Old Tantras (Kangyur Section)

  • Skt.: Pratantra
  • Tib.: rnying rgyud/

First published 2023. Last updated 6th Jun 2024.

Seventeen works representing a small selection of the many “inner” class tantras of the Ngagyur Nyingma (“earlier translation”) tradition (Toh 828-844).


This section of the Kangyur contains a small selection of the tantras belonging to the three main “inner tantra” categories and translated in the early, “imperial” period of translation in Tibet (i.e. the eighth and early ninth centuries, prior to the mid ninth century disintegration of the empire).

Although many such tantras had been propagated among a restricted circle of tantric practitioners associated with what later came to be known as the “ancient tradition of the early translations” (snga ’gyur rnying ma), they had not been as well documented as the large number of texts of other categories (sūtra, vinaya, dhāraṇī, eulogy, treatise, commentary, etc.) that had also been translated in the imperial period.

Most of the texts of these inner tantra categories were criticized as inauthentic by many of the scholars‍—mostly affiliated with the “later translation” traditions‍—who were responsible for compiling the various early Kangyurs and Tengyurs. Their objections regarding these tantras were in marked contrast to their acceptance of the early translations of all other textual genres. The limited selection in this section comprises those few tantric works that at least some later scholars (though not all) felt could be acknowledged as canonical.

Despite being far less numerous than the works collected independently of the Kangyur and Tengyur in the (varying) compilations known as the Nyingma Gyübum, the tantras in this section are at least partially representative of the three main divisions of the inner tantras according to the Nyingma doxography: Atiyoga, Anuyoga, and Mahāyoga.

History and Controversy During the Period of Canon Formation

Some hundreds of tantras are said to have been translated from Sanskrit and other languages during the early, imperial period under the royal patronage of King Tri Songdetsen and his successors. However, the texts of the three “inner tantra” categories were treated with great secrecy, and were not listed along with the other, non-esoteric texts translated at the time in official inventories such as the Denkarma (ldan dkar ma). Nor, because of their perceived sanctity, were they submitted to the process of terminological and orthographical standardization that took place in the early ninth century. When the imperial period with its centralized patronage of Buddhist institutions and scholarship came to an end in the mid ninth century (traditionally linked to the reign of Langdarma1), lineages of the transmission and practice of these tantras were maintained during the century of rebellions and general disintegration that followed, and have been maintained down to the present day. But in the eleventh century a different set of tantras‍—those to be found in India at the time‍—began to be introduced to Tibet by Rinchen Zangpo, Drokmi Lotsāwa, Marpa Lotsāwa, and others. These formed the beginning of the “new translations” (gsar ’gyur) of the “later spread” of the teachings (phyi dar) and became the tantric corpus of the new traditions that gradually emerged and developed over the centuries that followed.

The thirteenth and fourteenth century scholars responsible for the major initiative of the times‍—to establish canonical collections of the Tibetan translations by cataloguing, and then compiling, works of all genres into the systematic collections that became the Kangyurs‍—were scholars whose allegiance was to the traditions of the new translations, and they had much less interest in the tantras of the early period.

The early canons were compiled on the foundation of the inventories and dkar chags of canonical texts assembled and compared by Chomden Rigpai Raltri (1227–1305), his contemporary and student Üpa Losal Sangye Bum, and Butön Rinchen Drup (1290-1364). Regarding the Nyingma tantras, the opinions of these three key figures differed somewhat:

Chomden Rigpai Raltri’s inventory‍—which in any case was not an exhaustive survey of all existing tantras, whether old or new‍—from the old tantras included only the Guhyagarbha, not only because he had himself seen one Sanskrit manuscript and heard about another, but also because he had determined its content to be genuine on several criteria.2

Üpa Losal, who is thought to have compiled the dkar chags of the now lost “old Narthang” Kangyur and Tengyur is also thought to have formulated, for that Kangyur dkar chag, a short list of doxographically representative Nyingma tantras with relatively well verified provenance.3

Butön, despite having himself received many Nyingma transmissions, famously decided to remain agnostic and not to include, in his influential inventory of translated canonical texts, any Nyingma tantras at all.4

The overall tendency to exclude the old tantras from the Kangyurs was partly based on reasonable doubts regarding the authenticity of the Indian sources of some of them. Indeed it is no surprise that the new wave of Tibetan translators, visiting India and Nepal many centuries after the early period, and largely different regions, found little evidence of the tantric traditions encountered elsewhere by their forbears, and even if searching for such evidence had been of greater concern for them than it actually was, they might have been unlikely to find much.

Moreover, the criteria that proponents of the later tantra translations drew up to determine “authenticity” were themselves a set of parameters partly determined by the dialectic between the old and new traditions itself. Although in some respects objective and scholarly, they were not designed to take full account of the rather different approaches to textual and cultural adaptation that shaped the values of the earlier translations, and indeed even those of the Indian tantric communities from which the tantras, old and new, had been inherited.5

The scholars concerned with compiling canonical collections accepted wholesale the work of the early translators regarding all the textual genres other than the higher tantras; their reluctance to do the same regarding this category of tantra stands out in significant contrast. Overall, it would be difficult to conclude that their concerns were purely objective, and that they were not also influenced by partisan considerations, over-zealous reformist ideals, and even politics. In the struggles of the post-imperial period, edicts and arguments against the practice, the practitioners, and in many cases the very texts of the early tantras‍—some strikingly polemical‍—had been written by critical scholars including Lha Lama Yeshé’ö (947-1024) the king of Gugé in western Tibet, the translator Rinchen Zangpo (958–1055), Yeshé’ö’s grand-nephew the Tholing translator Zhiwa’ö (1016–1111), and the 11th century translator Gö Khukpa Lhetsé, all of whom were cited by Butön to support his decision.

Despite these strong influences disfavoring the Nyingma tantras, the limited selection of the most easily authenticated early tantras thought to have been first drawn up by Üpa Losal (see above) materialized as a three volume section included as part of the Kangyur that was compiled and printed at Tshal Gungtang, of which the original is lost but which is known as the Tshalpa manuscript. When the Tshalpa Kangyur was completed in 1351, Butön himself consecrated it, apparently raising no objections to the inclusion of these Nyingma tantras despite his own expressed preference for excluding them.

The Tshalpa Kangyur set one of the two major structural patterns for subsequent Kangyurs, and on it were modeled the xylograph Kangyurs of the Peking series beginning with early fifteenth century Yongle, the early seventeenth century Lithang Kangyur, the Choné, Narthang, Degé, and Lhasa Kangyurs.

However, while the other Kangyurs of predominantly Tshalpa lineage all included this limited section of Nyingma tantras,6 it is noteworthy that the Degé at first did not. The Nyingma tantra section was missing and is not mentioned at all in Situ Panchen’s dkar chag prepared for the first printing in 1733, and even when more texts were added in a later printing in 1762 the Nyingma section was left out. Only much later, at some point in the mid-eighteenth century‍—long after the death of both Situ Panchen and his successor‍—were these volumes added, apparently at the instigation of Jamyang Khyentse Wangpo (1820–1892).7 It is unclear why it was omitted in such obvious contrast to other existing Kangyurs, and for so long‍—especially in light of the fact that in 1794 the Degé printery had published a full, twenty-four volume Nyingma Gyübum collection (the only extant xylograph version, compiled and edited by Getsé Mahapandita on the basis of Jigme Lingpa’s Padma Öling dkar chag).

The Kangyurs of the other major lineage are those that follow the alternative pattern set by the Themphangma manuscript, produced in Gyantse in 1431 and probably derived from an earlier Kangyur compiled at Zhalu. These manuscript Kangyurs, among them the Stok Palace and Shey Kangyurs, do not include a Nyingma tantra section at all.

The Works in This Section

The works found in this Old Tantras section comprise only some of the principal texts representing the three classes‍—Mahāyoga, Anuyoga, and Atiyoga‍—into which the Nyingma tradition divides what it designates the inner tantras (nang rgyud), the equivalent of the Anuttara/Niruttara class of the Sarma traditions):


Atiyoga is represented here only by:

Toh 828, The All-Creating Sovereign (kun byed rgyal po), from the Mind Class (sems sde) which, of the three main subdivisions within Atiyoga, is the doxographically (and probably historically) inaugural stratum.


Anuyoga is represented by:

Toh 829, known best as The Sūtra That Gathers All Intentions (mdo dgongs pa ’dus pa) but here with another title derived from its root, Toh 831 (see below);

Toh 830, The Magnificent Lightning Wheel of Awareness (ye shes rngam pa glog gi ’khor lo), although in some classifications this is categorized as a Mahāyoga text;

Toh 831, The Sūtra of All-Gathering Awareness (kun ’dus rig pa’i mdo).


The remainder of the texts in the section, Toh 832-844, are a representative but limited selection of the much more numerous Mahāyoga tantras to be found in the complete Nyingma Gyübum collections. They are classified into tantras (rgyud sde) and means for attainment (sgrub sde):


The first group, the tantras, comprises a basic cycle of 18 tantras derived from the major corpus of tantra texts known as the Magical Net, the Māyājāla (sgyu ’phrul dra ba). To this group belong:

Toh 832, the twenty-two chapter version of the tantra The Secret Nucleus, the Guhyagarbha (rgyud gsang ba’i snying po), said to present all things as naturally manifesting, or to present mind and wisdom as naturally manifesting, and traditionally considered to be the primary text of the cycle (note that the Guhyagarbha, classified here as a Mahāyoga tantra, is also interpreted by some commentaries in Atiyoga terms or described from a hybrid perspective sometimes called “Mahā-Ati”);

Toh 833, The Magical Net of Vajrasattva (rdo rje sems dpa’i sgyu ’phrul dra ba), said to present the hundred deities of the maṇḍala and their details such as body color and hand implements;

Toh 834, called the “eighty chapter” version of The Secret Nucleus, the Guhyagarbha (rgyud gsang ba’i snying po), although it has eighty-two chapters; it is said to emphasize the buddha qualities, and differs from Toh 832 principally in its greater detail concerning the wrathful deities of the maṇḍala;

Toh 836, The Tantra of the Great Magical Net of the Goddess (lha mo sgyu ’phrul dra ba chen po), an explanatory tantra for the group;

Toh 837, The Tantra of the Great Guru, from The Secret Nucleus Definitive With Respect to the Real (gsang ba’i snying po de kho na nyid nges pa’i bla ma chen po), a tantra emphasizing empowerment.

Also placed here within the tantra category, but not part of the Māyājāla corpus, is:

Toh 835, The Noose of Methods (thabs kyi zhags pa), classified as a supplementary tantra. The version in the Degé and other Tshalpa Kangyurs is a reconstructed and incomplete version of the tantra, of which more complete versions are found in some peripheral Kangyurs.

Means for Attainment

The second group, the means for attainment, is classified according to the meditational deities whose practice is described in each text. First come the five deities of the supramundane:

Toh 838, The Secret Tantra of the Wheel of Mañjuśrī’s Four Activities (’jam dpal las bzhi ’khor lo gsang ba’i rgyud), on Mañjuśrī-Yamāntaka (body);

Toh 839, The Great Tantra of Aśvottama’s Display (rta mchog rol pa’i rgyud chen po), on Hayagrīva-Aśvottama (speech);

Toh 840, The Most Profound Secret Tantra of Śrī Heruka’s Compassionate Display (dpal he ru ka snying rje rol pa’i rgyud gsang ba zab mo’i mchog), on Śrīheruka (mind);

Toh 841, Eight Great Sections [starting with] the Quintessence of Great Accomplishment Whereby All Things Have the Nature of the Five Nectars (thams cad bdud rtsi lnga’i rang bzhin dngos grub chen po nye ba’i snying po bam po chen po brgyad pa), a text divided into eight distinct chapters or sections, on Vajrāmṛta (qualities);

here numbered Toh 841A but not cataloged in Toh, A Fragment from the Vajrakīlaya Root Tantra (rdo rje phur pa rtsa ba’i rgyud kyi dum bu), the fragmentary text representing the practice of Vajrakīla (activity), widespread in the Nyingma tradition.

Then the three deities of the mundane:

Toh 842, Tantra of the Flaming Ḍākinī (mkha’ ’gro ma me lce ’bar ba’i rgyud), on Mātaraḥ;

Toh 843, Root Tantra of Vajramantrabhīru (drag sngags ’dus pa rdo rje rtsa ba’i rgyud), on Vajramantrabhīru;

Toh 844, The Tantra of Lokastotrapūja (’jig rten mchod bstod sgrub pa rtsa ba’i rgyud), on Lokastotrapūjā.

Nyingma Tantras Elsewhere

Note that there are a few inner tantras translated in the early period found not here, in the Old Tantra section, but in the main Tantra Collection, as they are shared by the Nyingma and Sarma traditions (though often using different translations). They include the Mañjusrīnāmasaṃgīti (Toh 360), the Guhyasamāja (gsang ba sdus pa, Toh 442), and the Māyājāla (sgyu ’phrul dra ba, Toh 466).

The separate, much larger collection of tantras considered canonical by the Nyingma tradition, the Nyingma Gyübum, exists in several versions (and translations of its works will, it is hoped, be added to the 84000 collection).

Bibliography and Further Reading

Tibetan Sources

Chomden Rigpai Raltri (bcom ldan rig pa’i ral gri). bstan pa rgyas pa rgyan gyi nyi ’od . BDRC MW00EGS1017426 (modern computerized version) and MW1CZ1041 (scanned dbu med MS from Drépung).

Jigme Lingpa (’jigs med gling pa) (?). snga ’gyur rnying ma la rgol ngan log rtogs bzlog pa’i bstan bcos kun mkhyen ngag gi dbang pos mdzad pa [“The Omniscient Ngaki Wangpo’s treatise refuting negative statements and misunderstandings concerning the Nyingma tradition of the Early Translations”]. In gsung ’bum/ ’jigs med gling pa (a ’dzom par ma), vol. 6, cha (171 ff). BDRC MW7477.

Loter Wangpo (blo gter dbang po). rgyud sde rin po che kun las btus pa’i byung tshul dang bka’i bsdu ba ji ltar mdzad pa’i ’phros las brtsams te gleng ba zab don sgo brgya ’byed pa’i lde mig. In rgyud sde kun btus (sde dge par ma), vol. 30, a (398 ff). BDRC W21295.

Shamar VI (zhwa dmar chos kyi dbang phyug). li thang bkaʼ ʼgyur dkar chag. BDRC W1CZ881.

Sogdokpa (sog bzlog pa blo gros rgyal mtshan). gsang sngags snga ’gyur la bod du rtsod pa snga phyir byung ba rnams kyi lan du brjod pa nges pa don gyi ’brug sgra [“The Dragon’s Roar of Definitive Meaning: responses to earlier and later disputes in Tibet concerning the Early Translation’s secret mantras”]. In gsung ’bum/ blo gros rgyal mtshan, vol. 1, cha (171 ff). BDRC W8870.

Secondary Sources

Gyurme Dorje. The Guhyagarbhatantra and its XIVth Century Commentary Phyogs-bcu mun-sel. PhD thesis. University of London, SOAS (1987).

Smith, E. Gene. Among Tibetan Texts: History and Literature of the Himalayan Plateau (Studies in Indian and Tibetan Buddhism). Boston: Wisdom Publications (2001).

Imaeda, Yoshiro. “Note sur le Kanjur de Derge,” in (eds.), Tantric and Taoist Studies in Honour of R.A. Stein. Mélanges chinois et bouddhiques, vol. XX. Brussels: Institut belge des hautes études chinoises, 1981.

Dudjom Rinpoche Jigdrel Yeshe Dorje, tr. Dorje, G. and Kapstein, M., The Nyingma School of Tibetan Buddhism: Its Fundamentals and History. Boston: Wisdom Publications 1991.

Mayer, Robert. “Indigenous Elements in Tibetan Tantric Religion.” In Mongolo-Tibetica Pragensia ’14: Ethnolinguistics, Sociolinguistics, Religion and Culture, vol. 7:2, 2014.

‍—‍—‍—‍—. “Rnying ma Tantras.” In Brill’s Encyclopedia of Buddhism, vol. 1. Leiden: Brill, 2015.

Schaeffer, Kurtis R., and Leonard W.J. van der Kuijp. An Early Tibetan Survey of Buddhist Literature: the Bstan pa rgyas pa rgya gyi nyi ’od of Bcom ldan ral gri. Harvard Oriental Series. Cambridge (MA) and London: Harvard University Press, 2009.

Wangchuk, Dorji. “An Eleventh Century Defence of the Authenticity of the Guhyagarbha Tantra,” in Eimer and Germano (eds.), The Many Canons of Tibetan Buddhism (PIATS 2000). Leiden: Brill, 2002.

See also section 6, chapter 4 in: Jamgön Kongtrul (’jam mgon kong sprul blo gros mtha’ yas), shes bya kun khyab, Delhi: Shechen Publications (1997). Translated as Guarisco, E. and McLeod, I (trans.), The Treasury of Knowledge: Book 6, Part 4, Systems of Buddhist Tantra, Ithaca: Snow Lion Publications (2005), chapters 17 and 20.


The epithet “Langdarma” (glang dar ma), only used from the 11th century onwards, refers to the king whose actual name was Darma Tri Udumten (dar ma khri ’u’i dum brtan).
See Chomden Rigpai Raltri’s inventory, f. 37.a (modern edition) or f. 34.b (manuscript edition), and Schaeffer and van der Kuijp 2009, pp 46–50 and 181. His evaluation of the authenticity of the Guhyagarbha’s content is quoted verbatim in Sokdokpa vol. 1, cha, f. 132.b–133.b, and the passage as quoted almost identically by Dudjom Rinpoche is translated in Dudjom 1991, vol. 1, pp. 914–7.
This is according to snga ’gyur rnying ma la rgol ngan log rtogs bzlog pa’i bstan bcos, a much later text probably authored by Jigme Lingpa, p. 694. That list seems to have remained constant once it had been initially drawn up, whether by Üpa Losal or someone else. The oldest printed list of the corresponding texts may be the contents list of the three relevant volumes (dza–zha) in the dkar chag of the Lithang Kangyur, f. 16.b–17.a, written by the VIth Shamar, Chökyi Wangchuk, in the first decade of the seventeenth century.
See Butön, f. 79.b.
See Mayer 2014, pp.35-53, esp. pp.37-39; and Mayer 2015, pp.390-397, esp. p.395.
See, for example, the dkar chag of the Lithang Kangyur, f. 16.b–17.a.
See Imaeda 1981 and Loter Wangpo ff. 136.a–136.b (pp. 508–9).