The Aparimitāyurjñāna Sūtra (1)

There are two different but closely related versions of The Aparimitāyur­jñāna Sūtra in most Kangyurs. We have called them The Aparimitāyur­jñāna Sūtra (1) and (2), and in keeping with the tradition established by editors of all Kangyurs we have translated and published them separately, despite their similarity. They are found consecutively in the Tantra Collection, and according to Situ Panchen’s original catalog of the Degé Kangyur they 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 Toh 675, respectively.

Before discussing the complex differences between The Aparimitāyur­jñāna Sūtra (1) and (2), it should be mentioned here that The Aparimitāyur­jñāna Sūtra (1) itself also exists in two slightly different forms, the difference being a minor one in the dhāraṇī that is less significant than the differences between versions (1) and (2). In those Kangyurs that, like the Degé, have a separate compendium of dhāraṇīs (gzungs ’dus, dhāraṇī) section, a second version of (1) appears there; in the Degé it is Toh 849, and is essentially a duplicate of the first (Toh 674), except that the dhāraṇī differs in one word. In the middle part of the repeated dhāraṇī, where the Toh 674 version has aparimitāyur­puṇya-jñāna, the Toh 849 version reads aparimita­puṇye aparimita­puṇya-jñāna. This latter reading is also found in the versions in the compendium of dhāraṇīs in several Tshalpa Kangyurs;¹⁶ is the only version to be found in some other Kangyurs, notably those of the Themphangma group;¹⁷ and corresponds to the version in the Druptap Küntü, which will be mentioned in more detail below. Readers of this translation will see the version of the dhāraṇī corresponding to the version (Toh 674 or Toh 849) that they have selected.

To return to our discussion of the more significant differences between the two major variants, versions (1) and (2), the principal distinction lies again mainly in the repeated dhāraṇī, but in this case is a marked difference in its length and composition. In the present version, The Aparimitāyur­jñāna Sūtra (1), Toh 674 and 849, the middle part of the dhāraṇī comprises a phrase beginning “oṁ puṇye puṇye …” that is not present in the other version, The Aparimitāyur­jñāna Sūtra (2), Toh 675, 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 674 and 849) the “three oṁ” version (because the repeated dhāraṇī also contains two other phrases beginning with oṁ), while the other one (Toh 675) it calls the “two oṁ, no puṇye” version.¹⁸ In some other catalogs, e.g. the index to the Narthang Kangyur, the two texts are called respectively the “large and small Tsédo.”¹⁹

Another intriguing hint that the Kangyur provides us with regard to these two versions of the text is that the part of the dhāraṇī “present” in this version (1), Toh 674, but “missing” 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.²⁰ 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. Its use in arguments and counter-arguments in favor of the authenticity of each of the two versions of the sūtra is discussed in the introduction to Toh 675 at i.29–31.

The translation published here, the “three oṁ” version, represents the form of the text most widely recited and reproduced in modern compilations. That the “two oṁ” version, The Aparimitāyur­jñāna Sūtra (2), Toh 675, is also preserved in all Kangyurs as a separate text is an implicit recognition that both are authentic. The controversies that have nevertheless sometimes arisen about that question are discussed below and in the introduction to Toh 675, i.20-32.

Apart from the composition of the dhāraṇī, there are some other relatively minor differences in wording between Toh 674 and 675, as recorded in the notes. Among these differences, the most significant are as follows:

There are no translators’ colophons to either version of the text, in any of the different Kangyurs. Nor are there colophons for the many reproductions of the text in other Tibetan compilations‍—with the exception of the version in the Druptap Küntü, mentioned below. Outside the texts themselves, this absence of information about the translators is confirmed in most of the Kangyur inventories and catalogs (including the catalog of the Degé Kangyur, which explicitly states that the translators are unknown)‍—but with two exceptions.

These two exceptions are the catalog (dkar chag) of the Narthang Kangyur (regarding both versions) and the lineage record of transmissions received by Minling Terchen Gyurme Dorje (regarding this version, Toh 674, only). Both attribute the translation to the Indian paṇḍita Puṇyasambhava and the Tibetan translator Patsap Nyima Drak, who were active in the late eleventh and early twelfth centuries. This same attribution is made in the colophon of the sūtra 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).²⁴ Now, these two translators are widely agreed to be the translators of the text that follows immediately after the two versions of The Aparimitāyur­jñāna Sūtra in the Kangyur, The Dhāraṇī “Essence of Immeasurable Longevity and Wisdom” (Toh 676), which (as mentioned in i.15 above) is not closely related to The Aparimitāyur­jñāna Sūtra despite its title, and one possibility is that this attribution of The Aparimitāyur­jñāna Sūtra to those translators might have been an error brought about by the proximity of the texts and their confusingly similar titles.

In particular, the catalog of the Narthang Kangyur alone would seem to be an inadequate source of information here. The Narthang Kangyur was produced in the early 1730s, and reproduces versions of this group of texts, the equivalents of Toh 674 through Toh 676,²⁵ without colophons for either version of the sūtra. It is only in its catalog that the attribution to these two translators appears to be made. A look at the page in the catalog concerned reinforces suspicions of an erroneous attribution, as the carving of the catalog’s woodblocks²⁶ seems to have run into problems for this entry, which coincides with a folio break. Indeed, the catalog not only appears to attribute both versions of The Aparimitāyur­jñāna Sūtra to these translators, but also omits naming these (or any) translators for the following text, The Dhāraṇī “Essence of Aparimitāyur­jñāna” (Toh 676), which all the other Kangyur catalogs agree in attributing to them.

On the other hand, the evidence provided by Minling Terchen’s lineage record carries considerably more weight. As well as naming the same two translators mentioned above, it details the several different sequences of lineage holders through whom the text of the “three oṁ version” (i.e. Toh 674) came to Tibet, mostly including the sequence: Jetari; one or more of the masters known as Vajrāsana(pati); and Bari Lotsāwa Rinchen Drak (1040–1111), the second holder of the Sakya Throne. Abhayākaragupta (d. 1125, who may have been one of the Vajrāsana title holders) is also mentioned in one of these sequences.²⁷ Abhayākaragupta and Bari Lotsāwa are both associated with the transmission to Tibet of many Action Tantra practices, including several collections entitled The One Hundred Sādhanas. These were held and transmitted notably by the Sakya tradition, and if this text was among the many works brought to Tibet by this lineage and translated in the process, that would help explain how the attribution to the two translators was preserved outside the transmission of Kangyurs as such. It would also help explain why this text was classified as an Action Tantra, in common with all the other works in those collections.

Curiously, however, the attribution of this translation to Puṇyasambhava and Patsap Nyima Drak is complicated by several mentions in the commentaries of other translations. Writing in the early fifteenth century, Kunga Lekrin mentions a translation‍—apparently of this text‍—by Yarlung Lotsāwa Drakpa Gyaltsen, who lived in the late thirteenth and early fourteenth centuries, and a later fourteenth century one by the Jonangpa scholars Kunpang Chödrak Palzang, Sabzang Mati Paṇchen, and Jonang Lotsāwa Lodrö Pal, all three contemporaries and important disciples of Dolpopa.²⁸ In the early seventeenth century, Amé Zhap Ngawang Kunga Sonam mentions the same names,²⁹ and his contemporary Tāranātha brings up the latter two, as well as mentioning Patsap, in his discussion of the translation of the verbs used in their translation of 1.3.³⁰ These translations appear no longer to exist, but confirmation would require further research. The balance of the existing evidence, nevertheless, favors the idea that this version of The Aparimitāyur­jñāna Sūtra was indeed translated and introduced to Tibet by Puṇyasambhava and Patsap Nyima Drak in the late eleventh or early twelfth century.

We also know with little doubt (as discussed more fully in the introduction to The Aparimitāyur­jñāna Sūtra (2), Toh 675) that at least one version of the sūtra was translated during the early translation period. Both the early ninth century Denkarma (ldan dkar ma) and Phangthangma (phang thang ma) inventories of translated texts 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).³¹ Material evidence, too, 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 850,³² 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.

All the Dunhuang manuscripts that we have been able to examine contain the “two oṁ” version of the dhāraṇī.³³ 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.

An additional confirmation of this hypothesis can be found in commentarial works by later Sakya scholars, Kunga Lekrin (a disciple of Ngorchen Kunga Zangpo) in the fifteenth century, and Amnyé Zhab Ngawang Kunga Sönam in the seventeenth century. Both 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 (represented here). Following a leading statement attributed by Ngorchen Kunga Zangpo to his teachers, these authors are scathingly critical of the “two oṁ” version, The Aparimitāyur­jñāna Sūtra (2), Toh 675, 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 the mutual lack of trust between proponents of the “late” versus the “early” translation traditions regarding authenticity and lineage (which concerned principally the tantras rather than any of the other canonical genres).

Among other reasons they cite to justify this unusually harsh judgment, (for more detail, see introduction to The Aparimitāyur­jñāna Sūtra (2), Toh 675), Ngawang Kunga Sönam writes that Ngorchen Kunga Zangpo mentioned seeing three Sanskrit manuscripts of the sūtra, all of which were the “three oṁ” version. While there is little doubt that this “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 reports seeing, on a visit to Narthang, several Sanskrit manuscripts of the sūtra, of which one was clearly the “two oṁ” version. 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.³⁴ Ngorchen and his successors were also presumably unaware that in both of the two Chinese translations, too, the dhāraṇī is the “two oṁ” version.³⁵

Apart from Tāranātha’s, we have not been able to find writings of the period defending the “two oṁ” version from the disparagement of the authors mentioned, other than notes by catalog compilers to the effect that “both versions are authentic.” However, Amé Zhab’s writing on the subject in favor of the “three oṁ” version does present some of the arguments used, in order to refute them‍—somewhat unconvincingly. More details are to be found in the introduction to The Aparimitāyur­jñāna Sūtra (2), Toh 675, at i.21 et seq.

Both versions, therefore, 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.

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 between the two versions in the Kangyurs (the principal ones listed above at i.26 and others in the notes) are relatively so minor that 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 Puṇyasambhava and Patsap Nyima Drak by several centuries‍—shows that they almost certainly represent a Tibetan translation different from the “two oṁ” version (Toh 675) that has survived in the Kangyur.³⁶ It is therefore tempting to conjecture that the “two oṁ” version in the Kangyur could in fact be a back-adaptation of the later translation, edited 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 in the “two oṁ” version in the Kangyur. These elements are discussed more fully in the introduction to Toh 675, at i.15 et seq.