Peter Skilling’s notes on the Kanjur Translation Project


By Peter Skilling

Nonthaburi, March 11, 2009


This Dhamma that I have attained is profound, hard to see and hard to understand, peaceful and sublime, unattainable by mere reasoning, subtle, to be experienced by the wise.— Śākyamuni the Awakened One “The Noble Search”, in Ñāṇamoli Bhikkhu and Bodhi Bhikkhu (tr.), The Middle Length Discourses of the Buddha: A New Translation of the Majjhima Nikāya, Kandy: Buddhist Publication Society, 1995, p. 260.1

The production of the Tibetan translations that became the Kanjur and Tanjur was one of the greatest cultural exchanges that the world has ever seen. A “new” language and new conventions were established, and a vast body of knowledge was transferred not only in letter but in spirit to become an enduring monument of culture in all its forms. The translation project sponsored by the Dharma kings was certainly the greatest planned and sustained cultural exchange in early world history – over one thousand years before UNESCO and other international projects.

This makes the project to translate the Kanjur into English a bold revival of ancient ideals, in an entirely new age with new technologies and potentials. But in the end the project will depend on human capacities – on the translators. Translation is not just a matter of words. The translation of the Kanjur is a cultural transfer and a spiritual transmission. The goal is communication, which can be achieved through collaboration and consultation.

The Kanjur translations were produced by teams of Indian and Tibetan scholars. The Indians brought with them the vast knowledge and wisdom of the great Indian universities and the insight of their training and practice. Together they prepared translation manuals like the Sgra sbyor bam po gnyis pa (Madhyavyutpatti) and the Sgra bye brag tu rtogs par byed pa chen mo (Mahāvyutpatti) – tools, which we still depend on today. In their translations they were deeply concerned with the sciences of language, of grammar, and of etymology. They fixed Tibetan equivalents of Sanskrit roots, prefixes, particles, and so on. As a result, a new literary language was created, designed to convey the texts as accurately as possible.

Eventually, the translated texts were arranged and classified into collections which became the Kanjurs of we know today. The Kanjurs are precious repositories of Indian Buddhist texts, a large number of which no longer survive in Sanskrit. Therefore, the Kanjur is part of the cultural and spiritual heritage, not only of Buddhism but also of India and the world.

Kanjurs – I use the plural because the different editions are not entirely identical – are divided into several divisions, and we can study their contents through the catalogues (dkar chag) which give titles, sizes, and translators. The first European analysis of the Kanjur was published by the great Hungarian pioneer Alexander Csoma de Kőrös in Calcutta in 1836-1839. Csoma de Kőrös also translated the Mahāvyutpatti into English. He was followed by Léon Feer, who published two important works on the Kanjur – Analyse du Kandjour, recueil des livres sacrés du Tibet (Annales du Musée Guimet II, 1881, a translation and augmentation of Csoma de Kőrős’ work in 446 pages) and Fragments extraits du Kandjour (Annales du Musée Guimet V, 1883, 577 pages). Feer wrote that “by making known the vast sacred literature of Tibet, Csoma cast light on a part of the history of the human spirit which up to then was unknown”. Feer’s second work was equally important, since it was the first European translation of selected texts and passages of the Kanjur.

The divisions of the Kanjur include Tantra, Vinaya, Prajñāpāramitā, Buddhāvataṃsaka, Ratnakūṭa, and Sūtra. Today I will leave Tantra and Vinaya aside, and discuss the other divisions. Prajñāpāramitā (Sher phyin) is long and profound, and contains sixteen titles. Some translations into English have already been made, mainly by Edward Conze, pioneer of “Perfection of Wisdom Studies”. Buddhāvataṃsaka (Phal chen) is also very long (it takes up six volumes of the Peking Kanjur) but a complete translation from Śikṣānanda’s late seventh-century Chinese version has been published by Thomas Cleary.The Flower Ornament Scripture, 3 vols., Boston & London, Shambala, 1985.2 The Ratnakūṭa (Dkon brtsegs) is a collection of 49 sutras, some of which have been translated, especially from the Chinese.Garma C.C. Chang, General Editor, A Treasury of Mahāyāna Sutras: Selections from the Mahāratnakūṭa Sūtra, University Park and London, The Pennsylvania State University Press, 1983.3 Translations from the Sūtra collection (Mdo, Mdo sde, Mdo sna tshogs) – both long and short, both Śrāvakayāna and Mahāyāna – have been done by individual translators for various purposes, often for academic study. There are 762 sutras in the Peking Kanjur; there has been no sustained programme to translate them.

Tibetan tradition does not emphasize the study of individual sutras. Rather it stresses the Indian technical literature (śāstra, bstan bcos) through Tibetan commentaries illuminated by the living oral tradition. This is different from China and the Far East, where schools of thought and practice grew up around individual Mahāyāna sutras. As a result, very few Mahāyāna sutras have been translated into modern languages from within the Tibetan tradition. On the other hand, the publication of writings of the Tibetan masters has increased dramatically in the last decades, and has improved significantly in quality. Many important works of all four main traditions are now accessible.

In a sense the Kanjur translation project is turning over a new leaf by going directly to the sutras. We must recognize this fact when we embark on this exciting new twenty-first century project. Inspired by the precedent of the great mkhan po, paṇḍita, and lotsawa, we face many challenges.

I would like to present my ideas on this subject in three sections: planning, preparation, and product.


Planning means making decisions, and there are many to be made. These are some of them.

Source texts

What edition of the Kanjur should be used? At present we have access to several Kanjurs.

These may be divided into:
Kanjurs following the Tshal pa lineage;
Kanjurs following the Them spangs ma lineage;
Independent or local Kanjurs.

An individual sutra is not always the same in the different traditions – there can be significant variations (the translation may be older or newer, revised or unrevised, etc.). That is, Kanjurs are not uniform in contents or in textual tradition. The variation begins even with the titles. Here are three examples:

1. “Questions of Druma, the Kinnara King”Tshal pa mi ‘am ci’i rgyal po ljon pas zhus pa
Thems spangs ma mi ‘am ci’i rgyal po sdong pos zhus pa
2. “Sutra on Many Elements”/khams mang po pa’i mdoTshal pa (including Derge) dhātubahutaka
Thems spangs ma bahudhātukaThe Thems spangs ma tradition gives the correct Sanskrit title. The Tshal pa tradition gives a wrong title, presumably invented by the editors of the Kanjur at some point.
3. “Sutra on the similes of the young ones”/Gzhon nu dpe’i mdoAll Kanjurs: kumāradṛṣṭānta-sūtra.

The Sanskrit title is wrong in all Kanjurs. The correct title, known from Sanskrit texts, is Daharopama-sūtra.

These examples warn us that we cannot even take the title for granted. Obviously we must be cautious with every word of the text, down to the final colophon.

In general, scholars prefer the Derge xylograph edition, and for several reasons it seems advisable to appoint the Derge as primary source for the Kanjur translation project, especially because a new digital edition is under preparation. When that is ready is will be a marvellous tool. However, when the text is difficult it is useful to consult other Kanjurs, and in any case, translator(s) should learn as much as possible about the textual history of the sutra(s) they are working on. Other Kanjurs may be consulted online through the TBRC, through the meritorious work of Gene Smith. Furthermore, when critical editions of a sutra already exist, the translator(s) should certainly consult them. I wonder whether young scholars can be trained at one of the Tibetan institutes to do preliminary research on each sutra selected for translation – to collect all relevant materials.

Selection of texts for translation

What texts should be chosen to translate? That is a difficult question indeed. The project will not be accomplished overnight, and must be planned in several long-term phases (here we should compare the planning and the principles of selection of the BDK programme). Should sample texts be selected from all divisions of the Kanjur? The Tantras – about seven hundred in number – are difficult; they require notes and explanations, and in some case initiations. The Vinaya (’dul ba) texts (eight in number) are long and technical, but they are, of course, foundational for monasticism, for the saṃgha. Some sutras are very long – the “Perfection of Wisdom in One Hundred Thousand Stanzas” and the Buddhāvataṃsaka, for example. Beyond that, there are over three hundred sutras in the Ratnakūṭa and Sutra divisions. How to make a choice?

The relevance and interest of a sutra should be a determining factor for the first phase. I would suggest a selection of shorter or medium-length sutras, especially ones that are often referred to or cited in Indian and Tibetan śāstras. At the beginning it will be necessary to establish terminology and technique, so it may be useful to translate a text like the Arthaviniścaya-paryāya (Don rnam par nges pa zhes bya ba’i chos kyi rnam grangs), which has already been translated by N.H. Samtani along with its commentary (the latter only available in Sanskrit, although a different commentary is preserved in Tibetan). In the same volume one might consider including a śāstra, a Tanjur text, the Pañcaskandhaka of Vasubandhu, which is now available in Sanskrit and has several Indian commentaries in the Tanjur. These two texts would help establish a firm basis for the translation and understanding of terminology. Once terminology is established, many other translations can proceed smoothly.

Other suggestions:

“The Perfection of Wisdom in Eight Thousand Lines” (Aṣṭasāhasrikā Prajñāpāramitā, Brgyad stong pa): There is an English translation from the Sanskrit (Conze) and a recent French translation from the Tibetan.

“The Kāśyapa Chapter” (Kāśyapaparivarta, ’Od srung gi le’u): Sanskrit available. There is no translation so far from Tibetan or Sanskrit.

“The Śālistamba Sūtra” (Śālistamba-sūtra, Sā lu’i ljang ba): Sanskrit almost all available in quotation, for example in Candrakīrti’s Prasannapadā. Edition of Tibetan text by Schoening with edition of Kamalaśīla’s commentary. French translation.

“One Hundred [Stories about] Karma”(Karmaśataka): No translation.

Miscellaneous short texts on the merit of making or worshipping stupas – to include Adbhuta-sūtra, Kūṭāgāra-sūtra, Mahāraṇa-sūtra, Pratītyasamutpāda-sūtra, Caityapradakṣiṇagāthā, Prasenajit-gāthā.

For all of these texts, considerable research materials are available. The fact that translations already exist – whether from Sanskrit, Chinese, or, more rarely, Tibetan – does not mean we should not include a text. There is bound to be overlap. The Kanjur translations will be faithful renderings of the Tibetan versions as understood by the translation teams in eighth and ninth century Tibet. We might describe them as products of the flourishing Indo-Tibetan culture of the age.

We need not hesitate to translate sutras which have already been translated. On the contrary, we can learn from the translations – and furthermore, our goal is to represent the Indo-Tibetan translations of the Kanjur. For example, if we plan to translate sutras from the Tibetan that have already been translated from Chinese in the BDK project, our translators can compare to see how the ideas were expressed by the BDK translator. Of course the Chinese and Tibetan versions are often somewhat different but the comparison is always useful. And if the BDK translator knows Tibetan (perhaps a few do), he or she could be consulted for advice.

Vocabulary and style

To what degree should vocabulary be standardized? This is a big problem. Translators and scholars do not agree even on basic vocabulary – for example , the five aggregates (phung po lnga, pañcaskandha). We should be flexible, and leave the translators some leeway. The final rule should be internal consistency. I believe in guidelines, not inflexible rules.

With regard to vocabulary and style, we must consider the audience, the readers. Do we aim to translate for all intelligent beings, or only for those within the Tibetan teaching tradition? I believe the translations should have universal appeal, and for that reason I feel we must avoid trying to reproduce the Tibetan forms too literally. Should we use Sanskrit terms like arhat rather than artificial terms like “foe-destroyer”? Should we use the Sanskrit forms of names of persons and places, rather than translate them into English? These are important questions that must be decided carefully. If the translations are too artificial they will have limited appeal.

It will be necessary to draw up a cumulative glossary by looking at existing translations and glossaries, especially from Tibetan. It should give, for example, the English, French, and German (plus other languages, as far as is possible) equivalents of the terms with their source keyed to a bibliography. Common translations that are widely used could be described as “common”. The glossary can be a basis for shorter glossaries of relevant terms to be published at the end of each translation. Such a preparation is a massive lexicographical project but it seems to me necessary. Here again, perhaps young research scholars can be trained to do this.

Other questions will be how to deal with repetitions, common especially in longer sutras, and how to treat honorifics and titles. Should we preserve them in translation, or abbreviate and reduce them? My own tendency is to preserve them.

Many of the sutras are in mixed prose and verse. The prose includes narratives, dramatic stories that are often breathtaking in their scope and vision. The verse is poetry, often extremely beautiful, as for example the praises of the Buddha in the “Questions of Rastrapala” (Rāṣṭrapālaparipṛcchā-sūtra) or the “Exposition of Vimalakirti” (Vimalakirti-nirdeśa). There are lyrical passages on emptiness in for example the “King of Samadhis” (Samādhirāja). Can we translate the stories and poems in a way that they retain their beauty, that they inspire the readers? Let us reflect that Kumārajīva’s translation of the “Lotus Sutra” became a classic of Chinese literature. The Sanskrit Lotus Sutra was translated twice (into French and later into English) in the nineteenth century. Despite the fact that the Sanskrit is available and has been published in many editions, Kumārajīva’s Chinese version has stood its own, and has been translated many times into English, French, and other languages. (The Tibetan has never been translated.) This demonstrates the power of a good translation.


Translators should consult the existing literature on their text – critical editions, translations (including from Sanskrit and Chinese), studies, etc. Sometimes it may be useful to look at Dunhuang manuscript or other early versions. If possible they should compare the Sanskrit when available. However, we must remember that the Sanskrit will be a different text than the one translated into Tibetan, and not try to confuse the text lineages. The Tibetan should always be the primary text (except where we can identify clear-cut cases of mistranslation).

Translators should be familiar with other translations from Pali, Chinese, etc. A number of classical translations should be chosen for consultation. Above all, translators should study closely the translations of Étienne Lamotte – the “Heroic March” (Śūraṃgamasamādhi), the “Exposition of Vimalakirti”, in the original French or the English translations. Relatively recent translations include:

Jens Braarvig, Akṣayamati-nirdeśa
Paul Harrison, Pratyutpannabuddhasamādhi
Jan Nattier, The Questions of Ugra
Daniel Boucher, The Questions of Rāṣṭrapāla

The goal of these translations may be different from our goal, but we can learn a great deal about the questions that arise in translation.

Selection of translator(s)

Who should translate a chosen text? Individuals or groups? This should depend on circumstances. If someone who has seriously worked on a text as a thesis, a study, or a book, agrees to produce a version for the Kanjur project, they should be encouraged to join. This is, I believe, the BDK principle. Some sutras have already been fully translated for PhD programmes but have never been published, for example the Satyakaparivarta (Losang JAMSPAL 1993). Can we explore the re-edition and publication of works like this?

Editorial committee/board

There should be a committee to oversee the translations when they are submitted. The committee should be made up of scholars with several fields of expertise, including Sanskrit. Here too we must compare the BDK model.


2.1. Bibliographic survey

As I see it the first step is to compile a bibliographic survey of existing editions, translations, and studies. The bibliography should follow one of the Kanjur catalogues – the best may be the Otani Kanjur catalogue, which gives information about Chinese and Derge versions – giving information for each catalogue entry, as available (A Comparative Analytical Catalogue of the Kanjur Division of the Tibetan Tripitaka, Kyoto, 1930- 32; see online search at This must include translations from any language – Sanskrit, Tibetan, Chinese, Khotanese, Uighur, etc. It is necessary to find out what theses have been done recently (10-20 years) on sutras from Tibetan and have not been published, and, in some cases, to contact the authors.

The bibliography should cover reference material like Kanjur catalogues, dictionaries, indexes, etc.See here David Seyfort Ruegg, “Sanskrit-Tibetan and Tibetan-Sanskrit Dictionaries and Some Problems in Indo-Tibetan Philosophical Lexicography”, in Boris Oguibénine (ed.), Lexicography in the Indian and Buddhist Cultural Field, München, 1998 (Studia Tibetica Band IV).4 Peter Pfandt’s bibliography (Mahayana Texts Translated into Western Languages: A Bibliographical Guide, 1986) is now very much out of date. Some bibliographies must be available on the internet.


It is worthwhile to consider holding a number of in-depth courses or seminars during the period of preparation, that is in the next few years. Suggested topics include:

2.2.1. History of the Buddhist scriptures and scripture collections.

2.2.2. History of Buddhist translation

2.2.3. History of the Kanjur The courses should be conducted at an institute with a first-class library, for example with several editions of the Kanjur.


At a relatively early stage the editorial committee should decide on format, general design, standard contents (e.g. whether to have a glossary, how detailed the indexes should be, how to format bibliographies, etc.).

The series name is very important. It should be clear and straightforward. BDK is a bit confused sometimes, when they put more than one text in a single volume. The new series from Columbia is an example of what should not be done:

Loizzo, Joseph John (tr.) (2007). Nāgārjuna’s Reason Sixty with Candrakīrti’s Reason Sixty Commentary. New York: The American Institute of Buddhist Studies at Columbia University in New York/Columbia University’s Center for Buddhist Studies and Tibet House US.

The bibliographical reference is too complicated.

Making these decisions in advance will save a lot of trouble later. Translators should work with the same format and fonts, etc.

A note on other translation projects

BDK is a religious enterprise, connected to the Pure Land school, with (as far as I know) funding from a single source. The translation are non-sectarian, and in addition to Sutra or “word of the Buddha” include works by Indian, Chinese, Korean, and Japanese authors. We are fortunate that Prof. John McRae is here to explain how BDK works.

Pali canon
Almost all of the Pali canon” has been translated, starting in the late 1800s. Some of the ancillary texts and commentaries have been translated. New translations of important texts have been published, such as those by Bhikkhu Bodhi. The translation process never ends – knowledge changes, language changes, and new translations are needed.

Dharma Publishing
Some texts have been published by Dharma Publishing (Berkeley), including the “Fortunate Aeon” and the Sandhinirmocana-sūtra.

Sanskrit Buddhist texts
In general, very little has been translated. The “Clay Sanskrit Series” – mostly non- Buddhist texts – has published the Buddhacarita and the Saundaranandakāvya, both by the great poet Aśvaghoṣa, and several avadānas. These are exemplary translations.

Most recently, Andy Rotman’s Divine Stories, Divyāvadāna (Part I, Boston, Wisdom Publications, 2008) is a landmark in the translation of Buddhist narrative and of Mūlasarvāstivādin literature.

Other scholarly series to be consulted include the “Manuscripts in the Schøyen Collection” (MSC, three volumes published to date) and Gandhari Buddhist Texts (GBT, five volumes to date).

I hope my remarks are useful. Once again, I regret that I cannot be present and I send my wishes for the success of the conference.


BOUCHER, Daniel. 2008. Bodhisattvas of the Forest and the Formation of the Mahāyāna: A Study and Translation of the Rāṣṭrapālaparipṛcchā-sūtra. Honolulu: University of Hawai’i Press.

HARRISON, Paul. 1990. The Samādhi of Direct Encounter with the Buddhas of the Present: An Annotated English Translation of the Tibetan Version of the Pratyutpanna-Buddha- Saṃmukhāvasthita-Samādhi-Sūtra with Several Appendices relating to the History of the Text. Tokyo: The International Institute for Buddhist Studies. [Studia Philologica Buddhica Monograph Series V].

JAMSPAL, Lozang. 1991. The Range of the Bodhisattva: A Study of an Early Mahāyānasūtra, “Āryasatyakaparivarta”, Discourse of Truth Teller. PhD, Columbia University. Ann Arbor, Michigan: UMI.

KSHEMENDRA, La Liane Magique: Les hauts faits du Bodhisattva contés par le Bouddha pour expliquer la production interdépendante. Traduit du tibétain par le comité de traduction Padmakara. Saint-Léon-sur-Vézère: Éditions Padmakara 2001.

NATTIER, Jan. 2003. A Few Good Men: The Bodhisattva Path according to the Inquiry of Ugra (Ugraparipṛcchā). Honolulu: University of Hawai’i Press.

PADMAKARA TRANSLATION GROUP, THE. 2005. Nagarjuna’s Letter to a Friend with Commentary by Kyabje Kangyur Rinpoche. Ithaca, New York & Boulder, Colorado: Snow Lion.

ROTMAN, Andy. 2008. Divine Stories: Divyāvadāna Part I. Boston: Wisdom Publications.

TARANATHA. 2003. Le Soleil de la Confiance: La vie du Bouddha, tr. Le Comité de traduction Padmakara. Saint-Léon-sur-Vézère, France: Éditions Padmakara.

ZIMMERMANN, Michael. 2002. A Buddha Within: The Tathāgatagarbhasūtra. The Earliest Exposition of the Buddha-Nature Teaching in India. Tokyo: The International Research Institute for Advanced Buddhology, Soka University. [Bibliotheca Philologica et Philosophica Buddhica, vol. VI].

1 “The Noble Search”, in Ñāṇamoli Bhikkhu and Bodhi Bhikkhu (tr.), The Middle Length Discourses of the Buddha: A New Translation of the Majjhima Nikāya, Kandy: Buddhist Publication Society, 1995, p. 260.

2 The Flower Ornament Scripture, 3 vols., Boston & London, Shambala, 1985.

3 Garma C.C. Chang, General Editor, A Treasury of Mahāyāna Sutras: Selections from the Mahāratnakūṭa Sūtra, University Park and London, The Pennsylvania State University Press, 1983.

4 See here David Seyfort Ruegg, “Sanskrit-Tibetan and Tibetan-Sanskrit Dictionaries and Some Problems in Indo-Tibetan Philosophical Lexicography”, in Boris Oguibénine (ed.), Lexicography in the Indian and Buddhist Cultural Field, München, 1998 (Studia Tibetica Band IV).

Posted: 2 Jan 2010