Criteria for deciding if a text is an EBT

“At some point” is a rather crucial phrase there! Commentary on Prajnaparamita can hardly be considered ‘Early Buddhism’, right? So this would seem to be unrelated to the investigation of the Buddha and his community, wouldn’t you agree?

1 Like

It dates to the same period as many Pali texts do and agress with the Sarvastivada Vinaya, etc. It was translated 400 CE and quotes Agamas and Vinaya texts quite often.

It is, isn’t it? No one really knows anything about what happened during the time of the Buddha, since anyone can argue that anything is a later addition. It’s all subjective, especially with these texts were commonly edited and updated over time. It’s rather difficult to discern what’s “early” and what isn’t beyond correlating ideas with later Abhidharma and mythology. Beyond that, there’s lots of wishful thinking and guessing taking place.

1 Like

I’m lost - you’re talking about “the Commentary on the Prajnaparamita”, right? The earliest Prajñāpāramitā texts were composed around 1st century BCE, no? And presumably this commentary could be as late as 400CE. Whereas the Early Buddhist texts are from around 450BCE~300BCE, so far as I understand. I see no reason for taking Mahayana commentaries of such late origin, as evidence for the nature of buildings 5th century BCE.

Are you arguing against the whole Early Buddhist Texts concept? And the entire field of Early Buddhism? If so, I must disagree, since it’s an evidence-based discipline and has a vast amount of scholarship behind it.

1 Like

The Pali canon officially dates back to the 1st century BCE, as do the earliest fragments of texts discovered in Central Asia. So, these sources are all on equal footing in that regard.

The reason the commentary is relevant is because it quotes Sarvastivada texts verbatim. The Sarvastivada is an extinct early Buddhist sect that was a northern Indian equivalent to the Theravada today. So, yes, the quotations in a Mahayana text are valuable for the study of early Buddhism. It’s not any different than reading Pali commentaries that quote suttas. It’s the quotes that are potentially early texts, serving as witnesses to how those texts read when the commentaries were created.

No, but we don’t have texts that go back to the time of the Buddha. Or, to put it more precisely, we don’t know which ones do and which ones don’t. We can weed out the texts that are obviously later like I said, but we don’t know what happened between the 5th and 1st centuries BCE. We have to rely on stories that are told much later by sources like Buddhaghosa and the Vinaya histories. They’re valuable, but they may also have forgotten a great deal by the time they were written. Alot can happen and be forgotten in four centuries.

1 Like

Please correct me if I’m wrong, but do not the consensus of Early Buddhist scholars directly disagree with your view presented? If we are only talking about the Pāli Canon as we have it today, the date is many centuries later than that, so far as I understand. You seem to be talking about the time is was written down. But I am talking about neither of these things. When it was written down tells us nothing about either the date of its latest texts, or the dating of its earliest texts! It’s just one somewhat arbitrary point in the history of the collection. Furthermore, this is not the case with the Mahayana sutras, which were written texts from the start - written compositions, and thus categorically much newer than any of the oral traditions coming from the Buddha.

So far as I understand, the consensus of Early Buddhist scholars date the Early Buddhist texts to 300BCE and earlier, the 300BCE cutoff point being the time at which the schools separated.

So, should we not instead be quoting from that vinaya? We do have that in Chinese, right? So, should we not see what it has to say?

Yes that’s fair point. We also get such quotes in the Tibetan canon, which is important since they never even bothered translating a single āgama! Though, I think it’s logical to check with the vinaya itself also.

Also, I read in HARLES WILLEMEN, ‘Remarks about the History of the Sarvāstivāda Buddhism’ (2014):

During the reign of Kaniṣka (155–ca. 179 A.D.),15 about 160–170 A.D., a Sarvāstivāda synod was held in Kaśmīra. This synod established a shorter vinaya and a new abhidharma in Sanskrit.

So would this date their vinaya to earliest 160–170 A.D? If this is the only source for this information, then it would seem we cannot rely on it, as it’s a very late source, nearly 5 centuries after the ‘Early Buddhism’ period.

To quote Bhikkus Sujato and Brahmali in their wonderful book ‘The Authenticity of the Early Buddhist Texts’ [oh… copy and paste errors, I will correct the title at least but please forgive the rest]:

ARE THERE ANY AUTHENTIC BUDDHIST TEXTS? If so, what are they? These are questions of tremendous spiritual and historical interest, about which there is a range of opinions that often appear to be irreconcilable. Traditionalists insist that the texts were “spoken by the Buddha” in the most literal of senses, while sceptics assert that we cannot know anything about the Buddha for certain, and further, that the notion of authenticity is irrelevant or pernicious. Rarely, however, has the question of authenticity been systematically investigated. Seeing the lack of an easily accessible summary of the evidence, we decided to assemble this survey.

There are two main aspects to our argument: (1) there is a body of Early Buddhist Texts (󰡜󰠤󰧎s), which is clearly distinguishable from all other Buddhist scripture; (2) these texts originated from a single historical per- sonality, the Buddha. These two aspects are closely linked, so we have not tried to separate them or present them in sequence.

We consider the doctrinal and linguistic evolution of the texts, ground- ed in their social and economic context. No particular methodology is preferred; rather, we aim to be inclusive, as we believe that diverse per- spectives are useful, indeed essential. So we take into account the internal development and structure of the texts, as well as the results of compar- ative studies. Multiple independent lines of evidence from the 󰡜󰠤󰧎s con- verge on a point of origin geographically in Northern India and temporally around the 5th century 󰠤󰠷󰡜.

If you have not read it, I highly recommend it. i also admire the work of Bhikkhu Anālayo (I think you would be interested to hear his analysis of the Chinese terms for vitakka and vicāra by the way - I favour the translation ‘initial mental application’ and ‘sustained mental application’ myself, based on studying Pāli and Chinese texts in parallel with him (using English medium)). So far as I understand from interacting with these people and studying their work, they are all holding the view I have presented. Though please correct me if I’m wrong. I also find their reasoning to be thorough and valid. That is to say, we have a really quite clear picture of the teachings of the Buddhist community around 100 years after the Buddha died. And we have a good method of identifying Early Texts with a really rather comfortable degree of certainty. Though this does not mean that there aren’t some texts of which we are unsure whether they are early or not. Many however, we have that for. And the Āgamas play a huge role in that.

As for the likes of Buddhaghosa, personally I entirely ignore him when considering early Buddhism. Though perhaps once in a rare while I find the commentaries interesting when sharing their understanding of what earlier teachings meant. Such as their mention of some disagreement on whether bodhisatta’s satta meant Sanskrit’s satva or sakta. That was a rather tasty piece of evidence, and helps to support the mistaking of an adjective for an unrelated noun, in being the origin of the whole ‘bodhisattva’ concept. This is supported by other evidence, but the contribution from the commentaries certainly helps the case - it told us what people at least many centuries closer to the Buddha than us, were believing.


@cdpatton sorry I just added some text to my comment - I know you’re composing your reply, so, saying as you may not notice otherwise.

Right, but there are no actual texts that go back that far. The Chinese translations are probably better witnesses in some cases that today’s Pali canon (but not always), but they are still post 1st century BCE versions. What it boils down to is that we can’t look at a physical text from the 3rd century BCE and compare it to what we have today. Instead, we’re using conjecture and educated guesses as to what constitutes an early text and what doesn’t. So, for example, why does a text with mythological elements seem like a later text? It might actually be quite early, dating to when the myths about the Buddha’s time were first created. I mean, these are subjective things we’re basing our guessing on once we eliminate the texts that are clearly later. That’s all I’m saying.

This started because it was the text I remembered reading the story in. I know it’s verbatim from the Sarvastivada Vinaya because I checked it. I’m not saying that it’s the most important source that exists, but it’s not irrelevant because it’s found in a Mahayana text. That’s a misconception. We know alot about how the Sarvastivada Samyukta Agama, which was badly disordered in Chinese translation, was organized by reading Asanga’s description of it in the Yogacarabhumi. Mahayanists studied the Agamas and used them as scriptural sources.

1 Like

What has that got to do with anything?

Let’s examine your logic. I read " The oldest manuscripts from Sri Lanka and Southeast Asia date to the 15th Century, with few surviving examples." There’s also mention of " the oldest surviving Pali manuscript was discovered in Nepal, in the form of four folios from the Cullavagga dated to the 9th Century".

This means even from 15 century CE we don’t have the whole Pāli canon! So far as I understand your logic, we should therefore regard the late Mahayana sutras (which are all late anyway, but I mean the later of them, such as the Lotus Sutra) as far earlier sources than the Pāli suttas! Is this your view? I somehow doubt that it is.

Also, we must remember that even for the Chinese, knowing when a text was first translated into Chinese, is not the same as knowing what it actually said at the time, as that would be denying changes that occurred after that point, in the transmission of the texts up until the earliest physical documents we have of them, even though they may be earlier than the Pāli documents.

Let’s take a hypothetical example to explore the principle of your position, so far as I understand it. Suppose we somehow know when 2 texts were first written down. Suppose for both texts, this was 0CE. And, suppose one is clearly in written style - that is to say, it’s clear that it was composed as a written text (as is very clear from any Mahayana sutra I have ever read, for example). And suppose the other text, is clearly in oral style.

So what does this tell us about the two texts? Well, it tells us this:

  • The content of the first text was composed in exactly 0CE.
  • The content of the second text was composed 0CE or before.

The conclusion for each text is crucially not the same.

Now, supposing we have another 2 texts, in another language - let’s say Chinese. They are almost identical, but with some variation, to the above 2 texts. Both translations are known surely to have been first written/created in 700 CE. Furthermore, let’s say that we know for sure that the line of transmission (i.e. monastic lineage) of these Chinese has been separate from the 0CE oral text’s lineage, since 300BCE.

What does this tell us about our first 2 texts?

  • It gives no further information on the date of composition of the written-style text.
  • It proves that the date of composition of the oral-style text was 300BCE or earlier.

Do you agree with this logic? If not, please detail at which point you diverge.

Yes, I agree. My logic above shows why both the 0CE source and the Chinese 700CE source, in my hypothetical, both represent a 300BCE source. So I am in total agreement that in some cases, there are elements better preserved in the Chinese. This is why for our best understanding of Early Buddhism, we must rely on a comparative analysis of the various Early Buddhist texts. In particular, by analysing the parallel texts and passages. Furthermore, where we find no parallel, we must be skeptical as to whether we can trust the source! And, doing this actually allows us to see before the schools diverged, earlier than either source (which in my hypothetical would mean being able to reconstruct the composition to a form it had before 300BCE with greater accuracy than either text.

Specifically, since this gives us a vast wealth of material datable to pre-sectarian times, this is why this material is the material to be relied upon, in contrast to material that we cannot date to earlier than several centuries later than this!

Have you read Sujato and Brahmali’s text which I quoted from above? And perhaps Gombrich’s and Anālayo’s work? If so, I’m curious why to my ears you seem to have a rather different view on this than them. Did you find fault in their analyses? I’m not saying that there are no remaining ambiguities - but you seem to paint a picture that dismisses a wealth of evidence and as I said a really rather comfortable degree of clarity on the age of a wealth of material - unless I’m misunderstanding your position.

OK, that’s great. So we have that vinaya as the source.

Yes, having verified it as a vinaya source, that of course gives it the reliability of that vinaya in this case. However, would you agree as I asked above, that this puts this source therefore at earliest 160–170 A.D? If so, unfortunately, it would seem wise to discount the source for making any argument on buildings in the time of the Buddha.

I hope I don’t sound ungrateful! It’s wonderful to examine this topic, and potential sources :slight_smile: Thank you! I hope you do not mind how I treat potential sources. I try to be a little vigorous.

1 Like

I think we should probably ask the mods to split our discussion into a different thread as we’re going off-topic quite a bit.

That said, I think the problem is that I’m coming from an East Asian studies background and then looking at the Agamas and Nikayas. When a scholar studies the Pali as their primary background, there’s a different overall experience, and there are biases that arise from that. I don’t think it’s necessarily intentional, but sometimes it gets heated like it seems to be here.

So, when I study sutras in Chinese closely, I come to the conclusion that they changed quite a bit over time. I come to that conclusion objectively by comparing translations of the same text made repeatedly over the course of five or six centuries. With Pali studies, this isn’t an obvious thing because we don’t have an objective basis for comparing the versions of a Pali sutta as it existed between the same comparable period (2nd and 8th centuries CE). We’re just thinking, “Well, this seems like an early sutta. This doesn’t.” In Chinese, we can actually see examples of earlier and later Buddhist texts. The caveat, of course, is that they are all after 0 CE. Still, we can see the patterns.

From the outset, then, I’m looking for and seeing the same patterns of accretion and expansion that I see with say, the six translations of the Diamond Sutra or the Astasahasrika Sutra (and I have historical records telling me their dates). This is not something we can do with Indic texts. We can to some extent, but there’s more conjecture taking place.

The other thing Chinese studies makes clear is the heterogeneity of early Buddhist texts. The Pali canon is a single, uniform version, but in Chinese I can look at parallels from Sarvastivada, Dharmaguptaka, and other versions that are likely Mahasamghika, Mulasarvastivada, and Kasyapiya. There’s alot of variation between these canons.

However, some are closer to each other than others. So, for example, I can objectively say that the Sarvastivada and Theravada canons are very close to each other in canonical lineage because I can compare them and see how similar they are. I can also say objectively that the Ekottarika Agama is probably from a different canonical lineage than those other two because it’s quite different when I look at its parallels. I’m not guessing about it. I can see these things with actual texts.

The takeaway for me, then, is that there’s alot of conjecture and educated guessing happening in the Indic Buddhist field, and I can see that Buddhist texts evolved over time. I don’t see why that didn’t happen between the Buddha’s time until the time of the earliest texts that exist. To me, it’s a baked in assumption. They changed. But I don’t know how they changed because there are no texts to compare.


I think the topic of buildings in the early suttas definitely means we need to consider whether sources are early suttas or not. but I suppose we are giong into quite deep detail about this point!

Well, the Buddhist school to which I belong is Drukpa Kargyu, of Tibetan Buddhism. And my profession is teaching the music of a Japanese school of Buddhism (Fuke Shū). So, I do not view myself as bias in terms of not looking objectively in examining Buddhist history and the evolution of texts and doctrines. Many of my views in fact severely contradict the standard doctrine of my own school. But I have always preferred truth, so, I do not allow my school affiliation and my Drukpa Kargyu practice, to interfere with understanding the inauthenticity of the Mahayana texts (in terms of claimed authorship and history), the sectarianism in my own school of Buddhism, and so on. I try to just see everything as it is. And that’s not to say that I am all seeing! But, I would say that my bias, is the bias of wanting to see things as they are, rather than being bias towards any preconceived ideas. Hence I rely heavily on an evidence-based methodology.

Are we talking of repeated translations of Mahayana texts, or of Early Buddhist texts? The reason why I ask is that the evolutionary trajectories of these two categories may differ. In particular, the Mahayana texts were written compositions, and hence may have undergone different kinds of changes, and in some cases similar changes but to differing degrees, than the written compositions. So, I would be curious to know which you are referring to here.

Also, I would assume that some of the changes in earlier verses later versions in Chinese, would represent changes in the texts as they evolved in India, whereas some of the changes may represent a sinicization of the texts, or other changes from within China. I think you mentioned something on the vitakka vicāra topic which I found interesting, about the translation of these terms changing over time, for example, representing not a change in the source texts, but rather, a deepening of understanding within China of the Indic concepts within the texts.

Well, these things come to mind here.

  1. The element of change I mentioned above, of the Chinese texts changing due to improved understanding of the Indic, would not happen in the Pāli, since they maintained the Indic. That already makes them more stable.
  2. They were a single continuous tradition. In contrast, the Chinese had multiple sources, and didn’t even translate (so far as I know?) even the suttapitaka, of any school. It seems they took a bit from here and a bit from there. Some sources may have been more conservative, some less. And, having assembled this mishmash collection, there would likely be more tendency to change things, than if you had merely inherited a continuous tradition. For example, where monasteries are chanting the texts, how does this go for a single individual? Well, you are probably young when you become a monk, and you go into a room with a whole group of monks of various ages, who are all chanting in unison. And, anything you mispronounce, will be noticed. Of course if you entirely skip a word, let alone a phrase or paragraph, or put things in the wrong order, then in the activity of communal chanting, that is catastrophic.

So, through this circumstance, you eventually end up chanting just the way they all were when you arrived. Everyone is chanting in harmony, and it all works smoothly.

Eventually you become old. You’re still there chanting, in this same group, though many have now grown old and died, and many more have joined after you. But the group, has had continuous existence, and been continually communally chanting.

This circumstance makes it very hard to make changes. The majority of the group has to change at once! And not merely the majority - if even 5% of them don’t change, then it’s a huge drama! You need consensus really.

Now, very small changes can occur more easily, such as a gradual pronunciation shift. Also in smaller groups, sometimes similar passages can be switched over, or doubled etc. These things do occur. However, the situation is very different than for a written culture. And this applies even when that oral culture also writes down their teachings eventually. For if they continue the oral culture in parallel, having a strong chanting and memorisation tradition, this acts as a strong conservative force.

But if the text is only written, then for one text to change, all it takes is one scribe.

  1. Still regaring when you said “objective basis for comparing the versions of a Pali sutta as it existed between the same comparable period” - I’ve studied some Pāli suttas together with their Madhyāma Āgama parallels, and been blown away at how little change there was. I see music transmission in Japan shifting more in 100 years or even in a single generation, than in some of those parallels, which have been separate for… 2,300 years?

So, while we cannot know clearly how much the Pāli texts of today have changed between “2nd and 8th centuries CE”, what we can see, is how much difference there is between two texts that have been apart for let’s say 2,300 years. And where that change is almost none (in many cases), or however much that is, then we can assume that at least not more than that much change has occurred in the Pāli in the last 2,300 years. In some cases we might assume that the two traditions share the changes 50/50. Though it seems in some cases we can even identify which one has changed more, i.e. identify aspects within a sutta where one or other of the traditions has an earlier version. And this might differ even from portion to portion of a texts. But anyway, my point is that in cases where there is almost no difference, which are many, then this tells us that that Pāli underwent almost no change through the entire period that they were separate. Which of course includes your period of between 2nd and 8th centuries CE.

Ah, this answers my question. So yes, this is about change in post-Early Buddhism. Whereas really I am interested in Early Buddhism. I would totally expect more change in Mahayana texts. Some were even composed and arranged in China! The textual tradition is very different than the Pāli. Though the Theravadins did to some degree accumulate some new texts, of course.

Yes, well it’s a pity that there is no Indic canon translated into Chinese. To have complete canons to compare would tell us much more about how different or not they were from each other I think. But yes it’s great that we have parts of them that we can compare.

Another thing to consider I think, is the Indian schools that they received the Āgamas and individual texts from. For example, if they received them from a school which had fully embraced Mahayana, then the status of the Āgamas in that school would have been relatively different from a school which had not. If they were more focused on Mahayana teachings, they may not have even been communally reciting the Āgamas! And in such a case, their Āgams would therefore have been more susceptible to change. Also, the further the Mahayana doctrines diverged from the Buddha’s teachings, and the more their teachings came to contradict his, the more motivation there would have been to change them. So, the reliability of the Āgamas the Chinese were translating, would depend on their source for them, in this way.

This may go to explain why not all of the Chinese Āgamas are as reliable as each other. My sense is that the Madhyāma Āgama is the most reliable? And some have many bodhisattvayana influences, which are very clearly late. This can be explained potentially with processes I mentioned above.

I could propose a hypothesis:
Suppose school A splits in two, producing school B and C. So now there is only B and C. Then C splits into D and E. So there is only B D and E. Now E splits into F and G. So we have B D F and G. But suppose B splits into J and K. And K splits into L and M. So now we have L M D and E.

Now suppose E undergoes radical change, for example, they take on the Mahayana sutras, and make various revisions of their collection of earlier teachings, the Āgamas.

Now, Suppose we have the texts for L, D, and E. What we may expect to see, is that there are some minor differences between the Āgamas/Pitakas of L and D. But far more difference between them, and E.

Now if we are to then make the assumption that L and D therefore share canonical lineage, and the their lineage was further apart in time than they are from E, we would be wrong. Because, in this hypothetical, D and E actually have a far closer canonical lineage! They’re only one step apart! Whereas they diverged from L further back, and not even from L, but from L’s progenitor.

I’m not saying that that’s actually what happened. But if it is hypothetically possible (which I believe the above establishes), then unless we have evidence to discount it, I think it should be considered.

I also analyse the evolution of the music I teach. Partly through having studied all of the old extant oral traditions from the most reliable sources, but also through analysis of music notation dating back to the 1700’s, from the various traditions. In this context, where two schools which have been separate for along period, play very similar to each other, but different from other lineages, this shows that the form which they share in common is generally due to these being common to their common source. Which is to say, through commonality, we can understand their common source, in many cases even when the common source is unavailable! (This is also done in language research, and maybe other fields).

To give an example which I found rather interesting - I teach a number of repertoires within this field. One if associated with the Fudaiji temple. Through my analysis of the earliest notation a we have from that tradition, from the Meiji Period, I created a kind of ‘critical edition’, producing a way of laying the pieces by cross-analysing the different sources, to produce a style which I deemed to be a fair representation of the generation before these sources.

Then recently, I decided to do a comprehensive analysis of one piece of music, that exists in this Fudaiji tradition, but also several other traditions. I used… many thousands of data points to create 34 dimensional data to show in what position in 34 dimensional space each author (for notation dating back as far as the 1700’s) or performer (I transcribed recordings dating back to the early 20th century) was in relative to all the others. We can’t easily process 34 dimensional data, so I made a 2D projection of this data, which allows us to visually see the trajectory of evolution of the piece over time, and how far apart the different lineages have gone from each other over time. This can also reveal who learned from whom or from what lineage, when that data is unknown. And can reveal when and who makes significant changes. And track the rate of change in the different branches.

And, while I did not analyse my own performance (as that seemed inappropriate), I did put my student’s performance into the analysis. And, sure enough, he was plotted earlier than all of the notation I had from the tradition when making that arrangement, and earlier than any of my teachers. Which, having carried out this analysis long after I made that arrangement, and now with many more sources to consult, confirms that my method was valid, and that indeed one can look into the past with a good degree of accuracy by such comparative analysis. And I feel similarly with the Early Buddhists texts - though I have not done this kind of analysis with it! It’s rather time consuming!

Well in the above example of music I gave, I can see the evolution over time, but I can also understand that process. In rather fine detail! Admittedly it has taken many years and very deep analysis, but, my understanding relies very little on guesswork. And I always keep conscious of what is ambiguous and what is known. Even when I teach the music, I take care to pass on that differentiation.

Here you presumably mean the physical texts. Well, understanding a phenomenon and understanding why a phenomenon is so, are two different things. Regarding understanding it - for me, studying an early sutta in multiple versions that have been separate since not long after the Buddha’s death, but are still almost identical, gives me the understanding that in such cases, hardly anything changed in those texts over the last 2,300 years.

This has further implications. If they changed so little even when some of them had been transferred from school to school and undergone translation and textual transmission devoid of an oral component, then this brings the strong possibility that this tradition was not merely highly conservative for the last 2,300 years, but also highly likely to have been similarly conservative since before the divergence. This brings the strong possibility that at least a significant proportion of the material that they shared 2,300 years ago, was already established during the very long teaching career of the Buddha.

This is backed up by the fact that the Buddha taught in a very long and well established oral culture, in which preserving an oral canon of teachings was the norm. Furthermore, I think we can assume that the Buddha would have been aware shortly before his time, one of the two branches of Jainism had lost (mostly or completely?) their own canon, which was quite catastrophic. This would give him even more reason to establish a healthy body of oral teachings.

And this is not merely about the importance of the teachings carrying on after his dead, which would be a very natural concern. It’s also about teaching when he was alive. So far as the evidence shows, he had many disciples, who lived in wide distribution. He would travel from place to place but there would have been Buddhist communities that may not get to see him for a very long time! And his disciples would take on their own disciples. So he was not directly teaching everyone. There’s even a lovely story of the Buddha crashing in man’s barn, and there’s a Buddhist monk there already. They chat away, but this guy has never met the Buddha before, even though he’s a Buddhist monk! The Buddha doesn’t reveal his identity until the next day, so far as I remember.

So, the oral body of teachings would be passed around, kind of like a syllabus in a way! Or we can anyway say ‘reference teaching material’.

So it’s not surprising to find this being conservatively preserved in the continuous oral traditions as they eventually split apart from each other, and, eventually, wrote them down.

As for how they change - I covered some aspects above and I suppose we can all think of some more. I think it’s very natural evolution. But I note that the ancient oral tradition was, so far as the evidence shows, extremely conservative. As has also been established regarding the Vedic oral tradition - I cannot remember which scholar wrote on that.

It seemed like an opportunity to talk about it. Personally, I think our good moderator @ficus solved the original question expertly with archeology.

That sounds amazing. I have alot of respect for the creative arts of all kinds. Though I spend time these days being a (somewhat pedantic) translator, I’m actually a creative writer. So, I enjoy the mythology in these texts more than the dry philosophy, to be honest. When they start telling stories, I pull up a chair and listen.

When I compare Pali texts to their parallels (and sometimes it’s the reverse situation), I see the same patterns of expansion that I see with other texts. Granted, it’s usually two data points, not a half dozen. But I can quickly see which one has grown more than the other. The Diamond Sutra is just a convenient example.

When I look at repeated translations of Agamas in Chinese, it’s more complicated because they often aren’t from the same canon. So, for example, there’s four or five translations of the Parinirvana Sutra in Chinese, but only a couple match each other closely.

There are some translation issues, but by and large these were changes in the texts coming to China from Central Asia or India. This idea that (for example) Kumarajiva was a lazy translator who abbreviated and paraphrased is not a new one, and it’s been disproven by discoveries of early Sanskrit versions of the Diamond Sutra. The sutra really was simpler in 400 CE. Most scholars who worked with late Sanskrit texts assumed the same thing: The Chinese canon is full of bad translations. But they actually are pretty good translations, they just weren’t translating the same Sanskrit texts that Indic scholars were studying.

The Diamond Sutra is actually a good example of the growth of Buddhist texts over time for someone who doesn’t read Chinese. The Kumarajiva and Tibetan versions are available in English. These are the earliest and latest versions of the sutra. If you set them side by side, you’ll see how it evolved from about 400 CE to around 800 CE. The Tibetan version is very close to the last translation by Xuanzang.

Yes, this is probably the reason there’s an amazing correspondence between the Majjhima Nikaya and the Madhyama Agama. They come from closely related canonical lineages, meaning they split off from the same parent canon. Comparing only those two sets of parallels creates a false sense of similarity.

That said, though, it’s getting difficult to accept that there was a single, original canon. I’m basically in agreement with scholars like Salomon don’t think there was one. That’s because when we look at Pali, Chinese, Sanskrit, and Gandhari, they are really different. It seems more likely that Buddhists didn’t form canons until after regional evolutions had already changed the sutras quite a bit. The core teachings likely go back to the original disciples. Without older texts to look at, though, we’re guessing about these things.

I wanted to take time to give a more concrete example of why I think it’s a problem that we don’t have texts from earlier eras. It’s a bit academic, but it’s a good example of the problems I see in parallels.

So, one criteria for what’s an early Buddhist teaching might be whether it’s found universally in Agamas and Nikayas (and even Mahayana texts). So, we can quickly pull in a large core of teachings without worrying about the wording of the sutras. The four noble truths, eightfold path, five aggregates, six ayatanas, etc. Most of these basic teachings are found in all Buddhist texts in all languages, and there’s no variation beyond ambiguous readings here and there. They’re sometimes defined differently, but let’s set that aside.

There are universal teachings like this that vary from one source to the next. An example is the 10 powers of the Buddha. It’s found in all Buddhist canons from the Nikayas to late Mahayana texts. They don’t agree about all 10 powers, but they do agree about five of them.

I posted about this in an essay a few months back when I translated some of the parallels: The Tathagata’s Ten Powers (or Five).

It’s a puzzle that can’t really be solved easily without some evidence about what the EBTs actually looked like earlier. What appears to have happened is everyone adopted an expanded list of ten powers from an earlier list of five. I say that because almost all of the lists in the Agamas and Nikayas agree on the first two and last three powers (with a couple minor exceptions). Then they insert their own take on powers 3-7.

So, comparing the existing parallels doesn’t get us any closer to understanding what happened. All we can do is theorize.

There’s a sutra in the Samyukta Agama that could be a clue. It says arhats have the five powers from the 37 factors, and the Buddha has the 10 powers. What I wonder is: Did this sutra originally have two sets of five powers? It would seem a better fit for an oral tradition sutra; they liked parallelisms because they’re easier to memorize.

But today, this sutra has the ten powers like the other sutras. So, I can theorize about it. Originally it was five powers of the Buddha, not ten. One day, someone thought, “The Buddha is supreme and unique. Surely, he had more than five powers. What were his other powers?” Eventually, a sutra is being recited with ten powers instead of five. Word gets around from that region to other samghas, and they say, “Wait, our version of that sutra only has five powers. That must not be right; theirs has ten powers! Surely the Buddha did have more than five powers. We should adopt these ten powers.” Now, their sutras have ten powers, but they chose a different set of five more powers to add to the original list.

And so forth until at some point, the five powers version is erased from the Buddhist canon everywhere. And that’s the canon we have, after it had already changed.

It’s a nice story, but it’s just me daydreaming. It could be something else that happened. Maybe right after the Buddha’s parinirvana, the disciple who’s job it was to remember the 10 powers of the Buddha forgot five of them? Then everyone filled them in with different sets five powers to fix it. (A stretch, I know, but it’s another story we could tell ourselves.)

Without actual evidence, it’s a mystery. It’s a similar problem when it comes to parallels that are different but we can’t rule one out as later. Which one goes back to the Buddha’s time? It’s basically a guessing game.

1 Like

It seems arahants also have ten powers . The five powers of arahant are differents from arahants ten powers . Then the Tathagata ten powers appears differents from arahants ten powers .


AN 10.90

“Sir, a mendicant who has ended the defilements has ten powers that qualify them to claim: ‘My defilements have ended.’ What ten?

Firstly, a mendicant with defilements ended has clearly seen with right wisdom all conditions as truly impermanent.

Furthermore, a mendicant with defilements ended has clearly seen with right wisdom that sensual pleasures are truly like a pit of glowing coals.

Furthermore, the mind of a mendicant with defilements ended slants, slopes, and inclines to seclusion. They’re withdrawn, loving renunciation, and they’ve totally done with defiling influences.

Furthermore, a mendicant with defilements ended has well developed the four kinds of mindfulness meditation.

Furthermore, a mendicant with defilements ended has well developed the four right efforts …
the four bases of psychic power …
the five faculties …
the five powers …
the seven awakening factors …
the noble eightfold path.

十力 Ten Powers

SA 684


Furthermore , there are five learner powers and Tathagata ten powers .
And which are the learner powers ?
Namely confidence , effort , mindfulness , concentration and wisdom . And what are the Tathagata ten powers ?

1 Like

Wow … its a long and weary discussion. But, I grew an interest on what @cdpatton have notice

Is this means that the Dhamma we read or hear today is not 100% what the Buddha said? It looks like the Dhamma based on individual perception of those who interpret. Correct me if I am wrong.

Homage to the Precius Kagyu!

1 Like

Thanks for your thoughts here. Certainly some important things to consider. Out of interest the changes you have seen over time in the Chinese texts, are they major or minor? Is it a slight elaboration, or a drastic change in doctrine? I would also be interested to hear if you think we can safely say any teaching can with confidence be said to come from the Buddha?

Sure, it’s a great topic!

That’s interesting. @ficus’s input was some quotes from Wikipedia, based on scholarship from the 1960’s. I quoted from the source of the source that that was based on, and gave the much more recent 1996 counter-argument from Schopen, who seems to have written the most authoritative text on archeological research of Buddhist history. Perhaps you would like to explain what faults you see in Schopen’s analysis, over on that thread?

Thanks. And, I think being pedantic can be very useful for the sake of accuracy. When used appropriately, it’s a great tool!

That’s fine. However, when it comes to the discussion of history, then I think it’s crucially important to separate fact from fantasy, however much we may enjoy the fantasy. Otherwise we should maybe concern ourselves with mythology, rather than history.

In my own musical tradition for example, there are still people today who believe the Fuke sect came from China, from Fuke’s disciples. Despite this being thoroughly disproven in… I think it was the early 20th century when a Japanese man who believed in these myths, actually went in search of evidence, expecting he would find it. It turned out he did the opposite, to his surprise. So then we have the choice - emphatically reject the evidence, and cling to blind faith in pleasant fantasies; or embrace reality and see things ‘as they are’. As a Buddhist focused on freeing myself from ignorance, cognitive dissonance, and mental distortion, I am biased towards the latter.

When you say “two data points”, do you mean the Pāli and Chinese? Or do you mean two translations of the Chinese Āgama text, that were translated at different points in history?

So it would seem you’re talking about a single Chinese data point. In which case, you’re unable to see evolutionary change in the Chinese, just as you are unable to see it in the Pāli - so your point on this seems to only be about the Mahayana texts, which is perhaps therefore irrelevant to this topic. That is to say, you were giving this as a claim that you can see evolutionary change just by looking at the Chinese - but that’s not the case, when we’re talking about EBTs, it seems.

Not an EBT if you mean the Mahayana sutra! Or are you referring to an EBT?

As I stated in my previous comment, this does not inform us about the EBTs. In particular, because it’s an entirely different type of ‘text’, the Mahayana texts being written compositions, which I went into detail about.

By stopping at the very first sentence of that 4 paragraph hypothesis, you are completely ignoring the entire hypothesis. What is your reason for doing that? You are merely repeating your previous position and not even challenging or testing my hypothesis in any way.

I could take a similar approach to language evolution. I could look at French, Italian, Spanish, Portuguese, and Romanian, and say “They are different, so they could not possibly share a common origin!” (They all derive from Latin).

This statement rests on the assumption that “we don’t have texts from earlier eras”. Earlier than when? And by text, do you mean the contexts of a text? Or do you mean a physical document? If you mean physical document, then environmental conditions are more than enough to explain why we don’t have millenia old suttapitakas from India. It’s also enough to explain why the oldest documents we tend to have come from dry desert regions, such as Dunhuang.

I read somewhere about a certain period where they became obsessed with the number 10. I wonder if these changes might date to that period.

You’ve just contradicted yourself. You specifically elucidated what you believe appears to have happened, back then in the past. Which means you believe you are getting “closer to understanding what happened”, specifically by comparing the parallels.

Again you are showing how your study of parallels is, in your belief as you express it, getting you closer to understanding what happened.

Luckily, there is a whole community of Early Buddhist scholars, and there have been for generations. So, just like in science, where someone can come up with a cooky theory, this is balanced by the community, who review each others work, and provide counterarguments, refutations, replicate the test conditions and share their findings, and so on. Just so, ideas in the field of Early Buddhism are reviewed by peers, tested, proven or disproven with more current data or the continued uncovering of new data and analyses, and so on. And I see the evidence-based approach abound in this field, delightfully.

My answer: no. I suggest you study Bhikkhu Anālayo’s work. He had many articles and even books, studying suttas with their parallels. You may be surprised at how close they are to each other! I also recommend this book by Sujato and Brahmali, which you can freely download here:

I also recommend this wonderful version of the Dhammapada by Ānandajoti, ‘A Comparative Edition of the Dhammapada’ which uses more than 6 parallel sources:

Very nice! That happens to be a picture of my master.

1 Like

I thought you might have known Amtin-la! What an inspiration he was!


Yes, indeed! Becoming his disciple was quite life changing for me. Did you train under him also?

I spent several months practicing in Tashi Jong. I stayed next door to Popa Rinpoches’s house, where he lived. There wasn’t a translator available for most of that time, though. So I didn’t really receive many teachings from him, but the way he lived was enough of a teaching for me.

1 Like

Cool. Did Popa never translate for you? He was busy with his business (which relates to the vinaya discussion on the other thread!) but he did also translate for people…