An introduction to A Multitude of Buddhas, a celebrated Mahāyāna scripture in forty-five chapters exploring the nature of awakening through teachings delivered by Samantabhadra, Mañjuśrī, and other great bodhisattvas.
This celebrated Mahāyāna scripture, the Buddhāvataṃsaka, with its own characteristic perspective on the nature of the buddhas and their teachings, is presented as a single, very long sūtra (Toh 44), although many of its component texts or “chapters” have also been seen as independent works.
The forty-five chapters of which it is composed fill no less than four volumes of the Degé Kangyur. In its Tibetan translation, as well as in the two complete Chinese translations, it is presented as a single work described as a “greatly extensive sūtra” (mahāvaipūlyasūtra, shin tu rgyas pa chen po’i mdo). It may, however, have evolved as an encyclopedic coalescence of shorter works, many of which are known to have circulated independently in India; some were first translated into Chinese as independent sūtras. In Tibet, some of them are still seen as texts in their own right despite being treated formally as “chapters” of the larger work.
Nevertheless, as described below, all the “chapters” are closely related in their theme, style, terminology, and structure, and there is a degree of coherence to the way they are placed in order. A common feature is the view that all buddhas and bodhisattvas, wherever they may manifest, are emanations of the Buddha Vairocana. There are good reasons to think of these component parts of the whole as belonging to a “Buddhāvataṃsaka family”1 of texts, a notion that can be extended to cover some additional standalone sūtras and alternative Tibetan translations that are found elsewhere in the Kangyur and share the same general themes and perspectives.
Most of the narrative elements take place in the few weeks following the Buddha’s awakening. Tradition holds that he did not teach during this time, and in the account in this body of texts, too, as the Buddha remains silently in meditation he inspires the teachings that are mostly delivered by Mañjuśrī, Samantabhadra, and other bodhisattvas. Approximately the last third of the Buddāvataṃsaka is taken up by The Stem Array (Gaṇḍavyūha, chapter 45), one of the best-known texts of the family. Its story narrating the youth Sudhana’s spiritual quest and meetings with fifty-two teachers is depicted in detail in the stone carvings of Borobudur in central Java. It concludes with the dedication “Prayer of Good Action” (Bhadracaryāpraṇidhāna) recited widely in the Tibetan Buddhist tradition.
The Buddhāvataṃsaka has been highly influential in East Asian Buddhism as the foundational scripture of the Huayan tradition in China (Kegon in Japan, Hwaeom in Korea, Hoa Nghiêm in Vietnam). In Tibet, although less studied, it was greatly respected as representing the full flowering of Mahāyāna thought. The whole work is classified by Tibetan editors as belonging to the Buddha’s third turning of the wheel of Dharma.
The Tibetan translation, made from Sanskrit, has some unique features compared to the Chinese versions of the text.
The events that form the narrative framework of the text mostly take place (with the exception of the Gaṇḍavyūha) in the weeks that follow the Buddha’s awakening. He sits in silence under the Bodhi tree, and in other locations successively or simultaneously: the Trāyastriṃśa realm on the summit of Sumeru, Yāma and Tuṣita high above Sumeru, and Paranirmitavaśavartin, the highest level of the realm of desire.
In the period following the Buddha’s awakening as it is described in other, more biographical accounts, he remains alone, has not yet started to teach, and is still reflecting on how to communicate the profound and ultimately ineffable nature of the awakening he has discovered. Here, however, there is no sense in which his activity has been paused. Throughout the text, while he remains in meditation and does not himself speak, the nature of his awakening is explored and depicted against a backdrop of highly visual descriptions of the vast assemblies of bodhisattvas, gods, and other nonhuman beings who gather to witness and celebrate it, of the miraculous displays of lights and powers that he manifests through his samādhis, and of the whole vast cosmos itself in which each particle contains entire world systems and every detail manifests as an expression of enlightened activity. Everything that appears or is taught happens through his power.
The teachings that radiate from his presence, setting out the path to awakening, its nature and practice, and the culminating result of the qualities and activity of a tathāgata, are all expressed not by the Buddha himself but by the great bodhisattvas such as Samantabhadra, Mañjuśrī, and Vajragarbha, often prompted by questions put by other bodhisattvas. And just as the universe itself is described from the perspective of awakening, so too the teachings and explanations given by these great bodhisattvas—and even the questions that elicit them—are all the very expression of the Buddha’s awakened wisdom. Indeed, it is the awakening of the primordial Buddha Vairocana that the Buddha Śākyamuni has made manifest, and it is also due to Vairocana’s all-pervading presence that all bodhisattvas have found their way to the path to awakening.
Among the doctrinal viewpoints consistently expressed in A Multitude of Buddhas is that of the “single vehicle” (ekayāna), i.e. that the sole ultimate goal of the path is the bodhisattva vehicle culminating in the full awakening of a tathāgata, and that this is the vehicle on which śrāvakas and pratyekabuddhas will eventually embark having realized the limitations of their views and the incompleteness of their awakening. Another is the position that the apparent birth, awakening, activity, teaching, and parinirvāṇa of a tathāgata is an expedient manifestation conditioned by the needs of beings, while in reality the tathāgatas are primordially awakened, and their nature transcends all conditioned phenomena including passing away. These viewpoints on the single vehicle and on the transcendence of buddhas, sometimes termed “docetic Buddhism,” are shared by a number of Mahāyāna sūtras including the Saddharmapuṇḍarīka (“Lotus Sūtra,” Toh 113), the Mahāparinirvāṇasūtra (Toh 119), and the Mahāmegha (Toh 232).
The work has also been said to contain what may be early expressions of buddha-nature doctrine, i.e. that the nature or potential for awakening, the tathāgatagarbha, is inherently present in all beings and only needs to be freed of its adventitious, obscuring defilements to manifest as the full awakening of buddhahood. The term tathāgatagarbha itself is used only once in the whole sūtra, and in a sense that only approximates its meaning in the principal buddha-nature texts.
Yet in some ways A Multitude of Buddhas goes even further than the buddha-nature texts in the ways it pursues the implications of certain Mahāyāna doctrines to their ultimate point. It presents the dharmadhātu as the global realm in which all phenomena arise, and at the same time as the very essence of all phenomena. It is in its non-dual, unobstructed omnipresence that the tathāgatas arise. It is also the basis of the “sameness” of all phenomena, and indeed of all tathāgatas. Further implications of the entirety of the dharmadhātu being the very nature of everything include the notions that all phenomena are mutually interdependent, that they interpenetrate each other without obstruction, and that the part contains the whole and the whole the part. These and other points of doctrine and practice have been extensively documented and studied in the commentarial works of the Huayan tradition in China and its branches in other East Asian cultures (Kegon in Japan, Hwaeom in Korea, Hoa Nghiêm in Vietnam), but do not seem to have received detailed attention in Tibet.
Lofty though its perspective may be, the teachings set out in the sūtra are directed at guiding those who aspire to attain awakening, and the text is structured, despite its literary profusion, with a clearly didactic intent. Points are made clear and memorable with analogies and metaphors. In most of the chapters, passages of prose are followed by a recapitulation in verse. A recurring theme is that of the various ways in which the path can be seen as divided into sequential stages, and sets of ten—whether of stages, items, analogies, or categories—are among several recognizable features.2 The doctrinally rich presentations set out in the successive chapters are complemented in the final great section of The Stem Array (Gaṇḍavyūha), a rich narrative in which the understanding of the young layman Sudhana gradually unfolds as he encounters, one after the other, fifty-two different kinds of teachers, male and female, child and adult, human and divine, monastic and lay, and learns of their attainments. The whole work concludes with the widely known aspiration and dedication composed by the last of Sudhana’s teachers, Samantabhadra, the “Prayer of Good Action” (Bhadracaryāpraṇidhāna).
The earliest signs of the existence of at least some of the individual works that were later integrated into the Buddhāvataṃsaka come in the form of the dates recorded of their translation into Chinese, starting in the second century CE and continuing during the third and fourth centuries, made by Central Asian translators such as Lokakṣema, Zhiqian, Dharmarakṣa, and Faxian.
Citations and mentions in Indian treatises provide further evidence that some of these individual works circulated widely in Buddhist India from at least the fourth century. Authors who cite one or more of them include Asaṅga, Vasubandhu, Sthiramati, Śāntideva, Kamalaśīla, and Buddhaguhya. Possibly even earlier, but with far less certainty than these other mentions, are citations in the Sūtrasamuccaya, with its problematic attribution to Nāgārjuna.
The evidence from Sanskrit manuscripts is sparse. In the Nepalese Buddhist tradition, among the Buddhāvataṃsaka family of texts only the Daśabhūmika and Gaṇḍavyūha (chapters 31 and 45 in the Tibetan version of the whole sūtra) have been preserved in Sanskrit manuscripts, the oldest dating to the eleventh century CE, but both texts have been transmitted as independent works without reference to the Buddhāvataṃsaka as a whole. Some fragments in the Hoernle collection in London from Sanskrit manuscripts in Southern Turkestan Brāhmī, dating perhaps to the sixth century or later, have been identified as belonging to Buddhāvataṃsaka texts but, again, the fragments that are extant do not refer to the Buddhāvataṃsaka as such.3
However, one of a set of so far undated Sanskrit manuscripts from the Potala in Lhasa, of which editions and notes were published in 2010,4 is that of the Anantabuddhakṣetraguṇodbhāvana (chapter 37 in the Tibetan version of the whole sūtra, but also extant as two different Tibetan translations in the General Sūtra section of the Kangyur, Toh 104 and Toh 268), and the colophon to this Sanskrit manuscript is of particular interest as it specifies that the text belongs to “the extensive basket (vaipūlyapiṭaka), the Buddhāvataṃsaka.” It is thus the only clear sign in any Indic text that the Buddhāvataṃsakasūtra, as such, was a collection recognized in India, and was not an East Asian innovation.5
It is significant that in the colophon to the Sanskrit manuscript just mentioned, the Buddhāvataṃsaka is referred to as a collection or “basket” (piṭaka). This same term also comes up in a work translated into Chinese by the Indian monk Paramārtha (who calls the Buddhāvataṃsaka a bodhisattvapiṭaka), and in a Dunhuang manuscript of the Daśabhūmika in Tibetan.6 As noted below, it is also used in reference to the Buddhāvataṃsaka in Butön’s inventory and in several of the Kangyur catalogs, and as a term describing a particular grouping of chapters (12–27) within the work as a whole.
When, how, and where the individual members of the Buddhāvataṃsaka “family” of texts came to be seen as a piṭaka and eventually as a single “extensive sūtra” (mahāvaipulyasūtra) with that single title remains unknown.
In contrast to the paucity of early textual evidence, both Chinese and Tibetan historical traditions, in different ways, sustain the emic perception that the component parts have always formed a whole.
Chinese tradition relates that the original teachings, in the form of a vast Buddhāvaṃsakasūtra corpus in three versions, were preserved and guarded by the nāgas in their realm. The longest version was almost infinitely long. Nāgārjuna, after visiting the nāgas, was able to bring to the human world only the shortest of the three versions, which had 100,000 ślokas divided into forty-eight chapters.7 The legend closely resembles the story preserved in Indian and Tibetan sources of Nāgārjuna’s retrieval of the Prajñāpāramitā corpus, but the biography of Nāgārjuna on which this Chinese tradition is based does not seem to have been known in India, and the parallel narrative in Indo-Tibetan histories has survived only with regard to the Prajñāpāramitā.
Tibetan historical tradition—as recounted in Khepa De’u’s thirteenth century History, briefly by Butön, in more detail by by Tāranātha, and referred to in several of the Kangyur catalogs including that of the Degé Kangyur—also tells us that in India the Buddhāvataṃsaka existed as a collection and that it was originally a great deal larger, with one hundred chapters (according to the Kangyur catalog) or one thousand (Tāranātha), but gives a different account of why it is now incomplete. This narrative attributes the loss of most of a much longer version to an arson attack on the library at Nālandā, that reduced it to only thirty-eight surviving chapters. The date of this event, said to have been responsible for the decimation of many other scriptures, too (including much of the Ratnakūṭa), is placed sometime before the lives of Asaṅga and Vasubandhu, along with accounts of two other calamitous episodes during a period of political turbulence and unstable patronage for Buddhist institutions in India.8 The allusions to this story in the Kangyur catalogs that make this mention are included there to explain why their entries use the phrase “as much as is extant” (ci snyed pa); the same phrase is seen as part of the explicit (concluding statement) of the whole sūtra.9
Be that as it may—and with the important exception of the colophon to the Potala manuscript mentioned above—the earliest direct evidence we have of the Buddhāvataṃsaka being seen as a whole rather than simply as its individual component parts is in the form of the Chinese translations that present it as a single large work:
Both of the complete Chinese translations were made from manuscripts brought from Khotan, leading to some speculation that the Buddhāvataṃsaka might have been composed there or elsewhere in Central Asia rather than in India. However, there is enough evidence of the work’s circulation in India to conclude that it must have had an Indian origin.10
The Buddhāvataṃsaka was studied far more extensively in China than it ever was in Tibet, generating an extensive corpus of commentarial literature. On it is based a whole school of Chinese Buddhism, the Huayan, and it has also had a profound influence on Chan Buddhism.
In summary, the several earlier translations of individual component works known to have been made into Chinese, and possibly some into Tibetan, were later superseded by the only three full recensions of the whole Buddhāvataṃsaka that we have today: the two complete Chinese translations (Buddhabhadra’s fifth century 60-fascicle version and Śikṣānanda’s eighth century 80 fascicle one), and the ninth century Tibetan translation. Each successive recension added more material, but their content is, very broadly, similar. Some of the principal differences between them are noted below.
According to Chinese tradition with its several versions, the whole work is divided into thirty-four or thirty-nine chapters, grouped into eight or nine “assemblies” according to the places and occasions where the teachings they recount took place. Although these divisions could be discerned in the Tibetan translation, traditionally they have not been, and the way the Tibetan chapters are grouped (see below) is different.
Overall, the early ninth century Tibetan translation is of similar size and content to the late seventh century Chinese translation by Śikṣānanda, both being larger than the early fifth century Chinese translation by Buddhabhadra.
However, neither Buddhabhadra’s translation, nor Śikṣānanda’s, contain what constitutes the last section of the final Gaṇḍavyūha chapter in the Tibetan translation—the Bhadracārīpraṇidhāna (Tib. bzang spyod smon lam). This is despite the fact that the Bhadracārīpraṇidhāna had been translated by Buddhabhadra as a separate work, and was later translated twice more in the eighth century, first by Amoghavajra and then by Prajñā. Indeed the Gaṇḍavyūha reached its longest extent in Prajñā’s eighth century, 40 fascicle translation, which included the Bhadracārīpraṇidhāna as well as some new material.
The Tibetan translation also contains two chapters, 11 and 32, not found in any of the complete Chinese translations, although a Chinese translation by Śikṣānanda matching the second exists as an independent sūtra, Taishō 298.
In other words, we see a gradual augmentation and lengthening of the corpus over time, probably reflecting new versions of the Sanskrit texts reaching China (perhaps via Central Asia) and Tibet, and perhaps also the incorporation of texts that earlier had some level of independent existence—like the Bhadracārī-praṇidhāna, which in China at least seems to be the last section to have been added.
It is possible that Tibetans first encountered some of the major canonical collections, including the Buddhāvataṃsaka, through Chinese translations. There are a few intriguing mentions of there having been an early Tibetan translation of the Buddhāvataṃsaka made from the Chinese, although that possibility is backed up by little substantial evidence.11
The Tibetan Buddhāvataṃsakasūtra that is found in the Kangyurs available today is presented as a single large sūtra (mahāvaipūlyasūtra). It is almost certainly based on one or more Sanskrit manuscripts, and was translated in the early translation period; what is known of the translators is discussed below. However, signs that the work as a whole represents some degree of compilation and includes texts that once circulated independently remain evident in Tibetan sources that include text inventories, catalogs, and historical works, and it is not clear exactly when it began to be taken as a single work. Within the text itself of the Tibetan Buddhāvataṃsakasūtra, five chapters (1, 31, 43, 44, and 45) all start with their own nidāna (introductory setting) sections, although only the first starts with the classic “Thus did I hear...” formula, and three chapters (31, 32, and 43) conclude with the mention, typical of independent sūtras, that all present “rejoiced at what had been taught.”
Even after the component texts had seemingly been compiled into the single large sūtra found as such in Kangyurs today, some texts that belong to the Buddhāvataṃsaka family, but are not included in the large sūtra, were placed in the General Sūtra section of most Kangyurs, along with a few alternative Tibetan translations of texts that are included (both kinds of standalone works are listed below). We know little of what criteria or circumstances might have led to works in this group being excluded from the larger compilation, other than in the several cases of alternative translations.
If indeed some of the component texts of the Buddhāvataṃsaka reached Tibet as individual works before the complete compilation, we know little about where they might have come from. There are transmission lineage records indicating that Tibetans received transmissions of the Buddhāvataṃsaka as a complete mahāvaipulya or piṭaka from Chinese lineages (the two Chinese translators Buddhabhadra and Śikṣānanda are named in these Tibetan records) as well as from Indian ones,12 but if these lineage records shed some light on how particular Tibetans received transmission, they do not necessarily tell us how the texts themselves first came to Tibet. Whether this might have been through Khotan or Dunhuang, or via geographically more direct routes taken by lotsāwas and paṇḍitas, remains speculative. As mentioned below, the mdo sde snyan gyi gong rgyan preserved only in some Themphangma Kangyurs might possibly have come from Bruzha (the area of Gilgit and Hunza).
In the two surviving early text inventories, the Phangthangma and Denkarma (both dating to the early ninth century), the Buddhāvataṃsaka seems to be understood to be a category or class of texts (akin to the Prajñāpāramitā or the Ratnakūṭa) rather than treated as a single work, and both inventories list several items in that category.
The Phangthangma inventory13 lists five texts that are included in the “noble extensive great Buddhāvataṃsaka sūtra section” (’phags pa shin tu rgyas pa chen po sangs rgyas phal po che’i mdo sde), giving for each the numbers of fascicles (bam po) and ślokas (not reproduced here):
The Denkarma inventory14 is more ambiguous, starting with the statement that the “noble extensive great section, the Buddhāvataṃsaka” (’phags pa shin tu rgyas pa chen po’i sde sangs rgyas phal po che) includes forty-five chapters in 130 fascicles and 39,030 ślokas. It then goes on to list seven component texts (again, each with their length in fascicles and ślokas):
The fact that both ninth century inventories add to their lists the lengths of the individual texts in fascicles (bam po) and in ślokas reinforces the impression that at this period the texts were still considered independent works.
This impression is further reinforced by the fact that in the ninth century Mahāvyutpatti, a list of 104 texts (under the heading dam pa’i chos kyi ming, “names of the sublime Dharma”) includes, scattered throughout the list without any evident grouping together of these titles, not only the title Buddhāvataṃsakam (sangs rgyas phal po che) but also six individual titles that correspond to the component texts in the two inventories mentioned above.
From Chomden Rikpai Raldri’s survey of canonical texts15 compiled at Narthang three and a half centuries later (c. 1270 CE), it might seem that the forty-five chapters of the whole extensive sūtra as it is found today had by then been grouped into one (da lta sde gcig tu byas pa la le’u bzhi bcu rtsa lnga), although several interpretations of the word sde are possible. Rikpai Raldri mentions by name only the first chapter and the three last, without adding their lengths, and presumably if the other chapters were still being seen as independent texts he would have listed them. He states that the dkon mchog ta la la (the Ratnolkā, on which see below, also mentioned in the Phangthangma) should belong in the middle of the whole, but has been wrongly placed elsewhere as a dhāraṇī. He also points out that the various different locations and times in which the narratives of different sections are set suggest that the work is a compilation of many sūtras. To conclude, he alludes to the story of the loss in early times of parts of the text. This, he says, accounts for the apparent breaks in the text, and is the reason why the explicit (closing statement) of the sūtra says “... as much as there is, is here concluded (ci snyed pa rdzogs so).”
However, Butön in his History (1322 CE), gives much the same list as the Denkarma, and again lists the length of the component texts by fascicle and śloka as if they were independent works. He mentions the same description as used in the Sanskrit text from the Potala, discussed above, of the component works forming an “extensive collection” or piṭaka (shin tu rgyas pa’i sde snod).16
Of the various Kangyur catalogs (dkar chag), most start by distinguishing the first twenty-nine chapters as a group; some explain that these first twenty-nine chapters constitute what was originally known in the Denkarma and other inventories as the Buddhāvataṃsaka, and that the remaining chapters are known under that name by association. The Narthang catalog lists the individual chapter names even within this first group as part of its detailed volume-by-volume contents list, but the other catalogs name individually (if at all) only the subsequent chapters (30–45). Some catalogs (Degé, Urga, and Lhasa) follow Chomden Rikpai Raldri in alluding to the notion that the extant components of the Buddhāvataṃsaka are incomplete because of the destruction by fire of the library at Nālandā, as mentioned above.
The extant Tibetan text itself gives most of its component parts none of the trappings of independent sūtras, such as their own Sanskrit titles or the usual opening phrases, although there are introductory passages at the beginning of some chapters that place the setting of what follows in place and time in a way comparable to the introductory passage (nidāna) of many sūtras. By comparison, the Nepalese Sanskrit texts of the Daśabhūmika and Gaṇḍavyūha (chapters 31 and 45 in the Tibetan), preserved as independent works, both start with the classic “Thus did I hear ...,” as indeed do the equivalent chapters of Sikṣānanda’s Chinese translation. Unsurprisingly, this is also the case in an independent recension of the Daśabhūmika (a different Tibetan translation) that, as mentioned below, is found in the Kangyurs of the Themphangma line.
Just as it took time for the Buddhāvataṃsaka to evolve, first from a set of closely related but independent texts into a collection, and then into a single large work, it seems to have taken time too—at least in Tibet—for a stable title to become widely used for the whole work. Traces of alternative titles remain in the text itself, as well as in the inventories and other works. Some of these may have been used for individual components of the collection, and some extended to the whole.
The text as preserved in most Kangyurs (but not all) opens with the Sanskrit title Ārya-buddhāvataṃsaka-nāma-mahāvaipulyasūtra, and (in all Kangyurs) the Tibetan title ’phags pa sangs rgyas phal po che zhes bya ba shin tu rgyas pa chen po’i mdo, both of which we have rendered in English as The Noble Great Extensive Sūtra ‘A Multitude of Buddhas.’
The Sanskrit title Buddhāvataṃsaka has been interpreted in various different ways. As well as A Multitude of Buddhas, other possible English renderings would be Ornament(s) of the Buddhas, The Buddha’s Garland, and Garland of Buddhas. Its titles in both Chinese and Tibetan show signs suggesting that several interpretations have been considered.
In Classical Sanskrit, avataṃsa describes a garland, or an ornament in the form of a ring or circle. For example, karṇāvataṃsa means “earring,” and kusumāvataṃsa a “flower garland.” This is the interpretation followed by Thomas Cleary when he translates from the Chinese Huayan to call the Avataṃsaka “The Flower Ornament Scripture.”17
However, in Buddhist Hybrid Sanskrit, the idea of a garland or connected set of items is extended and generalized, so that avataṃsaka means “a great number,” “a multitude,” or “a collection,” and this interpretation is the one chosen by the Tibetan translators in the most widespread Tibetan rendering of the Sanskrit buddhāvataṃsaka as sangs rgyas phal po che (“A Multitude of Buddhas”). The Tibetan phal po che is also used elsewhere in the Kangyur to render other Sanskrit terms meaning “multitude,” such as nicaya and (in some contexts) mahat.
The word avataṃsaka is not very common, and even less so as a compound with buddha, so it is particularly significant that buddhāvataṃsaka is the term used in the Mūlasarvāstivāda narrative of the Buddha’s second main miraculous display in Śrāvasti to describe how he multiplied his bodily form to create a vast array of buddhas all the way up to the top of the god realms, a miraculous ability unique to buddhas—and clearly relevant to some of the main themes of the Buddhāvataṃsaka.18
The traces of other Tibetan titles that remain in some versions of the text and in the catalogs may record earlier ways of understanding the Sanskrit term buddhāvataṃsaka, or are perhaps based on interpretations via the Chinese:
The meaning and origin of the archaic form rma ga chad and its variant spellings is not clear. The several, relatively recent Tibetan lexical works that include it as an entry seem to refer only to its use as a name for the Buddhāvataṃsaka, and it may be misleading to read them as implying that it actually means “ear ornament” (snyan gyi gong rgyan) rather than simply recording its use as an alternative name for the text known by that name.19
The teaching given by Samantabhadra in Chapter 44 follows his immersion in a samādhi that is called sangs rgyas phal po che in all Kangyurs except for the Bathang manuscript, in which it is called sangs rgyas rmad ga cad.20 In combination with a reasonable interpretation of the differences between Kangyurs in the fascicle titles mentioned above, such that the archaic term rmad ga cad was at some point revised (in some Kangyurs) to phal po che, it appears that rmad ga cad and its variants may be more likely to have the same meaning as phal po che (“multitude”) than snyan gyi gong rgyan (“ear ornament”).
Interestingly, the use of snyan gyi gong rgyan (“ear ornament”) as a moniker for Buddhāvataṃsaka texts, even if superseded in the majority of canonical recensions of the larger work, is preserved in a text found in some Themphangma Kangyurs, including the Stok Palace, London (Shelkar), Shey, and Ulaanbaatar Kangyurs, as well as some of the Bhutanese manuscript Kangyurs. It appears to be an independent Tibetan translation corresponding to chapters 9 and 10 in the complete Tibetan Buddhāvataṃsaka, and has the initial title mdo sde snyan gyi gong rgyan (“The Ear Ornament Sūtrānta”), as well as a colophon title suggesting that it was also known as sangs rgyas kyi snyan gyi gong rgyan (“The Buddhas’ Ear Ornament”). It was translated by the lotsāwa Chetsan Kyé (ce btsan skyes), a ninth century translator from Bruzha (Gilgit-Hunza-Baltistan) who is known to have translated other texts from the Bruzha (Burushaski) language.
Modern scholars writing about the Buddhāvataṃsaka tend either to accept that it was translated by the Indian paṇḍitas Jinamitra and Surendrabodhi, and the Tibetan chief editor Yeshé Dé, or to report that attribution as a view held by Tibetan tradition.21
However, a completely clear and consistent attribution of the Tibetan translation cannot be made from the colophons of the versions extant in different Kangyurs. In many Kangyurs (including the Lithang, Qianlong, Cone, and London) the text has no translators’ colophon at all. In a substantial number of others (including the Narthang, Lhasa, Stok Palace, Toyo Bunko, and Ulaanbaatar) a brief colophon states that “the lotsāwa Vairocanarakṣita acted as chief editor and established [the text].” However, the colophons in the Degé, Ragya, and Urga Kangyurs say that the work was “translated, edited, and established by the Indian upādhyāyas Jinamitra and Surendrabodhi, the chief editor and lotsāwa Bandé Yeshé Dé, and others.” The Degé colophon is supplemented by a detailed appendix written by the eighteenth century Degé editor, discussed below.
Butön’s inventory states only that “Vairocanarakṣita acted as chief editor,” using the same phrase as in those colophons that mention Vairocanarakṣita, but without adding that he established the text.
The Kangyur catalogs are divided on the matter. The Narthang catalog says that the text was translated by the paṇḍita Surendrabodhi and the great lotsāwa Vairocanarakṣita; the Yongle catalog says that the paṇḍita Surendrabodhi and the great lotsāwa Vairocanarakṣita edited and established the text; others (Cone, Peking Kangxi, and the briefest of the two Lhasa catalogs) say that the translators were Jinamitra, Surendrabodhi, and Yeshé Dé; and the Degé, Urga, and the expanded Lhasa catalogs say that the text was translated by the Indian upādhyāyas Jinamitra and Surendrabodhi, and the lotsāwa Bandé Yeshé Dé, and was edited and established by Lochen Vairotsana.
The identity of the “Vairocanarakṣita” in these various mentions is presumably the early Tibetan translator Pagor Vairotsana. There are two later Vairocanarakṣitas, both Indian paṇḍitas, of the eleventh and twelfth centuries, respectively; not being Tibetans, they would not be described in colophons and catalogs as lotsāwa or Lochen. A complication worth noting here is that the eleventh century Indian Vairocanarakṣita is mentioned in lineage records as participating in the lineage through which Tibetans in the later period receive the transmission of the Buddhāvataṃsaka teachings.
Assuming that “Vairocanarakṣita” does indeed refer to the early lotsāwa Vairotsana, a potential inconsistency arises if Vairotsana and Surendrabodhi are understood to have worked together, since it is thought that Surendrabodhi only came to Tibet in the reign of King Tride Songtsen (also known as Senalek, who reigned c. 800–815), by which time Vairotsana would no longer have been alive. The mention of the two names Surendrabodhi and Vairocanarakṣita together in some of the catalogs is, presumably, related to mentions of the same two names in lineage records, which in turn is taken up by Tashi Wangchuk’s editorial note (see below).
Another puzzle is that The Dhāraṇī of the Jewel Torch (dkon mchog ta la la’i gzungs, Ratnolkādhāraṇī, Toh 145 and 847, see below), which contains all of Chapters 17 and 20 of the Buddhāvataṃsaka (along with some other material) is said in its own colophon to have been translated by Surendrabodhi and Yeshé Dé, yet is a rather different Tibetan translation. It seems unlikely that these two translators, working on the same passages in two different texts, would not have produced more similar translations—although conceivably the differences may be due to modifications made separately by later editors.
Unique to the Degé version of the Buddhāvataṃsaka is an informative editorial note,22 filling a little more than two folio sides and following the colophon at the very end of the last volume. According to its own colophon, it was written in 1722 by one Tashi Wangchuk, a monk scholar appointed by the Degé King Tenpa Tsering to edit the text that was (a decade or so later) published as part of the Degé Kangyur. Tashi Wangchuk is also the author of the Degé Kangyur’s colophon to The Perfection of Wisdom in Twenty-Five Thousand Lines (Toh 9), and appears have been a key figure in the editing and production of the Degé xylographs, at least of these very extensive and important works. However, it must be said that the dating of both this editorial note of Tashi Wangchuk’s, and of his colophon to The Perfection of Wisdom in Twenty-Five Thousand Lines, to 1722—seven years before Situ Panchen himself says he began the preparation of the Degé Kangyur—poses some important historical questions.
Tashi Wangchuk is mentioned in Situ Panchen’s account of the preparations for the production of the Degé Kangyur, in the Degé catalog, as having edited a version of the Kangyur by comparing a number of reliable old recensions of canonical works.23
A full, annotated translation of his editorial note on the Buddhāvataṃsaka can be seen at the end of the final chapter, The Stem Array (Gaṇḍavyūha, chapter 45), c.2–15. The points it makes are as follows.
This editorial note is particularly vauable in providing rare details of the editorial process implemented, and the sources used, by scholars like Tashi Wangchuk, tasked with producing improved versions of canonical texts as part of the prestigious production of a new Kangyur.
The list of chapters here is divided into the seven principal sections or groups of chapters that are distinguished in the earliest Tibetan sources starting with the Denkarma inventory, repeated by Butön and later authors, and listed as well as in the majority of Kangyur catalogs. The Tibetan titles of the forty-five chapters listed here are taken from the chapter colophons of the Degé Kangyur. The English renderings of still unpublished chapters should be considered provisional.
Chapters 1–29 are considered by the Degé Catalog to form a distinct group of chapters that, the catalog states, form a section that is said in the Denkarma and other inventories to be called phal po che (“the multitude”). The Catalog does not list the individual names of these chapters. The same name seems to be given by Butön. This group of chapters can therefore be presumed to correspond to the Denkarma’s de bzhin gshegs pa phal po che le’u (“the section on the multitude of tathāgatas,” *Tathāgatāvataṃsakaparivarta), which is also how it is named in Tashi Wangchuk’s editorial note. The first chapter opens with a nidāna (introductory setting) and the formula “Thus did I hear...”.
The sixteenth century scholar Pekar Zangpo, in his detailed survey of the Kangyur sūtras,25 divides most of this large group of chapters into two named sections (from which he omits the last two, chapters 28 and 29). For the first eleven chapters, he uses the name de bzhin gshegs pa rmad ga chad, which is a variant of the name the text gives for the eleventh chapter; for the possible interpretations of this name (which might include “the tathāgatas’ ear ornaments”), see above.
Pekar Zangpo names most of the second half of this first group of chapters, chapters 12–27, byang chub sems dpa’i sde snod (“the bodhisattva basket,” *Bodhisattvapiṭaka).
The last two chapters in the Degé Catalog’s phal po che group are considered by Pekar Zangpo to be included (along with Chapter 30) in his third section, which he names rdo rje rgyal mtshan gyi bsngo ba (“Vajradhvaja’s dedications,” *Vajradhvajapariṇāma). This is the same name as Tashi Wangchuk uses for the second of his seven groups.
Chapter 30 is considered by the Degé Catalog to represent a section on its own called rdo rje rgyal mtshan gyi le’u (“the Vajradhvaja chapter,” *Vajradhvajaparivarta), but as just mentioned Pekar Zangpo includes chapters 28 and 29 as well under a similar name. Whichever grouping one takes, this section presumably corresponds to the Denkarma’s byang chub sems dpa’ rdo rje rgyal mtshan gyis yongs su bsngo ba’i le’u (“the chapter on the bodhisattva Vajradhvaja’s dedications,” *Bodhisattvavajradhvajapariṇāmaparivarta). Tashi Wangchuk uses the same name but does not state whether he includes all three chapters 28–30 or only 30.
Chapter 31, as all the catalogs and inventories agree, should be treated as a subdivision on its own, and is also known to have circulated as an independent work; its Sanskrit equivalent the Daśabhūmikasūtra is so seen in the Newar tradition of Nepalese Buddhism. The Denkarma gives it the full title ’phags pa byang chub sems dpa’i sa bcu bstan pa (“The noble teaching on the ten bodhisattva bhūmis,” *Āryabodhisattvadaśabhūmikanirdeśa); Pekar Zangpo calls it mdo sde sa bcu pa (“the sūtra section of the ten bhūmis”); and Tashi Wangchuk calls it sa bcu bstan pa (“the teaching on the ten bhūmis”). The chapter opens with a nidāna (introductory setting) and concludes with the classic concluding formula stating how all present rejoiced at what had been taught.
Chapters 32–42 are considered by the Degé Catalog to form a group called kun tu bzang po’i spyod pa’i le’u (“the section on Samantabhadra’s practice,” *Samantabhadracarya). The group of chapters presumably corresponds to the Denkarma’s ’phags pa kun tu bzang po’i spyod pa’i bstan pa (“the noble teaching on Samantabhadra’s practice,” *Āryasamantabhadracaryanirdeśa) and Butön’s kun tu bzang po’i spyod pa’i bstan pa; both of these names match the name of the last chapter in the group. Pekar Zangpo takes the same set of chapters as a group and uses the name kun tu bzang po’i spyod pa.
Chapter 43 is listed on its own in the Denkarma, and by Chomden Rikpai Raldri, Butön, and Pekar Zangpo, all using the same title as the chapter‘s own colophon (de bzhin gshegs pa skye ba ’byung ba bstan pa, *Tathāgatotpattisambhavanirdeśa). The chapter is one of the few that opens with a nidāna (introductory setting) and concludes with the classic concluding formula stating how all present rejoiced at what had been taught.
Chapter 44 is also listed on its own in the Denkarma, and by Chomden Rikpai Raldri, Butön, and Pekar Zangpo, all using the same title as the chapter’s own colophon (’jig rten las ’das pa, *Lokottaraparivarta). The chapter opens with a nidāna (introductory setting) but does not conclude with the classic concluding formula stating how all present rejoiced in what had been taught.
Chapter 45 is also listed on its own in the Denkarma, which uses the title sdong po bkod pa (*Gaṇḍavyūha), as does Pekar Zangpo; this is the form of the title most used in later Tibetan treatises to cite the text. Chomden Rikpai Raldri and Butön instead use the same title as the chapter’s own colophon, sdong pos brgyan pa. In Tashi Wangchuk’s editorial note (see above) he mentions (c.10) that throughout the Buddhāvataṃsaka rgyan (usually meaning “ornament” or “adornment”) and bkod pa (“array,” “display”) are used as synonyms, presumably for the Sanskrit vyūha. The concluding section of the chapter with its long verse aspiration, the Bhadracārīpraṇidhāna (bzang spyod smon lam) is perhaps the most widely known of all Buddhāvataṃsaka texts. The chapter opens with a nidāna (introductory setting) but does not conclude with the classic concluding formula stating how all present rjoiced in what had been taught.
Several works that may be considered members of the Buddhāvataṃsaka “family,” but are found elsewhere in most Kangyurs, include:
This bibliography does not include details of the Buddhāvataṃsaka source texts themselves, nor 84000’s published translations of them; these are all listed and linked in “Contents of this Kangyur Section” above.
84000. The Third Well-Spoken Branch: An Exact Account of How All the Victorious One’s Teachings Extant Today in the Land of Snow Mountains Were Put into Print (Chapter 3 of the Degé Kangyur Catalog, Toh 4568). Translated by the Subhāṣita Translation Group. Online publication. 84000: Translating the Words of the Buddha, 2024.
84000. The Perfection of Wisdom in Twenty-Five Thousand Lines (Pañcaviṃśatisāhasrikāprajñāpāramitā, Toh 9). Translated by the Padmakara Translation Group. Online publication. 84000: Translating the Words of the Buddha, 2023.
Appleton, Naomi. Many Buddhas, One Buddha: A Study and Translation of Avadānaśataka 1–40. Sheffeld and Bristol: Equinox, 2020.
Butön Rinchen Drup (bu ston rin chen grub). chos ’byung (bde bar gshegs pa’i bstan pa’i gsal byed chos kyi ’byung gnas gsung rab rin po che’i gter mdzod). In The Collected Works of Bu-Ston, vol. 24 (ya), folios 1.b–212.a (pp. 633–1055). New Delhi: International Academy of Indian Culture, 1965–71. BDRC W22106.
Chimpa, Lama, and Alaka Chattopadhyaya (trs.), Debiprasad Chattopadhyaya (ed.). Tāranātha’s History of Buddhism in India. Atlantic Highlands, NJ: Humanities Press, 1981.
Chomden Rigpai Raltri (bcom ldan rig pa’i ral gri). bstan pa rgyas pa rgyan gyi nyi ’od. In gsung ’bum [Collected Works], vol. 1 (ka), pp. 96–257. Lhasa: khams sprul bsod nams don grub, 2006. BDRC W00EGS1017426.
Cleary, Thomas (tr.). The Flower Ornament Scripture: A Translation of the Avatamsaka Sutra. (First published 1984). Boston and London: Shambhala, 1993.
Denkarma (pho brang stod thang ldan dkar gyi chos kyi ’gyur ro cog gi dkar chag). Toh 4364, Degé Tengyur vol. 206 (sna tshogs, jo), folios 294.b–310.a.
Fiordalis, David. “The Buddha's Great Miracle at Śrāvastī: A Translation from the Tibetan Mūlasarvāstivāda-vinaya.” Asian Literature and Translation 2.3:1–33, 2014.
Hamar, Imre. “The History of the Buddhāvataṃsaka-Sūtra: Shorter and Larger Texts.” In Imre Hamar (ed.), Reflecting Mirrors: Perspectives on Huayan Buddhism (Asiatische Forschungen Vol. 151), pp. 159–61. Wiesbaden: Harrassowitz, 2007.
Hamar, Imre. “Buddhāvataṃsakasūtra.” In Brill’s Encyclopedia of Buddhism, vol. 1, pp. 115–28. Leiden: Brill, 2015.
Hamar, Imre. “Khotan and the Buddhāvataṃsaka-sūtra.” In The Buddha’s Words and Their Interpretations, pp. 129–138. Otani University Shin Buddhist Comprehensive Research Institute: Otani, 2021.
Hori, Shin’ichirō. “Sanskrit Fragments of the Buddhāvataṃsaka from Central Asia.” In Robert Gimello, Frédéric Girard, and Imre Hamar (eds.), Avataṃsaka Buddhism in East Asia: Origins and Adapatations of a Visual Culture (Asiatische Forschungen Vol. 155), pp. 15–36. Wiesbaden: Harrassowitz, 2012.
Khepa De’u (mkhas pa lde’u). rgya bod kyi chos ’byung rgyas pa. Lhasa: bod ljongs mi dmangs dpe skrun khang, 1987. (Translated in Martin 2022).
Khyungtrul Namkha Jigme Dorje (khyung sprul ’jigs med nam mkha’i rdo rje). “gangs can bod brda skad ming gzhi gsal bar ston pa’i bstan bcos dgos ’byung nor bu’i gter chen.” In mkhas dbang raghu wīra dang lokesha tsandra rnam gnyis kyis nyar tshags byas paʼi dpe tshogs, vol. 64, img. 377–564. BDRC MW1KG26281.
Martin, Dan (tr.). A History of Buddhism in India and Tibet: An Expanded Version of the Dharma’s Origins Made by the Learned Scholar Deyu. Library of Tibetan Classics. Somerville: Wisdom Publications, 2022.
Martin, Dan. 1987. “Illusion Web: Locating the Guhyagarbha Tantra in Buddhist Intellectual History.” In Christopher I. Beckwith (ed.) Silver on Lapis: Tibetan Literary Culture and History. Bloomington: The Tibet Society. 175-220. Internet version (accessed 09 Mar 2025).
Minling Terchen Gyurme Dorje (smin gling gter chen ’gyur med rdo rje). zab pa dang rgya che ba’i dam pa'i chos kyi thob yig rin chen ’byung gnas dum bu gnyis pa [“The Jewel Mine: A Record of Transmissions Received of the Profound and Vast Sublime Dharma, Part 2”]. In gsung ’bum / ’gyur med rdo rje, vol. 2 (kha), folios 1a–320a. Computer input edition. Dehra Dun: D. G. Khochhen Tulku, 1998. BDRC W22096.
Nattier, Jan. “The Proto-History of the Buddhāvataṃsaka: the Pusa benye jing and the Dousha jing.” In ARIRIAB 7 (March 2005): 323–60.
Obermiller, Eugéne, trans. and ed. History of Buddhism (Chos ḥbyung) by Bu-ston. Vol. 2, The History of Buddhism in India and Tibet. Materialien zur Kunde des Buddhismus 19. Heidelberg: O. Harrassowitz, 1932.
Osto, D.E. “The Supreme Array Scripture: A New Interpretation of the Title ‘Gaṇḍavyūha-sūtra.’” In Journal of Indian Philosophy vol. 37, pp. 273–90. Berlin and Heidelberg: Springer, 2009.
Ōtake, Susumu. “On the Origin and Early Development of the Buddhāvataṃsaka-sutra.” In Hamar, Imre (ed.), Reflecting Mirrors: Perspectives on Huayan Buddhism, pp. 87–108. Wiesbaden: Harrassowitz, 2007.
Park, Hyunjin. “The Bathang Manuscript of the Buddhāvataṃsakasūtra.” In Journalof India and Buddhist Studies, vol. 65 no. 3, March 2017.
Pekar Zangpo (pad dkar bzang po). mdo sde spyi’i rnam bzhag: bstan pa spyi’i rgyas byed las mdo sde spyi’i rnam bzhag bka’ bsdu ba bzhi pa zhes bye ba’i bstan bcos. Beijing: mi rigs dpe skrun khang (Minorities Publishing House), 2006.
Phangthangma (dkar chag ʼphang thang ma). Beijing: mi rigs dpe skrun khang, 2003.
Rotman, Andy, trans. Divine Stories: Divyāvadāna Part 1. Wisdom Publications, 2008.
Schaeffer, Kurtis R., and Leonard W. J. van der Kuijp. An Early Tibetan Survey of Buddhist Literature: The Bstan pa rgyas pa rgyan gyi nyi ’od of Bcom ldan ral gri. Harvard Oriental Series 64. Harvard University Press, 2009.
Sirisawad, Natchapol. “The Mahāprātihāryasūtra in the Gilgit Manuscripts: A Critical Edition, Translation and Textual Analysis.” PhD dissertation. Munich: Ludwig-Maximilians-Universität, 2019.
Situ Panchen Chökyi Jungne (si tu paN chen chos kyi ’byung gnas). dkar chag [Degé Kangyur Catalog, Toh 4568]. Degé Kangyur, vol. 103 (lak+Sh+mI), folios 1a–469a.
Skilling, Peter, and Saerji. “’O, Son of the Conqueror’: A note on jinaputra as a term of address in the Buddhāvataṃsaka and in Mahāyāna sūtras.” ARIRIAB 15 (March 2012): 127–30.
Skilling, Peter, and Saerji. “The Circulation of the Buddhāvataṃsaka in India.” ARIRIAB 16 (March 2013): 193–216.
Tāranātha. dam pa’i chos rin po che ’phags pa’i yul du ji ltar dar ba’i tshul gsal bar ston pa dgos ’dod kun ’byung (rgya gar chos ’byung from Degé xylographs), Tezu, A.P., India: Tibetan Nyingma Monastery (1974), folios 47a–48b. Translation in Chimpa, L. et al (trans.). Tāranātha's History of Buddhism in India, 140–43. Atlantic Highlands, N.J.: Humanities Press, 1981.
van der Kuijp, Leonard W. J. “Some Observations on the Buddhāvataṃsakasūtra in Tibet.” In Holly Gayley and Andrew Quintman (eds.), Living Treasure: Buddhist and Tibetan Studies in Honor of Janet Gyatso (Studies in Indian and Tibetan Buddhism). Somerville: Wisdom Publications, 2023.
Vinītā, Bhikṣuṇī (Vinita Tseng, ed., tr.). A Unique Collection of Twenty Sūtras in a Sanskrit Manuscript from the Potala: Editions and Translation. Sanskrit texts from the Tibetan Autonomous Region 7, 2 vols. Beijing: China Tibetology Publishing House / Vienna: Austrian Academy of Sciences, 2010.
Dr. John Canti was a founding member of 84000’s executive committee and editorial team and is now senior editor.
