In Nepal this spring, I spent a pleasant hour on a guesthouse rooftop in conversation with Dan Hirshberg, a Harvard PhD candidate. Dan has been studying the earliest biographies of Nyang ral, the great Nyingma master of the 12th century. Based on his reading of these biographies, Dan is developing the theory that the very earliest Buddhist termas (gter ma), the revealed ‘treasure texts’ for which the Nyingma school is so famous, could sometimes be based on the simple concrete recovery of actual old manuscripts. By contrast, Dan remarked, subsequent terma discovery depends more heavily on the more complex mystical processes of ‘remembering’ a teaching from a past life with which all students of terma are familiar. Unfortunately there was not time for me to hear at length how Dan himself sought to support his hypothesis, but I was able to briefly tell him that we had already found some possibly supporting evidence.
Last year, Cathy Cantwell established that a substantial component text from one of the most important early terma collections, Nyang ral’s famous Deshek Dupa (bde gshegs ‘dus pa), corresponds exactly to the Dunhuang manuscript IOL TibJ 331.III. While the Dunhuang version is anonymous, Nyang ral claimed the original author of his terma was Vimalamitra, a famous Indian scholar traditionally associated with the Dzogchen teachings and held to have come to Tibet in the late 8th century, at the time of Padmasambhava and Emperor Trisongdetsen.
IOL TibJ 331.III is a text that we studied in detail some years ago, in one of our earlier projects on Dunhuang manuscripts. Our earlier study established that this very same text also appears, albeit in slightly different order, as chapters 8-11 of a historically transmitted Nyingma tantra called the ‘Perfection of Activities Tantra’ (‘phrin las phun sum tshogs pa’i rgyud). In less exact replications, it also occurs in some other Nyingma tantras as well (ibid. 76-87). Clearly, this is a text with many ‘incarnations’.
Leonard van der Kuijp has established Nyang ral’s dates as 1124-1192. We know the Dunhuang texts are a lot older, and current palaeographic thinking attributes the handwriting style of IOL TibJ 331.III to the mid-tenth century, which is just before the start of the Sarma (gsar-ma) or New Translation period, and two hundred years before Nyang ral. We also know that Nyang ral was a very major scholar of the Nyingma tantras, although we do not yet know exactly which ones were studied by him or in which redactions.
What are we to make of Nyang ral’s terma discovery? Might he have extracted his terma from within the pages of canonical Nyingma tantras, such as chapters 8-11 of the ‘Perfection of Activities Tantra’? As far as we currently know, not very likely, since none of the versions we have so far found in any of the canonical tantras are as close to the Dunhuang text as is Nyang ral’s terma. According to current evidence, the most likely hypothesis is that Nyang ral found an old stand-alone manuscript corresponding to IOL TibJ 331.III, perhaps one that had been lost for some time, and put it back into circulation, supplying in the process a substantial commentary, and an attribution to Vimalamitra.
Cathy’s discovery that the text of Nyang ral’s terma existed at Dunhuang about 200 years before Nyang ral’s time raises many important questions about early terma, and about early Nyingma tantric literature as a whole. We are planning a substantial research project on these issues, all the more so since a detailed look at Nyang ral’s phurpa writings has long been on our agenda.
It is widely held within Tibetan Studies that whilst Bon terma were often simply conceived of as lost old manuscripts that eventually surfaced again, Nyingma terma are more typically associated with and largely dependent on additional visionary procedures, in which the text revealer ‘remembers’ the transmission of the teaching to him by Padmasambhava or some other great master in a past life. Clearly, as Dan Hirshberg suggested, and as some others have more tentatively suggested before him, a reliance on visionary procedures to produce the text need not always or on all occasions have been the only method. Whilst extant redactions of Nyang ral’s Deshek Dupa can contain such items as the terma punctuation mark (gter shad) and the Ḍākinī Symbolic Script (mkha’ gro brda yig), which nowadays are suggestive of the mystic processes of remembering, we must also conclude that at least one substantial stand-alone component part of the huge Deshek Dupa cycle looks like it was primarily based on a found manuscript.
In a subsequent posting, I will mention some other correspondences between Dunhuang manuscripts and Nyang ral’s writings, concerning early Nyingma traditions of Padmasambhava.
 Dunhuang is a Buddhist site in North Western China where a substantial library of texts was recovered intact in the early 20th century, undisturbed since their interment in the early 11th century. It contains a many manuscripts from an earlier date still.
 See pages 68-135 of our Early Tibetan Documents on Phur pa from Dunhuang.Vienna, The Austrian Academy of Sciences Press, 2008.