The Tibetan Buddhist Canon

The Tibetan Buddhist Canon

by Phillip Stanley

The Formal Buddhist Canonical Collections of Tibet

The canon of Tibetan Buddhism has often been described in academic literature as consisting of the Bkf 'gyur (pronounced Kangyur) or of the Bkl 'gyur and Bstan 'gyur together (pronounced Tengyur). The term Bka 'gyur means translations of the word of the Buddhas which includes the exoteric sutras, the

esoteric tantras, and the root texts on monastic discipline that are attributed to the Buddhas. The term Bstan 'gyur means utranslations of the treatises [on the word of the Buddhas]which includes the commentaries and treatises by Indian masters. Both collections are intended to include texts translated from

Indic languages. A few texts were translated from non-Indic languages, such as Chinese, Mongolian, and Uyghur, but these texts were presumed to have been translated from Indic originals. Treatises by Tibetan Buddhist masters were excluded from the Tengyurs, except for a small number of texts attributed to eighth ninth century Tibetan masters of the Early Translation Period (see below).

The first Kangyurs and Tengyurs were produced in the later thirteenth and early fourteenth centuries CE during the Later Translation Period. These collections are referred to in the plural because they were never formally closed and no two editions are exactly alike, with individual collections

sharing greater or lesser degrees of similarity or difference. Editors have made changes in the content and arrangement of these collections over the centuries, especially among the Kangyurs, but given the substantial similarities in the texts and textual categories included in the editions of both collections, one may speak of them as open but stabilized.

Some Tibetan masters of the New Schools (Tib. gsar ma) of the Later Translation Period have viewed the Kangyurs and Tengyurs as containing all the legitimate translations of Indic Buddhist texts. Other Gsar ma masters have accepted two additional textual collections of texts not included in the

Kangyurs and Tengyurs as translations of canonical Indian texts. These collections consist of esoteric texts (Skt. tantra; Tib. The Wiley Blackwell Companion to East and Inner Asian Buddhism, First Edition. Edited by Mario Poceski.

rgyud) of the Old School (Tib. rnying ma), whose roots go back to the Early Translation Period in pre-tenth century CE Tibet. The Old School views these two additional types of collections as having a canonical status equal to that of the Kangyurs, if not higher. They are (1) the

Rnying ma rgyud 'bums (Collected Tantras of the Old School) and (2) numerous collections of individual groups of gter ma (treasure texts). The former consist primarily of texts that they assert were translated from Indic originals during the Early Period. The latter consist of treasure texts they view as

having been hidden by the Indian master Padmasambhava or others during the Early Period一hidden physically or in the mind streams of their disciples一to be rediscovered centuries later in Tibet by reincarnations of those very disciples when the time was appropriate for their teachings to be spread. There are

Indian Mahayana exoteric sutras and esoteric tantras, such as the Pratyutpanna sutra and the Manj us rim ula kalpa tantra, that depict the recovery of such hidden texts or other objects, thus there are Indian roots for the Rnying ma tradition of rediscovering texts. The ongoing emergence of new forms of tantra

in India indicates that the canons of these Indian traditions were fundamentally open and dynamic, and that the open creativity of the Rnying ma treasure texts was thus a continuation of this “orthodox” activity

Broadly speaking, the masters who have accepted the Rnying ma collections as authentic have come from the range of larger and smaller traditions in Tibet, while masters who have not accepted them have tended to come from two of the larger traditions: the Sa skya and Dge lugs traditions. The Sa skya tradition

played a major role in the process leading to the formation of the earliest Kangyurs and Tengyurs in the later thirteenth and early fourteenth centuries and a number of Sa skya and other masters were explicitly critical of these Old School texts as apocryphal. They thus contributed to the virtual exclusion

of these texts from the Kangyurs. (A very small number are included in some Kangyurs.) The Dge lugs tradition, which was founded toward the later end of the fourteenth century, had a similar conservative orientation.

There is yet another controversial pair of formal canonical collections that belong to the Bon school, which is a heterodox mix of ancient religious traditions and Buddhist teachings. A prominent feature of these two Bon collections is that they consist almost entirely of rediscovered treasure texts.

That applies not just for their esoteric tantras like the Rnying mas, but even for their exoteric sutras and treatises. These Bon treasure texts began to emerge in the eleventh century一perhaps slightly before the first Rnying ma treasure texts of that century一and they coalesced into two Bon canonical

collections by the late fourteenth to mid-fifteenth centuries, following the model of the Buddhist Kangyurs and Tengyurs that had emerged a century earlier. The fact that these heterodox texts were almost entirely treasure texts may have intensified the suspicions of the Sa skyas and others regarding

the Rnying ma treasure texts during and after the period leading up to the formation of the first Kangyurs and Tengyurs. There has been controversy, then, over which formal Tibetan canonical collections of Indic texts are legitimate. All schools of Tibetan Buddhism accepted

the Kangyurs and Tengyurs, but for some these two collections could be used in a sectarian fashion if they were claimed to be the full extent of the formal canon, that is, they could be sectarian not for what they included but for what they excluded, namely, virtually all

of the texts in the two types of Rnying ma collections. (The Bon did not seek to have their texts viewed as suitable for inclusion in the Kangyur and Tengyur). It is thus incorrect to say that the Kangyurs and Tengyurs are “the formal c日non” of Tibetan Buddhism. Rather, it is accurate to make the more

limited statements that (1) the Kangyurs and Tengyurs are included in the formal canon of all Tibetan Buddhists and (2) they are the full extent of the formal canon only for some Tibetans.

There is still a further issue regarding the scope of the Tibetan Buddhist canon. The field of Buddhist studies has begun to explore the concept of a practical c日non” that differs in important ways from the concept of the formal c日non” that was presumed above. Practical canons consist of the texts

actively used in a specific sectarian tradition. Two major features distinguish a practical canon from a formal canon: a practical canon includes (1) only a select portion of the texts in a given formal canon, which is called a ucanon-within-the-canonM in the scholarship on the Christian canon, and (2) treatises written by regional Buddhist masters that are not included in the formal canons.

As with the New and Old Testaments of Christianity, the Kangyurs and Tengyurs contain doctrinal diversity and thus to create a more-or-less consistent orthodoxy a school must select such a canon-within-the-formal-canon. Different sectarian traditions that share the same formal canon will thus have

different practical canons, because they may have different canons-within-the-canon and, in any case, will have their own treatises that stand outside the formal canon and interpret that canon in distinctive ways. These interpretive treatises are crucial to the sectarian identity of a tradition and will be

among the most formative, authoritative, and revered texts of the tradition. While in theory they are subordinated to the sutras, tantras, and treatises of the formal canon, in practice they may vie with them in terms of the reverence and studious attention they receive on a regular basis.

Finally, one may also speak of an uinclusive c日non” that consists of all the texts known to a tradition that it accepts as Buddhist, even if it itself may not use them in its practical canon or include them in its formal canon. This is exemplified by the sympathetic acceptance of the treasure texts of the

Rnying ma school by the Tibetan Bkl rgyud school and by the Chinese imperial governments creation and repeated reprintings of the Peking edition of the Tibetan Kangyur and Tengyur without including them in their formal Chinese canonical collections. While inclusive canons are the most expansive type of

canon, they are nonetheless sectarian. For example, there are Sa skyas and Dge lugs pas who would not accept the Rnying ma treasure texts into their inclusive canon. In summary, it is proposed that understanding the ucanon of Tibetan Buddhism” requires a consideration of three types of canons: formal, practical, and inclusive canons.

Overview of the Translation of the Kangyur and Tengyur Texts

The creation of the Kangyurs and Tengyurs, with the equivalent of over 230,000 Peking pages,1 5,200 texts, and 300 volumes, is among the great literary achievements of humanity. The Rnying ma rgyud 'bums and the Rin chen gter mdzod (Treasury of

Precious Treasure Texts), a nineteenth-century collection of many of the Rnying ma treasure text collections, would add almost another 110,000 pages (equivalent to 72,000 or 31% more Peking pages) in 161 volumes. Though the transmission of Buddhism from India north to Tibet occurred quite late, by the

mid-seventh century, the formal Tibetan Buddhist canon became the largest of the three extant Buddhist canonical collections: the Pali (of Southeast Asia), Chinese, and Tibetan canons. Seven centuries earlier, in the first century BCE, Buddhism traveled along the Silk Route to the west of Tibet and into Central Asia north of Tibet, reaching China to the east by the first century CE (for the early history of Chinese Buddhism, see Mario Poceskfs chapter in

this volume). Buddhism arrived in areas southeast of Tibet, such as Myanmar and Thailand, in the third century BCE according to traditional sources. Early Tibetan sources depict king Srong btsan sgam po (r. 629-650 CE)as inaugurating the spread of Buddhism in Tibet following his marriage to two Buddhist princesses, one from China and one from Nepal. While the historicity of the Nepalese princess Khri btsun is uncertain, the Chinese princess Munsheng Kongco

(Ch. Wencheng Gongzhu), who married the king in 641, may well have encouraged the spread of Buddhism in Tibet. Buddhist masters may have entered Tibet before this period, but they left no discernible mark (for more on the early growth of Buddhism in Tibet, see James Apple's chapter in this volume). The introduction of Buddhism came at a time of increased contact with other cultures due to the territorial expansion of Tibet under Songtsen Gampo. He is

credited with inaugurating a cultural expansion within Tibet, ordering the creation of the written Tibetan language, the creation of the first set of written laws, and the translation of the first Buddhist texts. Though the tradition views him as Tibet's first Dharma King, there is no clear evidence that he had a personal commitment to Buddhism. Significant translation activity did not occur until over a hundred years later with the arrival of the Indian scholar-monk Santaraksita around 760 CE. He founded the first monastery in Tibet in 779, named Bsam yas (Inconceivable), with the royal patronage of king Khri srong Ide btsan (742-c. 797). This monastery became the center of translation activity during the remainder of the Early Period of Buddhism in Tibet. The spread of Buddhism in Tibet is traditionally divided into three periods: the snga dar (Early Spread), an intermediate period of decline, and the spyi dar (Later Spread). The first period spanned two hundred years, from c. 641 to the collapse of the Tibetan empire following the assassination of King Glang dar ma in 842 due to his repressive actions against Buddhism that included the closing of monasteries, the centers of translation activity. However, the great bulk of the early translations were completed in a mere sixty-three years at the end of this period, between 779 and 842. The intermediate period of

decline lasted about 144 years, from 842 to 986 CE, at which point king Ye shes 'od (c. 959-1036) of western Tibet exhorted his subjects to adopt Buddhism; most importantly for the history of translation, he supported the work of the great translator Rin chen bzang po (958-1055). The bulk of the later translations occurred in the three hundred years from 986 until the late thirteenth and early fourteenth centuries when the arrival of Indian Buddhist texts and translators began tapering off due to the Muslim conquest of northern and central India.

Three royal catalogues of the translations of Buddhist texts in Tibet were produced during the Early Spread, two of which are extant. The compilers of the catalogues arranged the growing body of texts into a single unified canonical collection. They used the Three Baskets (Tib. sde snod gsum; Pali: tipitaka;

Skt. tripitaka) for their overarching structure, arranged in the order of the Sutra Pitaka (Basket of the Discourses of the Buddhas), Vinaya Pitaka (the Basket of Monastic Discipline), and Abhidharma Pitaka (the Basket of Higher Teachings that systematized the teachings of the Discourses). The Three Baskets schema was to serve as a major organizing principle for the Kangyurs and Tengyurs of the Later Period as well.

This sequence of the Sutra, Vinaya, and Abhidharma Baskets belongs to the northern Sanskritic schools of Foundational Buddhism, called the “Lesser VehicleM (Tib. theg dman; Skt. Hinayana) by the followers of the uGreat VehicleM (Tib. theg chen; Skt. Mahayana). The sequence used in the single surviving school of Foundational Buddhism, the southern Pali school of the Theravada, reverses the order of the first two baskets: Vinaya, Sutras, and Abhidharma. The

Chinese and Tibetan traditions inherited the sequence of the northern Sanskritic tradition, though some later Tibetan collections altered this sequence, which then paralleled the order used in the southern Theravada tradition, though this was an indigenous Tibetan development with no evidence of influence from the Theravada tradition. The Three Baskets have been a major organizing principle for all these canons.

Of the 232,000 pages found in the Kangyurs and Tengyurs, a conservative estimate suggests that a prodigious 103,000 pages, or 44% of the total, were translated during the Early Spread, primarily in the sixty-three years between 779 and 842. The other 125,000 pages, or 56%, were translated during the main 300-year period of the Later Spread or else their period of translation is unknown.

There are significant differences in the types of texts translated in the early and later periods, which can be summarized as follows, stated in terms of their relative number of pages:

• There was a tremendous quantity of exoteric Kangyur sutras translated in the Early Period and virtually a complete absence of such translations in the Later Period.

• The quantity of esoteric Kangyur tantras translated in the Later Period was roughly twice the quantity translated in the Early Period.

• The quantity of exoteric Tengyur sutra treatises and commentaries translated in the Later Period was modestly larger (33% more) than the quantity translated in the Early Period.

• There was virtually a complete absence of esoteric Tengyur tantric treatises and commentaries translated in the Early Period but an explosion in the translations of such texts in the Later Period.

The only common characteristic between the Early and Later Translations Periods in terms of these four categories is that a large quantity of Tengyur sutra treatises and commentaries were translated during both periods. While commentaries, especially sutra commentaries, held a prominent place in the early translations, the central role

of commentaries only intensified in the later translations, especially for tantra commentaries. A prominent feature of Tibetan Buddhism is its devotion to writing Tibetan commentaries and treatises on both exoteric and esoteric Indic commentaries and treatises, in contrast to the Chinese Buddhist masters who tended to write treatises directly on exoteric Indic sutras. (The tantras had far less influence in China than in Tibet.) The Tibetan orientation to

writing commentaries on Indian treatises reflects the scholastic culture they inherited from their Indian masters. Unlike China, Tibet did not have a literary culture of its own when Buddhism arrived, thus Tibetan literature was deeply shaped by the inherited Indic tradition.

With regard to the small number of tantras and tantric treatises translated in the Early Period, the kings who supported Buddhism were oriented to its exoteric forms, which they used to support the state in their political struggles with the noble families. This is exemplified in two edicts issued by king Khri srong Ide btsan (r. 755-797 CE)to officially reestablish Buddhism and its moral order for the good of the land, after ministers had assassinated his

father and banned Buddhism. The early royal caution toward tantric Buddhism is reflected in a royal decree in the Sgra sbyor bam po gnyis pa ([Treatise on] Word-Combination in Two Bam po; Skt. Madhyavyutpatti) of c. 812, which stipulates that translators should receive royal permission before translating or even collecting tantras.

This restriction was not entirely effective since we have a number of tantras attributed to early translators in the later canonical collections that do not appear in the royal catalogues from the Early Period. The restrictions were effective in limiting the types of tantras included in the royal catalogues, which were overwhelmingly related to the less antinomian tantric form known as Action Tantra (Tib. bya rgyud; Skt. kriya). However, the rising

tide of Indic tantra texts and their related practices were enthusiastically embraced by the Tibetans over the subsequent centuries, achieving a prominent place in the public and private lives of the Tibetans, as well as in the formal canonical collections of the Kangyurs and Tengyurs. Structure of the Canonical Collections of the Early Period: The Royal Catalogues

As noted above, three royal catalogues were produced during the Early Period, of which two are extant: the Lhan dkar ma of c. 812 and the 'Phang thang ma of c. 843-881. They confirm that the Three Baskets in the northern Sanskritic order of Sutra Pitaka, Vinaya Pitaka, and Abhidharma Pitaka served as their

main overarching organizational principle. The compilers of these catalogues had a highly developed, multileveled structure in mind when they created their catalogues, but they left virtually all the meta-categories unnamed. Instead, they formally labeled only the lowest, most granular levels of their hierarchy and these granular levels were at different levels of the hierarchical structure. The full structures of the two extant catalogues are thus not self-evident and it has sometimes been said that they are poorly structured.

Understanding the structure of the royal catalogues involves the identification of two recurring hierarchical typologies:

(1) the dyad of Mahayana (Great Vehicle) and Hinayana (Lesser Vehicle), and

(2) the Three Wheels of Dharma (Tib. chos kyi 'khor lo gsum; Skt. tridharmacakra) with

(a) the First Wheel consisting of the teachings of the Hinayana,

(b) the Second Wheel consisting of the Mahayana teachings related to the Prajnaparamita sutras (Perfection of Wisdom Discourses) and the Madhyamaka school (the Middle Way school), and

(c) the Third Wheel consisting of the Mahayana teachings of the sutras related to the Yogacara school (Yogic Practice school) and the Buddha nature tradition.

The Three Wheels typology presents the Third Wheel as the highest teaching, but this was naturally contested by proponents of the other two wheels, and the royal catalogues as well as the Kangyurs and Tengyurs consistently place the Second Wheel texts before the Third Wheel texts.

To describe the higher level schemas of the structure in the Lhan dkar ma, the Sutra Basket is first, within which the exoteric Mahayana sutras precede the Hinayana sutras, which precede the esoteric Mahayana tantras and dharanis (texts which consist of long mantras). The esoteric Mahayana texts are followed

by miscellaneous types of texts of praise, aspiration, and auspiciousness that conclude the Sutra Basket. Then there is the short Vinaya Basket, followed by the Abhidharma Basket, within which Mahayana treatises are again first, then Hinayana treatises, followed by secular treatises on Logic, and finally

texts attributed to the Tibetan king Khri srong Ide btsan. At the very end are two categories of work in progress, namely, for translations that have not been fully edited and translations that are not complete. These last two categories indicate that this was a catalogue of an actual collection of texts. There are a range of other structural principles used at lower levels in the Lhan dkar ma structure:

• Within the Sutra Basket, Second Turning sutras precede Third Turning and miscellaneous Mahayana sutras, which precede First Turning sutras. Also, Hinayana and tantra treatises are placed directly after their related sutras and tantras, rather than being placed with the other treatises of the Abhidharma Basket, as would be the case later in the Tengyurs. Anomalously, treatises related to the Mahayana sutras are placed at the very beginning of the Abhidharma Basket in the Lhan dkar ma, not after their Mahayana sutras in the Sutra Basket.

• Within the Vinaya Basket, ancillary Vinaya treatises are placed after their seven Vinaya root texts, rather than with the other treatises of the Abhidharma Basket, as again would be the case later in the Tengyurs.

• Within the Abhidharma Basket, within its Mahayana treatises, commentaries directly on the Mahayana sutras precede other Mahayana treatises, and within those two subsections, Second Turning Mahayana treatises generally precede Third Turning and miscellaneous Mahayana treatises. All of these Mahayana treatises precede First Turning Hinayana treatises. Buddhist treatises precede non-Buddhist treatises (on logic), except for Buddhist texts by Tibetans.

• Within any given category, translations from Indian languages precede translations from Chinese. It is evident that the Lhen dkar ma was intensely structured, though there are occasional inconsistencies. These schemas formed a canonical continuum of interlocking hierarchical principles, with a gradually descending scale of authority and sacredness that included not just the sutras and tantras of the Buddhas, but also the treatises by Indian and even Tibetan masters. The 'Phang thang ma catalogue was even more structurally complex, with seventy-eight named categories compared with thirty in the Lhen dkar ma.

Canonical Collections of the Later Period, Part I: The Kangyurs and lTengyurs

A distinctive feature of the Kangyurs and Tengyurs is their bifurcation of the Three Baskets, with the Kangyurs containing the expanded Mahayana version of the Sutra Basket as well as a traditional Hinayana Vinaya Basket, and the Tengyurs containing the expanded Mahayana version of the Abhidharma Basket. This bifurcation retained the overarching structure of the Three Baskets, but it was now distributed across two distinct collections. One result of this

separation was that pious donors funded numerous editions of the Kangyurs, while they funded far fewer editions of the Tengyurs. After the collapse of the Tibetan empire, a large portion of the early translations were preserved, but the royal collections were broken up into their component parts, which then circulated separately in the Later Period, with new translations being added to them. The earliest Kangyurs and Tengyurs were made up out of these circulating groups of texts.

The Tibetan tradition holds that the first Kangyur and Tengyur were the Old Snar thang Kangyur and Tengyur created at Snar thang monastery in the decade after 1310 CE. However, the available evidence suggests that the first Kangyurs and Tengyurs were actually produced earlier, by the middle of the second half of the thirteenth century when the terms Kangyur and Tengyur first appear in texts (Schaeffer 2009: 9-32). There is a reference to a Kangyur at Sa skya monastery in 1278 and a reference to Tengyurs in the plural related to events of the late 1270s. There is an extantTengyur-style catalogue by Bcom ldan ral gri (1227-1305), who received extensive training at Sa skya monastery and lived at Snar thang monastery for much of his adult life, though this is

apparently not a catalogue of an actual physical collection. There is a reference to the lung (oral transmission) of an actual Kangyur and Tengyur at Sa skya monastery in 1304. This evidence suggests that the masters and institutions of the Sa skya school had a major role in the formation of the earliest Kangyurs and Tengyurs.

A number of conditions led to the development of the formal Kangyur and Tengyur collections in the late thirteenth century. As noted above, the arrival of Indian Buddhist texts and translators began tapering off in the late thirteenth and early fourteenth centuries due to the Muslim conquest of northern and central India, making it possible to create comprehensive, stabilized canonical collections. In addition, after four hundred years of political

decentralization that followed the assassination of King Glang dar ma in 842, the Sa skya tradition was granted political power over central Tibet in 1264 as representatives of their new Mongol overlords. The Sa skya rule lasted almost ninety years, until 1350. This centralization of political power brought increased wealth to the Sa skyas that they used to support large-scale textual projects. The Sa skyas also

had a sectarian motivation to create formal canonical collections at this particular time, namely to restrict the Kangyurs and Tengyurs to translations of what they considered to be legitimate Indic texts, systematically excluding the growing number of esoteric texts of the Rnying mas that they viewed as suspect, as questionable creations of Tibetan authors. As a result, they followed the conservative principle of only including texts in the Kangyurs and Tengyurs for which there was proof that they were translations of authentic Indian texts, such as the existence of the Indic language originals.

This sectarian bias against these Rnying ma (and Bon) texts is evident in the writings of the major early Sa skya figure of Sa skya Pandita (1182-1251), as well as others, such asDbon Shesrab 'byung gnas (1187-1241), Chagslo tsaba Chosrje dpal (1197-1265), and Bu ston Rin chen grub (abbreviated to bu ston) (1290-1364), the last of whom finalized the structure that was used in all five extant Tengyurs. However, a tiny number of such Rnying ma texts (two to

six) were included in three Sa skya catalogues of the thirteenth century, which included the important Gsang ba'i snying po (Tantra of the Secret Essence: Skt. Guhyagarbha tantra) of Mahayoga (see below), for which the original Sanskrit text is reported to have been seen (Schaeffer 2009: 50, 181). A group of eighteen or nineteen Rnying ma tantras were also included later in the Rnying rgyud (Old Tantras) section of the Kangyurs of the Tshal pa line and in the Ulan Bator manuscript Kangyur of the Them spangs ma line (see below).

This distrust of the creativity of the Rnying mas or “Old School” was ironic since the commentarial traditions of the Gsar ma or “New Schools” that based their practical canons on the Kangyurs and Tengyurs一e.g., the Sa skya, Bkl rgyud, Dge lugs, and Jo nang schools一displayed significant interpretative

creativity in their own divergent traditions, following the creation of the first Kangyurs and Tengyurs. It is as if the virtual closing of their formal canons led to a countervailing hermeneutical opening for creativity in their commentarial traditions that blossomed and matured during the fourteenth through mid-sixteenth centuries.

Structure of the Kangyurs and Tengyurs Compared with the Chinese and Pali Canons

Since no two Kangyurs or Tengyurs are the same, the analysis of their structures will focus on major shared principles. As noted above, the Three Baskets have been an important schema in all three major canons, namely the Pali, Chinese, and Tibetan canons. They are the explicit main structure for the canons of Foundational Buddhism. This is the case even for schools that add other baskets. The formal canons of the Mahasamghikas, Dharmaguptas, and Bahusrutiyas

had five baskets, adding two subordinate baskets: the Samyukta (Mixed), Dharani, and/or Bodhisattva Pitaka. While the Three Baskets of the Foundational schools contained only texts of Foundational Buddhism, the Chinese and Tibetan traditions expanded the scope of the Three Baskets to include both the exoteric sutra texts and esoteric tantra texts of the Mahayana, as well as some non-Buddhist texts on subjects deemed useful, such as logic and medicine.

Chinese catalogues often arranged their texts according to the historical sequence of the translators and dynastic periods, but when they arranged the texts by their

content, they used one of two schemas as the primary structure: (1) the Three Baskets in the northern order of Sutra, Vinaya, and Abhidharma, or (2) the two vehicles in the order of Mahayana and Hinayana (with the esoteric tantra texts spread among the exoteric sutra texts of the Mahayana), while the other structure was commonly used as the secondary structure within each of the primary categories. This produced the following structures in Chinese catalogues:

Paradigm 1:

I. Sutra Basket

A. Mahayana Sutras

B. Hinayana Sutras

II. Vinaya Basket

A. Mahayana Vinaya*

B. Hinayana Vinaya

III. Abhidharma Basket

A. Mahayana Treatises

B. Hinayana Treatises

Paradigm 2:

I. Mahayana

A. Sutra Basket

B. Vinaya Basket*

C. Abhidharma Basket

II. Hinayana

A. Sutra Basket

B. Vinaya Basket

C. Abhidharma Basket

The Mahayana Vinaya category was developed in China, but not in the Tibetan collections

As with the Tibetan royal catalogues, Kangyurs, and Tengyurs, these meta-categories were not necessarily explicitly stated in the Chinese catalogues but can be inferred from the sequence of stated categories. Chinese Paradigm 1 with the Three Baskets as it main schema is fully present in the Fajinglu mulu of 594 CE. The Lidai sanbao ji of 597 CE fully embodies Paradigm 2, with the Two Vehicles as its main schema (Storch 1995: 41-44; Stanley 2009: 377-78).

The Tibetan royal catalogues and many Kangyurs-Tengyur pairs followed Paradigm 1, while some Kanjurs followed a major variant of Paradigm 1. Chinese Paradigm 1 with the northern Sanskritic sequence of Sutra, Vinaya, and Abhidharma will be referred to as Tibetan Paradigm 1A. The major Tibetan variation

in some Kangyurs reverses the first two baskets: Vinaya, Sutra, and Abhidharma, as noted above. This will be called Tibetan Paradigm IB. For the present, these two paradigms will be summarized as follows:

Tibetan Paradigm 1A:

Kangyur I. Sutra Basket

II. Vin日y日 Basket: Hin日yNn日 Only

Tengyur: III. Abhidharma Basket

Tibetan Paradigm IB:

I. Vin日y日 Basket: Hin日yNn日 Only

II. Sutra Basket

III. Abhidharma Basket

It is straightforward to describe the structure of all five extant Tengyurs, since they follow the sequence of textual categories established by Bu ston Rin chen grub in his Zhwa lu Tengyur of 1334, though they vary somewhat in their total number of texts and their placement of texts within textual categories. It is not straightforward to describe the structures of the Kangyurs, since their compliers have continued to introduce

changes in their structure and content. That has produced further variations in both paradigms 1A and IB, but these will not be dealt with here. Kangyurs can be classified into three groups: (1) the stemma line from the Tshal pa Kangyur of 1347-1351, (2) the stemma line from the Them spangs ma Kangyur of 1431, and (3) independent local Kangyurs. The first and third groups include only handwritten manuscript editions. The second group includes all

xylograph editions (editions printed from carved woodblocks), as well as manuscript editions. The Tibetan tradition holds that all the Kangyurs of the Tshal pa and Them spangs ma lines stem from the Old Snar thang Kangyur of circa 1310. Close textual analysis of texts in these two lines has revealed that they could not have descended from the same source and it is unclear if one of them actually descended from the Old Snar thang, which is not extant.

The Tshal pa Kangyur line has two main sub-lines: (1) the Peking Kangyur sub-line that began with the Yongle Kangyur of 1410, which was the first xylograph Kangyur, and (2) the Thying ba stag rtse Kangyur line. The date of the original Thying ba stag rtse Kangyur is unknown, but its earliest known descendent, the Li thang xylograph Kangyur, was produced in 1608-1614. There is additional complexity in the Tshal pa Kangyurs because some of them have been “cont日

min日ted” by an edition from a different stemma line or sub-line. For example, the editors of the Sde dge Kangyur of 1733, from Tshal pls Thying ba stag rtse sub-line, consulted a seventeenth century Lho dzong edition of the separate Them spangs ma line. The five extant Tengyurs that descended from Bu ston's 1334 Zhwa lu Tengyur divide into two lines. Three of them stem from a mid-seventeenth century Thying

ba stag rtse Tengyur that is not extant: the Peking xylograph Tengyur of 1724, which is the oldest extant Tengyur, the Snar thang xylograph Tengyur of 1741-1742, and the Golden or Dga* ldan manuscript Tengyur of the mid-eighteenth century The other Tengyur line consists of the Sde dge Tengyur of 1737-1744 and its descendant, the Co ne Tengyur of 1753-1744. Four of these Tengyurs can be paired with Kangyurs from the same location, all of which are xylographs.

Stated according to the completion dates of their Kangyurs, they are: the Peking Yongle Kangyur of 1410 and Tengyur of 1724, which are the oldest pair, the Co ne Kangyur of 1721-1731 and Tengyur of 1753-1772, the Snar thangKangyur of 1730-1732 andTengyur of 1741-1742, and the Sde dge Kangyur of 1733 andTengyur of 1737-1744.

To return to Paradigms 1A and IB in light of the three groups of Kangyurs, while a category of Hinayana sutras is explicitly identified in the royal catalogues, they are not separately identified in the Kangyurs. However, nearly all of these sutras were placed in the Mdo sde or Collected Sutras category

along with miscellaneous Mahayana sutras. In the Kangyurs of both sub-lines of the Tshal pa line, these Mahayana and Hinayana sutras are strictly separated from each other, with all the Mahayana sutras placed before the Hinayana (Stanley 2009: 644—47, 666—70, 6 7 5). In the extant Them spangs ma Kangyurs and the independent Phug brag Kangyur, the Hinayana sutras are scattered widely among the Mahayana sutras in the Collected Sutra category, though their

distribution is weighted toward the end. The structural implications of this will become apparent shortly.

In addition, there is a major difference between Paradigms 1A and IB regarding where they place the tantras in the Kangyur. In 1A, the tantras are placed at the very beginning with the Vinaya texts at the very end, whereas in IB the position of the tantras and Vinaya texts are reversed. This results in following structures:

Tibetan Paradigm 1A: Tibetan Paradigm IB:

Kangyur I. Sutra Basket I. Vinaya Basket: Hinayana Only

A. Mahayana: Esoteric Tantras II. Sutra Basket

B. Mahayana: Exoteric Sutras A. Mahayana: Exoteric Sutras

C. Hinayana Sutras B. Hinayana Sutras

II. Vinaya Basket: Hinayana Only C. Mahayana: Esoteric Tantras

Tengyur: III. Abhidharma Basket III. Abhidharma Basket

A. Mahayana: Esoteric A. Mahayana: Esoteric

Treatises Treatises

B. Mahayana: Exoteric B. Mahayana: Exoteric

Treatises Treatises

C. Hinayana Treatises C. Hinayana Treatises

The royal catalogues follow lAin placing the Sutra Basket before the Vinaya Basket, though they have a similarity to IB in placing the esoteric tantras after both the exoteric Mahayana and Hinayana sutras, in that order, which reflects royal caution toward the tantras. In the Later Period, the tantras rose to preeminence. It is often stated that the tantras do not surpass the highest exoteric view within the Mahayana, but they are superior in their skillful means for realizing that view. This superior skillful means is apparently reflected in the lAKangyurs that place the tantras as their first category in both the Sutra Basket and the entire collection, where they are followed in descending order by the exoteric Mahayana sutras and then the Hinayana sutras that round out the Sutra Basket, which is then followed by the Hinayana Vinaya Basket. This descending structure of the 1A Kangyurs is well organized according the doctrinal principles of Tibetan Buddhism of the Later Period, with the Hinayana sutras and Vinaya placed together at the end.

Paradigm IB is not as well organized in terms of such doctrinal principles: the Hinayana Vinaya Basket is followed by the exoteric Mahayana sutras, which then revert back to the Hinayana sutras at the end of the Collected Sutras, which is followed by another reversal to the esoteric Mahayana tantras. (Sometimes the Collected Sutras category, with both sets of its miscellaneous Mahayana and Hinayana sutras, was moved from the end of the Sutra Basket

toward the front of the Basket, nearer the Vinaya Basket but still among other Mahayana sutras, which produced even more structural anomalies). The reason(s) why the compilers switched the placement of the Vinaya and tantra categories of the Kangyur is not clear, but placing the Vinaya first may

reflect a desire to stress the importance of the Vinaya and to situate tantra within its overarching ethical framework. The evidence from the Tibetan royal catalogues and both Chinese Paradigms 1 and 2 suggests that China and Tibet initially inherited a Sanskritic Indian paradigm that placed the Sutra Basket before the Vinaya Basket, which is reflected in Tibetan Paradigm

1A. If this is correct, then Tibetan Paradigm IB, which places the Vinaya first, would represent a later Tibetan development. Further support for this suggestion is provided by the structure of all five extant Tengyurs, which, after an initial auspicious category of texts of praise, place the tantric treatises before the exoteric Mahayana treatises, which are followed by the Hinayana treatises with the Abhidharma treatises first and the Vinaya treatises last.

To correlate paradigms 1A and IB with actual Kangyurs, the Peking sub-line of the Tshal pa Kangyurs follows Paradigm 1A precisely The Stog or London Kangyur of the Them spangs ma line follows 1A except for an anomaly in its placement of the Collected Sutras. The Thying ba stag rtse sub-line of the Tshal pa stemma (e.g. the sNar thang, Sde dge, Lha sa, and Li thang Kanjurs) follows Paradigm IB, except its Co ne edition follows Paradigm 1A. However, Co ne and Li thang also have an anomaly in their treatment of the Collected Sutras.

The oldest extant pair of Kangyurs and Tengyurs一the Yongle xylograph Kangyur of the Peking sub-line of the Tshal pa Kangyur stemma and the Peking xylograph Tengyur一exemplify Paradigm 1A exceptionally well. The original Tshal pa Kangyur of 1347-1351 was produced after Bu ston produced his Zhwa lu Tengyur in 1334, the latter of which supports 1A. It may not be an idle fact, then, that Bu ston was invited to the consecration of the Tshal pa Kangyur.

It is possible that the structure of the Tshal pa Kangyur consciously complemented the structure of Bu ston's Tengyur in a manner consistent with the inherited Paradigm 1A. If so, then the Peking sub-line of the Tshal pa Kangyur stemma that exemplifies 1A may reflect the structure of the original Tshal pa Kangyur which is not extant, while the structure of the Thying ba stag rtse sub-line that exemplifies IB would have been a later development.

Below is a table of the structure of the Peking Kangyur and Tengyur with Paradigm 1A fully implemented. It also shows how these collections continued to use many structural principles from the royal catalogues. In this table, levels of the structure that are unstated but inferred from the stated textual categories or from unnamed contiguous blocks of related texts are placed in square brackets: “[xxx].”

Canonical Collections of the Later Period, Part II: The Rnying mas and Bon

Controversy over the Rnying ma tantras began in the late tenth century, at the start of the Later Period. Initially, the controversy was over the perceived heterodox nature of the Rnying ma tantras that emerged from the prior udarkM period, which the Rnying mas presented as translations from Indian sources

carried out in the Early Translation Period. Their critics claimed that at least some portions of these texts were Tibetan fabrications. This controversy only increased when the Rnying mas began to discover cycles of treasure texts in the early eleventh century The Bon treasure texts began to appear then as well.

The Rnying mas began gathering these texts into their own collections in the twelfth century, before the emergence of the Kangyurs andTengyurs. The Vairo rgyud 'bum (The Collected Tantras of Vairo[cana]) was an early Rnying ma collection, possibly from the

Structure of the Peking Kangyur and Tengyur Explicit and Implicit Categories

Structure of the Peking Yongle Kangyur

I. Translated Word of the Buddhas]

A. [[[Sutra]] Basket]

1. [[[Mahayana]]]

a. Tantra

b. [[[Sutra]]]

i. [Second Wheel:] Prajnaparamita

ii. [Third Wheel:] Avatamsaka

iii. [Third Wheel & Miscell日neous:] Ratnakuta

iv. [Third Wheel & Miscel:] Collected Sutras I

2. [[[Hinayana]]: First Wheel: Collected Sutras II]

B. l. [[[Hinayana]]:] Vinaya [Basket]

2. [[[Aspiration]] Prayers and Auspicious verses]

Structure of the Peking Tengyur

II. Translated Treatises [on the Word of the Buddhas]

=C.= 'Abhidharma Basket

1. Praises

2. [[[Mahayana]] Treatises] a. Tantra Commentaries

b. [[[Sutra]] Treatises]

i. [Second Wheel Treatises]

(1) Prajnaparamita [Commentaries]

(2) Madhyamaka School

ii. [Third Wheel & Miscel. Mahayana Treatises]

(1) Comm, on Miscel. [& 3rd Wheel] Sutras

(2) Cittamatra School

3. [[[Hinayana]]: First Wheel]

a. Abhidharma Treatises

b. Vinaya Commentaries

4. [Miscellaneous Buddhist Treatises]

a. Birth Stories

b. Letters

5. [Miscellaneous Non-Buddhist Treatises]

a. [Four Outer Sciences]

i. Logic

ii. Grammar

iii. Medicine

iv. Arts and Applied Crafts

b. Ethics and Statecraft

6. Marvelous Treatises (Includes texts by Tibet

Explicit Tibetan Categories

Bk日，'gyur

Rgyud

Sher phyin

Phal chen

Dkon brtsegs
do sde I
[Mdo sde II]

Dul ba

[[[Smon lam]], bKra shis kyi tshigs su bea
d pa]
Bstan 'gyur

Bstod tshogs

Rgyud 4grel

Sher phyin

Dbu ma

Mdo tshogs 'grel pa

Sems tsam

Mngon pli bstan bcos 'Dul ba4 'grel pa

Skyes rab

Spring yig

Tshad ma

Sgra rig pa

Gso ba rig pa

Bzo rig pa

Thung mong ba lugs kyi bstan bcos

Ngo mtshar bstan bcos

twelfth century It is extant and consists of eight volumes, 196 texts, and 4,023 pages. The first Rnying ma rgyud 'bum was probably created by Kun spangs

grags rgyal between the 11th and 13th centuries, though his dates are uncertain. This collection is not extant. As with the Kangyurs and Tengyurs, texts explicitly written by Tibetans were generally excluded from the Rnying ma rgyud 'bums, though some uvisionary

transla-tions” by Tibetans, which were tantras authored by Rnying ma masters that appropriated new tantric teachings and techniques from translations of the Later Period by assimilating this new material with elements of their Rnying ma tradition, were included (Germano 2002). The Rnying ma rgyud 'bums also

generally excluded treasure texts, which were included in their own treasure collections. Another Rnying ma rgyud 'bum was made in the thirteenth century by 'Gro mgon nam mkha, dp al ba, the son of Nyang ral nyi ma4 'odzer (1136-1204), while another was made in the fourteenth century by Zur Bzang po dpal and another in the fifteenth century by Ratna gling pa (1403-1478) and son.

It is not clear how many texts these collections contained. The eighteenth-century Sde dge edition has 25 volumes, 447 texts, and 17,835 pages. The online catalogue of Rnying ma rgyud 'bums of the Tibetan and Himalayan Library, based on four editions, lists 1,017 texts and over 45,432 pages (equivalent to 30,125 Peking pages) and 50 volumes. The structure of these collections is based on the three highest classes of tantra of the distinctive nine vehicle

doxographical system of the Rnying ma tradition, which are arranged in these collections in the descending order of Atiyoga or the Highest Yoga (Tib. shin tu rnal 'byor, also known as rdzogs chen, the Great Perfection), Anuyoga or the Subsequent Yoga (Tib. rjes su rnal 厂)，and Mahayoga or the Great Yoga (Tib. rnal 'byor chen po).

Separate collections of treasure texts were also formed in the twelfth century, such as the collection of Zhang ston bkra shis rdo rje's (1097-1167). Treasure collections are typically devoted to the treasure texts of a specific treasure revealer or lineage. Nyang ral nyi ma 'od zer (1124/1136-1204) was

the first major treasure revealer. Treasure texts have continued to be revealed in the Rnying ma tradition to this day. Of the large number of treasure revealers that have appeared over the centuries, five are raised up as the gter ston rgyal po Inga, the “Five Kings [among] the Treasure Revealers^: Nyangralnyi ma 'odzer, Guru Chos kyi dbangphyug (1212-1270), Rdo rje gling pa (1346-1405), Pad ma gling pa (1450-1521), and 'Jam dbyangs mkhyen brtse4

dbang po (1820-1892), all of whom are considered emanations of King Khri srong Ide btsan. The first two of these five lived in the period of the twelfth through thirteenth centuries, before the formation of the first Kangyurs andTengyurs, while the third lived afterward, in the fourteenth century. There was an explosion of treasure texts from the fourteenth century onward. There is no single collection that includes all the treasure texts, but the largest composite collection is the Rin chen gter mdzod, mentioned above, which was compiled by 丁日m mgon kong sprul bio gros mtha, yas (1813-1899) and consists of 111 volumes and an estimated 63,746 pages (equivalent to 42,080 Peking pages).

As indicated above, to create an estimate of the Rnying ma tantras not included in the Kangyurs, one could add up all the volumes and pages of (1) the Rnying ma rgyud

'bums and (2) the Rin chen gter mdzod. If one converts the different page sizes to their equivalent in Peking pages based on average number of syllables per page, this yields 161 volumes and the equivalent of 72,205 Peking Kangyur pages. This is an underestimate of such Rnying ma texts, but they are still almost four times (3.9x) the size of all the tantras in the Kangyurs, which have the estimated equivalent of 30 Peking volumes and 18,402 Peking pages of tantras.

According to the catalogue by Nyi ma bstan 'dzin dbang sdud rgyal po (b. 1813), the Bon collection of the Bkf or the “Word [of the Buddha Shenrab]M had 113 texts in 175 volumes and the Bkf brten or “Th日t which Relies on the Word [of the Buddha Shenrab]M had 293 texts in 131 volumes, for a total of 406 texts in 306 volumes. The Buddhist Kangyurs and Tengyurs have over 320 volumes and some 5,262 texts with 231,802 Peking pages, but we do not have number of pages

and average number of syllables per page for the Bon collections to make a meaningful comparison with these numbers, other than to note that the Bon collections were clearly large. As with the Rnying ma tradition, the Bon tradition has continued to generate treasure texts; a new Bkl edition c. 1987 has 192 volumes, 1,262 texts, and 61,416 pages, while a Bkf brten edition of 1985-1988 consists of a remarkable 300 volumes, 8,246 texts, and 207,667 pages, for a total of 492 volumes, 9,508 texts, and 269,083 pages.

Role of Canonical Texts and Collections in Tibetan Society

Paradoxically, while the texts of the Kangyur are held in the highest esteem by Tibetan Buddhists, they are in fact rarely studied directly The primary practical role of the Kangyurs is as a revered ritual object. The texts of the Kangyurs are placed on shrines as objects of reverence, whether as

individual texts or as a complete collection. The Kangyur or, more rarely both the Kangyur and Tengyur, are recited in their entirety on special occasions, multiple volumes being rapidly and simultaneously recited for the purpose of accumulating merit. Under these circumstances, comprehension of the texts is not the focus.

This is not to imply the activity is done mindlessly. Such recitations are a ritual affirmation of the entire canonical continuum that extends beyond the Kangyurs to the Tengyurs, and especially to the practical canon of the monastery in which the recitation is occurring. When the monks of a Dge lugs

monastery engage in the ceremonial recitation of an entire Kangyur, the quintessential meaning of these sutras, tantras, and treatises is understood as having been set forth in the works of their founder, Tsong kha pa (1357-1419), and in the works of the subsequent masters of their tradition. These Tibetan

works are primarily oriented to expounding and adjudicating the meaning of select treatises in the Tengyurs, which, in turn, are understood as revealing the essential meaning of the sutras and tantras in the Kangyurs. Reciting a Kangyur thus serves as an affirmation of the entire worldview of the Dge lugs tradition.

During times of personal difficulties, such as illness or injury, families might offer donations to monks to recite one or more sutras, frequently from the Perfection of Wisdom sutras, or even an entire Kangyur, with the hope that the merit accumulated would alleviate their circumstances. A village community might sponsor a procession that carries the volumes of a Kangyur around their village and fields, again with the intent of accruing merit, whether to bring favorable conditions in a future life or to help alleviate or avert a communal calamity such as drought or armed conflict (Childs 2005: 41).

While these are typical roles of a Kangyur in Tibetan society, the texts of the Kangy-urs retain their capacity as inspiring scriptures, as illustrated in the creative writings of Dolpo pa Shes rab rgyal mtshan (1292-1361). In setting forth the “Empty of Other” (Tib. gzhan stong) tradition, Doi po pa went

back to the sutras and tantras in the Kangy-urs to provide numerous quotes as the scriptural basis for what some masters viewed as a new, controversial vision of Buddhism. Such a scriptural basis was essential for establishing Doi po pls tradition in the face of the sharp criticisms and resistance toward his work. However, his extensive use of texts from the Kangyur represents the exception rather than the rule in Tibetan Buddhism.

As Tibetans made Buddhism their own, their understanding was shaped and transformed by an expanding body of indigenous Tibetan commentaries and treatises (and practice texts) that developed into an array of distinct Tibetan traditions. With the emergence of successive layers of Tibetan commentaries, the

Indian texts receded ever more into the background (Wilson 1996: 128). The result of this increasing reliance on indigenous Tibetan texts has been that not only are the Kangyur texts rarely read, but even the reading of the Tengyur treatises has given way to the reading of indigenous Tibetan texts for all but the most advanced Tibetan scholars.

This is not to imply that the sutras, tantras, and Indian treatises are without ongoing importance to the scholastic traditions. In the logic of scholastic treatises and commentaries, they derive their authority from the teachings they explicate in these very sutras, tantras, and treatises. This is expressed

by quoting such texts. However, the act of providing such quotations does not necessarily entail studying the sutras and tantras themselves; the Indian treatises and their Tibetan successors provided a pool of essential, ready-at-hand quotations culled from the scriptures over centuries by learned masters. Content of the Kangyurs and lengyurs Compared with the Chinese and Pali Canons

To compare the content of the three extant formal canonical collections from a Tibetan perspective of (1) the texts of the Hinayana, (2) the exoteric texts of the Mahayana, and

(3) the tantra-related texts of the Vajrayana or uIndestructible Vehicle” that are considered to be the esoteric texts of the Mahayana, we see that the Tibetan Kangyurs and Tengyurs collections have the smallest number of pages of Hinayana texts, but have the largest number of pages of both the exoteric and esoteric Mahayana texts, and constitute the largest canon overall. The Pali formal canon of the Three Baskets is the smallest of the three canons. It consists solely of texts of Foundational Buddhism. The Chinese formal canon is second in overall size. It appears that it has more

Foundational Buddhist texts than the formal Pali or Tibetan canons, but the Tibetan canon has more exoteric Mahayana and esoteric Vajrayana texts. However, the Chinese canon has more exoteric Mahayana sutras, while the Tibetan canon has more esoteric tantras and more treatises of both exoteric and esoteric Mahayana Buddhism.

The Pali canon is the only complete surviving set of the Three Baskets from a single school of Foundational Buddhism. The Tibetan tradition did not translate a complete set of the Three Baskets of Foundational Buddhism. The Tibetan king Rai pa can (r. 815-838) decreed that translators should only translate Hinayana texts from the Mulasarvastivada tradition. The Tibetans translated all of the Vinaya Basket of this tradition in the Early Period. The treasure text of the Pad ma bkaf thang, which is a hagiography of Padmasambhava discovered by 0 rgyan gling pa (c. 132 3—13 60), states that the seven

texts of the Abhidharma Pitaka were translated during the reign of Khri srong Ide btsan (r. 755—797). However, only one of these seven Abhidharma root texts, the Prajnaptibhasya, is included in the royal catalogues and the Tengyurs, where it is divided into three texts: Lokaprajnapti, Karanaprajnapti, and Karmaprajnapti (see texts P5587-89 in the Peking Tengyur).

As for the Sutra Basket, the Pad ma bkaf thang also states that all four of its sutra collections called Agamas or Transmitted TextsM (Tib. lung) were translated during the same reign. The Ekottarikagama or Numbered DiscoursesM is listed as translated in the Lhan dkar ma and the 'Phang thang ma royal

catalogues (DK2 74 and PT242) and an important historical document from the Early Period, the Dba9 bzhed (Testimony of the Ba Clan), which is a royal narrative about the history of Buddhism coming to Tibet, reports that the Dirghagama or “Long Length Discourses^ was translated in the Early Period as

well. The Tibetans made a concerted effort to translate the Mahayana sutras during the Early Period, which is when a remarkable 9 7.4% of the 51,576 Peking pages of all the sutras contained in the Kangyurs of the Later Period were translated. It is thus quite possible that the early Tibetan translators had undertaken the translation of all four Agamas of the Mulasarvastivada Sutra Basket, but, if so, none survived into the Later Period.

The Chinese canon apparently has more pages of Hinayana texts than the Pali canon, but more research is needed on this point. Though the Pali tradition reports more texts in its Sutra Basket than does the Chinese tradition, this does not necessarily mean that these sutras have more total pages than the

Chinese sutras. And though the Chinese canon does not contain a complete set of the Three Baskets from a single school, it contains the equivalent of more than a complete set because the Chinese translated multiple Vinayas and multiple Abhidharmas from different traditions, and at least one of each of the Agamas of the Sutra Basket. In any case, the Chinese canon is far larger than the Pali canon because it contains large quantities of exoteric Mahayana and esoteric Vajrayana texts.

A more precise comparison is possible for the Tibetan and Chinese canons. Overall, the Tibetan Kangyurs and Tengyurs have about 42% more pages of translations of Indic texts than the large early twentieth century Chinese Taisho edition (excluding its indigenous Chinese, Japanese, and Korean

treatises).2 The Taisho contain 56% more pages of texts of Foundational Buddhism than the Tibetan collections (the equivalent of 55,760 versus 35,809 Tibetan Peking pages, respectively). This reflects the fact that the Chinese translation of Buddhist texts began over six hundred years before major translation work began in Tibetan, namely, in the mid-second century CE versus the later eighth century CE in Tibet. In the second century CE, multiple

traditions of Foundational Buddhism flourished along the Silk Route, and a number of early masters who reached China belonged to these traditions. Virtually all the Indian masters who came to Tibet were Mahayana and Vajrayana masters. Traditionally, texts of Foundational Buddhism are integral to studying the Mahayana and Vajrayana, thus the Indian masters supported the translation of some texts of Foundational Buddhism in Tibet, but to a more limited extent than occurred in China.

Both the Chinese and Tibetans translated more pages of exoteric Mahayana works than of Hinayana works. The first major translator of Mahayana texts in China, Lokaksema, worked between 168 and 188 CE, which was only twenty years after the first major Hinayana translator in China, An Shigao, who translated

between 148 and 170 CE. Thus the Chinese began translating both Hinayana and Mahayana texts six hundred years earlier than the Tibetans. However, the Tibetans ended up translating 29% more pages of Mahayana sutras and treatises than the Chinese (114,462 versus 88,809 Tibetan pages, respectively). As for the esoteric Vajrayana tantras and their related treatises, in the Taisho edition, Vajrayana texts account for only 12% (19,136 Tibetan pages) of

all its pages of Indic texts (163,705 total Tibetan pages). By contrast, in the combined Kangyurs and Tengy-urs, Vajrayana texts account for 35% (81,532 Tibetan pages) of the 231,802 total Tibetan pages in the combined Kangyurs and Tengyurs. The total number of pages of Vajrayana texts in the Taisho thus equals only 23% of total pages of such texts in the combined Kangyurs and Tengyurs.

Another distinguishing feature between the Tibetan and Chinese canonical collections is that 72% of the total pages of Indic texts in the Taisho are sutras and tantras and 28% are treatises and commentaries, whereas only 31% of the total pages of the Tibetan texts are sutras and tantras and 69% are treatises

and commentaries. The Chinese translated the equivalent of 64% more pages of sutras and tantras than the Tibetans (118,430 vs. 72,345 Tibetan pages, respectively), whereas the Chinese only translated around 28% of the pages of treatises and commentaries translated by the Tibetans (45,274 vs. 159, 458 in Tibetan pages, respectively).

Reflections on Canonicity, Scriptures, and Treatises in Buddhism

To summarize, formal Buddhist canons include not just the sutras and tantras attributed to the Buddhas, but also the treatises and commentaries by subsequent Buddhist masters. This is the case for even the Pali canon in which the Vinaya and Abhidharma Baskets are scholastic texts developed by Buddhist masters over centuries. The formal Tibetan and Chinese canons include treatises and commentaries by Buddhist Indian masters as well as some treatises by

regional masters. They even include useful treatises by non-Buddhist Indian masters. In the northern Sanskrit model of the Three Baskets, the texts of the formal canons are arranged in hierarchical order along a canonical continuum with the sutras (and tantras) above the treatises. Other hierarchical principles used to further subdivide this canonical continuum in the various Tibetan and Chinese collections include the principle that tantras may be

placed before sutras, Mahayana texts before Hinayana texts, Second Wheel texts before Third Wheel texts (before First Wheel texts), Buddhist Indian treatises before non-Buddhist Indian treatises, and so on. All of these treatises are included in the canonical continuum, though there is an increasingly attenuated sense of canonic-ity at the lower end.

This is a very different conception of the canon than found in Christianity, which has had at times an unconsciousness effect on scholars of Buddhist studies. Once the Christian canon was closed, perhaps as late as the fourth century CE, scripture and canon became virtual synonyms for that tradition.

This entailed a sharp divide between the scriptures in the canon and all the religious texts outside the canon. The canon was understood to consist of just scriptures, all the scriptures were understood to have been included in the canon, and all other religious texts were thus neither scriptures nor canonical.

The books included in the Hebrew Bible and the New Testament are understood as having been directly inspired by God. The complex historical process through which these collections were closed, or at least stabilized, is traditionally viewed as having been guided by God. The texts included in these two collections are thus considered to be intrinsically different and separate from all other Christian works. Compared with Buddhist traditions, the Christian traditions have a much more attenuated sense of cano-nicity for the portions of their practical canons and inclusive canons that are outside their formal canons.

There is a doctrinal basis for these two different views of canonicity in Christianity and Buddhism. The general view in Buddhism is that all beings, not just humans, can eventually attain the highest state of realization, whether that realization is understood in terms of Arhatship or Buddhahood. It is

proposed, then, that there is no fundamental doctrinal basis for drawing a sharp line between works attributed to the Buddhas and works by other Buddhist masters. Instead, there is a continuum of canonicity extending from the works of the Buddhas down through the works of Buddhist masters and even below them. By contrast, in the general Christian tradition, humans can aspire to be Christ-like, but they cannot become a Christ or a part of the Godhead that

consists of the Trinity This immutable distinction between humans and the divine is reflected in the sharp line between the Christian canonical scriptures inspired and selected by God and all other non-canonical religious texts.

This Buddhist doctrinal view is expressed in the classic Mahayana text of the Uttaratantrasastra attributed to Maitreya. He states that treatises by appropriately trained masters are to be revered as if they were the word of the Buddha himself (for the Tibetan: Sde dge Tengyur text D4024, vol. Phi/123, 72.2.5- 6):

Whatever has been explained by those with undistracted minds

Only in terms of the teachings of the Victorious Ones

[And that] accords with the path for attaining liberation,

That should be placed on the top of one's head as the word (bka") of the Sages [i.e. the Buddhas].

The reverence paid to the works of such Theravada masters as Buddhaghosa and Bud-dhadatta (fifth c.) that are outside the traditional Three Baskets of the Theravada indicates that the canonical continuum of the Pali tradition extends outside its formal canon as well. Their works are part of the Theravada^s practical canon.

In the Mahayanasutralamkara, Asariga articulates the expansive Mahayana view that its canon includes even non-Buddhist Indic works that are beneficial to beings. He states: “Without having striven in the five sciences/Even the supreme Noble Ones will not attain omniscienceM (for the Tibetan: D4020, vol.

Phi/123,176.1.3). That is to say, to become enlightened, one must master the five sciences, which consist of the four outer sciences of logic, grammar, medicine, and the arts and applied crafts, along with the inner science of Buddhism. The logical extreme of this inclusive view of canonicity is reached in

the Mahayana Sutra of the Adhyasayasamcodana, which states that uWhat-ever is well said (Skt. subhdsita; Tib. legs par smras ba), has been spoken by the BuddhaM (Gomez 1987: 535—36). This is a rephrasing of the traditional statement in Foundational Buddhism that Whatever is said by the Buddha is well saidM (Ahguttara Nikaya IV 163).

If the term uscriptureM is understood in the Christian sense of referring to the sacred texts inspired and chosen by God for inclusion in the Old and New Testaments, the analogous use of the term in Buddhism would be to apply it to the sutras and tantras attributed to the Buddhas. When scholars refer to the Kangyur as “the canon of Tibetan Buddhism,M it appears that they have implicitly accepted the Christian concept that the canon equals the scriptures and

that the term u scriptureM should be used in the restrictive sense of applying only to the texts attributed to the Buddha. However, the discussions above indicate that it is not appropriate to restrict the Buddhist canon solely to the texts attributed to the Buddha, since this leaves out the canonical treatises.

How then are we to use the three terms “scripture,” “treatise,” and ucanonM in Buddhist contexts? There appear to be two options: (1) retain the Christian practice of treating “scripture” and ucanonM as virtual synonyms, in which case “scripture” has to be redefined to include treatises as well as sutras and tantras, or (2) retain the Christian sense of the term “scripture” as referring solely, in Buddhist contexts, to the sutras and tantras attributed to the

Buddha, in which case “scripture” and ucanonM can no longer be treated as virtual synonyms. Option (1) is the more provocative approach because it includes treatises in the scriptures, and thus it might be more skillful in underscoring the differences between the Christian and Buddhist concepts of canonicity There is at present no consistent usage of the terms scripture and canon in the field, or even awareness of the issue. Time will tell if a consensus will emerge.

With regard to what traditional Tibetan vocabulary might be analogous to the English terms canon, scripture, and treatise, Bu ston Rin chen grub's History of Buddhism deals extensively with categorizing Buddhist texts and the broadest term he sets forth for the verbal expression of the Buddhist teachings is

gsung rab (highest speech; Skt. pravacana). This term has three synonyms: lung gi chos (udharma of transmitted words”； Skt. agamadharma), bshad pa'i chos (udharma of explanation^; Skt. desanadharma), and rjod byed (“the means of expression^; Skt. vacaka). He divides “highest dharmaM into two sub-classifications:

(1) the bkaf or “word” of the Buddhas (Skt. ajna) which includes the sutras and tantras and

(2) the bstan bcos or “treatises” of other Buddhist masters (Skt. sdstra).

If “scripture” is understood broadly, per option

(1), as virtually equivalent to “c日 non,” then it is suggested that “highest speech” or gsung rab could be used as analogous to the term “scripture,” while its synonyms such as udharma of transmitted words” and udharma of explanation,M could be used as analogous in meaning to ucanon.M As for the term “treatise,” this is the actual meaning of the Tibetan term bstan bcos. If “scripture” is used restrictively, per option

(2), as applying to just the sutras and tantras and not the treatises, then the term “word [of the Buddhas]M or bkaf could be used as analogous in meaning to the term “scripture.” In that case, “highest speech” (and its equivalents) could be used as analogous to the broader category of ucanonM that includes the treatises as well as the sutras and tantras.

Modern Developments: Exposure to Other Traditions and to Electronic Resources

The expansion of world travel and communication in the nineteenth and twentieth centuries brought increased contact among and within the world's religions. This manifested in dramatic fashion with the first Parliament of the World's Religious held in 1893 in Chicago with representatives from the Buddhist Zen

and Theravada traditions, as well from the Indian Jain and Vedanta/Yoga traditions. Though the impact has been slow, such contact has begun to affect how the Buddhist traditions understand their canons, and has even begun to change the content and structure of some editions. Such cross-tradition contact was

late in coming for the Tibetans, forced upon them by the Chinese invasion of Tibet in 1950 and the ensuing flight of many Tibetan masters from the country. By the late twentieth century, some Tibetan masters had become aware that Tibetan Buddhism was being criticized by East Asian Mahayana Buddhists and their

Western students for its reliance on commentaries rather than on the sutras. Khenchen Thrangu Rinpoche, who was the head abbot and leading scholar of the Karma Bkl brgyud^s monastic university in the Diaspora, at Rumtek monastery in Sikkim, argued in defense of the Tibetan orientation to the treatises, concluding “Thus the treatises are more important than the words of the Buddha ...” (Thrangu 2004: 95; 1992: 55-57). Khenpo Tsultrim Gyamtso Rinpoche, also

of the Rumtek monastic university, introduced the practice of directly studying sections of sutras in his teaching. The brilliant and unconventional master, Dge 'dun chos 'phel (1903-1951), translated the Dhammapada from Pali into Tibetan. Dharma Publishing then included his translation in its uNyingmaM edition of the Kangyur and Tengyur published in the United States in the early 1980s. This edition also demonstrated an increased interest in the Chinese canon by including nine Chinese texts: all seven root texts of the

Mulasarvastivada Abhidharma Basket, only one of which is included in the Tengyurs as discussed above, plus the related Mahavibhasa treatise and one Mahayana text, the Mahdprajndpdramitdsastra attributed to Nagarjuna.

In 1949, the Chinese monk Fazun (1902-1980) completed the monumental task of translating the Chinese translation of the Mahavibhasa into Tibetan, which was published separately in ten volumes in 2 011. Fazun worked primarily as a translator and teacher of Tibetan Buddhist texts in Chinese, including works from the Tibetan formal canons and from major Tibetan masters, such as Tsong kha pa and Klong chen pa (1308-1363). This illustrates a new Chinese interest in the Tibetan canon.

This nascent mutual interest between traditions may lead to direct expansions of their formal, practical, and inclusive canons. Such an expanded inclusive canon is already a physical reality in the university libraries around the world that collect the formal and practical canons of Buddhism. The research of

university scholars is also impacting how new editions of the canons are being formed. The compilers of the Chinese Taisho canon (1924-1934) used an innovative structure that placed Hinayana sutras first, before the exoteric Mahayana sutras, with esoteric tantras grouped separately afterward, based on a

new understanding of their historical sequence of development. The compilers also included texts from the Dunhuang caves in Central Asia initially discovered by Western explorers, and they also expanded the number of indigenous Chinese, Korean, and Japanese treatises, illustrating again the Buddhist conception of canonicity as an inclusive continuum.

The inclusive canon is also becoming a virtual reality through electronic projects for the Buddhist canons that are eclipsing the university libraries. For the Tibetan tradition, there are electronic projects for multiple editions of the Kangyurs, Tengyurs, Rnying ma rgyud 'bums, and texts by indigenous

Tibetan masters. See the Tibetan Buddhist Resource Center (http://www.tbrc.org), Catalogues of the Kangyur and Tengyur at the Tibetan and Himalayan Library (http://www.thlib.org/encyclopedias/literary/canons/kt/), Resources for Kanjur and Tanjur Studies (http://www.istb.univie.ac.at/kanjur/), and the Asian Classics Input Project (http://www.asianclassics.org/).

Such projects are also correlating their texts with texts across the canonical languages: Pali, Sanskrit, Chinese, Tibetan, Mongolian, and so on. This will make it possible to link all the canons together in a single catalogue. Discussions are underway to link these projects through a single Union Catalogue of

Buddhist Texts. It would then be possible to link all modern translations of these texts through such a catalogue. This collaborative work is the most expansive expression possible of the concept of an inclusive Buddhist canon.

There are also major projects to translate entire traditional canons into modern languages, especially English. The “84000: Translating the Words of the BuddhaM project has a 100-year vision that includes translating the entire Kangyur and Tengyur into English and other modern languages (www.84000.co).

There is a similar project for the Taisho (http://www.bdk.or.jp/english/index.html). Both projects are making their translations available online. Much of the formal Pali canon is already translated into English. This creates the possibility of a comprehensive virtual Buddhist canon in English for all three of these canons. This could then lead to comprehensive projects to translate the canons into other languages, perhaps even back into canonical languages

such as Tibetan and Chinese. The Tibetan Buddhist canon is entering a new, dynamic phase that may enlarge it, but also absorb it into a global Buddhist canon.

Notes

1 The Peking page data reported here are estimates based on combining all the texts found in six Kangyurs (Peking, Co ne, Sde dge, Urga, Snar thang, and Lha sa) and four Itengyurs (Peking, Co ne, Sde dge, and Snar thang), with all page sizes converted to Peking pages based on an average page ratio for all texts each edition shares with the Peking edition. The Peking Kangyur and Itengyur page size was selected as the unit of measure for all texts because they are the only pair of collections that use the same page size. It is not expected that adding other editions would substantially change the findings presented here, since it is likely that their additional unique texts would be small in number and size.

2 The comparative number and percent of pages of Indic texts translated into Chinese and Tibetan were created by calculating the average number of Peking Kangyur and Itengyur pages per page of the Chinese Taisho edition for texts shared between the two editions, namely, 5.319 Tibetan Peking pages per

1.0 Chinese Taisho pages. Then the sizes of these Taisho texts were converted into their estimated equivalent number of Tibetan Peking pages. There are Chinese translations not included in the Taisho and Tibetan translations not included in the Peking Kangyur and Itengyur, both of which are not included in the numbers reported here; thus, this data can be refined furthe匚

References

Childs, Geoff. "How to Fund a Ritual: Notes on the Social Usage of the Kanjur (Bka' 'Gyu】) in a Tibetan Village.^^ Tibet Journal 30/2 (2005):41-48. Germano, David. History and Nature of the Collected Tantras of the Ancient. Available at http://www.thlib.org/encyclopedias/ literary/canons/ngb/# wiki=/ access/wiki/ site/6a0a05fe-de5e-40f7-80f2-lccfa2ab7d 8e/history%20of%20ngb.html. March 25, 2002. Tibetan and Himalayan Library. Accessed March 17, 2008. Gomez, Luis 0. "Buddhist Literature: Exegesis and Hermeneutics.M in Mircea Eliade, ed. Encyclopedia of Religion. Vol. 2. New York: MacMillan, 1987, pp. 529-40.

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 Idan ral gri. Harvard Oriental Series. Cambridge, MA: Harvard University Press, 2009.

Stanley, David Phillip. "The Threefold Formal, Practical, and Inclusive Canons of Tibetan Buddhism in the Context of a Pan-Asian Paradigm: Utilizing a New Methodology for Analyzing Canonical Collections.M PhD dissertation, University of Virginia, 2009.

Storch, Tanya. "Chinese Buddhist BibliographyM PhD dissertation, University of Pennsylvania, 1995.

Thrangu Rinpoche. Tibetan Buddhist Meditation. Boulder, CO: Namo Buddha Publications, 1992.

Thrangu Rinpoche. Essentials of Mahamudra: Looking Directly at the Mind. Boston, MA: Wisdom Publications, 2004.

Wilson, Joe Bransford. "Tibetan Commentaries on Indian Shastras." In Jose Ignacio Cabezon and Roger R. Jackson, eds. Tibetan Literature: Studies in Genre. Ithaca, NY: Snow Lion Publications, 1996, pp. 125-37.

Source

The Tibetan Buddhist Canon