Tibetan translations of Buddhist scriptures, mostly Mahāyāna texts, began to be made in the seventh century C.E.; this is the beginning of the snga dar, the period of the first diffusion of Buddhism[page 73] in Tibet. Initially the production of these translations seems to have been a haphazard and irregular business, but significantly the central political authority soon moved to take control of the process. At the beginning of the ninth century, on the instructions of the Tibetan king, a group of Indian and Tibetan scholars devised a new set of guidelines and a new terminology for translating Buddhist texts, intended to be binding on all future translators. Some of the results of this remarkable attempt at literary standardization survive in the bilingual (later multilingual) glossary known generally as the Mahāvyutpatti,4 and in its accompanying volume, the sGra sbyor bam po gnyis pa (see Ishikawa). At the same time that new texts were being translated, previous translations were collected and revised by the committee, so that their wording could be brought into line with the new terminology. Lists of works so revised were made, one of which, the catalogue known as the lDan (or lHan) kar ma, has survived.5 The lDan kar ma provides no evidence that there was any move at this time towards setting limits to a Tibetan canon as such, presumably because no Mahāyāna or Vajrayāna canon existed in India. What it does show, however, is that even at this early stage Tibetans were beginning to classify Buddhist literature according to certain principles; and as we shall see, it is this attempt to order the scriptures, rather than to circumscribe them, which is most constitutive of Tibetan canon formation. Thus the lDan kar ma starts with sūtras, those of the Mahāyāna being followed by those of the "Hīnayāna." The Mahāyāna sūtras, which are much more numerous, begin with the Prajñāpāramitā texts, then the works making up the Avataṃsaka Sūtra, the Ratnakūṭa texts, various individual Mahāyāna sūtras, the Mahāsūtras, and lastly texts translated from Chinese. The sūtras are followed by a small number of treatises, then by tantras (gsang sngags kyi rgyud) and dhāraṇīs (gzungs), hymns of praise (stotra, bstod pa), prayers (praṇidhāna, smon lam) and auspicious verses (maṅgalagāthā, bkra shis tshigs su bcad pa). Next comes the Vinaya-piṭaka,6 followed by sūtra commentaries and treatises of various kinds, finishing up with works on logic and revisions and translations in progress. Anticipating subsequent developments, then, we could say that the lDan kar ma foreshadows the basic bKa' 'gyur/ bsTan 'gyur division of later times—that bka' (the sacred word) comes before bstan bcos (the treatises) is after all only logical—and that its "bKa' 'gyur section" follows the basic order Sūtra, Tantra, Vinaya.7 Within each category works are arranged according to length, with the[page 74] longer first. Over 700 titles are listed, testifying to the extraordinary level of activity at this time.

This efflorescence of scholarship, the precision and thoroughness of which has rendered the Tibetan translations so valuable to modern Buddhist scholarship, was eclipsed for some time by the political disturbances following the death of King Glang dar ma in 842 and the subsequent collapse of the Tibetan empire, but resumed eventually in the late tenth century with the translation work of Rin chen bzang po (958-1056) and others. Thus began the so-called second diffusion of Buddhism (phyi dar), which continued for many centuries, during which translations continued to be made, especially of tantric scriptures, which were still being produced in India. At the same time older versions from the snga dar period went on being copied and circulated throughout the greater Tibetan cultural sphere.