The Bon Ka ba nag po and the Rnying ma phur pa tradition

by Cathy Cantwell and Rob Mayer

Samten Karmay and others have described the Bon gter ma called the 'Black Pillar', or Ka ba nag po, as the earliest Bon phur pa tantra, according to Bon sources reputedly dating from the 11th century. As we will see, Buddhist sources place it in the 12th century. We have been interested in what we might learn by comparing the Ka ba nag po with Buddhist phur pa texts from broadly the same period.

Despite some claims to the contrary, the cultural category known as the phur pa or kila indisputably has an Indic background. A vast quantity of highly varied kila related materials survive in Indian literatures of all denominations and many genres, spanning many centuries (cf. Mayer: 1991). Perhaps the most direct antecedents of what became the Tibetan phur pa tradition can be found in Puranas (for example, the Garudapurana), Saiva tantras (for example, the Vmasikhatantra), and, above all, in Buddhist Vajrayana traditions such as the Guhyasamaja, which has quite a lot to say about Vajrakila as a personified three-bladed kila with the attributes of a wrathful male deity who removes obstacles (Cantwell & Mayer 2008: 15-31).

Yet there is a very great deal about the phur pa tradition as it developed in Tibet that seems indigenously Tibetan, rather than Indian. Above all, the tradition grew hugely in Tibet, achieving a complexity and magnitude undreamed of in India. Numerous and often voluminous tantric scriptures began to appear in post-imperial Tibet dedicated to the phur pa deity as the central figure in the mandala; yet so far we have discovered no evidence for such tantras in India, where a typical kila text was, as far as we can tell from the surviving literature, generally not much more than a subsidiary or peripheral component of some other deity's mandala.

The early Rnying ma phur pa tantras we have read so far were clearly compiled in Tibet, largely by the recombination of smaller text fragments. Their constituent fragments were partly translated from Sanskrit, including some passages found also in Guhyasamaja literatures such as the Pancakrama, while others were locally composed, but largely patterned

on the Indian model. The overall unifying conception of these early Rnying ma phur pa tantras was unambiguously that of Indian Vajrayana Buddhism, largely conceived in terms of a Mahayoga genre that included such scriptures famous in Tibet as the Rgyud gsang ba snying po and Thabs kyi zhags pa. Nevertheless the early Rnying ma phur pa literature was augmented by a modest number of indigenous Tibetan categories, so that we can clearly tell from these and other indicators that the texts were compiled in Tibet rather than

translated in toto from Sanskrit. The tradition also acquired a new take on its Indic name in Tibet: while the proper Sanskrit term is Vajrakila (rdo rje phur pa, rdo rje phur bu), from the tenth century until today Tibetans have consistently referred to it when using transliterated Sanskrit rather than Tibetan, as Vajrakilaya (still with the same

Tibetan equivalents, rdo rje phur pa, rdo rje phur bu). Even that arch Indophile and Sanskritist, the famous Sa skya Pandita Kun dga' rgyal mtshan (1182-1251), used the form Vajrakilaya rather than Vajrakila in his famous edition of the short Phur pa tantra that was included in the Kanjur. Hence when referring to the greatly expanded Tibetan branch of the tradition as opposed to the smaller Indian tradition, one should advisedly employ the Tibetan usage Vajrakilaya, rather than the Indian usage Vajrakila.

One of our interests is to try to account for the huge expansion of the Vajrakilaya tradition in Tibet. Dunhuang texts and other early literature suggest that Vajrakilaya's popularity accelerated in the period known by Tibetan historiography as the sil bu'i dus or the 'time of fragments', the politically turbulent period that extended from the mid ninth century Imperial collapse until the late tenth century. Despite the political disintegration, the time of fragments witnessed a remarkable flourishing of religion, and was

the period in which a considerable quantity of the Tantric literature nowadays known as Rnying ma pa first came into view. When looked at from the perspective of Buddhist propagation rather than political developments, this period has therefore recently been nick-named the 'bar dar', or the 'intermediate period of propagation', since it witnessed its own unique and distinctive Tantric dispensation, yet it lies in between the Imperially sponsored snga dar or early period of propagation proper, and the later phyi dar, when the Gsar ma tantras were introduced.

The Dunhuang manuscript Pelliot tibetain 44 tells us that the Buddhist authors of this intermediate period text closely associated the Vajrakilaya tradition with Padmasambhava, attributing to him not only the redaction and ordering of the Vajrakilaya tantras, but also the transmission of its practice lineages in Tibet, and the appointment of its protective deities at Yang le shod in Nepal. Hence the rise in popularity of Vajrakilaya seems connected with the rise in popularity of Padmasambhava (Cantwell & Mayer 2008: 41-68).

As we have proposed elsewhere (Cantwell & Mayer 2008: 15-31), Vajrakilaya proved uniquely suited to addressing the ritual needs of post-Imperial Tibet for a number of reasons. It is the only major Buddhist tantric cycle for which the central ritual comprises an elaborate and detailed re-enactment of a blood sacrifice, substituting dough effigies and Buddhist doctrinal categories in place of living victims. The shape of the kila itself refers meticulously to the Indian yupa or sacrificial stake, and it is likewise identified

with the cosmic mountain or Meru. We know from recent archaeological evidence as well as textual sources such as the Tang Annals that blood sacrifice was widespread among the populations of the Tibetan plateau, as also were powerful male mountain deities (Karmay 1998: 432-450; Wangdu & Diemberger 2000: 98), perhaps, one might speculate, permitting Vajrakilaya to serve doubly as a suitable Buddhist substitute. It was also deemed uniquely protective against enmity, encouraging of pure oaths and loyalties, and defender of territorial perimeters, all valuable qualities in times of civil strife.

Several interesting patterns emerged from the comparison of the Ka ba nag po with its roughly equivalent Rnying ma counterparts. Above all, it is clear that both represent a reasonably similar combination of elements drawn from Tibetan and foreign cultures, yet they achieve this combination in very different ways and proportions. The upshot is that although the Bon and Rnying ma phur pa traditions are very much the same in overall purport and structure, nevertheless they are perceptibly different in the majority of small

particulars. Above all, never at any stage do they share identical passages of text, nor do they exactly resemble each other in such crucial ritual minutiae as names of deities, mantras, mudras, or mandala arrangements. This differentiation has been sustained with sufficient rigour to make it well nigh impossible for any educated readers of the Ka ba nag po to be confused even for the duration of a single page about the religious affiliations of the text they are reading.

An investigation into the different ways in which the Rnying ma and Bon phur pa texts combine and represent their various foreign and Tibetan cultural categories is very instructive, and has the potential to raise many interesting questions. Most obviously, the Rnying ma texts celebrate Indianness while the Bon do not. Less obviously, Bon pos seem in several ways historically to have been more amenable to eclecticism than the Rnying ma pa. Could it then be that one strand within the complex entity that is Bon

represents the heritage of various elements within the Imperial and post-Imperial Tibetan population, who once liked to combine Buddhism with their ancestral religion? And could it be that Buddhist rhetorical rejections of any such eclecticism (other than on the completely unequal basis of subjugation) provided an additional impetus to meld such diverse elements into a more coherent non-Buddhist whole? Our comparative study so far of early Bon and Rnying ma phur pa texts certainly does not contradict such a scenario, and could be seen to support it.

However, in this paper, we will be taking a slightly different perspective. Phur pa is one of the few yi dams, and perhaps the most important one, that is fully shared by both Buddhism and G.yung drung Bon under the same name and sharing broadly similar iconography and ritual function. Nevertheless, in their current forms, the Bon and Buddhist phur pa traditions contrive to be sharply differentiated, with remarkably little demonstrable intertextuality and each bearing distinctive particular features that serve clearly to differentiate them from one another. In this paper, we will explore some of the still puzzling complexities of the origins of Bon phur pa, of how it was in some respects quite distinct from Buddhist phur pa, but in other respects dependent upon it.

We will discuss three interconnected strands of evidence: Firstly, the external historical circumstances; secondly the contents of the early Buddhist phur pa texts in general, and the particular case of the Phur bu Myang 'das; and thirdly, the contents of the early Bon phur pa text, the Ka ba nag po.

Let us begin with the external historical evidence. Samten Karmay suggested as long ago as 1975, that the Bon phur pa tradition was most probably based on a pre­existing older Buddhist phur pa tradition, and first introduced into the Bon tradition with Khu tsha zla 'od and his discovery of the Ka ba nag po with its associated explanatory texts (Karmay 1975: 198-200). Karmay arrived at the deduction of Buddhist origins in part by considering the nature of the central Vajrakilaya rite of sgrol ba, known as moksa in Sanskrit, in

which the effigy of a victim is killed or liberated in the simulacrum of a sacrificial rite, and their consciousness transferred forcibly to a higher realm. Karmay felt the structure and nature of this rite was typically Indian and unlikely to have originated with Bon. He seems increasingly likely to be correct, because subsequent research has shown that the complex and very specific procedures of this rite do indeed follow a typically Indian set of ideas. It is almost certainly based on an Indian Buddhist creative adaptation of Vedic, Saiva and Sakta sacrificial cults: for just one among numerous examples, we can cite the Saiva Vmasikhatantra, where a victim is slain in effigy by stabbing with a triangular kila, exactly as in the Tibetan rite (Cantwell & Mayer 2008: 17-20).

As Karmay further pointed out, Khu tsha zla 'od is believed to have been one of the comparatively few revealers who produced treasures for both the Bon and Rnying ma religions. He is also described as Ku sa sman pa, meaning the famous doctor from Ku sa. We are not sure what the earliest historical sources are for the identification of Khu tsha zla 'od as a revealer for both religions, or for his being a great doctor, but these important elements seem to be shared by both Bon and Buddhist sources alike.

His fame as a doctor is already mentioned in the biography of the great Rnying ma pa gter ston Gu ru Chos dbang (1212-1270) in a comment attributed to Chos dbang's father (Dudjom 1991: 765). It is noteworthy that Gu ru Chos dbang himself claimed to have discovered three cycles of gter mas for the Bon tradition (g.yu rung bon gyi skor gsum) as well as for Buddhism (Gyatso 1994: 282, 286 note 35; Gu ru Chos dbang Vol II, p.134), and moreover lived in Lho brag, where by some accounts Khu tsha was born, and also not very far from Khu tsha's region of gter ma discovery in Spa gro. Khu tsha may have lived not very much earlier than Chos dbang, so it is not impossible that Gu ru Chos dbang's grandparents might have had direct recollections of Khu tsha.

A biography of Khu tsha zla 'od is included in the Buddhist Rin chen gter mdzod, where Kong sprul tells us that Ku sa sman pa was a rebirth of Vairocana, born in Lho brag, and was a contemporary of G.yu thog pa (1126-1202), living until the first part of Guru Chos dbang's life, which would seem to put him squarely into the twelfth century, possibly living until the early thirteenth century (Kong sprul 418, 419). Outwardly a doctor and lay mantrin, he achieved accomplishment in lonely places, and became famed as Kun spangs

zla 'od and Sman pa padma skyabs (Kong sprul 418, 419). He was a great hidden yogin learned in numerous scriptural transmissions. He opened the treasure face in Spa gro cal gyi brag resembling a fighting scorpion, and withdrew four treasure caskets shaped like bells and so forth. From the bell shaped one he produced Buddhist dharma, from a gshang shaped one Bon texts, from a lancet shaped casket, medical texts, and from a divination-table shaped casket, calculation cycles (gter sgrom dril bu'i dbyibs sogs bzhi byung ba'i dril bu las chos/ gshang las bon/ gtsag cha'i sgrom bu las sman/ gab rtse las rtsis kyi skor rgya cher thon pa grags. Kong sprul 419). Although famous for these extensive

revelations, from the Buddhist side nowadays only the texts and oral transmission for his Dbang lag gi bcud len remain extant. Under the Bon rubric however, according to Kong sprul, the revelation of Khu tsha zla 'od's Black Phur pa flourished, and the transmissions of its empowerment, oral transmission, and sadhana are still widely disseminated and remain famous for their immense blessings (bon sde'i gras khu tsha'i phur nag 'gar ston la bka' babs pa da lta'i bar dbang lung bsnyen sgrub kyi rgyun sogs dar zhing byin rlabs

che bar grags/ Kong sprul 419). Kong sprul goes on to claim that Khu tsha zla 'od's treasure discovery was only the first portion of a larger three phased deposit, which secondarily yielded the famous Dkon mchog spyi'i 'dus of 'Ja' tshon snying po (1585-1656) and finally the Rtsa gsum gtso bsdus spyi 'dus snying thig of Mkhyen brtse dbang po (1820-1892) himself. Kong sprul also mentions a controversy as to whether Khu tsha zla 'od was the same person as G.yu thog yon tan mgon po, but we need not go into that here.

Kong sprul was not the only ris med pa to revere Khu tsha zla 'od: Mkhyen brtse dbang po claimed to be his reincarnation, so that depictions of Khu tsha zla 'od appear on lineage drawings and paintings of the Mkhyen brtse tradition. Thus Khu tsha zla 'od features in the Rnying ma liturgical lineage supplications of the latter. The story

concerning the mystical connection with 'Ja' tshon snying po looks rather like a creation of later times, perhaps a ris med gesture of support to the eighteenth century movement of New Bon, and what the ris med pas probably saw as the constructive sentiments expressed in Kun sgrol grags pa's (b. 1700) identification of himself as an emanation of the great seventeenth century Rnying ma gter ston.

It appears that no full length Bon biography of Khu tsha zla 'od survives, but short passages on his life exist elsewhere in the Bon literature. We have not yet had the opportunity to consult any of these fragmentary Bon sources on Khu tsha zla 'od. Dan Martin has read some, and he reports that Bon sources date his birth to 1024, which is a

good deal earlier than the Buddhist dates. However, Matthew Kapstein has recently pointed out that Khu tsha adopts in his Rdzogs chen commentary Mkhas pa mi bzhi'i 'grel pa the system of logic not current in Tibet before its introduction by Rngog Blo ldan shes rab (1059-1109). This consideration would seem to make the dates from the Bon sources too early, while making the later Buddhist dates seem quite plausible (Kapstein 2009).

Khenpo Tenpa Yungdrung, the learned abbot of Triten Norbutse Monastery in Kathmandu, has also read a number of fragmentary sources on Khu tsha. From an interview with him, we understand that the Bon traditions agree that Khu tsha revealed treasures for the Buddhists as well as for the Bon. They likewise agree that he was a famous doctor, and revealed medical texts. They say that after revealing his phur pa treasures, Khu tsha zla 'od gave the caskets containing it to Gar ston Khro rgyal. They add that he also discovered a

famous set of ten magnificent material phur pas, kept for centuries at the Bon monastery of Stag bro sa in Kongpo. One of these phur pas was eventually transferred as part of a tribute payment to the Dge lugs monastery of Sera, where it survives to this day as the famous Sera phur pa that is shown to the public in an annual blessing ceremony. However, we understand that the G.yung drung Bon sources make no mention of any special spiritual connection between Khu tsha zla 'od's treasures and those of 'Ja' tshon snying po or Mkhyen brtse dbang po.

We are not sure what conclusions to draw from these biographical details. From the start, there nearly always seems to have been some tendencies within Bon which enjoyed closer ties with the Rnying ma pa. In the early period there were the Bsgrags pa Bon traditions as described by Anne-Marie Blondeau, and those connected to the Rma clan that Henk Blezer has written about (for a discussion of both, see Blezer, 2013). As we have seen, the famous Guru Chos dbang, one of the most seminal figures in early Rnying ma history,

was also apparently a Bon gter ston and lineage holder (Gyatso 1994: 280-282, 286 notes 34 and 35). Later on, in the fourteenth century, there was Sprul sku Blo ldan (b. 1360), and a hundred years after him, Bstan gnyis gling pa G.yung drung Tshe dbang rgyal po (1480-1535) (Achard 2008: xxi). The famous early modern figure Shar rdza bkra shis rgyal mtshan likewise had strong connections with the Rnying ma pa (ibid). So the New Bon luminaries Sangs rgyas gling pa and Kun grol grags pa who are so famous for their close links

to Padmasambhava and the Rnying ma pa need not be seen as unparalleled in Bon history, either ancient or modern. Nevertheless, it is not yet very clear to us how exactly the question of Khu tsha zla 'od's dual affiliation has been envisaged by either tradition. Did he simply happen to reveal medical and alchemical texts useful for both traditions but only practise Bon as his spiritual practice? Or did he, like Gu ru Chos dbang, seemingly practise both traditions separately, without mixing them? Or, despite the contrary evidence from the Ka ba nag po, did he even envisage some kind of syncretism?

If a complete understanding of Khu tsha zla 'od's innermost intentions remains elusive, the internal evidence of some early Tibetan phur pa texts still survives. Let us therefore turn to the texts themselves to look at what evidence they might bring of the early relationships between the Bon and Rnying ma phur pa traditions. In this paper, we cannot consider this question in all its wider ramifications, which will have to await a subsequent paper. Rather, we will focus on one particular source that has come to our

attention in recent years. The Phur bu mya ngan las 'das pa'i rgyud chen po or the Kilaya Nirvana Tantra (henceforth, Phur bu myang ’das), is a famous Rnying ma canonical tantra long accepted as a fundamental source for both Rnying ma and Sa skya phur pa traditions. We have good reason to believe it is old because a text by this name is anathematized as a Tibetan-fabricated apocryphon by Zhi ba 'od in his polemic of 1092. It seems to have particularly annoyed Zhi ba 'od, because he puts it at the very top of his long list of

Phur pa tantras to be shunned (Karmay 1998: 17-41). Judging from its extant versions, we can identify at least one reason why he might have anathematized it: the Phur bu myang 'das's chapter nineteen contains a substantial section which looks as though it might be non-Indic, and even Bon. Although starting out with a conventional enough brief description of the Buddhist khro bo bcu, or Ten Wrathful Deities, chapter nineteen then moves on to a somewhat longer account of the rituals and activities of the gze ma (Cantwell & Mayer 2007: 196-203). The gze ma as a type of wrathful goddess are a well known hallmark of what John Bellezza usefully terms the lamaistic Bon tradition (Bellezza

2013: 5), but are also amply attested in the Gnag rabs, a text from the Dga'-thang bum-pa collection often said to be of old, pre-lamaistic provenance (Karmay 2009). In the lamaistic Bon phur pa tradition, the term gze ma is used not merely for a group of protective goddesses, but for a central group of emanations of the principal Ten Wrathful Deities of the mandala (which are themselves a parallel yet differently named and differently appearing set from the Buddhist group of ten). Yet as far as we currently know, the term gze ma occurs not at all in the Rnying ma phur pa sadhana and commentarial tradition, nor have we so far encountered it in any other Rnying ma phur pa tantras, nor indeed in any other Rnying ma literature of any kind. Much of this chapter nineteen has now also developed into a major grammatical crux, incoherent and stubbornly resistant to resolution, as sometimes happens when the sense of the text being copied is more than usually opaque to its copyists.

While the first part of the chapter seems to refer to the gze ma as a ritual implement, perhaps some kind of tripod, or perhaps more likely as a description of the phur bu itself, the passage seems to continue by describing gze ma as wrathful messenger deities, sometimes using language that can often be associated with the wrathful Rnying ma phur pa protectresses, such as 'three-legged' and 'iron' (Cantwell & Mayer 2007: 27-28, 81, 197-202). Yet these 'three-legged' and 'iron' protectresses of Buddhist phur pa traditions such as Shwa na (whose mounts can have nine heads) themselves seem to have a non Indic background, and similar deities are found, for example, in the gnag rabs text from the Dga'-thang bum-pa collection (Bellezza, forthcoming). However the Rnying ma scholars and lamas we consulted could not make any sense of the occurrences of the term gze ma in this text, and they did not seem to have much idea at all of what gze ma might be, either as ritual implement or as deity.

One possibility is that this chapter might describe a long-forgotten ritual, written at a distant time when Buddhists and Bon alike employed the indigenous term gze ma in their various ways. Another possibility, perhaps less likely but not out of the question, is that it might represent a garbled version of an originally Bon composition, lost as it were in translation and redaction from the Bon idiom to the Rnying ma. We cannot yet arrive at any firm conclusion because no one has so far been able convincingly to resolve

the crux or translate the chapter. All we can say with certainty is that it does repeatedly employ a technical term that we have otherwise seen only in the Bon context, where it is well known. The verses in question might even at one point specify the gze ma as fierce female messengers of the herukas who destroy obstacles, which is close to the Bon sense of gze ma, yet as far as we currently know, unattested elsewhere in Rnying ma literature under that name.

What are we to make of this occurrence of the term gze ma in the Buddhist Phur bu myang 'das? Samten Karmay long ago posited the Buddhist origins of the Phur pa tradition in Tibet, and a wealth of subsequent evidence has supported his suggestion by demonstrating older Indian antecedents. It would seem then that Vajrakila began its long trajectory of

development in Tibet as an Indian Tantric tradition; yet here in the Phur bu myang 'das, we find it has added this non-Indian terminology to its Indian inheritance, which moreover sounds quite Bon-flavoured. We cannot say at what date the gze ma terminology was introduced, although we can speculate it might have happened before Zhi ba 'od's vain attempts to ban the Phur bu myang 'das in 1092. By Kong sprul's and also Kapstein's calculation, that was probably a few decades before Khu tsha zla 'od 's revelation of the

first Bon phur pa treasures, which of course make numerous and very prominent references to the gze ma in many of its tantric cycles, including phur pa. Nevertheless it is highly likely that gze ma as a category were already prominent in other Bon cycles by 1092, and perhaps also by the time the Phur bu myang 'das was redacted. We can also say that this Bon-flavoured terminology was ignored by later Rnying ma and Sa skya traditions, both of which consistently cite the Phur bu myang 'das in their most famous commentaries, yet which find no role for gze ma anywhere in those works, nor as far as we can see in any other works.

Literary and historical analysis alike suggest that the early phur pa tantras of both traditions are equally the products of the same flourishing post-Imperial culture of indigenous Tibetan tantra writing, which has left as its legacy much of what is now contained in the Rnying ma'i rgyud 'bum, as well as many other works. The real question is largely one of ascertaining the degree to which in the eleventh and twelfth centuries Rnying ma and Bon perceived themselves as scrupulously segregated separate entities or even rivals, and the degree to which they perceived themselves as differentiated but interacting aspects of a single literary culture.

Until a great deal more work is done, we cannot know how many apparently Bon-flavoured categories might survive elsewhere within the large corpus of Rnying ma tantras, nor how many existed in the past but were edited out later. However, definite signs remain in several of the extant versions of these texts that the Rnying ma pas did not necessarily make huge efforts to domesticate their root tantras through subsequent editing, so that the surviving texts can sometimes be a little rough-hewn in various ways, very much like

their Indian counterparts. There might well have been some consciousness among the early Rnying ma pa that a truly authentic Indian-style root tantra was often somewhat rough around the edges, even to the point of including apparently non-Buddhist passages, so long as they are clearly subjugated to a Buddhist purpose. For example, several of the famous Buddhist Yoganiruttara tantras that developed in India at around the same time contained blatantly Saivite passages (Sanderson 2001). So it is not clear how many such passages would have been deemed extraneous and in need of being edited out, and the survival of one apparently non­Buddhist passage in such a prominent and closely-scrutinised text as the Phur bu myang 'das is in itself of interest.

The tiny number of Rnying ma tantras looked at so far have occasionally turned out to preserve passages of a clearly Tibetan provenance, sometimes with little or no attempt to conceal them. An editor's note to one early Rnying ma tantra in the prestigious Sde dge canonical edition goes so far as to explicitly advise against making any change to a

mantroddhara that offers unmistakeable evidence of Tibetan origins, despite the fact that such mantroddharas had long been seized upon with some hostility by critics of the Rnying ma pa (Mayer 1996: 135-147).

However, most of such Tibetan originated passages comprise categories that are nevertheless unambiguously Buddhist, whether by origin or by adaptation, such as the mantroddharas, or descriptions of deities tamed by Padmasambhava in Tibet. There are also Tibetan deities accepted by both Buddhists and Bon, such as the 'go ba'i lha lnga; local plants with ritual use, such as mtshe; Bon-originated rites, such as the separation of the good spirits from the victim before the rite of sgrol ba; or material terms such

as (gnam lcags) ur mo which, although laid claim to by later Zhang-zhung lexicons, merely designates meteoric iron. Chapter nineteen of the Phur bu myang 'das is the most likely candidate that we have found so far within the Rnying ma phur pa tantras that might look specifically or exclusively Bon-flavoured, something not elsewhere accepted by Buddhism, as opposed to merely Tibetan or Bon in origin but widely accepted by Buddhists.

By contrast, our third source of evidence, the Bon Ka ba nag po, at least in the form that we now have it, is immaculately conceived and executed, and there is nothing rough-hewn about it at all. It is conceptually complex, doctrinally sophisticated, ritually and iconographically detailed, and in all respects finely structured and arranged. Unfortunately, we have no way of knowing if its original form in the eleventh century was similar to the beautiful well-polished one we have today, and this lack of text-critical certainty lends an unwelcome but unavoidable speculative edge to some of what we are going to say next.

As we have already mentioned, historical evidence suggests that the Buddhist kila or phur pa tradition precedes the Bon phur pa tradition. We have also mentioned that the Ka ba nag po is considered probably the earliest of the Bon phur pa texts. The implication is that it could have had comparatively little textual precedent in Bon, even if many of the ritual components from which it was compiled may already have been established. If the textual tradition was largely unprecedented, and if the probably twelfth century redaction

of the Ka ba nag po resembled the extant versions to any great degree, then one must conclude that the Bon phur pa tradition was born, like the mythical garuda, fully formed and completely mature from the very start. In other words, if the original form of the Ka ba nag po resembles the form we have today, and if it is indeed amongst the earliest of Bon phur pa texts, then this leaves us with little evidence for a slow build up or development of the Bon phur pa tradition from simpler beginnings to a more complex maturity, at least in its textual form.

Yet the text we currently have is not in any way whatsoever merely a copy of a Rnying ma phur pa tantra. We do not find the Buddhist names simply changed into Bon names, in an act of plagiarism that could have been achieved by anyone. On the contrary, its constituent parts and a great many of its deities are drawn very considerably from indigenous Tibetan religion, a great many of which are attested in the dGa' thang Bum pa texts. It thus constitutes a truly comprehensive, painstaking and complex reworking and re-

imagining of the entire gamut of Rnying ma phur pa doctrines to produce an entirely new phur pa tradition found nowhere else, but one that is now composed to considerable degree of adapted elements of the indigenous religion. It is so thoroughly reworked that, as we have argued elsewhere (Cantwell and Mayer 2013: 95-98), it reveals no explicit trace whatsoever of lemmata, pericopes or text fragments taken directly from any known Buddhist work: everything has been thoroughly reworked, re-imagined, and re-expressed anew. We

are reminded of the words of David Snellgrove in his pioneering work of 1967, when he writes that the greater part of Bon literature "would seem to have been absorbed through learning and then retold, and this is not just plagiarism" (1980[1967]: 12). In this respect we consider the Ka ba nag po something of a literary tour de force, combining a deep and subtle doctrinal understanding with a brilliant and original visionary imagination. As far as we can judge from its literary coherence, although it no doubt embodies numerous already existent tantric categories, it could not have been created to any great degree by merging together some existing fragments of text, which is the way some Buddhist phur pa tantras were compiled.

What is significant for this discussion is that it seems to bear witness to a literary imagination working from the inside, an authorship with an exceptionally complete understanding of the ritual, doctrinal, and contemplative principles of the existing Tibetan phur pa tradition, which then laboriously recreates the entire system anew, using numerous indigenous building blocks. The high level of insider knowledge is revealed at every turn: the complex and subtle imagery and arrangement of phur pa mandalas and

deities, the rites of sgrol ba or forcible liberation, the subsidiary rites, and so on. Only extremely learned insiders could have produced it, a person or persons with a deep and comprehensive understanding of the already existent Tibetan phur pa tradition, who are here creating an entirely novel expression of it, but necessarily from the inside out. Such a re-creation is no doubt on a literary continuum with the revelations of the many Rnying ma treasure revealers who likewise re-imagined new gter ma versions of phur pa;

but here the re-imagining is considerably more comprehensive and pervades every minute aspect of the text, rather than just a few details. Once again then this might suggest that the earliest Bon traditions had enough close points of contact with Rnying ma for such a deep level of cultural transfer to have actually taken place, when someone so very deeply versed in the Tibetan phur pa tradition, which was initially Buddhist and Rnying ma, produced for the first time a phur pa tantra in the

idiom of Bon. Unless, of course, we are to find after subjecting it to more thorough text-critical analysis, that the received Ka ba nag po text has evolved considerably over time, and that its original nature was something much less informed by an interior view of Buddhism than our hypothesis currently assumes! Until the early Bon texts such as the Ka ba nag po receive the level of textual scholarship they so obviously require, much of what has so far been said about them will remain fraught with uncertainty.

References

Achard, J-L. (2008). Enlightened Rainbows: The Life and Works of Shardza Tashi Gyeltsen. Leiden: Brill. Bellezza, J. V. (2013). Death and Beyond in Ancient Tibet. Vienna: Osterreichische Akademie der Wissenschaften. Bellezza, J. V. (forthcoming). Straddling the Millennial Divide. A case study of persistence and change in the Tibetan ritual tradition based on the Gnag rabs of Gathang Bumpa and Eternal Bon documents, circa 900-1100 CE. Blezer, H. (2000). The ‘Bon' dBal mo Nyer bdun(/brgyad) and the Buddhist dBang phyug ma Nyer brgyad, a Brief Comparison. In Samten G. Karmay & Nagano Yasuhiko (Eds.), New Horizons in Bon Studies (pp. 117-178). Osaka: National Museum of Ethnology. Blezer, H. (2013). The Paradox of Bon Identity Discourse: Some Thoughts on the rMa Clan and on The Manner of bsGrags pa bon, ‘Eternal' Bon, New Treasures, and New Bon. Framing Identity Discourse in Buddhist Environments, And the Special Case of Nativist Strategies (36pp.). Leiden: Brill. Bushell, S.W. (1880). The Early History of Tibet. From Chinese Sources. Journal of the Royal Asiatic Society. 12, 435-541. Editions of the 'Bum nag: 1. Bdud 'joms bka' ma: Rnin ma Bka' ma rgyas pa Volume Tha, pp.215-557. 2. Phur pa 'bum nag and Phur pa'i 'grel chen bdud rtsi dri med (1976) (pp. 1-229). Gangtok: Gonpo Tseten. Cantwell, C. and R. Mayer (2007). The Kilaya Nirvana Tantra and the Vajra Wrath Tantra: two texts from the Ancient Tantra Collection. Vienna: Osterreichische Akademie der Wissenschaften. Cantwell, C. and R. Mayer (2008). Early Tibetan Documents on Phur pa from Dunhuang. Vienna: Osterreichische Akademie der Wissenschaften. Cantwell, C. and R. Mayer (2013). Neither the Same nor Different. In Brian Dotson, K. Iwao, & T. Takeuchi (Eds.), Scribes, Texts, and Rituals in Early Tibet and Dunhuang (pp. 87-100). Wiesbaden: Reichert Verlag. Chos dbang, Gu ru. (1979). Gu ru chos dbang gi rang rnam dang zhal gdams Vol.II (Rin chen gter mdzod chen po'i rgyab chos, Vol. 9). Paro: Ugyen Tempai Gyaltsen. Dudjom Rinpoche, Jikdrel Yeshe Dorje (1991). The Nyingma School of Tibetan Buddhism: Its Fundamentals and History. Gyurme Dorje & Matthew Kapstein (Trans. and Eds.). Boston: Wisdom Publications. Gyatso, Janet (1994). Guru Chos-dbang's Gter 'byung chen mo: An early survey of the treasure tradition and its strategies in discussing Bon treasure. In Per Kvaerne (Ed.), Tibetan Studies (PIATS 6, Fagernes) (pp. 275-287). Oslo: Institute for Comparative Research in Human Culture. Heller, Amy (2013). Observations on Painted Coffin Panels from Tibetan Tombs. In Christoph Cuppers, Robert Mayer & Michael Walter (Eds.), Tibet After Empire: Culture, Society and Religion between 850-1000 (pp. 117-168). Lumbini: LIRI. Kapstein, Matthew T. (2009). The Commentaries of the Four Clever Men: A Doctrinal and Philosophical Corpus in the Bon po rDzogs chen Tradition. In Samten G. Karmay & Donatella Rossi (Eds.), Bon The Everlasting Religion of Tibet. Tibetan Studies in Honour of Professor David L. Snellgrove. Eds. Spec. issue of East and West 59 (1-4), 107-130. Karmay, S.G. (1975). A General Introduction to the History and Doctrines of Bon. The Memoirs of the Research Department of the Toyo Bunko, 33, 171-218. Karmay, S.G. (1998). The Arrow and the Spindle. Kathmandu: Mandala Book Point. Karmay, S.G. (2009). A New Discovery of Ancient Bon Manuscripts from a Buddhist stu­pa in Southern Tibet. In Samten G. Karmay & Donatella Rossi (Eds.), Bon The Everlasting Religion of Tibet. Tibetan Studies in Honour of Professor David L. Snellgrove. Spec. issue of East and West 59(1-4), 55-84. Kun spangs zla 'od (2006). Slob dpon pad 'byung gi sman yig gces btus. Slob dpon rin po ches mnga' bdag la gdams pa'i bdud rtsi dbang lag gi bcud len (pp. 239-242). Beijing: Mi rigs dpe skrun khang. Kong sprul, 'Jam mgon Kong sprul blo gros mtha' yas (2007-8). Zab mo'i gter dang gter ston grub thob ji ltar byon pa'i lo rgyus mdor bsdus bkod pa rin chen baidurya'i 'phreng ba (gter brgya'i rnam thar). In Rin chen gter mdzod chen mo 2007-8 (Vol. Ka, 341-765). New Delhi: Shechen Publications. Mayer, Robert. (1991). Observations on the Tibetan Phur-pa and the Indian Kila. In Tadeus Skorupski (Ed.), The Buddhist Forum II (pp. 163-192). London: School of Oriental and African Studies. Mayer, Robert (1996). A Scripture of the Ancient Tantra Collection: The Phur-pa bcu-gnyis. Oxford: Kiscadale Publications. Pasar Tsultrim Tenzin, Changru Tritsuk Namdak Nyima & Gatsa Lodroe Rabsal, (2008). In Yasuhiko Nagano & Samten G. Karmay (Eds.), A Lexicon of Zhangzhung and Bonpo Terms. (Senri Ethnological Reports 76) Osaka: National Museum of Ethnology. Sanderson, A. (2001). History through Textual Criticism in the study of Saivism, the Pancaratra and the Buddhist Yoginitantras. In Francois Grimal (Ed.), Les Sources et le temps. Sources and Time: A Colloquium, Pondicherry, 11-13 January 1997 (pp. 1-37). . Pondicherry: Institut Frangais de Pondichery/Ecole Frangaise d'Extreme-Orient (Publications du departement d'Indologie 91). Sanderson, A. (2009). The Saiva Age — The Rise and Dominance of Saivism during the Early Medieval Period. In S. Einoo (Ed.), Genesis and Development of Tantrism. Tokyo: Institute of Oriental Culture, University of Tokyo. Snellgrove, D.L. (1980 reprint). The Nine Ways of Bon. Boulder: Prajna Press. Stein, R.A. (1971-1972). Annuaire du College de France. Tao Tong (2008). The Silk Roads Of The Northern Tibetan Plateau During The Early Middle Ages (From The Han To Tang Dynasty) As Reconstructed From Archaeological And Written Sources. PhD Dissertation, The Eberhard-Karls University of Tubingen. Wangdu, P. & H. Diemberger (2000). dBa' bzhed: The Royal Narrative Concerning the Bringing of the Buddha's Doctrine to Tibet. Vienna: Osterreichische Akademie der Wissenschaften. Xu Xinguo (Bruce Doar (Trans.) (1996). A Silver Casket from a Dulan Sacrificial Horse Trench. China Art and Archeology Digest 1(3), 37-49.