The Mahāsūtras

  • Skt.: mahāsūtra
  • Tib.: mdo chen po

First published 2024. Last updated 20th Aug 2024.

Overview

The Mahāsūtras, or “Great Discourses,” collectively comprise a corpus of texts extracted from the Āgama literature of the Sarvāstivādin and Mūlasarvāstivādin Vinaya, defined not so much by their subject matter, which is diverse, as by the way they were used‍—that is, they were recited for the protection conferred by the resounding of the Buddha’s words, perhaps representing the whole Āgama canon in a short anthology.

More generally, they constitute a set of works from a foundational layer of Buddhist literature and have many parallels across a wide range of canonical corpora in Sanskrit, Pali, Chinese, and other languages.

The epithet mahā (“great”) applied to them as a group and in their titles seems to denote their power, importance, and efficacity rather than their length; some are relatively short texts.

In all Kangyurs, there are ten sūtras whose titles identify them as belonging to the corpus of Mahāsūtras:

The Mahāsūtra of Illusion’s Net (Māyājālamahāsūtra, mdo chen sgyu ma’i dra ba, Toh 288), a teaching on perception and illusion.

The Mahāsūtra “Bimbisāra’s Going Out to Meet the Buddha” (Bimbisārapratyudgamanamahāsūtra, mdo chen po gzugs can snying pos bsu ba, Toh 289), an account of King Bimbisāra’s first audience with the Buddha after his awakening that includes teachings on the aggregates, non-self, and impermanence.

The [Lesser] Mahāsūtra on Emptiness ([*Cūḍa-]Śūnyatāmahāsūtra, mdo chen po stong pa nyid, Toh 290), a teaching to Ānanda on the experience of emptiness.

The [Greater] Mahāsūtra on [Great] Emptiness1 (Mahāśūnyatāmahāsūtra, mdo chen po stong nyid chen po, Toh 291), a teaching to monks in Kapilavastu on living in solitude, meditating on emptiness, and respecting the teacher.

The Mahāsūtra “The Crest Insignia” [1] (Dhvajāgramahāsūtra, mdo chen po rgyal mtshan mchog, Toh 292), a teaching to traveling merchants on the protective benefits of recollecting the Three Jewels.

The Mahāsūtra “The Crest Insignia” [2] (Dhvajāgramahāsūtra, mdo chen po rgyal mtshan dam pa, Toh 293), a teaching to monks on the protective benefits of recollecting the Three Jewels.

The Mahāsūtra on the Five and Three [Views] (Pañcatrayamahāsūtra, mdo chen po lnga gsum pa, Toh 294), a teaching to monks on the different views about what it is of a person that survives in future.

The Mahāsūtra, the Sūtra of the Great Assembly (Mahāsamājasūtramahāsūtra, mdo chen po ’dus pa chen po’i mdo, Toh 653/1062), an account of the many different kinds of gods, goddesses, and other supernatural beings who come to Kapilavastu to pay homage to the Buddha.

The Mahāsūtra, the Āṭānāṭīya Sūtra (Āṭānāṭīyasūtramahāsūtra, mdo chen kun tu rgyu dang / kun tu rgyu ma yin pa dang mthun pa’i mdo, Toh 656/1061), an account in which Vaiśravaṇa, the divine guardian of the North, provides the Buddha with a protective ritual text that monks and nuns should recite to prevent the various nonhuman beings who might harm them, particularly yakṣas, from doing so.

The Mahāsūtra on Entering the City of Vaiśālī, (Vaiśālīpraveśamahāsūtra, yangs pa’i grong khyer du ’jug pa’i mdo chen po, Toh 312/Toh 698/Toh 1093), the mantras and blessing verses the Buddha uses to dispel the great epidemic of Vaiśālī.

Of these ten texts, eight titles are listed in the Tibetan translation of the Vinayavibhaṅga (The Detailed Explanations of Monastic Discipline, ’dul ba rnam ’byed, Toh 3) as texts to be recited for protection. Since one of the eight titles in that list is the title of two different sūtras (5 and 6 above), the list in practice covers the nine texts (1–9). The same nine texts are listed as Mahāsūtras in the ninth-century Denkarma inventory, and all ten texts are known to have been translated into Tibetan in the early, imperial translation period.

The tenth, a slight outlier in the corpus, is not included in any lists but has affiliations with the others in terms of both its sources and its uses.2

In most Kangyurs, seven of the texts (1–7) are found grouped together in the General Sūtra (mdo sde) section. Two (8 and 9) are found in the Action Tantra (bya ba’i rgyud) section, and one (10) is duplicated in both the General Sūtra and Action Tantra sections.

Further copies of 8, 9, and 10 are also found in the Compendium of Dhāraṇīs (gzungs ’dus) section in those Kangyurs that include such a compendium.

Origins and History

Much of our current understanding of the Mahāsūtras comes from the work of Peter Skilling, who has published a two-volume study of this important early Buddhist corpus, including critical editions of the ten Tibetan texts along with their available Pali and Sanskrit counterparts.3

As Skilling notes, the Mahāsūtras are broadly defined, first of all, as a small, selected set of texts from the Āgamas. The Āgamas (and their Pali counterparts) are four (or five) collections of works considered to have been among those teachings of the Buddha that were recited at the First Council.4

The Sarvāstivāda and Mūlasarvāstivāda vinaya traditions each had their own selection of texts that they termed Mahāsūtras. Each tradition listed them in texts that prescribed their recitation under particular circumstances (set out below). The Sarvāstivāda lists eighteen texts in this way,5 while the Mūlasarvāstivāda lists six or eight.6 These lists mention only the titles, and we do not know what differences there may have been in the actual content of the texts in the two lists. Indeed, more generally, we do not know with certainty if there were differences between the Sarvāstivāda and Mūlasarvāstivāda Āgamas, since the Āgama collection of neither school has survived in full; the Chinese translations are of mixed or unknown origins.

References in Sanskrit to the Mahāsūtras, whether in the vinaya literature of the Sarvāstivādins or that of the Mūlasarvāstivādins, are lost. However, fragments of the Sarvāstivādin recensions of the Mahāsūtras have been found among a number of Sanskrit manuscript collections recovered in Central Asia.7

For the Sarvāstivādin Vinaya:

references are preserved only in Chinese translation (dated to the early fifth century): eighteen titles of Mahāsūtras are listed in the section on keeping the rains retreat.

For the Mūlasarvāstivādin Vinaya:

references are preserved in Chinese translation (dated to the very first years of the eighth century): six titles of Mahāsūtras are listed twice, in the sections on the taking of human life (the third pārājika) in the two Vinayavibhaṅga texts (the one for monks and the one for nuns).

references are also preserved in Tibetan translation (dated to the early ninth century): eight titles of Mahāsūtras are listed, also in the sections on the taking of human life (the third pārājika), also in the Vinayavibhaṅga (Toh 3), in passages under the rubric of ro langs (Skt. vetāḍa, “zombies”).

The Mahāsūtras were not translated into Chinese as a known corpus of texts, but a number of the texts are preserved in Chinese translation as independent works belonging to the various Āgamas.8

The eight titles mentioned in the Tibetan translation of the Vinayavibhaṅga match those of the nine Tibetan Mahāsūtra texts in the Kangyur (two of the nine, numbers 5 and 6 in the list above, have the same Sanskrit title, although rendered slightly differently in Tibetan). The six mentioned in the Chinese Vinayavibhaṅga match seven of the nine Tibetan Mahāsūtra texts (again because two of the latter have the same title); the two titles that appear in the Tibetan list but not in the Chinese are the Mahāsamāja and the Āṭānāṭīya.9

Based on the lists of Mahāsūtras in the Chinese and Tibetan translations of the Vinayavibhaṅga, the Tibetan affiliation with the Mūlasarvāstivādin pratimokṣa lineage to which this work belongs, and the integral role that the Kashmiri scholar Jinamitra played in transmitting the Mūlasarvāstivāda Vinaya to Tibet, Skilling has pointed out that the Mahāsūtra collection that Jinamitra, Prajñāvarman, and the Tibetan translator Yeshé Dé translated into Tibetan in the early ninth century probably represents a short corpus of discourses that were important to the Kashmiri Mūlasarvāstivādin textual community.

The Denkarma10 and Phangthangma11 royal Tibetan catalogs indicate that the set of nine Mahāsūtras was originally considered an independent corpus. The same set of nine is also mentioned in Chomden Rikpai Raldri’s thirteenth-century inventory of translated texts under the heading mdo chen po, but a tenth text with the title mdo chen po stong nyid brten cing ’brel par ’byung ba is added to the list; it is difficult to identify this title with any extant work.12 The tradition of classifying all the Mahāsūtras as discourses belonging to the “Lesser Vehicle” or Hīnayāna is in accordance with their Mūlasarvāstivādin provenance and is seen in Chomden Rikpai Raldri’s inventory as well as in the Tibetan scholar Butön’s (1290–1364) History of Indian Buddhism (rgya gar chos ’byung, ca. 1322–23).13 Nevertheless only a decade or two later, in the mid-fourteenth-century Tshalpa Kangyur, two of the Mahāsūtras began to be classified as Action (kriyā) tantras.

That the status of the Mahāsūtras as a set of important works was maintained during the centuries of canon formation in Tibet is reflected in their grouping‍—at least of the first seven‍—in a prominent position among the first works in the Hīnayāna subdivision of the General Sūtra section, following only the massive and important Application of Mindfulness of the Sacred Dharma (Saddharmasmṛtyupasthāna, dam pa’i chos dran pa nye bar gzhag pa, Toh 287).

It is Skilling who suggests including in the corpus the tenth text, The Mahāsūtra on Entering the City of Vaiśālī (Vaiśālīpraveśamahāsūtra, yangs pa’i grong khyer du ’jug pa’i mdo chen po, Toh 312/698/1093), although it is a slight outlier. It is not mentioned in the list of nine Mahāsūtras given in the Tibetan royal catalogs, but its title (like those of the other nine) includes the appellation mahāsūtra / mdo chen po. It also has a clear affiliation with the Mūlasarvāstivādin school and has uses similar to the other nine.

Uses

Many of the sūtras in the group are versions of texts widely recited in a range of Buddhist traditions for their efficacy in warding off evil circumstances and protecting what is good and positive (Skt. rakṣa or parītta, Pali paritta).

The likelihood that this specific corpus was put to similar protective use by monastic and lay communities affiliated with the Sarvāstivāda and Mūlasarvāstivāda traditions is supported by the contexts in which the lists of Mahāsūtras occur in their vinaya literature.

In the Sarvāstivādin Vinaya the list of Mahāsūtras occurs in the section on the rules governing the rains retreat vow that address various circumstances under which, for a period that is not to exceed seven days, a monk is allowed to break his vow to remain in the same dwelling for three lunar months. The passage lists the various discourses that are classified as Mahāsūtras by the Sarvāstivādins and then notes that a monk is permitted to break the rains retreat vow for up to seven nights if a learned layperson has requested him to teach or recite the Mahāsūtras.14

The Mūlasarvāstivādin Vinaya list of Mahāsūtras appears in a quite different context. In the section on the rains retreat equivalent to the Sarvāstivāda passage just described, there is a reference to the four Āgamas, but not to the Mahāsūtras. However, in a section of the Vinayavibhaṅga on the third of the serious offenses that a monk may commit entailing “defeat” (Skt. pārājika, Tib. pham pa), that of “taking a life,” one topic is a discussion of what happens when a monk reanimates a corpse as a zombie (Skt. vetāḍa, Tib. ro langs) and instructs it to kill a victim‍—an event somewhat esoteric to a modern audience but by no means unfamiliar in popular culture. Here the Vinayavibhaṅga lists the Mahāsūtras, almost incidentally, among a number of protective circumstances and measures that may have been taken by a potential victim that will cause the zombie’s murderous attack to fail:

“If some kind of protection measure has been taken‍—such as hanging garlands of beads made of forest flowers at the entrance door, setting out a vase filled with water, tying a cow or a calf, tying a sheep, setting out a mortar, laying a threshold slab at the door, or keeping a fire burning; if a victor, someone appointed by a victor, a universal monarch, a woman pregnant with a universal monarch, a bodhisattva, or a woman pregnant with a bodhisattva is present; or if someone recites, silently or aloud elaborately, the Prātimokṣasūtra, one of the four Āgama collections, or the great Mahāsūtras‍—the Cūḍaśūnyatāmahāsūtra, the Mahāśūnyatāmahāsūtra, the Pañcatrayamahāsūtra, the Māyājālamahāsūtra, the Bimbisārapratyudgamanamahāsūtra, the Dhvajāgramahāsūtra, the Āṭānāṭīyasūtramahāsūtra, and the Mahāsamājasūtramahāsūtra‍—the zombie, its mission having failed, might decide to kill the very monk behind it. Were the zombie to kill the monk, that monk would have committed a grave fault. Were the monk to kill the zombie, he commits a double grave fault: one because of killing the zombie, and the other because of his initial act.”15

The Mūlasarvāstivādin Bhikṣunīvinayavibhaṅga in its Tibetan translation does make brief mention of zombies in the summary for the equivalent section16 but does not contain a passage matching the citation above. The eighth-century Chinese translations of both Vinayavibhaṅga texts (for bhikṣus and bhikṣunīs) by Yijing, however, both include almost identical passages that closely match the Tibetan of the (bhikṣus’) Vinayavibhaṅga and mention the Mahāsūtras.

The Mūlasarvāstivādin parallel to the Sarvāstivādin passage mentioning the Mahāsūtras in the context of the rules for the rains retreat does not explicitly mention the Mahāsūtras.

Extrapolating from the way the Mahāsūtras are mentioned in two different contexts in these two related vinaya traditions, Skilling concludes that they were used as protective texts (rakṣā),17 probably in a wide variety of situations and contexts other than the occasions mentioned in these passages. Some of the Mahāsūtras (notably nos. 5, 6, 9, and 10 in the list above) have an explicit protective function, but others do not. It is likely that the use of the group as protection was one among other, wider applications of their function as a “practical canon” that represented in shortened form the entirety of the four Āgamas,18 which in turn represent the entirety of the Buddha’s words‍—the resounding of which would have fulfilled many applications including transmission, instruction, recollection, and protection.

Parallels and Related Collections

Representing as they do an ancient layer of Buddhist tradition and practices used widely across Asia, the individual Mahāsūtras all have numerous parallels. There are texts related to almost all of them in the Pali Theravādin tradition, as well as works in the Chinese Sarvāstivādin corpus and in some cases Sarvāstivādin texts (or fragments of them) that have survived in Sanskrit. Skilling’s Mahāsūtras volume II19 provides full lists for each text.

As a collection recited for protection, the Mahāsūtras are comparable in their uses (as well as, in part, in their content) to the book known in Sri Lanka as the Maha Pirit Pota (“Great Book of Protection”) or the Catubhāṇavārapāli (“Text of the Four Recitals”).20

Within the Kangyur are two other collections that have similar uses, one more explicitly formalized as a group than the other, and content that is comparable to varying degrees:

The “Thirteen late-translated sūtras” (Toh 31–43), a set of Theravāda texts translated from Pali, two of which, Toh 33 and 34, are the Theravāda equivalents of two of the Mahāsūtras, numbered 8 and 9 in the list above (Toh 656/1061 and Toh 653/1062) respectively; all are recited for protection.

The works related to the Pañcarakṣa, five protective dhāraṇīs and the goddesses personifying them, Toh 558–563 and 587–588 in the Action Tantra section. They are not defined as a distinct collection in Tibetan but are in widespread use and popular today in Nepal.

To these one could add the entirety of the Compendium of Dhāraṇīs; although of course its numerous texts have a considerably wider range of applications than the Mahāsūtras, it also represents a set of works compiled specifically for their practical and protective functions.

Bibliography

Vinayavibhaṅga (’dul ba rnam ’byed). Toh 3, Degé Kangyur vol. 5 (’dul ba, ca), folio 21.a–vol. 8 (’dul ba, nya), folio 269.a. Translation by 84000 forthcoming.

Bhikṣunīvinayavibhaṅga (dge slong ma’i ’dul ba rnam par ’byed pa). Toh 5, Degé Kangyur vol. 9 (’dul ba, ta), folios 25.b–328.a. Translation by 84000 forthcoming.

Denkarma (pho brang stod thang ldan dkar gyi chos kyi ’gyur ro cog gi dkar chag). Toh 4364, Degé Tengyur vol. 206 (sna tshogs, jo), folios 294.b–310.a.

Phangthangma (dkar chag ʼphang thang ma). Beijing: mi rigs dpe skrun khang, 2003.

Butön Rinchen Drup (bu ston rin chen grub). chos ’byung [History] (bde bar gshegs pa’i bstan pa’i gsal byed chos kyi ’byung gnas gsung rab rin po che’i gter mdzod). In The Collected Works of Bu-Ston, vol. 24 (ya), folios 1.b–212.a (pp. 633–1055). New Delhi: International Academy of Indian Culture, 1965–71. BDRC W22106.

Chomden Rikpai Raldri (bcom ldan rig pa’i ral gri). bstan pa rgyas pa rgyan gyi nyi ’od [Ornamenting Sunlight of the Teachings’ Spread]. In gsung gsung thor bu, 2:3–158 (manuscript). BDRC MW1CZ1041. See also Schaeffer and van der Kuijp 2009. 

Blackburn, Anne. “Looking for the Vinaya: Monastic Discipline in the Practical Canons of the Theravāda.” Journal of the International Association of Buddhist Studies 22, no. 2 (1999): 281–310.

Herrmann-Pfandt, Adelheid. Die lHan kar ma: ein früher Katalog der ins Tibetische übersetzten buddhistischen Texte. Vienna: Verlag der Österreichischen Akademie der Wissenschaften, 2008.

Lévi, Sylvain. “Sur la récitation primitive des textes bouddhiques.” Journal Asiatique (May–June 1915): 401–47.

Pemaloka, Kotawila Sri Nayaka Thera. The Great Book of Protection: The Text of the Four Recitals (Catubhāṇavārapāli): Sinhala – Maha Pirit Pota, with Translation into English. Colombo: Samayawardhana, 2018.

Schaeffer, Kurtis R., and Leonard W. J. van der Kuijp. An Early Tibetan Survey of Buddhist Literature: The Bstan pa rgyas pa rgyan gyi nyi ’od of Bcom ldan ral gri. Harvard Oriental Series 64. Cambridge, MA: Harvard University Press, 2009.

Skilling, Peter (1994–97). The Mahāsūtras: Great Discourses of the Buddha. 2 vols. Oxford: Pali Text Society, 1994–97.

‍—‍—‍—(2007). “Zombies and Half-Zombies: Mahāsūtras and Other Protective Measures.” Journal of the Pali Text Society 29 (2007): 313–30.

Notes

The Sanskrit title and especially its rendering in Tibetan has‍—in at least one commentary‍—seemingly allowed one of the two instances of mahā / chen po (“great” or “greater”) to be interpreted as referring to the greatness of emptiness, although the most common and likely interpretation is that it refers to the text’s length (compared to that of the preceding text in the list). See Skilling Mahāsūtras II, p. 368.
Its inclusion is suggested by Skilling (see below).
Skilling, Mahāsūtras II, pp. 9–12. For Skilling’s complete list of Mahāsūtras that were translated into Tibetan, see p. 12. For charts of the variations of Mahāsūtras across Buddhist textual lineages, see pp. 54–61. For charts of the distribution and mention of the Mahāsūtras in Tibetan sources, see pp. 209–20.
The Āgamas (Pali: nikāyas) of the Mūlasarvāstivāda are divided into four sections: the Long Discourses (Dīrghāgama), the Middle-Length Discourses (Madhyamāgama), the Connected Discourses, (Saṃyuktāgama), and the Numbered Discourses (Ekottarāgama). Skilling (Mahāsūtras II, p. 17) notes that “the sūtras in question were transmitted orally for several centuries before being written down; during both the oral and written periods they would have been amplified, edited, and revised any number of times, translated from an original Prakrit or Prakrits into a ‘Buddhist’ and finally a more standard Sanskrit, and then further revised.”
Skilling provides a list of the Sarvāstivādin Mahāsūtra titles in Table 1, Mahāsūtras II, p. 54.
See n.­3 and Skilling, Mahāsūtras II, pp. 4–8.
Skilling, Mahāsūtras II, p. 18. Skilling notes that six of these texts correspond to the Mūlasarvāstivādin list and ten correspond to the Sarvāstivādin list.
Skilling, Mahāsūtras II, pp. 105–9.
See Skilling, Mahāsūtras II, pp. 13–14.
Denkarma, folio 300.a.3. See also Herrmann-Pfandt 2008, p. 128, no. 241.
Phangthangma (2003), p. 20.
Chomden Rikpai Raldri, folio 16.a–b; his inventory is thought to have been composed around 1270. See also Schaeffer and van der Kuijp 2009, pp. 135–36.
Butön, History, folio 144.a.
Skilling, Mahāsūtras II, p. 20. Skilling provides an English rendering of Sylvain Lévi’s French translation of the original Chinese (Lévi 1915). The passage opens with the Sarvāstivādin list of Mahāsūtras and then reads, “If this layman has not yet learned them and wishes to do so, or if he has learned them and then forgotten them, and wishes to recite them, and if he sends someone to a monk (bhikṣu) with this message: ‘The bhadanta is very learned, well versed in the Great Sūtras... I have not yet learned them and wish to do so, or I have forgotten them and wish to recite them. May the bhadanta come so that I may receive instruction, study, recite, and inquire about the meaning’: in this case, he [the monk] should accept and go for [up to] seven nights.”
Vinayavibhaṅga (Toh 3), folio 143.a, here translated from 84000’s forthcoming translation. The passage is also translated and discussed in Skilling 2007.
Bhikṣunīvinayavibhaṅga (Toh 5), folio 51.b. 84000 translation forthcoming.
Skilling, Mahāsūtras II, pp. 63–88.
Skilling, Mahāsūtras II, p. 27. Skilling writes, “A complete set of the four Āgamas would be somewhat unwieldy, difficult both to master and to prepare in manuscript form: it may be that the Mahāsūtras were extracted in order to make a collection that would be easier for diligent monks, nuns, or lay-followers to study. This might explain their arrangement according to the sequence of Āgamas, with the internal sequence of some sūtras intact (at least Mahāsūtras 2, 3, and 4 for the Dīrghāgama; Mahāsūtras 8 and 9 for the Madhyamāgama; and the first three sections of the Saṃyuktāgama).” Skilling does not employ the formal/practical canon distinction here, but the phenomenon he describes here certainly fits this approach to Buddhist “canonicity”; see Blackburn 1999.
Skilling 1994–97, vol. 2, p. 442.
See Pemaloka 2018.