Now let us turn our attention to Tibet. Sūtra commentaries were among the early translations into Tibetan. We know this from early catalogues such as the Lhan (or lDan) kar ma (LKM), which is preserved in the bsTan 'gyur, "translated treatises," which constitutes one half of the Tibetan Buddhist canon (the other half is the bKa' 'gyur, "translated word [of the Buddha]"; see Harrison and Martin, in this volume). This catalogue, compiled in a Dragon year such as 800, 812, or 824 C.E., after approximately one hundred and fifty years of Tibetan translations of Buddhist texts, is an inventory of treatises stored in the Lhan kar ma Palace in Tibet. Lalou, who has transcribed and indexed the LKM, records 736 titles2 in thirty sections. Section twenty (nos. 514-564) contains the "Commentaries[page 114] on Mahāyāna Sūtras"; section twenty-one (nos. 565-572) contains the "Sūtra Commentaries Translated from Chinese" (318). Of these sixty recorded in the LKM, approximately3 half have been preserved in the bsTan 'gyur while the other half have been lost. Thus, fifty percent of the sūtra commentaries recorded in the LKM did not survive during the dark ages (ca. 840-1040 C.E.) between the early and later propagations of Buddhism in Tibet.

Eventually, Tibetan savants preserved translated sūtra commentaries in the bsTan 'gyur. The original Old sNar thang bsTan 'gyur dates back to the early fourteenth century. Bu ston Rin chen grub of Zhwa lu Monastery copied and expanded the bsTan 'gyur in 1335. All of the extant bsTan 'gyurs are descended from the Zhwa lu Monastery bsTan 'gyur and all of them have divided the sūtra commentaries into two sections: Prajñāpāramitā (Sher phyin), containing commentaries on the Prajñāpāramitā sūtras, and Sūtra Commentary4 (mDo 'grel), containing commentaries on non-Prajñāpāramitā sūtras. Although the LKM did not divide the Mahāyāna sūtras into these same two sections, it did place the Prajñāpāramitā commentaries first among sūtra commentaries. Likewise, the LKM placed the Prajñāpāramitā sūtras before all other sūtras, a tradition continued in many of the extant bKa' 'gyurs.

Each of these two sections of the bsTan 'gyur contains about forty sūtra commentaries. However, not all sūtra commentaries are found in the Prajñāpāramitā and Sūtra Commentary sections; seven more can be found in the Tantra (rGyud),5 Cittamātra (Sems tsam),6 and Miscellany (sNa tshogs) sections. Three of the four sūtra commentaries in the Miscellany section are by Tibetans, for this section is reserved for writings of ancient Tibetans, and the fourth lists no author.7 The compilers of the LKM included four or five (see the previous note) of these seven texts among the sūtra commentaries, but the editors of the bsTan 'gyur decided to place them in these other sections. Their placement in the Tantra and Cittamātra sections highlights the occasionally arbitrary nature of the classification of treatises as commentaries of sūtra, tantra, or Cittamātra treatises. For the most part, the Peking and sDe dge bsTan 'gyurs have the same sūtra commentaries, with some minor differences as to placement and total number. When the thirty sūtra commentaries lost since the compilation of the LKM are added to the ninety preserved in the bsTan 'gyur, we get a total of 120. Thus, of the more than one hundred sūtra commentaries translated into Tibetan, fewer than one hundred still exist.[page 115]

One-tenth of the sūtras in the bKa' 'gyur, a mere thirty-four, have extant commentaries in the bsTan 'gyur. Eight Prajñāpāramitā sūtras have extant commentaries (a ninth whose commentary is lost is recorded in the LKM)8; approximately twenty-five non-Prajñāpāramitā sūtras have commentaries. The non-Prajñāpāramitā sūtras include four spells (dhāraṇī, gzungs),9 three cherished recollections (anusmṛti, rjes su dran pa),10 one verse (gāthā, tshigs su bcad pa) entitled Ekagāthā, one prayer (praṇidhāna, smon lam) entitled Bhadracaripraṇidhāṇarāja, and sixteen sūtras proper, for a total of twenty-five. Thus, sūtra in this context seems to mean "the word of the Buddha" (buddhavacana) rather than the genre of sūtras that have prologues, introductions, lengthy discourses, and conclusions. Seven sūtras that received one-third of the extant commentaries include some of the most famous, popular, or important. These are the Hṛdaya (with seven commentaries), Vajracchedikā (three), Saddharmapuṇḍarīka (one), Bhadracaripraṇidhāna (six), Laṅkāvatāra (two), Saṃdhinirmocana (five), and Aṣṭasāhasrikāprajñāpāramitā (six). Although all of these sūtras have been translated into Western languages, only some of these sūtras' commentaries have been analyzed with the results published. One example is Donald Lopez's study of Indian and Tibetan commentaries on the Heart Sūtra in which he summarized the seven Indian commentaries and translated two Tibetan commentaries.

Now let us take a closer look at the sūtra commentaries themselves. They range in length from several volumes (Haribhadra's Pañcaviṃśatisāhasrikāprajñāpāramitā, vols. ga to ca) to less than a folio (Asaṅga's Dharmānusmṛtivṛtti); some are in verse (Śālistamba[ka]kārikā [SJT]) while most are predominantly prose (Kamalaśīla's Śālistambaṭika [SJGG]); some discuss several immense sūtras (Smṛtijñānakīrti's*Śatasāhasrikāpañcaviṃśatisāhasrikāṣṭādaśasāhasrikātrayasamānārthāṣṭabhisamayaśāsanā), others only a single verse (Vasubandhu's Ekagāthābhāṣya). Some comment upon entire sūtras (any of the SJD commentaries) and others only on parts of a sūtra such as the prologue (Śākya'i blo's *Daśabhūmisūtranidānabhāṣya) or a chapter (Ye shes snying po's *Saṃdhinirmocanasūtre ṭryamaitreyakevalaparivartabhāṣya). Thus, the commentaries are not homogeneous.

One sūtra commentary has been the subject of more commentaries than any one of the sūtras themselves. The Abhisamayālaṃkāra, a systematic exposition in verse of the Mahāyāna path of deliverance based on the doctrines of the Prajñāpāramitā sūtras [page 116] (in particular, on the Pañcaviṃśatisāhasrikāprajñāpāramitā, according to ṭrya Vimuktisena) has inspired at least twenty commentaries. Tradition includes the Abhisamayālaṃkāra, which has been translated into English by Edward Conze, as one of the Five Treatises of Maitreya, a heavenly bodhisattva, but many scholars attribute the work to Asaṅga, fourth-fifth century C.E. The text has eight chapters, one for each of its eight subjects, which also become the organizing principle for most of its commentaries. The first and dominant subject is the Buddha's omniscience. Because the treatise is very concise, it is difficult to understand without its commentaries, not unlike the ritual sūtras of the non-Buddhists. In fact, it has more features in common with the ritual sūtra genre than with other Buddhist sūtra commentaries: Stcherbatsky describes the Abhisamayālaṃkāra as descriptive, summarizing Prajñāpāramitā doctrine and its practice; concise, requiring commentary to be understood; and mnemonic in arrangement (vi, viii). It has also had the most lasting impact of any sūtra commentary; it serves as a gateway for the study of Prajñāpāramitā sūtras by Tibetan Buddhists of all schools, whose savants have amply added over the centuries to the number of its commentaries. One noteworthy example is gYag ston Sangs rgyas dpal's (1348-1414) eight volume gYag Tik for the study of the Prajñāpāramitā.

The other sūtra commentaries exhibit various commentarial techniques. (Indigenous Tibetan typology of commentary includes, but is not limited to, the tshig 'grel, mchan 'grel, don 'grel, and dka' 'grel; see Wilson, in this volume.) Versifications such as the SJT summarize their sūtras and require commentaries to explain both sūtra and versification. Prose commentaries invariably explain the words and phrases of their sūtras, again to lesser and greater degrees. Kamalaśīla's SJGG comments upon the opening phrase of Buddhist sūtras, evaṃ mayā śrutam ekasmin samaye:

In that [connection], by the expression "THUS" ('di skad; evaṃ), the compiler, having been supplicated, indicates all the contents of the sūtra that come below, in order to avoid disparagement (skur pa; *apavāda) and false attribution (sgro 'dogs pa; *samāropa).

These two [words], "I HEARD" (bdag gis thos pa; mayā śrutam), indicate that I directly heard [the sūtra from the Buddha] and did not understand [its meaning]; I myself heard but [what was heard] is not hearsay coming through a lineage from one [person] to another. [It] was merely heard and not understood, because it is impossible that another besides the Buddha [could][page 117] understand a matter such as this. That also is a cause for inducing belief; otherwise, if an impossible matter were stated, it would not be believed.

"ON ONE OCCASION" (dus gcig na; *ekasmin samaye) is joined to the above "heard"; "occasion" [means] either "time" or "gathering [of] the retinue," because of the great difficulty to hear such a precious sūtra anytime, anywhere. Also, "on one occasion" is joined to the following "the Blessed One resided"; this indicates that for the sake of infinite disciples, at other times the Blessed One resided at other [places]. (146b)

The next level of organization is for a commentary to follow its sūtra's chapter arrangement or a set of topics for its organizing principle. A twofold example of this is Haribhadra's Aṣṭasāhasrikāprajñāpāramitāvyakhyābhisamayālaṃkārālokā, which includes the eight subjects from the Abhisamayālaṃkāra and follows the thirty-two chapters from the Aṣṭasāhasrikāprajñāpāramitā. More than thirty commentaries are organized along similar principles.

Many sūtra commentaries employ five terms in order to introduce their exposition: the "purpose" (prayojana, dgos pa), the "text" (abhidhāna, rjod pa), the "subject matter" (abhidheya, brjod par bya ba), the "connection" (sambandha, 'brel pa), and the "purpose of the purpose" (prayojanaprayojana, dgos pa'i dgos pa). Broido characterizes these terms as describing "the connection between the whole work and the general purposes for which it was written and is to be studied" (6). As far as he knows, the Indians had no single word for these terms whereas the Tibetans called them dgos 'brel ("purpose-connection") (6). Any number of the five terms may be found in a sūtra commentary, and they can be found in twenty-five of the commentaries, most often using four of the terms.

The relation of the five terms to the four anubandhas, which Huparikar describes as the four requisites at the beginning of a text that explain its purpose, may be quite simple. The Buddhists use the five introductory terms called dgos 'brel and certain non-Buddhists use the four anubandhas in order to introduce a text and its purpose. Three terms are similar: subject matter (viṣaya [non-Buddhist], abhidheya [Buddhist]), connection (sambandha), and purpose (prayojana) (121-122). Not surprisingly, in connection with the five terms, no Tibetan translation of the term anubandha is found in any of the sūtra commentaries.

Four of the five terms are used in the SJGS, a commentary to both the SJT and the SJD. After quoting and commenting on the[page 118] verses (kārikās) as well as on many of the sūtra's passages, it interprets both texts according to Yogācāra doctrine, thus bringing into question its traditional attribution to Nāgārjuna, who is credited with founding the Madhyamaka in approximately the second century C.E. The SJGS, whose organizing principle is the quoted verses from the SJT, is, however, encyclopedic in its descriptions of the Eightfold Path and its antithesis, the various realms, their inhabitants, the many localities of rebirth, the five aggregates, the Four Noble Truths, and so forth.

The SJGS gives us more information about its four introductory terms than most of the other commentaries that use them. It discusses at some length these four: the connection, the purpose, the text, and the subject matter. The commentary can be said to have a "connection" because it will explain the SJD and its kārikā; also, it is "connected" with the Buddha and not the works of non-Buddhists. Its "purpose" is—by understanding the meaning of causes and conditions, by realizing that persons and the factors of existence are selfless, and by realizing the absence of grasped and grasper—to become free of the obscurations of defilement and knowledge and so attain the supreme, truly complete buddhahood. Its "text" is the Śālistamba, which uses the example of a young rice plant (śālistamba, sā lu ljang pa) to link inner and outer dependent arising. Its "subject matter" is dependent arising, which is devoid of an agent and so forth, the understanding of which leads to the abandonment of defilement, the arising of wisdom, and the attainment of the Dharma Body (dharmakāya, chos kyi sku) (21b-22b). The omitted term is the "purpose of the purpose." It might also be translated as the ultimate purpose. It is the deeper purpose of the work and, according to Broido, is often more important than the purpose, though dependent upon it (7). However, the SJGS appears to combine the "purpose" with the "purpose of the purpose," since the stated "purpose" is so long and concludes with the attainment of buddhahood, a typical "purpose of the purpose."

Another commentarial system is explained in Vasubandhu's Vyākhyāyukti (NR), a treatise on how to explain and comment upon sūtras. He sets out five components to be included in a sūtra commentary: the purpose (prayojana, dgos pa), concise meaning (piṇḍārtha, bsdus pa'i don), meaning of the words (padārtha, tshig gi don), connections (anusaṃdhi, mtshams sbyar ba), and objections and answers (codyaparihāradvaya, brgal ba/dang lan gnyis) (30b). The[page 119] "purpose" points to the goal or result of the treatise, the "concise meaning" to the meaning and subject of the treatise, the "meaning of the words" explains the concise meaning and so forth, the "connections" explains the order of the words, and the "objections and answers" uphold the treatise's logical and internal consistency. Even though Vasubandhu composed a number of sūtra commentaries, Kamalaśīla (late eighth century C.E.) is the author who most explicitly follows Vasubandhu's instructions. The best example is the SJGG, in which Kamalaśīla introduces the treatise according to the NR's five components. He organizes the commentary according to a sevenfold concise meaning that conforms to Vasubandhu's directives in the NR. Eleven commentaries in all either mention or actually employ this fivefold method. Kamalaśīla wrote three of them: the SJGG, the Avikalpapraveśadhāraṇīṭīkā, and the Vajracchedikāṭīkā.

As recorded by the Tibetan tradition, the authors of the sūtra commentaries include the greatest luminaries of India: Maitreya, Nāgārjuna, Asaṅga, Vasubandhu, Dignāga, and Śāntideva. However, the authenticity of the authorship of the first two authors is not accepted unequivocally, making Asaṅga the most venerable of the sūtra commentary authors credible to most modern scholars. The next oldest author, and most prolific in this category, is Vasubandhu, with nine commentaries. Some of the other authors of sūtra commentaries have only a single surviving work: ṭrya and Bhadanta Vimuktisena, Dharmakīrtiśrī, Dharmamitra, Kumāraśrībhadra, Jaggatatālar gnas pa, Praśāstrasena, Śrīmahājana, Jñānadatta, Guṇamati, Śīlabhadra, Nyi ma grub, mDzes bkod, rGyan bzang po, and Yuan ts'e (Wen tshegs). Little is known about them. The authenticity of the attribution to later figures from the eighth and ninth centuries C.E. such as Kamalaśīla, Haribhadra, and Vimalamitra, who could have been alive when their works were translated into Tibetan, is more likely.

The LKM clearly identifies eight commentaries as translations from Chinese (332). Of these eight texts, only three survive in the bsTan 'gyur: the Saṃdhigambhīranirmocanasūtraṭīkā (= Lalou 565 according to Steinkellner [234]), Saddharmapuṇḍarīkavṛtti (= Lalou 567), and Laṅkāvatāravṛtti (= Lalou 568). Oddly, neither of the authors of the first two commentaries is Chinese: the first is Korean, Yuan ts'e (613-696 C.E.), according to Inaba (105), and the other, Pṛthivībandhu, Sinhalese, according to the colophon.11 Steinkellner observes that these two treatises display the analytical system used[page 120] by Tibetans of all epochs to structure their texts, the "divisions" or "sections" (sa bcad), a technique he has not been able to find in treatises of Indian origin; he concludes they are of Chinese origin (235).

According to the sDe dge catalogue, important translators of the sūtra commentaries include dPal brtsegs rakṣita and Ye shes sde (ca. 812) from the early spread of Buddhism in Tibet. Important translators of the Prajñāpāramitā sūtra commentaries include rNgogs lo tsā ba bLo ldan shes rab (1059-1109) and Rin chen bzang po (958-1055) from the later spread. More than forty paṇḍitas and translators translated sūtra commentaries.