontinent. Thus the Pali Canon appears to be aware of an Indo-Hellenic kingdom that succeeded Alexander and therefore could not be from the Pre-Mauryan era i.e. before 300 BCE.
The aramaic and kharosthi therefore represent the very origins of writing in India. Ashoka has Gandharan edicts in Aramaic and Greek as well apart from Gandhari. Gandhari therefore goes back to the late 4th century BCE if not earlier (as it is known to precede Aśoka). No kind of Pali goes back that far on the attested record.
Epigraphic Middle-Indic of the Aśokan era is neither Pali nor Prakrit, rather Pali is based on it, Pali is a later (phonetically and grammatically standardized/reformed) version of it. So it is wrong to call the phonetically and grammatically way more diverse Epigraphic Middle Indic as “just Pali”.
Incorrect. India has an exception only as regards the Vedas. It cannot be extrapolated to all other literature. None of the Indo-Aryan literature apart from the Vedas was transmitted orally, they are all written literature.
The Early Buddhists had no identical scholastic mechanism to transmit the Pali canon as the Vedic oral transmitters had, there is no evidence to claim that they did have such a scholastic apparatus. None of the classical sanskrit texts have been orally transmitted although traditionally sanskrit scholars memorize written texts orally as part of teaching them to others and make their students memorize it so their memories act effectively as RAM (Random access memory) to fetch parts of the text at will as they study it in detail - this memorization of the text from written sources is not oral transmission.
The early Buddhists might have similarly memorized written texts to study them, but that does not qualify as an exclusively oral transmission. The early Buddhists did not need oral transmission as they were aware of writing from the earliest period of canonical Buddhism. The earliest Buddhist canon therefore was not in Pali but in what we now consider Gandhari. Pali canon shows evidences of being created from a Gandhari original, including several spelling mistakes that make sense only in Gandhari/Kharosthi script.
Besides Aśoka names suttas and he as the ruling monarch would not have been part of any Buddhist oral transmission to recollect sutta names from memory, he was obviously naming written suttas that were read out to him. Suttas only got named when they were compiled and canonized in writing, not when the Buddha was speaking or before the written record.
Oral transmission virtually guarantees that a text wont stay undistorted. That is why there are 21 rescensions of the Rgveda, 101 rescensions of the Yajurveda, 1000 rescensions of the samaveda etc. All the variant rescensions were deemed authoritative by their own oral traditions and were passed down orally.
The text of the Pali canon however is extremely regular and fixed and is very unlike an orally transmitted corpus of texts.
I do find Bhante’s comment curious. If he meant it as a shortcut that Gandhari canon is later than Pāli canon, I still can’t make sense of that.
Even if we agree that Pāli might’ve been written down in 3rd bc, and with Gandhari scrolls having Prajñaparamita texts, and even assuming that Prajñaparamita tradition is later than Agamic tradition — even then, we don’t actually have evidence that Pāli texts written in 3rd bc are the same texts we have today.
Pāli tradition points to an earlier tradition from Prajñaparamita? Maybe. But for example, is a tradition that closes with Prajñaparamita, earlier or later than an Agama tradition which incorporates commentaries written some centuries laters? And so: Pāli texts being earlier than Gandhari texts? That seems a stretch.
Yona derives from Ionian/Yaunā, a designation already used in the Achaemenid world before Alexander. Kamboja referred to a northwestern population or region with Iranian associations, but it does not mean “the Persian Empire” or necessarily a Persian-ruled kingdom. Gandhāra itself was listed as an Achaemenid territory around 520 BCE.
The names could fit a pre-Alexandrian northwestern context just as easily. Consequently, they provide no terminus post quem of 300 BCE.
The term Yona can refer to actual Hellenistic Greeks in Asoka’s inscriptions, but that does not mean every occurrence of Yona in every textual layer must originate after Alexander.
This claim needs named examples and publications. Without them it is an assertion, not evidence. Although if it’s qn original research of yours, I’m sure it would shake the ground.
The Mahābhārata has (allegedly, I’m little versed in jt) deep bardic and oral-formulaic dimensions, and scholarship overwhelmingly accepts an early oral phase for Buddhist literature.
I think I understand that you’re trying to separate Oral Transmission from Oral/Written Memorisation.
Even then, the sangha is one proof for the pedagogical apparatus. In fact, how almost all the important Patimokkha rules are the same across the different schools, and how their backstories change so widely, is a curiosity. It seems composition was a part of the culture.
In fact many Buddhists would deny this, but if we treat the canon as a fictionalised “play of formulas” as suggested by Shulman, we can see the very pedagogical apparatus in action.
Pārāyanavagga, Vinaya has ample episodes of Brahmins beings involved with the recitation of materials. The suttas are abundantly intimate with the Vedic culture, linguistic training, both the pedagogical corpus and methodology.
It’s not inconceivable that a disenchanted brahmin intelligentsia, coupled with the esthethics and sensibilities of samana movements, would result in a perfect oral community, the most successful example being the Buddhist project.
Just when do you place the birth and death of the Buddha, then, if you think writing was always a part of it?
There are suttas that are clearly referred to by name by other suttas, and Buddha mentions other suttas, examples, similes, etc.
Now, of course, I do take this with a healthy grain of suspicion. However, the earliest record of “Oral Period” is the attestation of the suttas themselves. Yes, to “prove” beyond a doubt, there would need to be other documents confirming it, ideally; still. It seems just as outrageous to ignore it completely and somehow simply assert that suttas were only named in the written period, with no proof to this assertion.
It’s kind of unfortunate that the thread with the original question is locked. Oh, well. Anyway, I don’t think Pali is the original language of Buddhism, but I also don’t think that it is particularly late, either, as the OP of the other thread was arguing. My experience of working with Sanskrit, Pali, and Chinese Buddhist EBTs is that they all are more-or-less of equal dating.
Well, except for Sanskrit - that language in my opinion is a later adoption by Buddhists when they decided to standardize their Prakrit texts in the north. But otherwise, all of the material has a range of dates, some earlier and some later, inside each corpus. Indian literature throughout is like that. They weren’t originalists the way Westerners are originalists. They progressively added to and developed their works over time, which bedevils attempts to determine their history. It’s pretty obvious if a person spends the time comparing versions of a given Indian work over time when we have the witnesses to do it with.
Actually, I shouldn’t single out Indian literature. It’s a feature of ancient literature in other cultures too - I’m reminded of the studies of the Dao De Jing in China. There’s the version everyone reads today, but scholars find different versions of it at different times in history.
In my understanding Mahāyāna and Prajñāpāramitā literature take the Āgama corpus as their precedents and are therefore a few centuries later than the Āgama/Nikāya Buddhism.
But between the Sanskrit/Gandhari/Pāli Āgamic corpus collections that used to exist historically, it is my understanding that the people who composed those texts did not view them as independent languages the way we think of them today - but as different ways to write the same text (using different written conventions/scripts etc). This is clear if we compare different copies of the Aśokan edicts which are word for word the same edict written following different transliteration conventions/scripts:
The Middle-Indic (seemingly proto-pali) form found at Girnar (western India) in Brahmi script of Ashoka’s rock edict 1:
iyaṃ dhaṃma-lipī Devānaṃpriyena Priyadasinā rāña lekhāpitā
idha na kiṃci jīvaṃ ārabhitpā prajūhitavyaṃ
na ca samājo katavyo
bahukaṃ hi dosaṃ samājamhi pasati Devānaṃpriyo Priyadasi rājā
The same edict in Middle-Indic at another location (Kalsi, northern India):
iyaṃ dhaṃma-lipi Devānaṃpiyena Piyadasinā lekhitā
hidā no kici jive ālabhitu pajohitaviye
no pi ca samāje kaṭaviye
bahukā hi dosā samājasā Devānaṃpiye Piyadasī lājā dakhati
The same edict in middle-Indic (gandhari) form found written in Kharosthi script at Shahbazgarhi in the Gandhara area (in the north-west):
aya dhrama-dipi Devanapriasa raño likhapitu
hida no kici jive arabhitu prayuhotave
no pi ca samaja kaṭava
bahuka hi dośa samayaspi Devaṇapriye Priadraśi raya dakhati
The same edict in Middle-Indic (Gandhari, kharosthi script) at another location in Gandhara (Mardan):
ayi dhramadipi Devanaṃpriyena Priyadraśina rajina likhapita
hida no kichi jive arabhitu prajohitaviye
no pi ca samaje kaṭaviye
bahuka hi doṣa samajasa Devanaṃpriye Priyadraśi raja dakhati
In standard sanskrit, it would be written as (which is not very different from the language used by Aśoka or even from Canonical Pāli):
iyaṃ dharmalipiḥ Devānāṃpriyeṇa Priyadarśinā rājñā likhitā
iha na kiñcid jivaḥ ālabhya prahotavyaḥ
nāpi ca samājaḥ kartavyaḥ
bahavo hi doṣāḥ samāje Devānāṃpriyaḥ Priyadarśī rājā paśyati
These are not different languages but different systems of writing the same language. The words used are the same in each case, written using different phonetic and written conventions.
If you can understand the grammatical and phonetic properties of these, you will notice at once that the Gandhari here is simultaneously more Prakritic as well as more Sanskritic - than any kind of Canonical Pali. This is the historical evidence for it, one cant claim in good faith and knowledge that Pali precedes this stage of historical middle-Indic. Pali seems to be a far more smoothened and sanitized (later) version of Middle-Indic that we see here historically.
The Gandhari script (Kharosthi) due to its similarity and derivation from the Imperial Aramaic script of the Achaemenid Persian empire has a long pedigree before Aśoka. So based on evidence, Gandhari is attested well before Pali but they are surface level registers whose underlying language was the same (and that language was probably Buddhist Hybrid Sanskrit or something very close to it).
Even Aśokan epigraphic Middle-Indic is older than Pali (and Pali appears to be a standardized derivative of the epigraphic middle-Indic register). Pali in its current canonical form is not found attested outside Buddhism as far as I am aware, so I am pretty confident that it is a register forged by early Buddhists based on a pre-existing natural language and not a pre-existing natural language by itself.
Not really, Yona and Kamboja in the EBTs are described as janapadas (Indic kingdoms) and are not referring to either the Persian/Greek empires outside India or their non-Indian territories. There were no Greek states in or close to India, before Alexander’s time.
Many Classical Sanskrit texts that were likely coeval to the Pali canon mention yona and kamboja along with the Gandhara people showing knowledge of their relative locations in NW India.
Mahabharata 12.200.40.2 yauna-kāmboja-gāndhārāḥ kirātā barbaraiḥ saha
Mahabharata 5.19.21.1 sudakṣiṇaśca kāmbojo yavanaiśca śakaistathā (mentions the śakas in addition to the persians and greeks. śakas are connected etymologically to the buddha’s dynasty, śākya = belonging to the śakas )
Besides Pali has many weird spellings of popular names such as calling the mountain ‘Meru’ or su-Meru (great Meru) as sineru (which makes no sense even in Pali). Similarly the name of the king ikṣvāku, which in Gandhari becomes iṣmahu, becomes okkāka in Pāli (which again has no phonetic, etymological, or semantic sense even within Pali) These weird spelling-mistakes point to a written manuscript tradition rather than an oral transmission tradition where such meaningless spelling mistakes would not arise.
The bards of the Mahabharata did not claim to be oral transmitters of the epic, rather they were public reciters/performers/storytellers of the Mahabharata back in a time (i.e. in Ashoka’s time and earlier) when Sanskrit used to be spoken widely and the Mahabharata was well understood when recited in the original.
There was however no oral transmission of the Mahabharata of the kind that existed for the Vedas, and nobody has suggested that the Vedic oral transmission was used for classical sanskrit texts to my knowledge. Theoretically speaking however the Vedic oral transmission could be made to work for the Mahabharata which is almost entirely in verse (like the Vedas which are also predominantly in verse). However that method cannot be applied to prose texts either in Sanskrit or in other languages.
Besides that, the Vedic oral tradition is intimately intertwined with the grammatical and phonetic traditions of sanskrit scholarship i.e. the science of grammatical analysis was invented for Sanskrit as a consequence of the need to preserve the Vedas perfectly. It would not work in a non-sanskrit environment as the pedagogy of vedic oral transmission involves learning the vedangas (i.e. grammar, etymology, phonetics, metric versification etc) which are all linguistic tools that support the oral transmission tradition. Pali did not have such linguistic traditions until much later and could not have simply copied the Vedic oral transmission without such pedagogical apparatus
The sangha is not by itself a pedagogical apparatus, your idea of pedagogy and its use in oral transmission appears to be very different from any form of pedagogy that the historical vedic people employed, so there is no comparison. The patimokkha rules can be the same in a manuscript tradition, that has nothing to do with any pedagogy of oral transmission.
Despite not demonstrating the existence of any historical pedagogical tools or apparatus or methods about how exactly that oral transmission took place?
It is not whether Buddhism was popularized by a set of brahmins disgruntled with the Vedic religion or not that is the issue, it is whether oral transmission is possible for texts like the Pali canon even by Brahmins, and if it is possible, whether we can find out how exactly they did such an oral transmission for centuries. To me as I see it with what I know, neither is there any evidence of the possibility of oral transmission for the Pali canon, nor is there any evidence of there being any historical oral transmission of it.
So to me, the Pali canon (and other Agama literature) have their beginnings not from the mouth of the Buddha, but from the pen of their first writer. We cannot recover any canonical text from its pre-written stage, because such a stage does not exist for the canon. The canon has always been a written canon from the beginning, and that is why it employs a uniform language for the most part (at least for the prose). The only parts where oral transmission could be in contention are the wholly versified suttas, such as those in the Sutta Nipata. The prose suttas cannot have had any oral transmission.
Sure, and that’s because they existed as named suttas in prior written texts. Some suttas have multiple names for example the Buddha suggests at the end of DN1 that the Brahmajāla sutta can be called either Brahmajāla, or as the Dhammajāla, or as the Atthajāla, or as the Diṭṭhijāla, or as the Anuttara-saṅgāma-vijaya. So if other texts called it the brahmajāla-sutta consistently, rather than as the dhammajāla, or as the atthajāla or diṭṭhijāla etc (each reciter in an oral transmission would pick a different title if it were transmitted orally) - it is because brahmajāla was the title frozen in written manuscripts.
I am not ideologically opposed to a Buddhist oral transmission, I simply want evidence of it. What I find from most academic sources are merely hand-waving theories like these that attempt to superimpose the vedic oral tradition on Buddhism without evidence:
That doesn’t make sense. The Chinese translations are almost entirely from the early centuries of the common era, most of them post-dating the Sarvāstivādin scholar Aśvaghoṣa who had before him almost the entire Sarvāstivāda canon along with their commentaries and the garangutan Abhidharma Mahāvibhāṣā in classical sanskrit.
So the Chinese & Tibetan Āgama translations are way younger than almost all Indic transmissions of the canonical texts. Between the two however the Chinese started translating centuries before the Tibetans.
There were Gandhari canons probably before any other Indic versions of the canon were prepared as Gandhari had the earliest written traditions. From Gandhari to Pali & BHS is not a huge jump linguistically, it is likely therefore that Pali & BHS represented the 2nd stage of canonical redaction.
Buddhist Hybrid Sanskrit in some ways represents the Buddha’s own mother tongue.
The Sarvastivadins however used standard Sanskrit for the most part for their canon, and they were the most historically influential of all the EBT sects.
It is true that Middle-Indic (not Prakrit) versions of the canon were eventually Sanskritized - however that is not the whole story. Middle-Indic itself originates in the early Buddhist period from Sanskrit which is Old-Indic (there is no evidence of pre-Buddhist middle-Indic), and the transition from Old-Indic to Middle-Indic is partly triggered by the introduction of writing and the linguistic disruption it created in the early centuries of Buddhism. People who had been speaking Old-Indic all-along now had to comprehend it being represented non-phonetically in writing. This gave rise to plenty of linguistic confusions heralding the start of a new linguistic stage called Middle-Indic. However the Buddhist sects that sought to put their canon into Sanskrit would have known that Sanskrit was etymologically their ancestral language and probably the spoken language of the Buddha and his Brahmin contemporaries so they may have chosen to take it back to Sanskrit once Sanskrit writing had evolved enough to represent it phonetically accurately. Phonetic writing was likely not widespread in the early centuries of writing so what we call Middle-Indic in the Ashokan era is basically non-phonetically (and/or ungrammatically) written old-Indic. This is clear from the study of Gandhari.
I dont know if you can see what I am trying to explain above, the historical picture is certainly much more nuanced than someone deciding on a whim to use one or the other language. Firstly there were no multiple Indo-Aryan languages to choose from in India in the 3rd century BCE. Pali & BHS did not exist as independent languages before the Pali canon and BHS canons were written for the first time. They were not considered independent languages even when the Pali canon and BHS canons had been written down. Pali and BHS were merely canonical registers. Their linguistic reality is only to the extent of their shared characteristics with Old-Indic. No modern Indo-Aryan language has evolved directly from Pali (or retains the characteristic sound shifts introduced by Pali), as it is an artificial canonical literary language, not a spoken dialect. It’s not that Pali and Gandhari are more original from a Buddhist POV as people seem to assume.
It doesn’t seem too hard to imagine a scenario where first, smaller units of the Buddha’s teaching were transmitted orally, and later a fixed, written down corpus of teaching,“the Pali canon” was created. Which, of course, would not need an oral tradition memorization apparatus at that stage.
I believe something somewhat similar existed with what we know as the Illiad and the Odyssey. The bards that recited the later, fixed, written down versions of these poems were doing something different from the early, oral , more improvisational tradition. And a distinction needs to be made from transmitting fragments purely orally vs. memorizing a larger text from a written document.
If I memorize a chunk of the Pali canon, I’m not participating in an oral tradition, I’m using a written text to learn how to recite something, just as I might learn lines of a Shakespeare play.
Something that doesn’t seem mentioned here is that perhaps there was a motivation in keeping the Vedic texts oral (i.e. so they could not be read by just anyone and were therefore somewhat secret, even suppressing a written form when writing was available), while this motivation does not seem to exist for the Buddhist texts.
P.S. Why was the thread where this discussion better belongs, “date of Pali texts” locked?
I think repetitive structures and formulaic passages can be seen as evidence for an oral phase. If the canon had been composed in writing from the very beginning, it would be difficult to explain such a heavy waste of effort and writing materials.
Writing does not ipso facto mean phonetic writing. Vedic is a tonal language, and the tone placements shift dynamically within words, and were very difficult to capture in writing. Writing early vedic hymns down non-phonetically in the Ashokan/Early-Buddhist eras would have killed both the archaic language and its comprehensibility, making the texts sound gibberish when read out. Even Ashoka complains in one of his edicts that the scribes were sloppy in writing down his commands accurately on the edicts. So in the early centuries phonetical accuracy was not a feature of most written texts, people followed several shortcuts simplifying complex syllables to make words easier to write in the early scripts. As the writing grew more and more phonetic in later centuries, texts began to be written phonetically accurately (i.e. as in standard sanskrit). This is not a switch from Middle Indic to Sanskrit, rather Middle Indic itself was malformed sanskrit, at least that is how people would have seen it back then.
So whatever the motivations were for not writing it down, I am pretty confident the Vedas falling into the wrong hands was not probably the main or only reason. Eventually there were systems developed to mark the tones and the script itself largely evolved and became capable of phonetic representation, and the Vedas were in fact written down.
I dont think so, the formulaic passages and repetitions help the scribe to correct spelling mistakes and patch damaged/illegible portions of the palm leaf manuscripts (by referring to cognate sections in other suttas). If anything the repetitions are even more secure as evidences of a written tradition. We see this historically in Gandhari, the evidence is fragmentary, and by reference to repeated passages in similar texts we are able to partly reconstruct a fragmentary/damaged/lost text.
One does not repeat what one says over and over in natural speech, that’s not how any real oral transmission tradition preserves texts. It would only confuse both the reciter as well as the listener to see the same sentences popping up again and again at different points of the recitation so they would lose track of the place they were reciting or listening to. The Vedas which were transmitted orally dont contain such repetitions.
the lack of any mention of reading, writing, or written transmission in the extensive Nikaya corpus which mentions nearly every other social practice or phenomenon of equal importance multiple times.
the fact that the social world accurately depicted in these texts is that of the early Second Urbanization, which was a period of very rapid change (so it’s easier to pin chronologically) at least 200 years before your terminus post quem.
We can grant that they didn’t do oral transmission using the exact same methods as those by which the Vedas were transmitted, but it’s a leap to infer from this that they had no means of maintaining an oral transmission. And besides there is such evidence, circumstantial perhaps like the two points I made above, but it’s not the case that there is literally no evidence.
It’s obtuse to restrict our view to the Pali canon when the hypothetical tradition of oral transmission we’re discussing is also the source of the Āgamas and the Gandhari texts you mention, and indeed the entirety of all Buddhist canons is written at least in reference to this tradition and its fruits. If we don’t restrict our view this way, there is ample evidence of exactly the kind of variation you are denying, for example the variations Analayo has pointed out in Early Buddhist Oral Tradition (2022) comparing Chinese to Pali texts.
This is a popular misconception. Secular topics were forbidden by the Buddha for the monks and therefore talking about them isn’t attested much in the canon, but still there are suttas like these (Ud3.9) which mention vocations in the Buddha’s time - of which that of a writer is one.
We know from independent sources that writing was already being introduced at least since the Achaemenid Persian invasion of Gandhara in the 520s BCE, and the Buddha lived after that so writing did exist in his time though it may not have been widespread until after his time.
So even if the EBT canonical texts dont mention writing or reading at all (they do, but even if they didnt), it is not conclusive evidence for the lack of writing as there is other evidence of writing being in use in the Buddha’s era.
Even if you think this sutta Ud3.9 is not from the Buddha’s time, the fact that writing is mentioned in a canonical text is conclusive proof that the canonical authors were aware of writing from the very start.
I don’t get your point clearly, but the gap between the Buddha’s death and Ashoka’s birth was more like 100 years, perhaps lesser. The mahaparinibbanasutta speaks about the Moriyas of Pipphalivana laying claim to the Buddha’s corporeal relics. Moriya (Skt. Maurya) is the name of the dynasty and empire that Ashoka ruled, and they are already mentioned in the canon at the point of the Buddha’s demise. Where do you get the idea that of it being “at least 200 years before your terminus post quem.”?
Those who claim that there was an oral tradition usually try to superimpose the Vedic oral tradition on Buddhism uncritically. Nobody to my knowledge has claimed or shown evidence that there was a different kind of oral transmission in Early Buddhism unconnected to the Vedic approach.
It is not my intention to exclude the Sanskrit/Gandhari Agamas from the purview of oral transmission and restrict it to the Pali canon only, but the existence of several canonical linguistic registers complicates the picture rather than simplyfying it. How did oral transmission work with multiple linguistic registers all simultaneously claiming to preserve texts from the same single source?
It’s as if the works of Shakespeare were available in English, Frisian and German from the beginning, all claiming to be from the mouth of Shakespeare himself. How does that work?
What is this evidence? I read a 2020 paper by Analayo supposedly about the Early Buddhist Oral Tradition, where, apart from the same uncritical superimposition of the Vedic oral tradition on Early Buddhism, there is precious little actual evidence of any EBT oral transmission.
So if Analayo in his 2022 book is positing a non-Vedic model of oral transmission (which doesnt seem to be the case from this paper), and if he has evidences for it, what exactly does he posit?
I’d like to thank everyone for all the interesting information they have provided about this topic. To elaborate, I know Sanskrit and I’m just starting to learn Pali. I’m acquainted with the dramatic Prakrits, but I very much doubt that they reflect the way anyone actually spoke. Things like the inscriptions of Asoka are more realistic. I’m inclined to agree with srkris here. We are not talking about separate languages but dialects of a language, with Classial Sanskrit as the elite dialect, the equivalent of BBC English. In fact it’s striking how close some of the Asokan inscriptions are to Classical Sanskrit. You just make a few phonetic changes and you’ve got it! There is very little syntactic difference. Sure, there are a few changes like the loss of the dual, and some survivals from Vedic, but the differences are not great.
Anyway, to return to the Pali canon, I can certainly see that the contents reflect the world before the Nanda and Mauryan empires. The language, however, shows all the signs of being a constructed language from various dialects. My guess is that the original literature was in various dialects and translated into Pali, for communication needs. It’s like the way a person from Yorkshire and one from Cornwall may not be able to understand each other’s dialect, but both will understand BBC English. You got a similar situation with Koine Greek in the Mediterranean area after Alexander. The society had high communication needs and so a dialect that everyone could follow was necessary. Does that sound plausible?
Are you saying that all or most Early Buddhists (from all or most janapadas) that lived between the Buddha’s period and Ashoka’s period settled on Pali as their common lingua franca? If so, at what point would they have abandoned whatever they were speaking in and switch to Pali as a lingua franca? Also why would all the non-Buddhists, from different social strata and geographies, who (as per the Pali canon are reported as speaking canonical Pali to the Buddha and his contemporaries) have chosen Pali? Did they all have to work out Pali’s linguistic idiosyncracies before having those conversations with early-Buddhists? Having chosen Pali as their common lingua-franca (going by your supposition), why would most of the EBT sects (except the Theravadins) abandon Pali equally quickly and move towards Sanskritizing their canons? There were Gandhari canons in existence, as well as Hybrid Sanskrit and Classical Sanskrit, their vocabularies are way more morphologically and phonetically diverse and cannot plausibly be grammatically derived from Pali - so how can Pali be considered as the original Buddhist Lingua Franca par excellence?
In my understanding, the early Buddhists (and the Buddha) could have only used a widely spoken pre-existing pre-Buddhist vernacular with the wider world (a language plausibly linguistically ancestral to all the Indic canonical registers in which canonical EBTs are attested in), and not a ‘Buddhist’ register like Pali.
Right, that’s the dating of the Chinese translations. The EBTs that were translated to Chinese are of more-or-less equal dating as the ones we have in Sanskrit and Pali in terms of their content - which involves looking at their relative development, the degree of added material, the verbosity that generally increased in Buddhist texts in India as time went on, etc. You have to think in terms of the content, not just the date of the physical manuscripts because EBTs throughout their existence have moved from one language to another. The development of the content itself is another issue aside from their translations. Generally speaking, when Buddhist texts were translated to Chinese they stopped being developed. The EBTs in particular did not continue to change beyond the problem of errors that crop up when they are copied over and over to maintain them.
I’m still doubtful about this idea that Sarvastivadins had converted their entire canon to Sanskrit as early as Asvaghosa. I can see that there are assertions by Indic scholars about this, but my experience working with the Agama translations from the 4th and 5th c. CE is that they were at best in early BHS. The proper names have the features of a Prakrit language like Gandhari. It may be that some Sarvastivadins did it earlier than others, I suppose. The Agamas that the Chinese received and translated were certainly not in standard Sanskrit.
Now, when we look at the Mahavibhasa and Abhidharma or the Mulasarvastivada Vinaya, those works were clearly in Sanskrit when they were translated, and their translators (Xuanzang and Yijing) tell us they were in Sanskrit. But they are 7th c. CE witnesses, not 2nd c. CE. So, I would like to know what the earlier dated Sanskrit fragments for these texts are.
Now it’s my turn to say, “That doesn’t make sense.” You are thinking the Buddha was speaking in Sanskrit, but his disciples translated it to Prakrit and then back to Sanskrit again? That doesn’t pass the Occam’s Razor test for me. I think they were speaking vernacular Prakrit from the start and resisted Sanskrit precisely because it was associated with brahmins and their elitism. Buddhists prioritized meaning over wording and pronunciation as far as I can tell, being a more vernacular and informal tradition in the early centuries. Hence, we see so much variation and change in their EBTs.
This doesn’t sound plausible to me, either. Human languages are always evolving in specific regions, becoming dialects that eventually diverge too much and function like distinct languages. That would have been true in ancient India as much as it is today. That we don’t have direct evidence of those languages at the time doesn’t mean they didn’t exist. It’s like saying different species of animals didn’t exist unless we have fossils of them. I’m sure many are not represented in the fossil record that existed at any give time in the past.
To me, this argument also applies to the “There was no Oral Transmission apparatus / tradition then”. From Jains, we know that at least one other group also had some sort of a oral transmission.
Considering for example, the time of the Jesus and just how many “Mesiahs” were there at the time. The Buddhist canon also remarks on multiple different teachers at the time. In a society without writing, oral pedagogy would’ve been the only tool at their disposal - there’s nothing else to teach with. And so the easiest explanation is that the Buddhist Project represents the most successful version of this post-Vedic oral traditions.
This doesn’t mean that all the material in EBTs come from oral tradition. The metion of writing in places like Udāna does point to a long range of date when suttas were composed.
Also, for a tradition that’s painstakingly detailed with regards to everything, if writing existed from the time of the Buddha, I’m sure we’d have mountains of suttas and rituals and such on how, when and what to write and so on.
The Abhidharma Mahavibhasha is dated to the middle of the 2nd century CE (150 CE). This is also in Ashvaghosha’s lifetime. The Abhidharma Mahavibhasha presupposes the existence of the entire Sarvastivadan canon as well as the canons of the other EBT sects.
I can see this too about the names, the transliterations of the names in Chinese don’t yield Sanskrit pronounciations consistently. It’s because the proper names are a category by themselves, it is not easy to translate them accurately when the text is Sanskritized, because most of the names are historical fossils that needs philological knowledge to Sanskritize accurately, knowledge that the translators didnt have. People translating the name of the monk vapa from Gandhari (or vappa from Pali) centuries after the monk lived may not have known that it must be vāṣpa/bāṣpa originally in Sanskrit, so they might have left it as vapa in the BHS translation, which the Chinese captures (as in MA12). This is a constant issue in most Buddhist Hybrid Sanskrit, it is also an issue in most Pali as well (the authors were at a loss about how best to transliterate proper names to Pāli). This does not mean however that Buddhist Sanskrit and/or Pali were independent languages of their own. Buddhist hybrid sanskrit is basically half-baked Sanskritization (not an independent language by itself) but in the eyes of the people translating to BHS, they were most certainly trying to translate to standard sanskrit and viewed their output as Sanskrit, not as middle-Indic (that they were translating from). Besides even in Gandhari (and in Ashokan epigraphic middle-Indic and Pali), one sees the reverse i.e. Sanskritic forms that could not be perfectly put in Middle Indic for example purandara (an epithet of Indra/Śakra) becomes a meaningless ‘purindada’ in pali as a result of misreading the gandhari script, kakuda kātyāyana becomes pakudha kaccāna in Pali, the name pārāśarya has 4 different spellings in Pali (pārāsariya, pārāpariya, pārāsiviya & pārāsiriya) and so on. So Pali has pretty much the same issue of problems in transliterating proper names correctly as the Chinese has, except in reverse from Sanskrit.
I’ve also mentioned in another topic how R.C.Childers when speaking about Pali sandhi points out Sanskritic words that remain phonetically fossilized in the Pali canon without satisfying middle-indic grammatical constraints.
These issues don’t however mean that the Pali canon is not in Pali, it simply means that it was not converted to Pali 100% perfectly, and still retains the earlier Sanskritized spellings that the Gandhari original might have had. So BHS having middle indic vocabulary does not mean it is not Sanskrit, but just that it is imperfectly Sanskritized. These imperfections reduced with time but they were not 100% eliminated in canonical texts. I must know as I see the same problems when translating suttas into Sanskrit at every turn.
So whether the Chinese translation of the suttas was from standard Sanskrit or Hybrid Sanskrit doesnt matter beyond a point.
As I’ve tried to explain above, they are not different languages in the beginning but different (non-phonetic) conventions to write the same spoken language. Phonetic representation in writing would yield what we call Sanskrit, whereas Gandhari and Pali and Ashokan middle-indic are non-phonetic ‘written’/literary registers of the spoken language. This is my understanding. There were no multiple spoken languages to translate from or to. Gandhari and Pali and BHS did not have a name (neither did Sanskrit) in the Ashokan era to distinguish one from the other as independent languages.
If they were different languages, they were just beginning to diverge away from Old-Indic. No form of Middle-Indic had a lengthy pedigree of being a spoken language at that point, and their representation in the written record was non-phonetic in the Ashokan era i.e. their actual pronunciation and their written representation are two different things. This is the same situation as today’s english - thought (pronounced like thot), though (pronounced like thow), through (pronounced like threw), trough (pronounced like truff) are written similarly but pronounced differently by the reader. This was the case in all forms of Middle-Indic, including Pali. The writing is not a phonetically perfect representation of pronunciation. Modern readers (even scholars) dont appreciate these nuances well enough and they try to superimpose Standard Sanskrit’s phonetic accuracy standards on the Middle-indic registers uncritically. This also is one of the reasons why they misdiagnose the identity of these registers and dont realize that the early forms of Middle-Indic hadnt diverged sufficiently enough in the Ashokan era to be considered independent languages. There was no translation from the Buddha’s mouth to Middle-Indic the way you posit it. Rather it was the Old-Indic itself that the Buddha spoke that was written non-phonetically and with grammatic and phonetic idiosyncracies resulting therefrom, which we call Middle-Indic. So in the eyes of the Buddha, he didnt find several independent languages to translate his teachings to or from, in the regions he visited, he rather found janpada nirutti (i.e. different nouns used to refer to similar things in different places, among people speaking the same language - think of automobile vs car, lift vs elevator etc).
Sanskrit was not associated exclusively with brahmins in the Buddha’s era (if that were the case, Brahmins would be depicted as speaking in Sanskrit in the Pali canon), nor were Buddhists trying to avoid Sanskrit (this is a mistranslation of the vinaya rule sakāya niruttiya. Meaning depends on wording, and cannot be isolated from wording, so prioritizing meaning over wording makes no sense. The fact that we have similar wording - syntax, grammar and vocabulary in the suttas redacted by almost all EBT sects shows that they did prioritize the wording and did not consider themselves free to paraphrase the Buddha however they liked. Also the Indic Buddhists did not keep on translating their canons to later-evolved Indo-Aryan and non-Indo-Aryan languages, they had a notion of Pali/Gandhari/Sanskrit as being their source canons and there are no translated canons in later vernaculars. Even the Sri Lankan commentaries in the Pali tradition are in Pali, which was not Sri Lanka’s spoken Indo-Aryan language in the time of Buddhaghosha. So the idea that early-Buddhism did not care about wording is wholly untrue.
They were not as fastidious as the brahmins were about pronunciation because pronunciation played little role in a written tradition, unlike the Vedic tradition where orality was central. But words did matter a great deal to Early Buddhists.
Yes, linguistic diversity is a fact of life, there is no exception as regards India, you’re on the dot there. However, we are talking about identifiable and attested Buddhist registers here, not about linguistic diversity in general. We dont have independent evidence that the EBT traditions (or their coeval non-Buddhist contemporaries) considered Buddhist registers like Pali or BHS as being independent languages in their own right. The fact that canonical Pali and BHS have no independent existence on the attested record outside the Buddhist canon militates against your supposition that they were pre-made readily available languages that Buddhists were trying to translate into or out of. This is why I say that linguistic reality of these ecclestiastical Buddhist registers must be examined critically with much nuance. Things are not what they appear to be at the surface level.
I guess what I’m thinking of is that dialects that may have been mutually intelligible in Asoka’s time, or before, were no longer mutually intelligible by (let’s assume) the first century BCE. So a standardisation may have been necessary. I don’t know if anything like this did happen, but it seems possible, given the parallels with Koine Greek. May I ask, what is your view of the development of Pali, which certainly seems like a constructed dialect, for which the most obvious reason would be to facilitate communication.
Why would that be uniquely a Buddhist problem though, for It to have a Buddhist solution like Pali?
That Pali is a constructed register based on Ashokan middle Indic is not in doubt, however what I dont see the evidence for is about it being a spoken dialect, or that it was used to facilitate communication.
Besides such a standard (if one was required) did already exist in the form of classical Sanskrit, and it had widespread acceptance not only within Buddhism but among Jains, early Hindus, late-Vedic religionists and others. It is not the case that Sanskrit had a resurgence in later periods, it was the case that Sanskrit has had a steady decline from the earliest periods, the early Buddhist period was more Sanskritic than in the later Buddhist period. This might sound paradoxical for modern scholars who are used to the refrain that early Buddhists were anti-Sanskrit but historically the Buddha lived in the Old-Indic period, the Nanda and Maurya and Śunga eras leading to the start of the common era is a period of transition to Middle-Indic triggered by the Indic discovery and adoption of (non-phonetic) writing originating from the Achaemenid Persians, which gradually became phonetically accurate (Sanskritic) over the centuries.
So the early Buddhists didnt intend to construct a dialect for communication, rather they were putting into a new script (Brāhmī) the pre-Existing EBT canon in Gandhari (that was written in Kharoṣṭhī). The peculiarities of the Kharoṣṭhī script were that it didnt distinguish long vowels from short, nor did it consistently depict complex syllables etc. So the canon written in Gandhari/Kharoṣṭhī was was not phonetic either. These were partially fixed in the Brāhmī representation taking the output closer to Sanskrit in some ways, a phonetically and grammatically standardized version of this Brāhmī representation is what survives today as Pali.
The Pali spellings of proper names are not attested anywhere outside the Pali canon. There is no evidence for example that the brahmin gotra patronymic Pārāśarya (descendant of Parāśara) was called in different regions or dialects by the names pārāsariya, pārāpariya, pārāsiviya & pārāsiriya (all of which are attested in the Pali canon as lexical variants of the same name). They are merely spelling mistakes, they are not phonetically accurate representations of actual dialectal pronunciations.
If Pali were a Middle Indic spoken standard, one would find Pali phonetics and grammar in later Prakrits and in modern Indo-Aryan languages at least among Indo-Aryan populations significantly influenced by Pali Buddhism like the Sinhalese.
Any oral teaching/explanation is not ipso facto oral transmission of a pre-existing text. The goals of the two are quite distinct. In an oral explanatory teaching the goal is to make the student understand, in an oral transmission the goal is to preserve the text without change and the transmitters are merely instruments that ferry the texts from place to place and time to time. Teaching ordinarily needed no transmission specially when the Buddha and his audience were in most cases face to face. Transmission is a different beast.
The suttas are normally narrated in third person which implies that the narrator was not there as an eyewitness, when the Buddha was speaking. He recites aloud from a text that was formerly recited aloud by someone else, by saying “evam me sutam”. Even the very first reciter was not therefore an eyewitness, he merely reads out what is written, else he would’ve been saying evaṃ me diṭṭham (thus I’ve seen) not evam me sutam (thus I’ve heard).
In the Buddha’s time there are numerous suttas where the Buddha’s teaching is shown to be misrepresented by someone, and then that person is called to the Buddha’s presence where he admonishes them for misrepresentation. If there was an oral transmission, they would not have needed the Buddha’s intervention, they could have simply said your teacher has not taught you correctly, and/or your teacher’s teacher didnt teach your teacher correctly and so on i.e. there would be a source to the mistake and how it arose.
But in the EBTs we dont see transmission errors like that being recognized or discussed, rather what we find are errors in comprehending the teachings, thus there is no oral transmission nor even a mention of problems that would plusibly arise in such an oral transmission tradition.
Besides nobody in the canon appears to be interested in transmission of the teachings across generations, that worry seems to have arisen towards the end of the Buddha’s life or after his demise.
The Buddha himself in DN13 on the other hand refers to the Vedic oral tradition explicitly (22.8 Yepi kira tevijjānaṁ brāhmaṇānaṁ pubbakā isayo mantānaṁ kattāro mantānaṁ pavattāro, yesamidaṁ etarahi tevijjā brāhmaṇā porāṇaṁ mantapadaṁ gītaṁ pavuttaṁ samihitaṁ, tadanugāyanti, tadanubhāsanti, bhāsitamanubhāsanti, vācitamanuvācenti…) without recognizing that such a oral transmission tradition existed for his own teachings.
Very interesting points about the variations in Pali being due to writing in Kharosthi. It seems plausible that something like this was happening in part.
Two points I would like explained that seem like evidence against a wholesale adaption from texts.
First, the Vinaya has rules for how to chant, how to preserve certain sounds etc. while there are no rules on how to write the suttas.
Second, I’ve heard it said that the Pali canon first came to be written down in Sri Lanka because the monks who could remember it were too few. My understanding is this story comes from much later than the events, but even so: why would they claim an oral tradition and the need to write it down if there was already a written one?
It has been pointed out that the Pali Canon was first written down in places far from where the suttas are set, namely Sri Lanka and Gandhara. Why this is the case is open for debate.
There seems to have been an oral tradition prior to the writing down, but it may not have matched what we have today, written down as “the Pali Canon.” It seems hard to determine what the oral tradition consisted of.
