Summary

The Buddha, while at the Jetavana in Śrāvastī, tells Mañjuśrī of a buddha realm far above the world, in which lives the Buddha Aparimitāyur­jñāna. He states that those who recite, write, hear, and so on, the praise of this buddha, or make offerings to this text, will have numerous benefits, including a long life and a good rebirth. As vast numbers of buddhas recite it, the mantra, or dhāraṇī, of this buddha is repeated numerous times. This is the lesser known of the two versions of this sūtra in the Kangyur, but possibly represents the earlier translation.

Acknowledgements

The sūtra was translated from Tibetan by Peter Alan Roberts with reference to related Sanskrit versions. Tulku Yeshi of the Sakya Monastery, Seattle, was the consulting lama who reviewed the translation. The project manager and editor was Emily Bower. The proofreader was Ben Gleason. The introduction was compiled by the 84000 Editorial Team, and incorporates material by Peter Alan Roberts.

The translation was completed under the patronage and supervision of 84000: Translating the Words of the Buddha.

Introduction

Overview

The Aparimitāyur­jñāna Sūtra is among the many canonical works in which a particular buddha in another buddhafield is invoked along with the benefits of recalling his name and reciting his dhāraṇī. Associated as it is with longevity, this is one of the most widely read texts in the Kangyur, and Aparimitāyur­jñāna (“Immeasurable Longevity and Wisdom”’) is one of the most frequently portrayed and well-known buddhas in the pantheon of Tibetan Buddhism.

The sūtra is commonly referred to as the Tsédo (tshe mdo, “Sūtra of Longevity”) or Tsézung (tshe gzungs, “Dhāraṇī of Longevity”), and contains a dhāraṇī that is repeated in the text twenty-nine times. It is included in many Tibetan liturgical compilations, and its recitation, usually with a specified number of repetitions, is often advised to people in poor health or facing other difficulties, or is commissioned on their behalf in monasteries.

Although its title identifies it as a sūtra, it is placed in all Kangyurs with the Action Tantras (bya ba’i rgyud, kriyātantra). In common with many other works classified as Action Tantras, there is nevertheless little in the text to identify it as a tantra. The inclusion of a long, repeated dhāraṇī in Sanskrit is presumably one criterion for this classification, although there are many other canonical works with a similar structure that are placed with the sūtras. Other criteria may have been its classification and line of transmission in India, before it was taken to Tibet, or the fact that it has also formed the basis for a wide range of tantra practices, particularly among the higher levels of tantra, in the form of sādhanas of Aparimitāyus.

The text has also survived in a large number of Sanskrit manuscripts (mostly later Nepalese ones); in two Chinese translations; and in a slightly different‍—perhaps earlier‍—Tibetan translation, represented by most of the very numerous manuscripts found in the caves of Dunhuang, where a Khotanese manuscript (probably the oldest surviving version) was also found.

All Kangyurs include two major versions of the sūtra, similar in most respects but differing mainly in the presence or absence of one phrase in the repeated dhāraṇī. The background of the existence of these two versions is discussed below. The version translated here is the lesser known of the two, is much less widely used, and differs from Sanskrit source texts brought to Tibet in the later translation period to a greater extent than the other version.

The present version appears to be derived from the same translation as the other major version, The Aparimitāyur­jñāna Sūtra (1), Toh 674, but the dhāraṇī it contains is the shorter one, matching those found in the Chinese translations, the Dunhuang manuscript in Khotanese, and the earlier Tibetan translation represented by the majority of the very numerous Dunhuang manuscripts in Tibetan.

This sūtra, to the extent that it may represent the translation available in the late eighth century, is one of the set known as the “ten royal sūtras,” thought to be so called either because they represent distillations of the most profound scriptures, or because according to traditional histories they were recommended to King Trisong Detsen for his daily practice by Guru Padmasambhava. As a result of practicing them, the king is said to have extended his life by thirteen years.

In a similar vein, the fact that so many manuscript copies of this text have been found in the Dunhuang caves is due to their production by scribes there having been commissioned on behalf of Trisong Detsen’s grandson, Ralpachen (who reigned in the early ninth century) in order to ensure for the king the longevity that the text itself promises.

Aparimitāyus, Amitāyus, and Amitābha

Aparimitāyur­jñāna (“Immeasurable Longevity and Wisdom”), the buddha who is the subject of this sūtra, despite being described in this text as dwelling in a realm situated in an upward direction from this world, i.e. toward the zenith, has been identified to a varying extent in both Tibet and East Asia with Amitābha, buddha of the realm called Sukhāvatī in the west. Both are often referred to by the shortened name Amitāyus in Sanskrit, Tsépamé (tshe dpag med or, in full, tshe dpag tu med pa) in Tibetan, but are nevertheless likely to have originally been seen as distinct. The confusion that has been caused by these partially overlapping identities is discussed in detail in the introduction (i.9–i.16) to Toh 674, The Aparimitāyur­jñāna Sūtra (1).

The Tibetan Versions of the Sūtra in the Kangyur, Their Differences, and Their Translation

Differences in Content

There are two different but closely related versions of The Aparimitāyur­jñāna Sūtra in most Kangyurs. In keeping with the tradition established over the centuries by editors of all Kangyurs, we have here translated and published them separately, despite their similarity, and have labeled them The Aparimitāyur­jñāna Sūtra (1) and (2). In the Kangyur the two versions are found consecutively in the Tantra Collection, and according to Situ Panchen’s original catalog of the Degé Kangyur both are placed with the Action (Kriyā) Tantras in the subdivision that corresponds to the principal deity (rigs kyi gtso bo) of the Padma (lotus) family. Their Degé recensions are cataloged as Toh 674 and 675, respectively.

The principal distinction between the two versions lies in the length and composition of the repeated dhāraṇī. Compared to the present version (2) of the text, the dhāraṇī in the other version, The Aparimitāyur­jñāna Sūtra (1), Toh 674 and 849, is longer and contains an extra phrase in the middle beginning “oṁ puṇye puṇye…” that is not present here, although the rest of the dhāraṇī is almost the same. It is on that basis that the catalog of the Degé Kangyur distinguishes the two texts by calling this one (Toh 675) the “two oṁ, no puṇye” version (because the repeated dhāraṇī also contains two other phrases beginning with oṁ), while the other one (Toh 674 and 849) it calls the “three oṁ” version. In some other catalogs, e.g. the index to the Narthang Kangyur, the two versions (1) and (2) are called respectively the “large and small Tsédo.”

It is the other version, the “three oṁ” version, The Aparimitāyur­jñāna Sūtra (1), Toh 674 and 849, that is the best known and most widely recited form of the text. It is also the one reproduced in most compilations of dhāraṇīs and anthologies of practice texts, whether ancient or modern. By contrast, the present “two oṁ” version is only preserved in the Kangyur, as here, and is very little known or used. Nevertheless, the fact that it has been included in all Kangyurs as a separate text is an implicit recognition that both versions are authentic. The background to the existence of these two different versions, and the controversies that have sometimes arisen about their origins and authenticity, are explored in the introduction to The Aparimitāyur­jñāna Sūtra (1), Toh 674 and 849 (i.21 et seq) and are set out in slightly more detail here.

One intriguing hint that the Kangyur provides us with regard to these two versions of the text is that the part of the dhāraṇī that is “added” in version (1), Toh 674, but “missing” here in version (2), Toh 675, is included in all Kangyurs, almost identically, but on its own, as the dhāraṇī that forms the very short content of another text in this group, The Essence of Aparimitāyus, Toh 673a. There is no explanation in the various Kangyur catalogs for its presence, but the term “essence” (hṛdaya, snying po, sometimes rendered “heart mantra”) in its title identifies it as a mantra used in at least one tradition of the practice of Aparimitāyus. The only other mention of this mantra in the Kangyur appears to be in the Sarva­durgati­pariśodhana, known best in Tibetan by its shortened title sbyong rgyud (the “Purification Tantra”), a tantra of the Yoga class, in which this same mantra is given as the “essence vidyā-mantra” (hṛdayavidyā, snying po’i rig pa) of the tathāgata Aparimitāyuḥ­puṇya­jñāna­sambhāra­tejo­rāja, the principal figure in a secondary maṇḍala. What is confirmed by this mantra’s mention in the tantra, as well as its presence as Toh 673a, is at the very least that it is a potentially independent stand-alone mantra phrase, making it easier to understand that it might have been either added or removed at some stage in the evolving transmission of the dhāraṇī in the sūtra.

Apart from the composition of the dhāraṇī, there are a few other differences between Toh 674 and 675, mostly minor. Overall, the two versions appear to be much more closely related than they would be if they represented two different translations made entirely independently of each other. Some of the minor textual differences between the present version of the sūtra and Toh 674 are flagged in the notes, but among the more significant are the following:

• The name tshe dang ye shes dpag tu med pa (Aparimitāyur­jñāna), which occurs a number of times here in Toh 675, is given its shorter form in Toh 674: tshe dpag tu med pa (Aparimitāyus).

• The Tibetan rendering of the longest form of the name Aparimitāyur­jñāna­suviniścita­tejo­rāja here (in Toh 675) is tshe dang ye shes dpag tu med pa shin tu rnam par gdon mi za ba’i rgyal po, while in Toh 674 it is tshe dang ye shes dpag tu med pa shin tu rnam par nges pa’i gzi brjid kyi rgyal po.

• The name of Aparimitāyur­jñāna’s realm in Toh 674 is yon tan dpag tu med pa sogs pa (Aparimita­guṇa­saṃcaya, “Accumulation of Immeasurable Qualities”), but here in Toh 675 at the first mention of it (see 1.3) it is just yon tan dpag tu med pa (Aparimitaguṇa, “Immeasurable Qualities”).

• Its second mention here in Toh 675 (see 1.8), however, not only gives it in full, as in Toh 674, but also prefixes it with the words bde ba can, which can be rendered as “blissful” but also as the realm name Sukhāvatī, somewhat confusingly identifying this realm with that of the “other” Amitāyus who is Amitābha, and perhaps confirming that the conflation of these two buddhas (see above) occurred at an early date in Tibet.

• In the three paragraphs in which comparisons using analogies are made of the amounts of merit to be obtained through‍—in Toh 674‍—reciting the sūtra, here in Toh 675 (1.56, 1.58, and 1.60), the merit seems to be obtained through the tathāgata Aparimitāyur­jñāna himself rather than through reciting the text.

• Finally, Toh 674 has a curious concluding line that is not present here in Toh 675: in addition to the beings in the world being overjoyed and rejoicing at the Buddha’s words, the Buddha himself is said to be pleased or delighted (dgyes pa). This appears to be the result of the Sanskrit āttamanās being translated twice, as descriptive for both the world and the Buddha. The ending here in Toh 675 matches similar instances of this standard ending formula.

Differences in Origin and Transmission

In essence, the other, best-known, “three oṁ” version of the sūtra, Toh 674, was brought to Tibet and translated in the later, post-imperial period of transmission. The origins of the present “two oṁ” version are less clear, but it is likely to represent sources known and translated in the early, imperial period.

We have set out the evidence that the translation preserved in the Kangyur of the “three oṁ” version (Toh 674) may be attributed to Puṇyasambhava and Patsap Nyima Drak in the late eleventh or early twelfth century‍—a widely accepted belief‍—in the introduction to that version of the text. It is based primarily on a mention in the lineage record of transmissions received by Minling Terchen Gyurme Dorje, mentions in Tāranātha’s commentary to the sūtra, and the colophon appended to the version of the text as reproduced in the Druptap Küntü (sgrub thabs kun btus), a collection of sādhanas of the Sakya tradition compiled by Jamyang Loter Wangpo (1847–1914). There is also evidence that other, later translations of the “three oṁ” version were made. Here we will concentrate in particular on what may be inferred of the origins of the present, “two oṁ” version, Toh 675.

The available sources of information about the origins of the two versions do not include the most usual ones for Kangyur texts, colophons and Kangyur catalogs. There are no translators’ colophons to either version of the text in any of the different Kangyurs. Most of the Kangyur inventories and catalogs merely list the titles and distinguish them using epithets such as “three oṁ” and “two oṁ,” or “long” and “short.” Some, including the catalog of the Degé Kangyur, explicitly state that the translators are unknown. The one exception is the catalog (dkar chag) of the Narthang Kangyur, which appears to attribute both versions of the sūtra to Puṇyasambhava and Patsap Nyima Drak. As well as being inherently unlikely, this is at odds with all other sources of information. A look at the folios of the catalog concerned reinforces suspicions of an erroneous attribution, as the carving of the catalog’s woodblocks seems to have run into problems for this entry, which coincides with a folio break. Indeed, the Narthang catalog’s confusing attributions at this point do not reflect the titles or colophons in the body of the Narthang Kangyur itself.

Despite this dearth of direct canonical records, what we do know is that at least one version of the sūtra was translated during the early translation period. Firstly, both the early ninth century inventories of translated texts, the Denkarma (ldan dkar ma) and Phangthangma (phang thang ma), mention, respectively, texts entitled The Dhāraṇī of Aparimitāyus (tshe dpag tu med pa’i gzungs, 110 ślokas in length) and Aparimitāyus (tshe dpag tu med pa, 120 ślokas), both probably referring to a version of this text; in both inventories it is placed in the category “miscellaneous long and short dhāraṇīs” (gzungs che phra sna tshogs).

Secondly, very tangible evidence of the existence of translations in the early period is provided by the very large number of manuscript copies of the sūtra, in Tibetan, found among the Dunhuang manuscripts, a large group of which can be dated to between 830 and 850, most having apparently been made on the orders of King Ralpachen (r. 815–41)‍—presumably to create the meritorious results that the text itself describes. Crucially, all the Dunhuang manuscripts that we have been able to examine contain the “two oṁ” version of the dhāraṇī. This fact, combined with the evidence from Minling Terchen, makes it very likely that the present version of the text, the “three oṁ” version, is the translation dating from the later translation period, while Toh 675, the “two oṁ” version, is more closely related to the one originally made in the early translation period.

Thirdly, an early period translation of the “two oṁ” version is the target of some openly dismissive remarks made in several works by later Sakya scholars, including Ngorchen Kunga Zangpo and his disciple Kunga Lekrin in the fifteenth century, and Amé Zhab Ngawang Kunga Sönam in the seventeenth century. Both the latter authors discuss the various theories raised to account for the existence of the two versions, and go on to confirm the validity only of the later translation, The Aparimitāyur­jñāna Sūtra (1), Toh 674. Following a leading statement attributed by Ngorchen Kunga Zangpo to his own teachers, they are both scathingly critical of the “two oṁ” version, and of the “earlier translators” responsible for it, whom they accuse of deliberately omitting part of the dhāraṇī to express dissatisfaction with their stipend. Such comments have to be seen in the context of arguments regarding authenticity between proponents of the “late” versus the “early” translation traditions (which concerned principally the tantras rather than any of the other canonical genres), but they do at least confirm that the “two oṁ” version of the text was widely recognized as having been translated in the early, imperial period.

Debates About Difference and Authenticity

The writings of Ngorchen, Kunga Lekrin, and Amé Zhab go to some lengths to throw doubt on the authenticity of the “two oṁ” version, with Amé Zhab’s, in particular, using arguments that in some cases refute points made in favor of it by opponents in the debate who remain invisible and unidentified. For apart from Tāranātha’s commentary, we have not been able to find actual writings of the period defending the “two oṁ” version from the disparagements of the authors mentioned, other than general notes by catalog compilers to the effect that “both versions are authentic.” The substance of such writings or statements can only be inferred, from arguments put forward in order to neutralize them by those in favor of the “three oṁ” version. The following points represent the principal arguments that these writers deploy to justify their unusually harsh judgment.

Absence of Indic source texts

Amé Zhab writes that Ngorchen Kunga Zangpo personally examined three Sanskrit manuscripts of the sūtra, all of which were the “three oṁ” version, implying that there were no Sanskrit texts attesting to the “two oṁ” version. While there is little doubt that the “three oṁ” version of the text is indeed authentic to its source, Ngorchen’s claim that it is the only authentic version is explicitly refuted and dismissed by Tāranātha, who in his commentary says there were many Sanskrit manuscripts of both versions, and relates in his autobiography how, on a visit to Narthang, he was himself shown several Sanskrit manuscripts of the sūtra, of which one was clearly the “two oṁ” version.

From today’s perspective, no surviving Sanskrit versions of the text from Tibet have been discovered, and the only known Sanskrit manuscripts are Nepalese ones dating from much later centuries. However, the veracity of Tāranātha’s report can be indirectly confirmed by Sten Konow’s 1916 comparison of two versions, one in Sanskrit based on Nepalese manuscripts and one a Khotanese manuscript (probably centuries earlier) found in the Dunhuang caves by Sir Aurel Stein. The Nepalese Sanskrit has the “three oṁ” version of the dhāraṇī while the Khotanese has the “two oṁ” version. Ngorchen and his successors were also presumably unaware that in both of the two Chinese translations, too, the dhāraṇī is the “two oṁ” version.

Different versions might refer to different buddhas

The argument assumed here is based on the two more or less distinct buddhas to whom the shortened name Amitāyus, or Tsepamé (tshe dpag med) in Tibetan, might refer. Tibetan authors took pains to distinguish between, on the one hand, “Amitāyus of the zenith” (steng phyogs kyi tshe dpag med) and “Amitāyus of Akaniṣṭha” (’og min gyi tshe dpag med), both likely to be references to Aparimitāyur­jñāna; and on the other, “Amitāyus of Sukhāvatī” (bde ba can gyi tshe dpag med), also sometimes known as “Amitāyus of the Drum of Immortality” (’chi med rnga sgra’i tshe dpag med, Dundubhisvara-Amitāyus), both of whom can be identified with Amitābha.

Conceivably, it could therefore be proposed that while one version of the text described the existence of one of these buddhas and the benefits to be obtained from remembering his name and honoring him, the other might be legitimately different because it focused instead on the other buddha. There are indeed some minor differences between the two versions concerning the names used at certain points in the text, as mentioned above at i.14. It is not clear who might have used this notion as an explanation for the existence of variant dhāraṇīs, but in any case both Kunga Lekrin and Amé Zhab are emphatic in insisting that both versions are clearly focused on the tathāgata Aparimitāyur­jñāna, whose buddha realm is in the zenith.

One hundred and eight names are mentioned in the sūtra itself

Both versions of the sūtra mention (in 1.5, 1.6, and 1.8) that the benefits of an increased lifespan will come to beings who hear, remember, or write “the one hundred and eight names of Aparimitāyur­jñāna.” Kunga Lekrin and Amé Zhab interpret the one hundred and eight names as a reference to the dhāraṇī, the “three oṁ” version of which does indeed have 108 syllables if the euphonic rules of Sanskrit are applied to elide the two instances of final and initial a, or if the final svāhā is omitted. The “two oṁ” dhāraṇī, they point out, only has seventy-seven syllables, is therefore incomplete, and must be incorrect.

At first sight, this is perhaps the most convincing argument in favor of the “three oṁ” version of the text. On closer examination, however, some of its initial attractions seem less clear. To see the dhāraṇī as comprising one hundred and eight names requires a stretch of the imagination. The dhāraṇi contains semantically coherent words and phrases of which few are actual “names,” and even if all the many compound words are divided into their irreducible units their number still remains less the number of syllables. Of the many “hundred-and-eight-name” or “hundred-name” texts to be found in the Kangyur and elsewhere, most actually do contain distinct lists of names just as their titles suggest, and there do not seem to be other instances of a dhāraṇī being referred to as a “one-hundred-and-eight-syllable appellation,” as this phrase might possibly be interpreted. Indeed, Tāranātha points out in his commentary that to account for the fact that the one hundred and eight names are also mentioned in the “two oṁ” version of the text we have to assume that it refers to a separate text on the one hundred and eight names elsewhere. Such cases, he says, are not unknown; he gives the example of the hundred names of Śrī Heruka. In fact, however, no text enumerating the hundred and eight names of Aparimitāyur­jñāna appears ever to have been identified.

From the perspective we can take today, historically and geographically wider than was possible for those scholars, the strength of this argument using the hundred and eight names to promote the exclusive authenticity of the “three oṁ” version is much diminished. That not only the Khotanese version (probably the oldest manuscript of all) but also the many Dunhuang texts and both Chinese translations all contain the “two oṁ” dhāraṇī, with its seventy-seven syllables, yet still make mention of “one hundred and eight names” adds considerable weight to Tāranātha’s inference that this is not necessarily a reference to the dhāraṇī itself.

The longer dhāraṇī may include a phrase added from elsewhere

Amé Zhab writes that “certain later scholars” have claimed that the “two oṁ” version is correct but that the dhāraṇī of the “three oṁ” version has been expanded by adding to it the mantra of Aparimitāyus from the sbyong rgyud, i.e. the Sarva­durgati­pariśodhana, a tantra, widely used for funerary rites, of which there are two Tibetan translations in the Kangyur made in the early and later translation periods, Toh 483 and 485 respectively.

It is true that in the “three oṁ” version the centrally placed phrase starting “oṁ puṇye puṇye mahāpuṇye…”‍—the phrase whose presence or absence marks the difference between the dhāraṇīs of the two versions‍—is very similar to a mantra found in the Sarva­durgati­pariśodhana as the “essence” (hṛdaya, snying po) mantra in one of the secondary maṇḍalas in the text, that of of a tathāgata whose full name is given as Aparimitāyuḥ­puṇya­jñāna­sambhāra­tejo­rāja, but who is referred to as Amitāyus in the several commentaries. This mantra is also strikingly similar to the one introduced as the “essence” mantra of Aparimitāyus in the very short text that precedes the two Aparimitāyur­jñāna Sūtra s in most Kangyurs, and of which we have published a translation under the title The Essence of Aparimitāyus (Toh 673a).

Amé Zhab, however, dismisses the argument as inadmissible, writing that in the Sarva­durgati­pariśodhana mantra the words āyuḥpuṇya and sarvacayaṃkari [sic] are found that are not in the dhāraṇī of the “three oṁ” version; these scholars, he says, have simply not consulted the tantra properly. But his objections, valid though they presumably seemed from the copies of the relevant books he had to hand, are more easily dispelled than sustained. The first of the words he cites, āyuḥpuṇya, is present in one of the two only slightly differing variants of the “three oṁ” dhāraṇī preserved in most Kangyurs, the version in the Tantra Collection (rgyud ’bum), Toh 674. The second, sarvacayaṃkari, is indeed not in any version of the present dhāraṇī but, even assuming that Amé Zhab had intended the somewhat different spellings upacayakāriṇi or upacayakāraṇi actually found in the Sanskrit and Tibetan respectively of the Sarva­durgati­pariśodhana, it is present only in the version of that tantra as translated in the later translation period (Toh 485), and is absent in the earlier translation (Toh 483), which reads oṁ puṇye puṇye mahāpuṇye aparimita‑ayūḥpuṇya-jñāna-saṃbhāropacite svāhā.

In other words, a close relationship between this Aparimitāyus mantra in the Sarva­durgati­pariśodhana and the dhāraṇī of the “three oṁ” version of the sūtra does seem eminently possible, and would not seem particularly surprising given the cross-text interrelationships between mantras and dhāraṇīs of related buddhas and deities. It does seem to suggest that the phrase in question could be a potentially independent and detachable component of the dhāraṇī, but gives us no particular hint as to whether it is more likely to have been added to one version, or removed from another.

Differences and Mixture: Possible Conclusions

The story of these two different versions may seem already complex enough, but there is more to be explored. Apart from the differences in the dhāraṇī, the other differences (listed above at i.14 and in the notes) between the two versions in the Kangyurs are relatively minor. These two translations seem most unlikely to have been made independently, even allowing for the possibility that the Sanskrit texts they were made from were very similar. The most obvious explanation might be that the later translation was based on the earlier, adapting it to conform to a slightly different Sanskrit original. However, an examination of the wording and terminology of the Dunhuang manuscripts‍—which predate the work of Puṇyasambhava and Patsap Nyima Drak by several centuries‍—shows that they almost certainly represent a Tibetan translation different from the present “two oṁ” version (Toh 675) as it has survived in the Kangyur. It is therefore tempting to conjecture that this “two oṁ” version in the Kangyur might in fact be a back-adaptation of the later translation, edited at some stage to conform to the alternative “two oṁ” dhāraṇī of the earlier translation but not otherwise reflecting that earlier translation fully. This hypothesis is supported by the fact that a few significant elements of the later “three oṁ” version that are not present in the “two oṁ” Dunhuang manuscripts are present here. Among the most significant in this regard is the similar rendering, common to both Kangyur versions, of what we may presume to be the standard, modular phrase in the Sanskrit manuscripts, eva hi tiṣṭhati dhriyate yāpayati, as bzhugs so/ /tshe ’dzin cing tshe mthar phyin par bzhed de in 1.3, which Tāranātha identifies as a feature unique to the work of Patshap and the other later period translators.

Whatever the case in this confusing blend of textual variation and admixture, there is no single clinching argument in favor of accepting one version over another. Both versions, surely, can be considered authentic; and although the compilers and editors of the many Kangyurs do not seem to have noted their reasons, there must have been enough consensus on this point‍—despite all the controversies‍—for both to have been preserved in all Kangyurs.

Sanskrit and Chinese Versions of the Text

There are many surviving manuscripts of the text in Sanskrit, but none that can be reliably dated much before the early ninth century, the period when it was first translated into Chinese, and probably into Tibetan for the first time (see below). Most of the Sanskrit manuscripts are Nepalese and are dated considerably later.

The oldest known Indic version of the text may be one from east Turkestan in what came to be called “Khotanese,” the old Iranian dialect of that region during the later period of the time when Buddhism was prevalent there. It is written in the Upright Gupta script, and probably dates to the seventh or eighth century. The manuscript was discovered in the Mogao caves in Dunhuang by Sir Aurel Stein in 1907, and in 1912 Ernst Leumann made a short comparison of the Sanskrit of this sūtra’s opening sentences with a few fragments of the Khotanese text. In 1916, two Sanskrit editions saw the light independently. One, by Sten Konow, compared an edition of a Nepalese Sanskrit version with a complete edition of the Khotanese fragments, along with the first English translation. The other, by Max Walleser, was based on a Nepalese manuscript and included a German translation. Walleser’s German translation has been translated into English by Richard K. Payne in his paper on this sūtra. Jonathan Silk has made an English translation from Walleser’s edition of the Sanskrit, and attests in his unpublished paper, “The Most Important Buddhist Scripture? The Aparimitāyur­jñāna and Medieval Buddhism,” to the relatively large number of extant Sanskrit manuscripts‍—well over one hundred‍—either in isolation or compilations, indicating how popular this sūtra was in Buddhist practice.

As noted above, the Nepalese Sanskrit manuscripts are all of the “three oṁ” version of the text, and therefore correspond more closely to version (1), Toh 674. The Khotanese, on the other hand, contains the “two oṁ” version of the dhāranī throughout, and is thus closer to the present version and to the many Dunhuang manuscripts in Tibetan.

Two translations of the sūtra into Chinese were made, one by Facheng (Taishō 936) in the early ninth century, and the other by Fatian (Taishō 937) in the late tenth century. Both Chinese translations contain the “two oṁ” dhāraṇī, and are therefore closer to the Khotanese, the present Tibetan version of the text, and the Dunhuang manuscripts than they are to the Nepalese Sanskrit manuscripts and to version (1) in Tibetan, Toh 674.

The Sūtra in Buddhist Practice

As mentioned above in the overview of this introduction, recitation of the Tsédo a specified number of times has historically been‍—and is still nowadays‍—prescribed as a practice to people in poor health or facing other difficulties, and to practitioners more generally in order to ensure longevity, and so on. Recitation by the monks or nuns in a monastery is also commissioned for the same reasons. According to the Padma Kathang, the eighth century Tibetan king, Trisong Detsen, was advised to recite this text daily (along with the other works known as the “ten royal sūtras”), as a result of which the king’s life is said to have been prolonged by thirteen years beyond the limit predicted by astrological reckoning.

The sūtra itself particularly emphasizes the beneficial effects of writing it out or causing it to be written out, and there is ample evidence that this recommendation was taken seriously in the form of the very large number of commissioned copies, mostly in Tibetan and Chinese and dating to the eighth and ninth centuries, found in the Dunhuang caves by Stein and Pelliot in the early years of the twentieth century. Many of them appear to have been commissioned in the name of the Tibetan king Ralpachen, who reigned in the early ninth century and was the grandson of Trisong Detsen. Among the bundles acquired by the two explorers for the British and French governments, there are over a thousand copies of the sūtra now in the British Library in London, and a similar number in the Bibliothèque Nationale in Paris. Some five hundred further copies remain in libraries in China, many have found their way to Japan, and there are some two hundred in St. Petersburg. As noted in i.19, a large majority of these manuscripts, and probably all of those from the period concerned, contain the equivalent of the “two oṁ” dhāraṇī

In the context of tantric practice, the Tibetan canonical literature contains a number of sādhanas of Aparimitāyur­jñāna, particularly among the higher levels of tantra. One example is the liturgy composed by the tenth-century Jetāri, one of the principal teachers of Vikramaśīla Monastery, which was known for its promulgation of higher tantras. That practice was introduced into Tibet by Bari Lotsāwa Rinchen Drak (mentioned above), and thereby became a part of the Sakya tradition.

There are also five Aparimitāyur­jñāna texts in the Tengyur that were composed by Siddharājñī, a female guru active in India in the beginning of the twelfth century, from whom Rechungpa, pupil of the famous Milarepa, received transmissions that Marpa had not been able to bring back to Tibet. At least three if not all of these Siddharājñī texts were translated into Tibetan by Varacandra, another of Rechungpa’s teachers, together with the Tibetan Lenchung Darma Tsultrim (glan chung dar ma tshul khrims), about whom little is known other than this translation work with Varacandra. Rechungpa introduced the practice into Tibet, where it spread from his own lineage, the Rechung Kagyü or Rechung Nyengyü, to other Kagyü traditions. These tantric Aparimitāyur­jñāna practices are based upon the five-family system of the higher tantras, and they involve an elaborate visualization of oneself as a red Aparimitāyur­jñāna, wearing the saṃbhogakāya costume and holding a vase of amṛta, with an entourage of deities within a palace, and the visualization of channels within the body. As Aparimitāyur­jñāna is auspicious for long life, his empowerment is given as a long life blessing.

Within the indigenous Tibetan literature, a very large number of Aparimitāyur­jñāna sādhanas have been created over the centuries within all the lineages and schools. The Nyingma tradition of rediscovered treasure texts (gter ma), too, has produced many Aparimitāyur­jñāna revelations, from Nyangral Nyima Ozer (nyang ral nyi ma ’od zer, 1136–1204) onward. The importance that Aparimitāyur­jñāna assumed in tantric practice may be one of the reasons why this Aparimitāyur­jñānasūtra was classified within most Kangyurs as a tantra rather than as a sūtra.

By far the most widely used version of the text for recitation in recent times has been the other version (1), Toh 674 and 849, and within that “three oṁ” version it is the slightly different dhāraṇī of Toh 849 that most readers will find corresponds to practice texts in use in Tibetan Buddhist monasteries and Dharma centers. Nevertheless, the present “two oṁ” version has been preserved over the centuries in all Kangyurs, too, and readers familiar with the Chinese versions of the sūtra will find it more familiar than the other version. The text in this version is mentioned little in the literature of recent centuries but, interestingly, the Druptap Küntü collection, already mentioned above, contains two liturgical texts, both composed by Jamyang Khyentse Wangpo in the late nineteenth century, in which he specifies that it is the “two oṁ” dhāraṇī that is to be recited, both from the tradition brought to Tibet by Rechungpa. One is a longevity practice of Four-Armed White Amitāyus, extracted from the texts of Riwoche, and the other a White Aparimitāyur­jñāna sādhana from the lineage that Rechungpa received from Siddharājñī.

We are delighted to have translated and published this version of The Sūtra of Aparimitāyur­jñāna, little known yet present in all Kangyurs, in the hope that it will stimulate the interest of both practitioners and scholars.

Summary

The Buddha, while at the Jetavana in Śrāvastī, tells Mañjuśrī of a buddha realm far above the world, in which lives the Buddha Aparimitāyur­jñāna. He states that those who recite, write, hear, and so on, the praise of this buddha, or make offerings to this text, will have numerous benefits, including a long life and a good rebirth. As vast numbers of buddhas recite it, the mantra, or dhāraṇī, of this buddha is repeated numerous times. This is the lesser known of the two versions of this sūtra in the Kangyur, but possibly represents the earlier translation.

Acknowledgements

The sūtra was translated from Tibetan by Peter Alan Roberts with reference to related Sanskrit versions. Tulku Yeshi of the Sakya Monastery, Seattle, was the consulting lama who reviewed the translation. The project manager and editor was Emily Bower. The proofreader was Ben Gleason. The introduction was compiled by the 84000 Editorial Team, and incorporates material by Peter Alan Roberts.

The translation was completed under the patronage and supervision of 84000: Translating the Words of the Buddha.

Introduction

Overview

The Aparimitāyur­jñāna Sūtra is among the many canonical works in which a particular buddha in another buddhafield is invoked along with the benefits of recalling his name and reciting his dhāraṇī. Associated as it is with longevity, this is one of the most widely read texts in the Kangyur, and Aparimitāyur­jñāna (“Immeasurable Longevity and Wisdom”’) is one of the most frequently portrayed and well-known buddhas in the pantheon of Tibetan Buddhism.

The sūtra is commonly referred to as the Tsédo (tshe mdo, “Sūtra of Longevity”) or Tsézung (tshe gzungs, “Dhāraṇī of Longevity”), and contains a dhāraṇī that is repeated in the text twenty-nine times. It is included in many Tibetan liturgical compilations, and its recitation, usually with a specified number of repetitions, is often advised to people in poor health or facing other difficulties, or is commissioned on their behalf in monasteries.

Although its title identifies it as a sūtra, it is placed in all Kangyurs with the Action Tantras (bya ba’i rgyud, kriyātantra). In common with many other works classified as Action Tantras, there is nevertheless little in the text to identify it as a tantra. The inclusion of a long, repeated dhāraṇī in Sanskrit is presumably one criterion for this classification, although there are many other canonical works with a similar structure that are placed with the sūtras. Other criteria may have been its classification and line of transmission in India, before it was taken to Tibet, or the fact that it has also formed the basis for a wide range of tantra practices, particularly among the higher levels of tantra, in the form of sādhanas of Aparimitāyus.

The text has also survived in a large number of Sanskrit manuscripts (mostly later Nepalese ones); in two Chinese translations; and in a slightly different‍—perhaps earlier‍—Tibetan translation, represented by most of the very numerous manuscripts found in the caves of Dunhuang, where a Khotanese manuscript (probably the oldest surviving version) was also found.

All Kangyurs include two major versions of the sūtra, similar in most respects but differing mainly in the presence or absence of one phrase in the repeated dhāraṇī. The background of the existence of these two versions is discussed below. The version translated here is the lesser known of the two, is much less widely used, and differs from Sanskrit source texts brought to Tibet in the later translation period to a greater extent than the other version.

The present version appears to be derived from the same translation as the other major version, The Aparimitāyur­jñāna Sūtra (1), Toh 674, but the dhāraṇī it contains is the shorter one, matching those found in the Chinese translations, the Dunhuang manuscript in Khotanese, and the earlier Tibetan translation represented by the majority of the very numerous Dunhuang manuscripts in Tibetan.

This sūtra, to the extent that it may represent the translation available in the late eighth century, is one of the set known as the “ten royal sūtras,” thought to be so called either because they represent distillations of the most profound scriptures, or because according to traditional histories they were recommended to King Trisong Detsen for his daily practice by Guru Padmasambhava. As a result of practicing them, the king is said to have extended his life by thirteen years.

In a similar vein, the fact that so many manuscript copies of this text have been found in the Dunhuang caves is due to their production by scribes there having been commissioned on behalf of Trisong Detsen’s grandson, Ralpachen (who reigned in the early ninth century) in order to ensure for the king the longevity that the text itself promises.

Aparimitāyus, Amitāyus, and Amitābha

Aparimitāyur­jñāna (“Immeasurable Longevity and Wisdom”), the buddha who is the subject of this sūtra, despite being described in this text as dwelling in a realm situated in an upward direction from this world, i.e. toward the zenith, has been identified to a varying extent in both Tibet and East Asia with Amitābha, buddha of the realm called Sukhāvatī in the west. Both are often referred to by the shortened name Amitāyus in Sanskrit, Tsépamé (tshe dpag med or, in full, tshe dpag tu med pa) in Tibetan, but are nevertheless likely to have originally been seen as distinct. The confusion that has been caused by these partially overlapping identities is discussed in detail in the introduction (i.9–i.16) to Toh 674, The Aparimitāyur­jñāna Sūtra (1).

The Tibetan Versions of the Sūtra in the Kangyur, Their Differences, and Their Translation

Differences in Content

There are two different but closely related versions of The Aparimitāyur­jñāna Sūtra in most Kangyurs. In keeping with the tradition established over the centuries by editors of all Kangyurs, we have here translated and published them separately, despite their similarity, and have labeled them The Aparimitāyur­jñāna Sūtra (1) and (2). In the Kangyur the two versions are found consecutively in the Tantra Collection, and according to Situ Panchen’s original catalog of the Degé Kangyur both are placed with the Action (Kriyā) Tantras in the subdivision that corresponds to the principal deity (rigs kyi gtso bo) of the Padma (lotus) family. Their Degé recensions are cataloged as Toh 674 and 675, respectively.

The principal distinction between the two versions lies in the length and composition of the repeated dhāraṇī. Compared to the present version (2) of the text, the dhāraṇī in the other version, The Aparimitāyur­jñāna Sūtra (1), Toh 674 and 849, is longer and contains an extra phrase in the middle beginning “oṁ puṇye puṇye…” that is not present here, although the rest of the dhāraṇī is almost the same. It is on that basis that the catalog of the Degé Kangyur distinguishes the two texts by calling this one (Toh 675) the “two oṁ, no puṇye” version (because the repeated dhāraṇī also contains two other phrases beginning with oṁ), while the other one (Toh 674 and 849) it calls the “three oṁ” version. In some other catalogs, e.g. the index to the Narthang Kangyur, the two versions (1) and (2) are called respectively the “large and small Tsédo.”

It is the other version, the “three oṁ” version, The Aparimitāyur­jñāna Sūtra (1), Toh 674 and 849, that is the best known and most widely recited form of the text. It is also the one reproduced in most compilations of dhāraṇīs and anthologies of practice texts, whether ancient or modern. By contrast, the present “two oṁ” version is only preserved in the Kangyur, as here, and is very little known or used. Nevertheless, the fact that it has been included in all Kangyurs as a separate text is an implicit recognition that both versions are authentic. The background to the existence of these two different versions, and the controversies that have sometimes arisen about their origins and authenticity, are explored in the introduction to The Aparimitāyur­jñāna Sūtra (1), Toh 674 and 849 (i.21 et seq) and are set out in slightly more detail here.

One intriguing hint that the Kangyur provides us with regard to these two versions of the text is that the part of the dhāraṇī that is “added” in version (1), Toh 674, but “missing” here in version (2), Toh 675, is included in all Kangyurs, almost identically, but on its own, as the dhāraṇī that forms the very short content of another text in this group, The Essence of Aparimitāyus, Toh 673a. There is no explanation in the various Kangyur catalogs for its presence, but the term “essence” (hṛdaya, snying po, sometimes rendered “heart mantra”) in its title identifies it as a mantra used in at least one tradition of the practice of Aparimitāyus. The only other mention of this mantra in the Kangyur appears to be in the Sarva­durgati­pariśodhana, known best in Tibetan by its shortened title sbyong rgyud (the “Purification Tantra”), a tantra of the Yoga class, in which this same mantra is given as the “essence vidyā-mantra” (hṛdayavidyā, snying po’i rig pa) of the tathāgata Aparimitāyuḥ­puṇya­jñāna­sambhāra­tejo­rāja, the principal figure in a secondary maṇḍala. What is confirmed by this mantra’s mention in the tantra, as well as its presence as Toh 673a, is at the very least that it is a potentially independent stand-alone mantra phrase, making it easier to understand that it might have been either added or removed at some stage in the evolving transmission of the dhāraṇī in the sūtra.

Apart from the composition of the dhāraṇī, there are a few other differences between Toh 674 and 675, mostly minor. Overall, the two versions appear to be much more closely related than they would be if they represented two different translations made entirely independently of each other. Some of the minor textual differences between the present version of the sūtra and Toh 674 are flagged in the notes, but among the more significant are the following:

• The name tshe dang ye shes dpag tu med pa (Aparimitāyur­jñāna), which occurs a number of times here in Toh 675, is given its shorter form in Toh 674: tshe dpag tu med pa (Aparimitāyus).

• The Tibetan rendering of the longest form of the name Aparimitāyur­jñāna­suviniścita­tejo­rāja here (in Toh 675) is tshe dang ye shes dpag tu med pa shin tu rnam par gdon mi za ba’i rgyal po, while in Toh 674 it is tshe dang ye shes dpag tu med pa shin tu rnam par nges pa’i gzi brjid kyi rgyal po.

• The name of Aparimitāyur­jñāna’s realm in Toh 674 is yon tan dpag tu med pa sogs pa (Aparimita­guṇa­saṃcaya, “Accumulation of Immeasurable Qualities”), but here in Toh 675 at the first mention of it (see 1.3) it is just yon tan dpag tu med pa (Aparimitaguṇa, “Immeasurable Qualities”).

• Its second mention here in Toh 675 (see 1.8), however, not only gives it in full, as in Toh 674, but also prefixes it with the words bde ba can, which can be rendered as “blissful” but also as the realm name Sukhāvatī, somewhat confusingly identifying this realm with that of the “other” Amitāyus who is Amitābha, and perhaps confirming that the conflation of these two buddhas (see above) occurred at an early date in Tibet.

• In the three paragraphs in which comparisons using analogies are made of the amounts of merit to be obtained through‍—in Toh 674‍—reciting the sūtra, here in Toh 675 (1.56, 1.58, and 1.60), the merit seems to be obtained through the tathāgata Aparimitāyur­jñāna himself rather than through reciting the text.

• Finally, Toh 674 has a curious concluding line that is not present here in Toh 675: in addition to the beings in the world being overjoyed and rejoicing at the Buddha’s words, the Buddha himself is said to be pleased or delighted (dgyes pa). This appears to be the result of the Sanskrit āttamanās being translated twice, as descriptive for both the world and the Buddha. The ending here in Toh 675 matches similar instances of this standard ending formula.

Differences in Origin and Transmission

In essence, the other, best-known, “three oṁ” version of the sūtra, Toh 674, was brought to Tibet and translated in the later, post-imperial period of transmission. The origins of the present “two oṁ” version are less clear, but it is likely to represent sources known and translated in the early, imperial period.

We have set out the evidence that the translation preserved in the Kangyur of the “three oṁ” version (Toh 674) may be attributed to Puṇyasambhava and Patsap Nyima Drak in the late eleventh or early twelfth century‍—a widely accepted belief‍—in the introduction to that version of the text. It is based primarily on a mention in the lineage record of transmissions received by Minling Terchen Gyurme Dorje, mentions in Tāranātha’s commentary to the sūtra, and the colophon appended to the version of the text as reproduced in the Druptap Küntü (sgrub thabs kun btus), a collection of sādhanas of the Sakya tradition compiled by Jamyang Loter Wangpo (1847–1914). There is also evidence that other, later translations of the “three oṁ” version were made. Here we will concentrate in particular on what may be inferred of the origins of the present, “two oṁ” version, Toh 675.

The available sources of information about the origins of the two versions do not include the most usual ones for Kangyur texts, colophons and Kangyur catalogs. There are no translators’ colophons to either version of the text in any of the different Kangyurs. Most of the Kangyur inventories and catalogs merely list the titles and distinguish them using epithets such as “three oṁ” and “two oṁ,” or “long” and “short.” Some, including the catalog of the Degé Kangyur, explicitly state that the translators are unknown. The one exception is the catalog (dkar chag) of the Narthang Kangyur, which appears to attribute both versions of the sūtra to Puṇyasambhava and Patsap Nyima Drak. As well as being inherently unlikely, this is at odds with all other sources of information. A look at the folios of the catalog concerned reinforces suspicions of an erroneous attribution, as the carving of the catalog’s woodblocks seems to have run into problems for this entry, which coincides with a folio break. Indeed, the Narthang catalog’s confusing attributions at this point do not reflect the titles or colophons in the body of the Narthang Kangyur itself.

Despite this dearth of direct canonical records, what we do know is that at least one version of the sūtra was translated during the early translation period. Firstly, both the early ninth century inventories of translated texts, the Denkarma (ldan dkar ma) and Phangthangma (phang thang ma), mention, respectively, texts entitled The Dhāraṇī of Aparimitāyus (tshe dpag tu med pa’i gzungs, 110 ślokas in length) and Aparimitāyus (tshe dpag tu med pa, 120 ślokas), both probably referring to a version of this text; in both inventories it is placed in the category “miscellaneous long and short dhāraṇīs” (gzungs che phra sna tshogs).

Secondly, very tangible evidence of the existence of translations in the early period is provided by the very large number of manuscript copies of the sūtra, in Tibetan, found among the Dunhuang manuscripts, a large group of which can be dated to between 830 and 850, most having apparently been made on the orders of King Ralpachen (r. 815–41)‍—presumably to create the meritorious results that the text itself describes. Crucially, all the Dunhuang manuscripts that we have been able to examine contain the “two oṁ” version of the dhāraṇī. This fact, combined with the evidence from Minling Terchen, makes it very likely that the present version of the text, the “three oṁ” version, is the translation dating from the later translation period, while Toh 675, the “two oṁ” version, is more closely related to the one originally made in the early translation period.

Thirdly, an early period translation of the “two oṁ” version is the target of some openly dismissive remarks made in several works by later Sakya scholars, including Ngorchen Kunga Zangpo and his disciple Kunga Lekrin in the fifteenth century, and Amé Zhab Ngawang Kunga Sönam in the seventeenth century. Both the latter authors discuss the various theories raised to account for the existence of the two versions, and go on to confirm the validity only of the later translation, The Aparimitāyur­jñāna Sūtra (1), Toh 674. Following a leading statement attributed by Ngorchen Kunga Zangpo to his own teachers, they are both scathingly critical of the “two oṁ” version, and of the “earlier translators” responsible for it, whom they accuse of deliberately omitting part of the dhāraṇī to express dissatisfaction with their stipend. Such comments have to be seen in the context of arguments regarding authenticity between proponents of the “late” versus the “early” translation traditions (which concerned principally the tantras rather than any of the other canonical genres), but they do at least confirm that the “two oṁ” version of the text was widely recognized as having been translated in the early, imperial period.

Debates About Difference and Authenticity

The writings of Ngorchen, Kunga Lekrin, and Amé Zhab go to some lengths to throw doubt on the authenticity of the “two oṁ” version, with Amé Zhab’s, in particular, using arguments that in some cases refute points made in favor of it by opponents in the debate who remain invisible and unidentified. For apart from Tāranātha’s commentary, we have not been able to find actual writings of the period defending the “two oṁ” version from the disparagements of the authors mentioned, other than general notes by catalog compilers to the effect that “both versions are authentic.” The substance of such writings or statements can only be inferred, from arguments put forward in order to neutralize them by those in favor of the “three oṁ” version. The following points represent the principal arguments that these writers deploy to justify their unusually harsh judgment.

Absence of Indic source texts

Amé Zhab writes that Ngorchen Kunga Zangpo personally examined three Sanskrit manuscripts of the sūtra, all of which were the “three oṁ” version, implying that there were no Sanskrit texts attesting to the “two oṁ” version. While there is little doubt that the “three oṁ” version of the text is indeed authentic to its source, Ngorchen’s claim that it is the only authentic version is explicitly refuted and dismissed by Tāranātha, who in his commentary says there were many Sanskrit manuscripts of both versions, and relates in his autobiography how, on a visit to Narthang, he was himself shown several Sanskrit manuscripts of the sūtra, of which one was clearly the “two oṁ” version.

From today’s perspective, no surviving Sanskrit versions of the text from Tibet have been discovered, and the only known Sanskrit manuscripts are Nepalese ones dating from much later centuries. However, the veracity of Tāranātha’s report can be indirectly confirmed by Sten Konow’s 1916 comparison of two versions, one in Sanskrit based on Nepalese manuscripts and one a Khotanese manuscript (probably centuries earlier) found in the Dunhuang caves by Sir Aurel Stein. The Nepalese Sanskrit has the “three oṁ” version of the dhāraṇī while the Khotanese has the “two oṁ” version. Ngorchen and his successors were also presumably unaware that in both of the two Chinese translations, too, the dhāraṇī is the “two oṁ” version.

Different versions might refer to different buddhas

The argument assumed here is based on the two more or less distinct buddhas to whom the shortened name Amitāyus, or Tsepamé (tshe dpag med) in Tibetan, might refer. Tibetan authors took pains to distinguish between, on the one hand, “Amitāyus of the zenith” (steng phyogs kyi tshe dpag med) and “Amitāyus of Akaniṣṭha” (’og min gyi tshe dpag med), both likely to be references to Aparimitāyur­jñāna; and on the other, “Amitāyus of Sukhāvatī” (bde ba can gyi tshe dpag med), also sometimes known as “Amitāyus of the Drum of Immortality” (’chi med rnga sgra’i tshe dpag med, Dundubhisvara-Amitāyus), both of whom can be identified with Amitābha.

Conceivably, it could therefore be proposed that while one version of the text described the existence of one of these buddhas and the benefits to be obtained from remembering his name and honoring him, the other might be legitimately different because it focused instead on the other buddha. There are indeed some minor differences between the two versions concerning the names used at certain points in the text, as mentioned above at i.14. It is not clear who might have used this notion as an explanation for the existence of variant dhāraṇīs, but in any case both Kunga Lekrin and Amé Zhab are emphatic in insisting that both versions are clearly focused on the tathāgata Aparimitāyur­jñāna, whose buddha realm is in the zenith.

One hundred and eight names are mentioned in the sūtra itself

Both versions of the sūtra mention (in 1.5, 1.6, and 1.8) that the benefits of an increased lifespan will come to beings who hear, remember, or write “the one hundred and eight names of Aparimitāyur­jñāna.” Kunga Lekrin and Amé Zhab interpret the one hundred and eight names as a reference to the dhāraṇī, the “three oṁ” version of which does indeed have 108 syllables if the euphonic rules of Sanskrit are applied to elide the two instances of final and initial a, or if the final svāhā is omitted. The “two oṁ” dhāraṇī, they point out, only has seventy-seven syllables, is therefore incomplete, and must be incorrect.

At first sight, this is perhaps the most convincing argument in favor of the “three oṁ” version of the text. On closer examination, however, some of its initial attractions seem less clear. To see the dhāraṇī as comprising one hundred and eight names requires a stretch of the imagination. The dhāraṇi contains semantically coherent words and phrases of which few are actual “names,” and even if all the many compound words are divided into their irreducible units their number still remains less the number of syllables. Of the many “hundred-and-eight-name” or “hundred-name” texts to be found in the Kangyur and elsewhere, most actually do contain distinct lists of names just as their titles suggest, and there do not seem to be other instances of a dhāraṇī being referred to as a “one-hundred-and-eight-syllable appellation,” as this phrase might possibly be interpreted. Indeed, Tāranātha points out in his commentary that to account for the fact that the one hundred and eight names are also mentioned in the “two oṁ” version of the text we have to assume that it refers to a separate text on the one hundred and eight names elsewhere. Such cases, he says, are not unknown; he gives the example of the hundred names of Śrī Heruka. In fact, however, no text enumerating the hundred and eight names of Aparimitāyur­jñāna appears ever to have been identified.

From the perspective we can take today, historically and geographically wider than was possible for those scholars, the strength of this argument using the hundred and eight names to promote the exclusive authenticity of the “three oṁ” version is much diminished. That not only the Khotanese version (probably the oldest manuscript of all) but also the many Dunhuang texts and both Chinese translations all contain the “two oṁ” dhāraṇī, with its seventy-seven syllables, yet still make mention of “one hundred and eight names” adds considerable weight to Tāranātha’s inference that this is not necessarily a reference to the dhāraṇī itself.

The longer dhāraṇī may include a phrase added from elsewhere

Amé Zhab writes that “certain later scholars” have claimed that the “two oṁ” version is correct but that the dhāraṇī of the “three oṁ” version has been expanded by adding to it the mantra of Aparimitāyus from the sbyong rgyud, i.e. the Sarva­durgati­pariśodhana, a tantra, widely used for funerary rites, of which there are two Tibetan translations in the Kangyur made in the early and later translation periods, Toh 483 and 485 respectively.

It is true that in the “three oṁ” version the centrally placed phrase starting “oṁ puṇye puṇye mahāpuṇye…”‍—the phrase whose presence or absence marks the difference between the dhāraṇīs of the two versions‍—is very similar to a mantra found in the Sarva­durgati­pariśodhana as the “essence” (hṛdaya, snying po) mantra in one of the secondary maṇḍalas in the text, that of of a tathāgata whose full name is given as Aparimitāyuḥ­puṇya­jñāna­sambhāra­tejo­rāja, but who is referred to as Amitāyus in the several commentaries. This mantra is also strikingly similar to the one introduced as the “essence” mantra of Aparimitāyus in the very short text that precedes the two Aparimitāyur­jñāna Sūtra s in most Kangyurs, and of which we have published a translation under the title The Essence of Aparimitāyus (Toh 673a).

Amé Zhab, however, dismisses the argument as inadmissible, writing that in the Sarva­durgati­pariśodhana mantra the words āyuḥpuṇya and sarvacayaṃkari [sic] are found that are not in the dhāraṇī of the “three oṁ” version; these scholars, he says, have simply not consulted the tantra properly. But his objections, valid though they presumably seemed from the copies of the relevant books he had to hand, are more easily dispelled than sustained. The first of the words he cites, āyuḥpuṇya, is present in one of the two only slightly differing variants of the “three oṁ” dhāraṇī preserved in most Kangyurs, the version in the Tantra Collection (rgyud ’bum), Toh 674. The second, sarvacayaṃkari, is indeed not in any version of the present dhāraṇī but, even assuming that Amé Zhab had intended the somewhat different spellings upacayakāriṇi or upacayakāraṇi actually found in the Sanskrit and Tibetan respectively of the Sarva­durgati­pariśodhana, it is present only in the version of that tantra as translated in the later translation period (Toh 485), and is absent in the earlier translation (Toh 483), which reads oṁ puṇye puṇye mahāpuṇye aparimita‑ayūḥpuṇya-jñāna-saṃbhāropacite svāhā.

In other words, a close relationship between this Aparimitāyus mantra in the Sarva­durgati­pariśodhana and the dhāraṇī of the “three oṁ” version of the sūtra does seem eminently possible, and would not seem particularly surprising given the cross-text interrelationships between mantras and dhāraṇīs of related buddhas and deities. It does seem to suggest that the phrase in question could be a potentially independent and detachable component of the dhāraṇī, but gives us no particular hint as to whether it is more likely to have been added to one version, or removed from another.

Differences and Mixture: Possible Conclusions

The story of these two different versions may seem already complex enough, but there is more to be explored. Apart from the differences in the dhāraṇī, the other differences (listed above at i.14 and in the notes) between the two versions in the Kangyurs are relatively minor. These two translations seem most unlikely to have been made independently, even allowing for the possibility that the Sanskrit texts they were made from were very similar. The most obvious explanation might be that the later translation was based on the earlier, adapting it to conform to a slightly different Sanskrit original. However, an examination of the wording and terminology of the Dunhuang manuscripts‍—which predate the work of Puṇyasambhava and Patsap Nyima Drak by several centuries‍—shows that they almost certainly represent a Tibetan translation different from the present “two oṁ” version (Toh 675) as it has survived in the Kangyur. It is therefore tempting to conjecture that this “two oṁ” version in the Kangyur might in fact be a back-adaptation of the later translation, edited at some stage to conform to the alternative “two oṁ” dhāraṇī of the earlier translation but not otherwise reflecting that earlier translation fully. This hypothesis is supported by the fact that a few significant elements of the later “three oṁ” version that are not present in the “two oṁ” Dunhuang manuscripts are present here. Among the most significant in this regard is the similar rendering, common to both Kangyur versions, of what we may presume to be the standard, modular phrase in the Sanskrit manuscripts, eva hi tiṣṭhati dhriyate yāpayati, as bzhugs so/ /tshe ’dzin cing tshe mthar phyin par bzhed de in 1.3, which Tāranātha identifies as a feature unique to the work of Patshap and the other later period translators.

Whatever the case in this confusing blend of textual variation and admixture, there is no single clinching argument in favor of accepting one version over another. Both versions, surely, can be considered authentic; and although the compilers and editors of the many Kangyurs do not seem to have noted their reasons, there must have been enough consensus on this point‍—despite all the controversies‍—for both to have been preserved in all Kangyurs.

Sanskrit and Chinese Versions of the Text

There are many surviving manuscripts of the text in Sanskrit, but none that can be reliably dated much before the early ninth century, the period when it was first translated into Chinese, and probably into Tibetan for the first time (see below). Most of the Sanskrit manuscripts are Nepalese and are dated considerably later.

The oldest known Indic version of the text may be one from east Turkestan in what came to be called “Khotanese,” the old Iranian dialect of that region during the later period of the time when Buddhism was prevalent there. It is written in the Upright Gupta script, and probably dates to the seventh or eighth century. The manuscript was discovered in the Mogao caves in Dunhuang by Sir Aurel Stein in 1907, and in 1912 Ernst Leumann made a short comparison of the Sanskrit of this sūtra’s opening sentences with a few fragments of the Khotanese text. In 1916, two Sanskrit editions saw the light independently. One, by Sten Konow, compared an edition of a Nepalese Sanskrit version with a complete edition of the Khotanese fragments, along with the first English translation. The other, by Max Walleser, was based on a Nepalese manuscript and included a German translation. Walleser’s German translation has been translated into English by Richard K. Payne in his paper on this sūtra. Jonathan Silk has made an English translation from Walleser’s edition of the Sanskrit, and attests in his unpublished paper, “The Most Important Buddhist Scripture? The Aparimitāyur­jñāna and Medieval Buddhism,” to the relatively large number of extant Sanskrit manuscripts‍—well over one hundred‍—either in isolation or compilations, indicating how popular this sūtra was in Buddhist practice.

As noted above, the Nepalese Sanskrit manuscripts are all of the “three oṁ” version of the text, and therefore correspond more closely to version (1), Toh 674. The Khotanese, on the other hand, contains the “two oṁ” version of the dhāranī throughout, and is thus closer to the present version and to the many Dunhuang manuscripts in Tibetan.

Two translations of the sūtra into Chinese were made, one by Facheng (Taishō 936) in the early ninth century, and the other by Fatian (Taishō 937) in the late tenth century. Both Chinese translations contain the “two oṁ” dhāraṇī, and are therefore closer to the Khotanese, the present Tibetan version of the text, and the Dunhuang manuscripts than they are to the Nepalese Sanskrit manuscripts and to version (1) in Tibetan, Toh 674.

The Sūtra in Buddhist Practice

As mentioned above in the overview of this introduction, recitation of the Tsédo a specified number of times has historically been‍—and is still nowadays‍—prescribed as a practice to people in poor health or facing other difficulties, and to practitioners more generally in order to ensure longevity, and so on. Recitation by the monks or nuns in a monastery is also commissioned for the same reasons. According to the Padma Kathang, the eighth century Tibetan king, Trisong Detsen, was advised to recite this text daily (along with the other works known as the “ten royal sūtras”), as a result of which the king’s life is said to have been prolonged by thirteen years beyond the limit predicted by astrological reckoning.

The sūtra itself particularly emphasizes the beneficial effects of writing it out or causing it to be written out, and there is ample evidence that this recommendation was taken seriously in the form of the very large number of commissioned copies, mostly in Tibetan and Chinese and dating to the eighth and ninth centuries, found in the Dunhuang caves by Stein and Pelliot in the early years of the twentieth century. Many of them appear to have been commissioned in the name of the Tibetan king Ralpachen, who reigned in the early ninth century and was the grandson of Trisong Detsen. Among the bundles acquired by the two explorers for the British and French governments, there are over a thousand copies of the sūtra now in the British Library in London, and a similar number in the Bibliothèque Nationale in Paris. Some five hundred further copies remain in libraries in China, many have found their way to Japan, and there are some two hundred in St. Petersburg. As noted in i.19, a large majority of these manuscripts, and probably all of those from the period concerned, contain the equivalent of the “two oṁ” dhāraṇī

In the context of tantric practice, the Tibetan canonical literature contains a number of sādhanas of Aparimitāyur­jñāna, particularly among the higher levels of tantra. One example is the liturgy composed by the tenth-century Jetāri, one of the principal teachers of Vikramaśīla Monastery, which was known for its promulgation of higher tantras. That practice was introduced into Tibet by Bari Lotsāwa Rinchen Drak (mentioned above), and thereby became a part of the Sakya tradition.

There are also five Aparimitāyur­jñāna texts in the Tengyur that were composed by Siddharājñī, a female guru active in India in the beginning of the twelfth century, from whom Rechungpa, pupil of the famous Milarepa, received transmissions that Marpa had not been able to bring back to Tibet. At least three if not all of these Siddharājñī texts were translated into Tibetan by Varacandra, another of Rechungpa’s teachers, together with the Tibetan Lenchung Darma Tsultrim (glan chung dar ma tshul khrims), about whom little is known other than this translation work with Varacandra. Rechungpa introduced the practice into Tibet, where it spread from his own lineage, the Rechung Kagyü or Rechung Nyengyü, to other Kagyü traditions. These tantric Aparimitāyur­jñāna practices are based upon the five-family system of the higher tantras, and they involve an elaborate visualization of oneself as a red Aparimitāyur­jñāna, wearing the saṃbhogakāya costume and holding a vase of amṛta, with an entourage of deities within a palace, and the visualization of channels within the body. As Aparimitāyur­jñāna is auspicious for long life, his empowerment is given as a long life blessing.

Within the indigenous Tibetan literature, a very large number of Aparimitāyur­jñāna sādhanas have been created over the centuries within all the lineages and schools. The Nyingma tradition of rediscovered treasure texts (gter ma), too, has produced many Aparimitāyur­jñāna revelations, from Nyangral Nyima Ozer (nyang ral nyi ma ’od zer, 1136–1204) onward. The importance that Aparimitāyur­jñāna assumed in tantric practice may be one of the reasons why this Aparimitāyur­jñānasūtra was classified within most Kangyurs as a tantra rather than as a sūtra.

By far the most widely used version of the text for recitation in recent times has been the other version (1), Toh 674 and 849, and within that “three oṁ” version it is the slightly different dhāraṇī of Toh 849 that most readers will find corresponds to practice texts in use in Tibetan Buddhist monasteries and Dharma centers. Nevertheless, the present “two oṁ” version has been preserved over the centuries in all Kangyurs, too, and readers familiar with the Chinese versions of the sūtra will find it more familiar than the other version. The text in this version is mentioned little in the literature of recent centuries but, interestingly, the Druptap Küntü collection, already mentioned above, contains two liturgical texts, both composed by Jamyang Khyentse Wangpo in the late nineteenth century, in which he specifies that it is the “two oṁ” dhāraṇī that is to be recited, both from the tradition brought to Tibet by Rechungpa. One is a longevity practice of Four-Armed White Amitāyus, extracted from the texts of Riwoche, and the other a White Aparimitāyur­jñāna sādhana from the lineage that Rechungpa received from Siddharājñī.

We are delighted to have translated and published this version of The Sūtra of Aparimitāyur­jñāna, little known yet present in all Kangyurs, in the hope that it will stimulate the interest of both practitioners and scholars.