The Original Bodhicaryāvatāra

Shantideva The Bodhicaryāvatāra or “Way of the Bodhisattva” as it is often translated, is one of the most read texts in the Tibetan Buddhist tradition.* This is hardly surprising, as it’s author Śāntideva managed the feat of encapsulating the vast expanses of Mahāyāna Buddhism in a single elegantly and often movingly written text. What is surprising is that the text as it has been passed down in the Tibetan tradition, and as it is read in translation today, is not the earliest version, and is quite different from it. We know this largely thanks to the excellent work of Akira Saito, which is undeservedly difficult to get hold of. During the 90s Saito worked on four Dunhuang manuscripts of the Bodhicaryāvatāra, showing how the early version that they contained differed quite radically from the familiar later version. Here I’ll summarize his conclusions and add some of my own thoughts. The four manuscripts are actually three:


(1) A complete copy comprising IOL Tib J 628 and Pelliot tibétain 794.


(2) A copy with several missing pages, now representing about 60% of the text (IOL Tib J 629).


 (3) A copy of the last chapter, on the dedication of merit (IOL Tib J 630).

So, what are the differences between this Dunhuang version of the Bodhicaryāvatāra and the one that was passed down through the centuries in Tibet and is still being taught today? Most strikingly, it is significantly shorter, containing roughly 700 verses instead of 1,000. Also, chapters 2 and 3 of the longer version are combined together, so that the Dunhuang version has nine chapters instead of ten. Most of the missing verses come in the chapters on meditation and wisdom, perhaps the most read and discussed chapters in the Bodhicaryāvatāra. Further confirmation that the Dunhuang Bodhicaryāvatāra was the earlier version is found in the early 9th-century library catalogue, the Ldan dkar ma, which records the existence of this text, with 600 verses (Saito suggests that it is given as 600 rather than 700 because the verses were estimated at 300 verses per volume, and the text was in two volumes). Saito’s close study of the differences in the wisdom chapter lead him to conclude that the familiar longer version of the Bodhicaryāvatāra is “an enlarged version” of the Dunhuang manuscript version, with many verses added on criticism of other systems of thought, including that of a supreme deity (Īśvara) and the metaphysics of the Sāṃkhya. On the other hand, some of the verses on non-self have been cut from the older version.

*  *  *

 The existence of different versions of the Bodhicaryāvatāra was known in Tibet. Buton was aware that a shorter version was listed in early library catalogues like the Ldan dkar ma, and he wrote:

Though this [Bodhicaryāvatāra] is described the three great catalogues as comprising 600 verses, it actually has 1,000. Many state that this [Bodhicaryāvatāra] is not the same as the one with nine chapters said to be written by Akṣayamati. However, apart from the difference arising from separating the chapter on the confession of sins, and differences in the earlier and later translations, I would say that they are the same [text].

Another major difference between the two versions, which Buton mentions, is the name of the author, which in the colophon to the complete Dunhuang manuscript (IOL Tib J 629) is not Śāntideva but Akṣayamati. Saito quotes from two sources that suggest this was an alternative honorary name for Śāntideva. Here’s that colophon: ITJ629 colophon

Note also that this colophon gives the names of two translators, an Indian, perhaps Sarvajñādeva (Sa ra bad nya de ba) and a Tibetan, Bande Paltseg. If we look at the colophon of the longer version of the text, we can see that this is the first stage of its translation history:

The Indian master Sarvajñādeva and the Tibetan translator Paltseg edited and finalized [a translation] based on a text from Kashmir. After that, the Indian master Dharmaśrībhadra, the great translator Rinchen Zangpo, and Shakya Lodro completed an amended translation by combining a text and its commentary from central India. Furthermore, in a later period, the Indian master Sumatikītri and the monk translator Loden Sherab completed a correctly amended translation, which was excellent.

Thus there is a gap of some two centuries between the work of the first translation team in the late 8th century and the second in the early 11th century. It is not surprising that a significantly different manuscript version was in circulation by the time Rinchen Zangpo was travelling in India in search of books to translate. (Incidentally, the surviving Sanskrit versions of the Bodhicaryāvatāra are all the longer version.)

Writing much later still, in the 17th century, the Tibetan scholar Tāranātha claimed that there were three versions of the Bodhicaryāvatāra: a version from eastern India in 700 verses, and two different versions from Kashmir and central India in 1,000 verses. Tāranātha then tells a story of two monks being sent to Śāntideva to ask which was the correct version, to which the author replied that it was the one found in central India. The same story is told by Buton in his history of Buddhism. These look very much like post facto justifications that the version already accepted in the Tibetan tradition was indeed the correct version.

*  *  *

I’m  inclined to agree with Saito in his assessment that the version found in  Dunhuang is earlier, and the longer ones that were passed down through the Tibetan tradition were based on Indian manuscripts that had been supplemented with new verses between the 8th and 11th centuries. Since the first translation was done quite soon after the time of Śāntideva himself (who lived in the 8th century) the shorter version may represent something close to what the author first circulated. The popularity of the text in India meant that it would have been copied multiple times, and that it could have been adapted to the needs of the people using it, resulting in multiple variously expanded versions circulating in India. As we saw, in the longer version there is new material in the chapter on wisdom, disputing theists and metaphysical dualists, and this would have been useful in religious debates.

I’ve written a little bit in the past here on the concepts of “originality” and “authenticity” in Buddhist texts. It is difficult to hold on to the idea of an original and authentic text when one looks at the multiple versions on offer in printed and manuscript versions. In manuscript culture in particular, it is clear that texts were constantly trimmed, augmented and supplemented. Commentary added in interlinear notes might be inserted into the main text in a later copy. Whole sentences might disappear due to a copyist’s eye’s skipping too far down the page.

At one end of the spectrum there are the scriptural texts that are preserved as much as possible from the innate malleability of manuscript culture, as they represent the authority of the tradition, and are not actually used much. At the other end there are those works that are found to be useful in a variety of ways, spread far and wide, and are much changed in the process. The latter was surely the case with the Bodhicaryāvatāra in India. Later on in Tibet, as a prestigious text translated from the home of Buddhism, it became important to designate and preserve an original version. Though as we have seen, due to the quirks of history and politics, the version that came to be accepted as authentic in Tibet was probably further from the original form than the version that was forgotten.

Not that the shorter version is the “original” either. Unless a Sanskrit manuscript turns up, signed by Śāntideva himself, we’d better not worry about that. As for the “authentic” Tibetan version: if pressed I would go for the one that has been used with great success by generations of teachers and students.

*  *  *

* The full title is Bodhisattvacaryāvatāra, literally meaning something like “Engaging in the activity of a bodhisattva.”

*  *  *


The above is based on the following works by Saito:

Akira SAITO. 1993. “A Study of Akṣayamati (=Śāntideva)’s Bodhisattvacaryāvatāra as Found in the Tibetan Manuscripts from Tun-huang. Report of the Grant-in-Aid for Scientific Research published by Miye University, Japan.

Akira SAITO. 2000. “A Study of the Dūn-huáng Rescension of the  Bodhisattvacaryāvatāra.  Report of the Grant-in-Aid for Scientific Research published by Miye University, Japan.

Saito’s conclusions were discussed in:

Kate Crosby and Andrew Skilton (trans.). 1996. Śantideva: The Bodhicaryāvatāra. Oxford: Oxford University Press.

The image at the top is from a 19th-century painting: see here.

19 thoughts on “The Original Bodhicaryāvatāra

  1. As for the authenticity of the various text, what do you think about “Terma”.
    Is it possible to consider as a means of verification how far current version of teaching has deviated from the original?
    And whether those texts found among the Dunhuang manuscripts?

  2. Thank you! Fascinating. Seems like we have the advantage of added scholarly commentaries over the centuries, but the disadvantage of lost passages.
    The Tibetans had proclaimed the reincarnation of Shantideva as a Tibetan monk decades ago (Mandala Magazine/FPMT: apologies, I don’t remember the issue number). I wonder if anyone has ever consulted him on this.
    In any case, I hope there will be an English translation available of the Dunhuang text, as all versions if this wonderful work contain the opportunity for insight.

  3. PS: yes, the comment above on Terma is interesting. It could very well be the case that Shantideva had wanted to do a bit if editing, either inspiring Terma or as an actual manuscript copier in a later life!
    Being in the process of writing a large tome, myself, I have needed to do several rewrites just to accommodate changing times. & public perceptions!

  4. … Found it: March 1999 issue. Interview with Shiwa Lhasa (8th recognised reincarnation of Shantideva, at Sera Je monastery in India). Not the scholarly approach, perhaps, but an interesting option!

  5. Sorry! iPhone auto-correcting texting mishap: his name of course is Shiwa Lha ( not Lhasa!)

  6. About twelve years ago I received some teachings on the Bodhicaryāvatāra from a Nyingma Khenpo in Nepal. When introducing the text, he said that the two members of Śāntideva’s audience who wrote the text down had each heard it differently. Thus, one produced a 700 verse edition and the other a 1000 verse addition, and both could be considered ‘authentic.’

    Just another interpretive tradition to throw into the mix!

  7. Dear Geoff and Sam,

    That wouldn’t have been the Nyingma Khen po Kunpal (Kun-dpal)? His commentary on the text is available in English in a PDF on the net (I at least downloaded it long ago), which reads at p. 59:

    “A few of the learned panditas in attendance later wrote down what they had heard,
    arriving at different versions of the Bodhisattva‐caryavatara. The Kashmiri panditas are
    said to have compiled a version with seven hundred stanzas in nine chapters, while the
    panditas from the Central Land produced a version of one thousand stanzas in ten
    chapters. These significant differences led to many doubts among the scholars.”

    If you want to see the whole work, the whole title is

    “Dzogchen Khenpo Choga Rinpocheʹs Oral Explanations of Khenpo Kunpal’s Commentary on Shantidevaʹs Bodhisattvacaryavatara (Entering into the Conduct of the Bodhisattva)”

    so it should be Schmooglable.

    Do you think the Khenpo-la had heard about the research of Saito-san?


  8. PS: The story continues on the next page:

    “In order to definitively resolve these doubts, as well as to question Shantideva about the Shiksha‐samuccaya and the Sutra‐samuccaya, two panditas from Nalanda journeyed to meet him. Shantideva told them that the version of the Bodhisattvacaryavatara held by the panditas from the Central Land, composed of ten chapters and one thousand stanzas, was in fact the correct one. In response to their questions about the Shikshasamuccaya and the Sutra‐samuccaya, Shantideva said they would find both texts hidden between the rafters in his old room at Nalanda. Then he gave them the reading transmission and the instructions for these texts. This concludes the second of the seven amazing episodes of Shantideva’s life story.”

    So yes, Павел, it was a hidden text.

  9. Dear Pavel, I can’t think of a case where a terton has discovered an earlier version of a text already in circulation, which would have raised questions about the authenticity of that text. But my knowledge is limited, so perhaps somebody else will come up with an example. In a looser sense of ‘terma’ we can consider the Dunhuang manuscripts to just that, but perhaps the most important difference from terma proper is that they lack a transmission lineage.

  10. Dear Kerrie, I agree that Shantideva could have produced more than one version in his own lifetime – why not? However I’m not sure it’s the most likely explanation – see my reply to Geoff and Dan. Oh, and I believe an English translation of the Dunhuang version is in the works!

  11. Dear Geoff and Dan,
    I guess you are both talking about the same Khenpo, but the stories are actually different. In the published commentary it sounds like he was drawing on Buton and Taranatha’s story, rather than Saito, whereas in the version remembered by Geoff the difference becomes a matter of differences in reception among the audience, a familiar trope in Mahayana Buddhism and a much earlier version of reader reception theory. The first version always ends up justifying one version as authentic, with the author’s own blessing, while the second allows both to be authentic (I hope it wouldn’t be considered out of line to say that this is a characteristically Nyingmapa approach!). I find it interesting that even in the case of non-scriptural texts like this, a need was felt to authenticate one version by going back to the author and having him state ‘this is it’ (in story if not in fact). I don’t know if there are any similar stories for other Indic texts.

  12. As to whether both versions were written by Shantideva – this is possible but from the point of view of manuscript studies, it doesn’t seem the most likely option. Firstly, because all accounts suggest that the different versions were circulating in different parts of India, which would mean variations evolved locally. Second, because of the two centuries that elapsed between the translation done by Kawa Paltseg and his colleague, and the one done by Rinchen Zangpo and his colleagues. As we know from many other Sanskrit Buddhist texts, it is very common for them to grow in size over a period like that. But let’s remind ourselves that this is not a scriptural text, so how much does it really matter if some parts of the BCV didn’t come from the pen of Shantideva?

  13. Thanks for this very interesting and “generous” post. I have been working a lot on the concept of “originality” in South Asia (including Tibet), especially in its connection to textual reuse, but I am afraid I missed your articles elaborating on this topic. Could you tell me where to look for?

  14. Again, Michael…great thanks to Akira Saito, you and all correspondents. In the late ’70’s Rato Khyongla Rinpoche used S. Batchelor’s translation for our classes in Little India, Manhattan. Very glad to learn of this fundamental work and appreciate your beautiful efforts to spread the knowledge. Best! Sande Green

  15. Dear Sam, thank you for such an interesting read! Do you know if Saito prepared a transliteration of the Tibetan which is publicly accessible? I’d like to give a reading and translating a try, but I’m not familiar working with manuscripts such in the Dunhuang digital library. I hope you could help me out :) have a nice day all, Justin

  16. Justin, sorry that I missed your post when it first came through. Unfortunately Saito’s transliteration doesn’t seem to be easy to get hold of. However the manuscript IOL Tib J 629 is in quite clear dbu can and is available on the IDP website (see the link in the post). It shouldn’t be too hard to read once you get used to the style.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s