Short essay on Dharmacakra sūtras in Chinese

I wrote this short essay about six years ago, after translating SA 379 and reading through the other versions of the Dharmacakra Sūtra in Chinese. I had more details in the original, but today I cut it down to just the essentials. Maybe at some point I will do a proper rewrite and expansion, but I'm not sure it would be so useful.

In this essay, the Chinese editions of the Dharmacakra Pravartana Sūtra will be outlined. Various differences will be hilighted, as well as their implications on the understanding of basic doctrines in Buddhism, and their place within several of the Indian monastic sects.

I. Chinese editions of the Dharmacakra Pravartana Sūtra

The Chinese Buddhist canon as preserved in the Taishō Tripiṭaka includes a number of different translations of the Dharmacakra Pravartana Sūtra, each made separately over the course of approximately five to six hundred years. Each of these translations is dateable within twenty-five years, and every translator is identifiable. These versions come in three different forms: translations made specifically for the sūtra, translations in the āgamas, and translations incorporated into the vinayas. There are almost certainly other versions in the Chinese canon in addition to the following, but these are the Dharmacakra sūtras that have been located by the author:

  1. An Shigao, 148-170 CE, in a separate Dharmacakra Pravartana Sūtra.
    Taishō volume 2, number 109 《佛說轉法輪經》
  2. Dharmanandi and Saṃghadeva, 397 CE, in the Ekottarika Āgama.
    Taishō volume 2, number 125, sūtra 24.5. 《增壹阿含經》
  3. Buddhayaśas and Fonian, 405 or 408 CE, in the Dharmaguptaka Vinaya.
    Taishō volume 22, number 1428, fascicle 32. 《四分律》
  4. Vimalākṣa, 413-? CE, in the Sarvāstivāda Vinaya.
    Taishō volume 23, number 1435, fascicle 60. 《十誦律》
  5. Buddhajīva, 423-424 CE, in the Mahīśāsaka Vinaya
    Taishō volume 22, number 1421, fascicle 15. 《彌沙塞部和醯五分律》
  6. Guṇabhadra, 435-443 CE, in the extant Saṃyukta Āgama
    Taishō volume 22, number 99, sūtra 379. 《雜阿含經》
  7. Yijing, 710 CE, in a separate Dharmacakra Pravartana Sūtra
    Taishō volume 2, number 110. 《佛說三轉法輪經》

Of the various translations, the version in the Saṃyukta Āgama looms large over the rest, since this āgama can be considered a parallel collection to the Theravādin Saṃyutta Nikāya, and the Dharmacakra Pravartana Sūtra has traditionally been located in this collection. There is also a version of the text that has been incorporated into the Ekottarika Āgama as part of a larger sūtra.

Scholars have long considered the possibilities of comparing the Pali and Chinese language collections, in hope of learning more about pre-sectarian Buddhism, yet the vast majority of the āgama teachings have not been translated into English. This is presumably due to the size and difficulty of such a task. These, along with all other large Chinese collections, remain in waiting for a later era.

The versions of the Dharmacakra Pravartana Sūtra contained in the vinayas are not complete sūtras with the proper framing, but instead are the main contents, which have been incorporated into a larger narrative. These are reasonably complete and valuable as well because they can be very clearly identified with one particular school. By including these, we can have examples of the Sarvāstivāda, Dharmaguptaka, and Mahīśāsaka versions from the fifth century CE.

II. A translation from the Saṃyukta Āgama

(Removed. Now on SuttaCentral.)

What we have in this Saṃyukta Āgama version is not simply a slight variation, but almost a completely different sūtra. The three turnings and twelve motions of the Dharma Wheel are mentioned in the Pali as well, yet they seem to correspond more with the contents of this Chinese version. In this Dharmacakra teaching from the Saṃyukta Āgama, we have three clearly-divided sections of the discourse, with principles for the Four Noble Truths in each of the three sections.

III. A survey of Dharmacakra doctrines

To clarify the matter further, ideally each of the Chinese editions of the Dharmacakra would be translated into English and provided in example, along with detailed analysis of each section of the text. However, as a simpler first step we may begin by simply providing an overview of the contents, which should suffice and serve others well who simply wish to understand a basic outline of the differences.

From this overview, others can then approach these versions of the Dharmacakra with a general understanding of their respective doctrines. The criteria used for presenting the various differences centers around the essential principles taught in the Dharmacakra Pravartana Sūtra, as they are commonly accepted in the Pali version. The existence of five main doctrines that will be checked for in each of the extant sūtras, are the following:

  1. Two Extremes: The pursuit of sensual pleasures, and the unhelpful asceticism of self-mortification.
  2. Middle Way: The way of avoiding the two extremes, that has been realized by the Buddha.
  3. Four Noble Truths: Mention of the names of the Four Noble Truths.
  4. Definitions: Explanations of the Four Noble Truths (suffering as birth, aging, disease, death, etc.).
  5. Noble Eightfold Path: Mention of the Noble Eightfold Path as the way to achieve realization.

The contents of the extant Dharmacakra texts include the following doctrines (this was originally a table but has been converted into a list):

  1. An Shigao: Two Extremes, Middle Way, Four Noble Truths, Definitions, Eightfold Noble Path
  2. Mahīśāsaka Vinaya: Two Extremes, Middle Way, Four Noble Truths, Definitions, Eightfold Noble Path
  3. Dharmaguptaka Vinaya: Two Extremes, Middle Way, Four Noble Truths, Eightfold Noble Path
  4. Ekottarika Āgama: Four Noble Truths, Definitions, Eightfold Noble Path
  5. Saṃyukta Āgama: Four Noble Truths
  6. Sarvāstivāda Vinaya: Four Noble Truths
  7. Yijing: Four Noble Truths

An examination of these various texts shows that the only constant doctrine found through each of them is the Four Noble Truths. The versions fall roughly into two camps: those that contain the Four Noble Truths to be contemplated, and those that contain many extra doctrines and explanations. Of the two main groupings, the versions that contain the additional doctrines vary more widely in their contents.

An Shigao’s translation of the text is remarkably similar to that of the Pali, showing that in the second century CE, a relatively common formula linking the five doctrines had already been compiled. This version also has several other features, such as the existence of 1000 bhikṣus and a multitude of devas. After hearing the Buddha speak the Dharmacakra discourse, all 1000 bhikṣus become arhats. This is the only translation that includes any mention of 1000 bhikṣus and their attainment of arhatship. The Mahīśāsaka version is very similar to that of the An Shigao translation, which shows that An Shigao would have been in doctrinally close proximity to the Mahīśāsaka school, which is likely given that An Shigao was considered Parthian. However, the Mahīśāsaka version has no mention of 1000 bhikṣus or of arhatship. It seems likely that a detail such as this would have been more likely to have been added in An Shigao’s independent edition of the text, simply due to the nature of the transmission of short, independent texts.

The Dharmaguptaka version contains most of the five doctrines, but lacks any definitions for the Four Noble Truths. Instead, it contains explanations in the vein of “This is the Noble Truth of Suffering: a teaching which has never been heard before, and which is to be contemplated.” It also contains instructions for contemplating the Noble Truths to sever knowledge or awareness of other truths such as the Noble Truth of Suffering. The Ekottarika Āgama version has content similar to that of the Mahīśāsaka Vinaya, but without any mention of the Two Extremes or the Middle Path.

The versions of the Dharmacakra Pravartana Sūtra in the Saṃyukta Āgama, the Sarvāstivāda Vinaya, and the Yijing translation, are largely consistent in contents. Each starts with the simple statements shared with those of the Dharmaguptaka school, and then enter into discussion of contemplation of each of the truths to achieve realization. The text is generally cryptic and repetitive, and these repetitions of a common refrain break the text into several sections. When the occurrences of the truths to be contemplated are counted, they number twelve, and may be divided very cleanly into three phases of the teaching. This corresponds well with the following text that mentions three turnings and twelve practices or aspects of the Dharma Wheel. In these editions there is no mention of Two Extremes, a Middle Path, Four Noble Truth definitions, or a Noble Eightfold Path.

IV. Bodhi and Nirvāṇa

Although we have found that the Theravādin version of the Dharmacakra contains many elements in common with the An Shigao version, we do find differences unique to the Theravāda school. Most significantly, we find that the common refrain is:

... produces vision, knowledge, tranquility, understanding, Bodhi, and Nibbāna.

Few would second-guess this sequence, since it seems to conform to the basic understanding we have of Buddhism. However, it is notable that vision, knowledge, understanding, and Bodhi all reflect mental realization; however, tranquility and Nirvāṇa are more often considered states. Suspicions of these two items are found to be correct when we examine all seven Chinese versions, where these two are not present at all. Instead we find four levels of realization:

... produces vision, knowledge, understanding, and Bodhi.

In all extant Chinese texts from the other early Buddhist schools, the Buddha is teaching the five bhikṣus that the final result of the Four Noble Truths is realization of Bodhi. The overwhelming evidence from the Chinese editions shows that they were being taught the way to Bodhi.

This is an important matter that impacts our basic understanding of Buddhism, and on assumptions made about the basic goals of early Buddhist practice. In these statements we can see some divergence between the Theravāda school and the schools prevalent in India. Why was Nirvāṇa added as a separate item?

V. Āgama assignment implications

It has been conventionally thought that the extant Saṃyukta Āgama belongs to the Sarvāstivāda school. Since the versions of the Dharmacakra Pravartana Sūtra match between the Saṃyukta Āgama and the Sarvāstivāda Vinaya, the evidence seems to support a very close relationship between the two, and most likely a common origin in the same school. Because Yijing translated numerous texts of the Mūlasarvāstivāda school, we can tentatively place his translation in with the Mūlasarvāstivāda or Sarvāstivāda schools. If the text was indeed Mūlasarvāstivādin in origin, it is evidence that this school did not add anything significant in this fundamental discourse as it had been preserved by the Sarvāstivāda school at least 300 years earlier.

The Ekottarika Āgama has in the past often been attributed to the Sarvāstivāda school. However, in more recent years an origin in the Mahāsāṃghika branch has become a more favorable explanation. The contents of the Dharmacakra doctrine that this āgama includes are in fact different from that of the Sarvāstivāda school, and also markedly different from those of the Mahīśāsaka and the Dharmaguptaka. Unfortunately we cannot clearly assign it to any school yet, and the extant Mahāsāṃghika Vinaya does not appear to contain any Dharmacakra Pravartana Sūtra, unfortunately. We can therefore tentatively state that the text may yet have a Mahāsāṃghika origin, but almost certainly does not belong to the Sarvāstivāda school.

VI. Issues of Indian composition

In other cultures and literary traditions, there are varying ways that texts are modified. However, in Buddhist works authored in India, there tend to be two main patterns that are easily observable when comparing various Chinese translations with extant Sanskrit manuscripts. One is the tendency to add more sections of text and explanations. This means that Buddhist works tend to grow over time, sometimes to enormous proportions. The other is to change relatively little in the process of adding the text. Some transitions therefore appear quite jarring, and others seem to make little sense. However, it does appear that they were added for various reasons, often in response to questions or disputes.

What is somewhat rare in Buddhist works, is for sections of text to simply be removed. Therefore, we can reasonably rely on understanding the differences in Buddhist texts as being the product of additions, combinations, and minor modifications.

There is no real reason why the Dharmacakra should be exempt from the general patterns of Indian Buddhist composition in being substantially cumulative. In light of differences between the various recensions, there is little choice but to assume that the common material is pre-sectarian, and that differences largely reflect later additions.

VII. Conclusions

As for the Dharmacakra Pravartana Sūtra itself, it does seem to the author that the version as it exists in the Sarvāstivāda recensions is probably closest to a hypothetical "earliest" version. Of the variations that exist, it is the most doctrinally independent, the simplest, and the most internally consistent.

VIII. A few notes from 2016

I know that Bhikkhu Analayo has written at length about these texts, but I had not read his articles when I first wrote this. After reading the articles, though, I did not find some of his ideas to be convincing with regard to the shorter and simpler versions. It seems that whenever genuine differences came up, he would say that while the material was missing from the simpler versions, it could be found in other texts (usually later texts). Therefore, the smaller (older) versions were likely just "extracts" from the larger (later) versions.

I cannot agree with this line of reasoning. If we were to take that route, we would simply assume that the longest texts in the Buddhist canon were the earliest, and the shorter versions were simply extracts from the longer "originals." Yet we know that almost the opposite was the case. The Saṃyukta is now widely recognized as the earliest collection. Without good reason, we should take not assume that the smaller texts therein are simply "extracts" of the longer and more complex texts of later collections.

I am probably fixating too much on this one point, but at least this was the big thing that really stuck out in my mind. Regardless, I hope you found at least the above short summary of doctrines between the different versions to be interesting. :smile_cat:


Thanks for this. A couple of questions, if you don’t mind.

So the Mahīśāsaka school was strong in Parthia? Do you have any references for this off-hand?


Thanks so much. This is a valuable contribution, and I look forward to more!

We do have it in Sanskrit, from the Mahasanghika-lokuttaravada:

It would be interesting to compare this with the EA version. It’s actually very similar to the Pali text. For example, compare the refrain:

cakṣukaraṇīyā jñānakaraṇīyā upasamasaṃvartanikā nirvidāye virāgāye nirodhāye śrāmaṇyāye saṃbodhāye nirvāṇāye saṃvartati

cakkhukaraṇī ñāṇakaraṇī upasamāya abhiññāya sambodhāya nibbānāya saṃvattati

So not only does it include the two terms you identified as being absent from all Chinese versions, it adds some more, (nirvidāye virāgāye nirodhāye śrāmaṇyāye) while omitting abhiññāya. My initial guess would be that here, as elsewhere in this text, it is based on a version very similar to the Pali, with minor elaborations, mostly additions of stock phrases and the like.

It is of course curious as to why the Mahasanghika version should have more similarities to the Pali than other Sthavira texts. While we should take agreement between versions as significant, agreement between Sthavira and Mahasanghika texts is always of special interest.

This is mostly true, but far from universal. In fact there are numerous cases of texts being extracted or divided up. One example is the Mahasudassana Sutta. This is present in a vestigal form in the Pali Mahaparinibbana, but the bulk of the text has been made into a separate sutta. In the Sanskrit, however, it is retained within the Mahaparinirvana itself. That version is substantially shorter than the Pali, so it seems that it was removed as it grew to large for its container, so to speak.

That’s not to dispute your basic point, which is that texts mostly tend to grow. But if this is to be established in any particular case, we should consider whether there are any other factors that support one thesis or the other.

In the case of the Dhammacakka, I think we need to consider it in context. From the earliest times it would have been embedded in a narrative context, however slight, representing the actual event when the Buddha taught. In different versions there is a fluidity between the content found in the text, and that separated out into subsequent texts.

Even the Anattalakkhana Sutta should be considered as part of this; in fact, that text is best understood as an expansion and elaboration of the line pañcupādānakkhandhā dukkhā in the Dhammacakka.

Not only that, but the Vinaya versions tell us that the Buddha was giving extra explanations to the monks at the time, although these are not recorded. In the Dharmaguptaka Vinaya, there are various Vinaya instructions given here.

Given all this, it may be fruitful to look, not just specifically at “the” Dhammacakka, but at the textual environs, considering the evolution and relation between this constellation of texts as a whole.


Nothing about Parthia specifically, I’m afraid, but the Mahīśāsaka were known to be in the Northwest, like in Gandhara and some other areas. What I meant was just that since we know the Mahīśāsaka could be found in the Northwest, a Parthian monk might possibly be Mahīśāsaka.

On this point, apparently some of An Shigao’s works are more related to the Sarvāstivāda, but I think the whole matter is still a bit unclear because the Sarvāstivādin versions of the Dharmacakra are completely different from An Shigao’s version.

If I remember correctly, there is also some evidence for Mahīśāsaka in southern India, and in Sri Lanka. At least the Mahīśāsaka Vinaya was found at Abhayagiri Vihara, so there must have been some significant contact.

After looking at the Dharmacakra versions, and the strong similarities between An Shigao / Mahīśāsaka / Theravāda versions, I even played around with the idea that maybe the Theravāda were once Mahīśāsaka – but I think that’s highly unlikely. It seems impossible that nobody would have realized such a thing since both vinayas are available and the Mahīśāsaka had an ancestor that they venerated. What may be more likely is just that they shared some common textual ancestry and were geographically close in western India at some point before the Theravāda went to Sri Lanka.

Also, although only tangentially related to the Mahīśāsaka, I am reminded of one interesting point: according to a few sources in Chinese, the Mahīśāsaka wore blue robes. In the later Nikāyasaṅgraha, there are some strange passages about monks with blue robes who distort the Triple Gem into prostitutes, liquor, and love. It’s hard to know what to make of such statements, but at least two possibilities pop out in my mind. [1] The first is that maybe blue-robed Mahīśāsakas in Sri Lanka were demonized by their opponents at the Mahavihara, and there was still some distant memory of “bad monks” in blue robes. [2] The second is that maybe they were Mahīśāsakas who had turned to tantric practices and who actually were engaged in all sorts of shocking taboos as part of these practices.


Thank you, and I agree with you for the most part. It was probably making too much of one text. If I were to look at the situation again, I would probably start with Bhikkhu Analayo’s papers, then look at Choong Mun-keat’s book to examine the doctrines and start comparing the saṃyuktas. Then finally document the differences in more detail, including the original Chinese passages.

I am not sure it is really such a big matter, though. Rather than considering one individual text across different versions, it makes more sense to me to look at the saṃyuktas and compare the doctrines that way, since the collection consists of groups of topics, which are then taught in various ways, usually in very small pieces.

In any case, though, the basic idea typically found in (pretty much) all versions is that the Buddha first taught five monks the Four Noble Truths.

1 Like

I think this is very likely. Of all the theories on the mahisasaka the one that makes sense to me is that they’re based in Mahissati, which would place them right where Mahinda stayed before going to Sri Lanka. It would also place them on the Dakkhinapatha, which was a trade route to Andhra (where the Mahavihara had a branch) and thence to Sri Lanka.

Of course they would likely have branches elsewhere, and texts could be carried further still. If Chinese monks could carry manuscripts to China, surely the Indians could carry them between monasteries in India!

I’ve always been sceptical about the various colors proposed for the different sects, as the identity of Buddhist monastics wearing ochre robes is so deeply ingrained. However one thing I noticed from my time in Thailand was that, since it’s assumed that robes are in the “robe-ish” color spectrum, it becomes normal to talk about variations on that spectrum as different colors, even if it’s only a tint. Very dark robes might be called “black” for example. Or there’s a particularly ugly color called “horse-poo green”, which is actually this color!

Sorry about that.

But anyway, by “blue” perhaps we are not to think of bright blue, but merely a blueish or greenish tinge to the robes.

1 Like

Well I first saw that picture, and only later started reading post. Context. :slight_smile:

Two classical sources describe the colors of monastic robes with these lists. One source claims the Sarvāstivāda wore red and the Dharmaguptaka wore black. The other claims the opposite. However, the lists do largely agree, and the purpose of these colorings seem to be to distinguish between the different sects.

In the later period of Indian Buddhism, we see that the vinaya lineages have been subsumed under four groups. Each group had two of the eight auspicious symbols sewn on the back of their robes. However, the purpose would be the same as the different colors: identifying their place within a system. The same reason that Catholic officials wear different clothing.

What strikes me in all this sort of material is that it would require real communication between the different monastic groups, and the forging of political bonds. I think that this is what was happening in some Indian chronicles which go out of their way to recognize other sects, acknowledge their ideas, compliment them, and give them a meaningful place within the Buddhist fold. Groups like the Sarvāstivāda and Mahāsāṃghika were reaching out and acknowledging others, and by doing so, going for a leadership position.

There must have been a certain amount of diplomacy going on between groups, and we don’t always know how that happened, but I think it must have occurred. We know about the Buddhist Councils, but I think the communication must have continued on a large scale after that as well. Mahāyāna texts also functioned in this way to some extent, absorbing ideas about interpretations of the Dharma, and being passed between members of different monastic groups.

Unfortunately, not many paintings from India survived, and the ones that did tend to focus on the Buddha himself and his earlier life as a bodhisattva, or as a royal prince. Most paintings from India show the Buddha in a deep red colored robe, but these are mostly depictions from places where the Sarvāstivādins and Mūlasarvāstivādins would have been dominant anyways. Occasionally the Buddha is shown in white robes or even blueish or greenish robes, but that seems to be more of a cosmic depiction.

Given that Dharmaguptaka Chinese monks are often depicted in black, and the Central Asian and Tibetan monks often wore red, it seems to fit the general narrative about Dharmaguptaka wearing black and Sarvāstivāda wearing red. Depictions of Central Asian monks with robes of various colors give us the sense that maybe the use of colors for designating monastic affiliations was either an old practice, or had assumed other localized forms.

Either way, it is an interesting subject, and I tend to take the accounts about robe colors literally because the robes of Central Asian and Chinese monks were sometimes quite colorful. It seems unlikely that Buddhism in Central Asia, so affected by the Dharmaguptaka, Sarvāstivāda, Mūlasarvāstivāda, and other Silk Road trends for centuries, would be going in a completely different direction from Indian Buddhism.

In any case, there are some cool old murals depicting Central Asian monks, and so I made a new thread with pictures. It’s not exactly early Buddhism, but somewhat related.

1 Like