[DRAFT VERSION ONLY - DO NOT CITE WITHOUT PERMISSION]
[This paper was written for the South Asian Anthropologists' Group Annual Conference, London, September 1995.]
The title of this session ("Hegemony and Dissent") raises an implicit question, at any rate in looking at aspects of Asian societies in the pre-modern period: are these appropriate terms in which to analyse such societies? One could of course add: how far are these useful terms for analysing any societies?
I take it that the session organisers intended "Hegemony and Dissent" as an indication of a general problematic, rather than necessarily involving any particular way of theorizing that problematic - such as, in particular, Gramsci's - but the terms nevertheless do carry a substantial weight of theoretical baggage with them, and it is perhaps worth starting with a short examination of this issue.
The contemporary usage of hegemony and dissent derives, of course, in large part from the work of Gramsci. He had a particular kind of society in mind, that in which he lived, the Italy of the first half of this century. More specifically, he was thinking of a society in which the sphere of knowledge-production, if we can call it that, is overseen and limited, if not directly controlled, by the State. In such a society we can speak of hegemonic knowledge as that which is consonant and supporting of the established regime, which we take for granted, I think, is a centralized state of some kind, and dissenting knowledge as knowledge which in various ways contests the terms of that regime and which exists in a grey area, tolerated or persecuted, on the margins of state power.
The conditions which legitimated Gramsci's model became, of course, in many respects, more applicable in the decades following his death, and many modern Asian states (one might think, for example, of China, Indonesia, or Singapore) provide situations all too close to that of Gramsci's starting point, Italy during the Fascist period.  Whether the terms of the analysis apply very well to modern South Asia, particularly India, is another question, if not one I shall address directly in this paper. For that matter, the new technologies of the Internet, satellite TV, and other fast and efficient international information channels would seem to be bringing about on a global scale an information regime which the Gramscian framework in its classic form fits rather poorly. Access to such technologies is of course open to only a tiny fraction of the Asian population at present, but there is a sense in which even states such as China or Indonesia are fighting a temporarily effective but ultimately doomed struggle against full incorporation into the new global regime.
If this new situation fits rather poorly into the Gramscian model in its classic form, rival models - Bourdieu or Foucault might come to mind - do not seem obviously much better. It would seem that rather than hegemonic knowledge being the knowledge which is supported by and produced with the encouragement of a particular state (in retrospect, a formulation which in any case suggests a little too much unity on the part of the state) we are now talking about a situation in which there are multiple forms of knowledge, with varying claims to authority and varying degrees of backing from powerful institutions. This perhaps lends the case-study I discuss in this paper a somewhat wider interest than it might otherwise have, because the decentralized social and political space of Tibet, with its multiple sources of political and religious authority, none of them ever dominant for more than a short period over more than a fraction of the space at issue, has certain formal similarities to the postmodern world, where any claims at hegemonic knowledge are likely to fall victim to the corroding impact of the fax machine or the World Wide Web.
I spoke in my title of Indian Vajrayana as a dissenting tradition which was partially hegemonized in Tibet. One could equally speak of Indian monastic Buddhism as a hegemonic tradition which was partially de-hegemonized in Tibet. Both statements refer to a process which I believe occurred as Buddhism was transferred from India to Tibet from the 7th and 8th century onwards and as it developed and transformed within the very different social, cultural and intellectual environment of Tibet. This was a transition from something resembling a series of centralized states in 8th to 12th century India - in which "hegemonic knowledge" could have a certain kind of meaning - into something on the Tibetan side which, at least from the mid-9th century onwards, had very little resemblance to any kind of centralized state - so that bodies of knowledge, to the extent that they survived, did so as part of a very different context. Maybe, as I said before, this transition has some analogies to contemporary transitions, though I wouldn't push the comparison too far. At any rate, it is this transition and its consequence which is at the centre of this paper.
As for the transition itself, I have discussed aspects of this process elsewhere at length, if from a somewhat different perspective (Samuel 1993), and I shall try to be brief here. A useful starting point is provided by Reginald Ray's important recent book, Buddhist Saints in India (Ray 1995), essentially a study of kinds of Buddhist practice in Indian society up to, say, the 6th or 7th centuries. Ray, whose work is informed by the anthropological studies of contemporary Theravadin societies, in particular the writings of Stanley Tambiah (1976, 1984) and Michael Carrithers (1983), creates a plausible model of Indian Buddhism throughout much of its history as analysable into three components. Two of these are associated with groups of religious practitioners; the first located primarily in large urban monasteries, oriented towards scholarship and the ordered routine of the monastic life; the second ascetic and yogic in approach, associated with forest hermitages and pilgrimage sites and with an altogether less settled way of life. Ray's third component, lay Buddhism, has a complex relationship with these two.
While this is not a major issue for Ray, one might also add that the urban monastics are, almost by definition, more closely linked to centres of state power; dissent is much more likely to be found on the margins. At any rate, this was to become the typical relationship within Theravadin Buddhist societies in more recent centuries, where one certainly can speak of Buddhism as a state religion (e.g. Tambiah 1976, 1984). I have suggested elsewhere (Samuel 1993, passim) that the emphasis we find among the urban monastics of these societies on scholarship and the propriety of monastic life as opposed to yoga and magical power makes sense in the context of a state religion. The modern Theravadin countries offer a series of classic examples of this situation, with the magical, ecstatic miracle-working style of Buddhist practice never exactly eliminated - it has too much to offer for that - but routinely marginalized and, whenever it threatens state power, forcibly suppressed in one of the "purifications of the Sangha" that have happened periodically in South East Asian history (Tambiah 1984). This is a pattern that has by no means come to an end, as the recent history of the Thai sangha demonstrates.
The scholarly, karma-oriented Buddhism of the urban monastics in countries such as Thailand today is, in other words, arguably part of a hegemonic order, and there is a sense in which it has been so for many centuries. A key issue here is that the magical power associated with Buddhism has to be both kept under control, and reserved to the State - in other words, in the pre-modern period at least, the monarchy. Incidentally, I am deliberately using the term "magical power" here rather than "charisma" or the like, since I want to emphasize that we are talking about something seen as powerful and effective in the real world - one could refer, for example, to Neils Mulder's discussion of the significance of power in everyday life in Thailand (Mulder 1992: 15-58). Lesser states and regional authorities within what Tambiah referred to as the "galactic polity" each have their share of this magical power, associated with relics of the Buddha and of great saints of the past, and with powerful Buddha-images such as the famous Emerald Buddha in Bangkok - and private individuals may have limited access via amulets and miracle-working monks - but no private lay individual can be allowed to acquire magical power sufficient to rival centres of state power, and even ascetic forest monks who do so are in a somewhat suspect and ambiguous position. Thus in such societies, Ray's forest ascetic style becomes marginal if by no means entirely insignificant.
It would be going a little too far to call the urban monasteries and their monks simply agents of the State, but they nevertheless have reached a compromise with the State, consciously or unconsciously, and this compromise both gives them a position of high status but sets strict limits on their freedom of action. Typically, these societies have for many centuries had Supreme Patriarchs or the like appointed by the State - or subject to State approval - and the hierarchical structure is quite clear; the Sangha - the monastic order - is subject to the State. Whether the relationship was of quite the same kind and extent within Indian Buddhism is perhaps hard to say, since we are talking about a long period of time and a variety of different states. I would imagine that it would have been similar at the height of state patronage of Buddhism under the Pala empire in present-day Bengal and Bihar, for example, when the great monastic universities of Nalanda, Vikramasila, Somanatha and so on were flourishing. This is not to say that these Buddhist monastic institutions were supported exclusively by state patronage - we know that the wealthy commercial sectors of ancient Indian cities, with their large merchant classes, also played a major part in supporting the monasteries - but it seems unlikely that establishments of the scale and significance of Nalanda or Vikramasila could have existed and prospered has they not reached an accommodation with the state. Indeed, many of these great monastic universities were initially royal foundations, and they seem to have acted as major centres of secular education and learning rather than being purely monastic institutions.
At other times and places, however, rulers, wealthy merchants and landowners, might mainly patronize Saivite, Vaisnavite or Jain institutions, and Buddhism might have a less significant role to play in relationship to the state. In addition, marginal spaces and territories such as the great pilgrimage sites, some of them within the domains of the Pala empire and similar states, but other in more peripheral areas, provided a context where other styles of Buddhism could flourish. At this point we need to take a closer look at the second component of Ray's threefold model - the forest monastic tradition of ascetic renunciates - and more specifically to its representatives in the period we are mainly concerned with, the eighth to twelfth centuries C.E. These are the Tantric practitioners known as the siddhas, and it is they rather than the urban monasteries who provided a critical component of the corpus of Buddhist knowledge taken over by Tibet: Vajrayana or Tantric Buddhism.
The Vajrayana originated within - or at least was transmitted to Tibet from - a group of Indian practitioners known, in Sanskrit, as the siddhas, conventionally dated from around the 7th or 8th up to the12th centuries. Ray does not include the siddhas directly in his paradigm of the forest saints, since he is concerned with a somewhat earlier period, but it is clear enough from his analysis that they represent a later development of the classic Buddhist forest saints whom he does discuss. They share their association with the pilgrimage sites and the itinerent ascetic lifestyle. 
They also seem in many respects precursors of the sadhus of modern India, though modern Indian sadhu are of course unlikely to identify themselves as Buddhist. Buddhist siddha practitioners nevertheless continued to be a significant part of Indian society for many centuries after the collapse of state Buddhism in India - we know this in particular from the writings of the 16th century Tibetan historian Taranatha (Templeman 1997) - and seem to have formed a sub-group among the Nath yogins who are still a major component of the sadhu scene. Certainly, contemporary practitioners, whether Nath yogins found through much of North India, or the Bauls of Bengal, discussed in another paper in this session by Jeanne Openshaw, or the Cittars (i.e. Siddhas) of Tamilnadu still today practice yogic techniques very similar to those of the Vajrayana Buddhist siddhas of the 8th to 12th centuries, combined at times with an itinerant way of life focussed on the festivals at the major pilgrimage shrines.
To return to the 8th to 12th centuries, a number of features of the siddha lifestyle are worth considering here, some of them no doubt common to earlier versions of the forest practitioners, others probably not:
I begin with the complex relationship to urban, monastic Buddhism revealed in the siddha biographies; respectable academic urban monks are depicted as running off to join the siddhas; siddha practice is seen as an inner, more profound and more effective version of regular monastic practice (so the Tibetan saying, "outwardly Hinayana, inwardly Mahayana, secretly Vajrayana"). The siddha takes vows additional to the Pratimoksha vows of the Hinayana practitioner and the Bodhicitta vows of the Mahayana practitioner -though whether these supersede the earlier vows is a complex and vexed question, both in theory and in practice.
Next, we note the siddha's rejection of monastic celibacy - often linked to the significance of forms of sexual sadhana or yogic practice. Siddhas, both male and female, regularly have sexual partners - though it goes without saying that the relationship is for spiritual purposes, a point often underlined in the stories of the siddhas, who tend to be challenged on this point and to respond with a display of magical powers sufficient to convince their critics that they they are the genuine article. You may note however that this rejection of monastic celibacy means that one can be a layman or laywoman (in the sense of a married householder) and a siddha, and it also means that siddhas can have children who can inherit their family property.
Another important point is the extensive overlap between the siddhas and Saivite practitioners of the Kapalika variety (cf. Lorenzen 1972). Several major siddhas (e.g. Kanha = Krsnacarya) are important in both Buddhist and Saivite (Nath, etc) teaching and initiation lineages. It is now clear that this goes beyond the Bhairava-like nature of fierce Tantric Buddhist deities to actual sharing of deities and texts (e.g Mahakala; cf. Samuel 1993: 422) and to substantial textual borrowings from Hindu Tantric scriptures being incorporated into central Vajrayana Buddhist Tantric texts (Sanderson 1991).
This leads on to another issue, the "shamanic" aspect of Vajrayana practice in India.  The assumption of divine identity through ritual is a critical part of the Vajrayana, and the implication is that it confers divine power on the practitioner. This divine power is the basis of the siddha's magical ability, and there are indications that some siddhas at some periods at least made a living as itinerant magical practitioners, travelling sorcerors one could say - sections of major Tantric texts such as the Hevajra consist of collections of spells to be used for various purposes, to defeat an enemy army, compel a woman's love, and the like.
This brings us to the question of the specifically antinomian character of aspects of the siddha's way of life. Vajrayana siddhas adopted the way of life of the Kapalika ascetics, complete with bone ornaments, skull bowls and human bone trumpets. The central rituals of the Indian Vajrayana, festive gatherings called ganacakra, ideally took place in dangerous and marginal sites and times such as cremation-grounds at night. They involved an explicit transition into an alternative and non-normal mode of existence in which the participants and their surroundings were transformed through "pure vision" into the Tantric deities and the sacred space of the mandala.
Needless to say, such behaviour was viewed with some ambivalence by the surrounding society. The dakini-s, female deities and their human representatives, who were (and still are today) for Vajrayana Buddhists exalted beings associated with the revelation of the Tantric teachings, were for respectable Brahmanical society much more dubious figures; the modern derivatives of the term in North Indian languages generally means "witch".  The Kapalika siddhas feature in elite literary productions of theperiod alternatively as figures of fun or as evil magicians - the latter being a stereotype still operative in modern India.
Despite all this, we notice the insistence throughout the siddha material on bodhicitta - the central motivation of Mahayana Buddhism - as a defining aspect of the siddha lifestyle.  The siddhas saw themselves as rejecting conventional behaviour and morality, where appropriate for their purposes, in order to embrace a higher morality.
It is these features that suggest that we might reasonably refer to the siddhas in India as bearers of a dissenting tradition. It is worth noting that, if the siddha biographies and the Vajrayana lineage histories are correct, by no means all siddhas were wandering sadhus and other marginal people. They included kings, princes, brahmins and others of high social status. Siddha practice, in other words, could be part of a "dissenting" life-style at fairly exalted political levels. This should perhaps be seen in terms of the political instability of the period, where conventional, monastic state Buddhism was on the retreat through much of India, the new Hindu and Islamic state ideologies were still in process of formation, and the bases for conventional morality must at times have seemed particularly fragile and unconvincing. The siddha path offered Enlightenment (bodhi), but it also offered power, and it seems likely that power was at least part of the attraction.
The story of the coming of Buddhism to Tibet has been often told and only a brief summary is called for here. It seems important however to differentiate between an earlier period in which Buddhism was, among other things, being imported in its monastic form as a State ideology, and a later period in which this was not the case. Tibet from around 640 to 842 CE was an expansionary empire, absorbing first the older state of Zhang-zhung to its west, then substantial slices of regions then under the control of Chinese, Nepalese and other surrounding states. Towards the end of this period, the first Tibetan monastery was founded, at bSam-yas, under royal patronage.
A central mythic narrative of Vajrayana Buddhism in Tibet relates to this occasion. The local deities, opposed to the introduction of Buddhism, destroyed at night what was built during the day, until Santaraksita, an Indian monastic who had been imported to serve as the first abbot of the new monastery, counselled that the great Tantric siddha Padmasambhava be summoned from India to tame the local deities and bind them to the service of the Buddhist teachings. Padmasambhava's journey through the Tibetan landscape, overcoming and binding a long succession of named deities at specific places, is a central episode in the legendary accounts of his life, and laid down the terms on which Tibetan Buddhists today deal with those deities. He is for the older traditions of Tibetan Vajrayana the most important siddha - in fact, he is a kind of second Buddha - and the central figure in the lineages through which the power of the siddhas has been transmitted to the present day.
In 842, the Tibetan state collapsed, and monastic Buddhism more or less disappeared. Vajrayana or Tantric Buddhism however survived and gradually prospered in the form of mostly hereditary lineages of Tantric practitioners who over time seem to have taken over a substantial part of the ritual life of village communities throughout Tibet. What remained of older religious traditions were gradually made over in Buddhist forms. At the same time, however, Tantric Buddhism began to become something subtly different from its Indian forms. In part this was a question of "cleaning up" those elements surviving from the old antinomian, cremation-ground context - though they never disappeared entirely. More particularly for our present topic, it seems to me that by the 13th or 14th centuries we can no longer, in any meaningful sense, describe Vajrayana Buddhism in Tibet as a dissenting tradition. Padmasambhava in India would have been a representative of a dissenting form of Buddhism. His lineage successors in Tibet were practitioners of Buddhist Tantric ritual in a field in which there was no effective state or established state-sponsored tradition to dissent against.
During the following four centuries there were at most local centres of power, one or two of which made efforts to re-introduce monastic Buddhism. Monasticism gradually did become re-established towards the end of this period, but by this time it was inextricably tied up with Tantric Buddhism, in a quite different relationship to that described earlier for India. Monasticism in Tibet could not provide the hegemonic order against which Tantra might dissent, because monasteries were themselves significant as centres of Tantric magical power rather than being merely the centres of state-sponsored, karma-centred scholastic Buddhism we find in modern Theravada societies. Political power in Tibetan societies became a question of fragile and shifting alliances between local monastic centres, lay aristocratic lineages, and rich merchant families. In time, the monastery-based regime of the Dalai Lamas became dominant over somewhat less than half the Tibetan cultural area, though I have argued elsewhere that it cannot really be seen as imposing detailed control on everyday life throughout much of that territory, and other centres of local power, including rival monastic orders, continued to exist within and without the Dalai Lama's realm (cf Samuel 1993).
This account is drastically simplified. Complicating factors include the involvements of foreign powers, mostly Mongol, Chinese and Manchu, the gradual development of scholarly and purist forms of Buddhist practice that attempted to restrict the more powerful Tantric techniques to an elite of advanced practitioners, and various counter-movements from the yogic or shamanic side. It is these complications that suggested the phrase "partial hegemonization" in my title. However, rather than describing them in detail, I would like to turn to the overall nature of the field of knowledge which developed in Tibet during this period and prevailed until modern times.
Some underlying assumptions of this field, it seems to me, changed only slowly over the entire Buddhist period. These are assumptions about the nature of magical power, and I have suggested elsewhere (Samuel 1992) that they are similar to those found in other societies of the mid-Himalayas (Tamang, Gurung, Magar) and the South-East Asian highlands. Power is thought of as implicit in the landscape, particularly in certain places such as mountain-peaks and lakes, or sites sacralized through the activity of culture-heroes such as Padmasambhava. It should ideally be channelled and directed to bring about the welfare and prosperity of local communities and to protect them from harm. In this, magical power is no different from any other kind of power. Its channeling and directing involves the maintenance of harmonious relationships with the local deities or spirit-forms who inhabit the landscape, and the proper exploitation of the blessing-power (byin rlabs) inherent in sacred sites such as those associated with holy men of the past.
The stress on the unsatisfactoriness and ultimate unreality of everyday life in Mahayana Buddhist thought relativizes these concerns somewhat and lends them a certain prevailing sense of irony (Lichter and Epstein 1983), as well as providing trans-worldly goals for a minority. Concerns for the proper use of magical power for this-worldly ends nevertheless remained at the basis of the Tibetan approach to reality, constantly reiterated in domestic ritual, communal village ritul and monastic ritual on behalf of local communities, and Vajrayana Buddhism provided the primary technology for dealing with them.
Certain people have special access to power, through family connections or innate ability, and it is necessary to form relationships with them. These may be straightforward exchange relationships, as when one pays a village shaman to find the cause of an illness, or a group of monks to recite auspicious texts on the occasion of a wedding, but more interesting for our purposes are ongoing hierarchical relationships, which in Tibet usually have something of a contractual and patron-client character. (Even local deities swear vows and are bound to observe the terms they have agreed to.) If a lama or monastery fail to provide effective magical services, one can always turn elsewhere. I hvae suggested elsewhere that elements of Tibetan Buddhism even in modern times can be understood quite well in terms of centres of quasi-shamanic power competing for custom (Samuel 1993).
A significant question then becomes: on what grounds can one claim to be an effective supplier of magical power? Here I will list some of the possible bases for such claims:
being the possessor of important Tantric ritual teaching lineages, particularly where handed down by heredity (e.g. the 'Khon family of Sa-skya) or reincarnation (e.g. the 'Bri-gung hierarchs);
being the descendant or recognized reincarnation of a past lama known for magical power and ability; power is thought of as inhering in the lineage (gdung rabs);
being head of a large and disciplined monastic community (this does not necessarily mean that one is oneself an ordained monk);
having established a personal reputation for spiritual development (and thus magical power);
having access to direct revelation from Padmasambhava, the tantric deities, etc. (this is likely to become the basis of a body of personal ritual);
having access to important holy sites or relics containing power (byin rlabs).
The role of myths and quasi-historical narratives in supporting such claims is probably already clear. Such material was - and still is - propagated in all sorts of ways, ranging from ritual performance (such as the famous Black Hat ceremony by which the rGyal-ba Karma-pa lamas indicate their identity to the founder of their reincarnation series, Dus-gsum mkhyen-pa), hagiographies of lamas (rnam thar), histories of monasteries and their teachers (lo rgyus, gdan rabs, etc) and pilgrimage guides (gnas yig) as well as through all kinds of oral transmission in formal and informal contexts. In these ways lamas and monasteries in Tibet established claims as channels for magical power on behalf of th wider community. These claims historically formed the basis of their political role; secular landowners and local aristocratic lineages reached accommodations with them (including networks of marriage alliance over several generations); foreign powers from the Mongols in the 13th century to the Manchus in the 17th to 19th centuries accepted their claims and set up alliances with them. Such stability as Tibetan politics maintained until the 20th century depended on these networks of alliance between monastic and lay power and the willingness of local populations to give adherence to them. 
In terms of hegemony, this is a rather weak situation.  The alliances tended to be fragile, their power (in secular terms) was limited, and their control over the everyday lives of the Tibetan population, even in the more centralized areas, was limited to taxation and control of labour power rather than encompassing the ideological mechanisms we generally associate with hegemony. At the same time, an established idiom for claims to power did gradually develop, and it is in this sense that we can speak of the partial hegemonization of Vajrayana Buddhism in Tibet. This idiom was made up of elements such as the recognition of reincarnate status or of hereditary descent from significant historical lamas; of the ceremonies of initiation into and transmission of valued Vajrayana teachings; and the monastic ritual dances and life-empowerment rituals, in which monastic power is explicitly trasferred to the lay population. These elements, along with the oral narratives and textual material mentioned before, outlined a discursive space within which claims to magical power (and hence religious status) could validly be made. As I have suggested, this discourse incorporated and extended earlier Tibetan understandings of the nature and functions of magical power, and formed the basis for the widespread acceptance of Tibetan Buddhism by the Tibetan population. It seems that this discourse had considerable ability to reshape and absorb dissenting elements. Here I briefly give two examples:
In Civilized Shamans , I discuss the late 15th century "yogic reaction" to the increasing formalization and monasticisation of Tibetan religion (Samuel 1993: 518-22). This reaction was rapidly absorbed and incorporated within parts of the monastic tradition. Within a couple of generations, the "dissenting" and anti-structural songs of the radical yogins, with their inbuilt critique of monastic practices, were being recited as part of monastic liturgy (1993: 522).
More recently, David Germano has sketched the developmentof the non-Tantric meditation tradition of rDzogs-chen Sems-sde ("Mind Series"). In its earlier stages in Tibet (10th-11th centuries) this seems to have been the basis of a series of non-Tantric contemplative exercises, based on a textual discourse with a "constant antinomian tendency... a rhetorical lawlessness asserting a primordial dimension that is neither accessed by, nor governed by, law-abiding patterns" (Germano 1994: 240). It did not include deity visualization nor traditional initiation and transmission rituals. The gradual incorporation of rDzogs-chen into Mahayana discourse can be traced through the works of a series of important Tibetan lamas, including kLong-chen rab-'byams-pa (1308-63) and 'Jigs-med gling-pa (1730-98). By the early 19th century, rDzogs-chen had become fully incorporated into a characteristically Vajrayana series of preliminary Tantric practices ad initiations (Germano 1994: 275-6). 'Jigs-med gling-pa himself became the origin of a series of reincarnate lineages, particularly those deriving from 'Jam-dbyangs mkhyen-rtse'i dbang-po (1820-92), who dominated the monasteries of Eastern Tibet by the late 19th and early 20th century, forming an effective counter to the dGe-lugs-pa monasticism prevailing in Central Tibet and the Chinese borderlands (A-mdo, Tre-hor). It is interesting for our purposes here both that rDzogs-chen retained a heterodox edge as a higher-level "reframing" of conventional Vajrayana, and that it only became the basis of monastic traditions competitive with the established Vajrayana orders by adopting their basic ritual paraphernalia and conceptual structure.
This paper is, as is clear, no more than a sketch of some of the underlying parameters of what I have called the Tibetan field of knowledge. However, it may suggest some features of how this strongly decentralized and apparently non-hegemonic field operated. In particular, the gradual evolution of a common set of assumptions underlying the field, which determined what might be viable and potentially legitimate claims, is of considerable interest. Such a set of assumptions does not arise out of nowhere. It grows from ideas and concepts already pervading the field at a previous stage, and it develops in response to the social and political circumstances prevailing during the period of formation. It would be possible to trace this evolution in some detail (cf. Samuel 1993: 447-50 for some examples), although I have not attempted this here.
Likewise, the field is vulnerable to drastic changes in social and political circumstances. Among Tibetan Buddhist polulations today, the kinds of claims I have discussed here have not entirely lost their meaning, but they have a tense and complex relationship to new claims based on assumptions derived from a new global context: the Dalai Lama as leader of a nation-state; high lamas as figures in a quasi-ecclesiastical hierarchy ; Buddhist philosophy as an academic subject making knowledge-claims comparable to those of Western philosophy or physics, and the like; and above all, the need to defend against Chinese political claims. The field is unstable and changing, and it will be interesting to see how it develops in years to come, and what kinds of space are left within it for "dissent," whether religious or political.
 In fact, it is arguable that a state such as China in 1995 is a poor example of the Gramscian role of ideology, since nobody any longer appears to believe the ideology, and power is actually maintained by the use or threat of use of physical force. [Return]
 It may be noted that Tambiah regards those forest saints in Thailand associated with magical power as representing a more "tantric" style of practice than that characteristic of Theravadin monks (Tambiah 1984). [Return]
 I use "shamanic" in the same sense as in my books Mind, Body and Culture (Samuel 1990) and Civilized Shamans (Samuel 1993). [Return]
 One should perhaps be wary of taking this picture as uniformly true for the entire siddha period. Vajrayana practitioners had probably gained considerable legitimacy in some areas by the end of the period, as is suggested by the high status of the Vajracarya priests in present-day Newar society and of their equivalents, the pedanda bauddha, in modern Bali. It seems that in 12th-century Java, both Vajrayana and Pasupata practitioners had acquired state support and high social status (Becker 1993), though it is not clear how far these were at the antinomian end of the overall spectrum of Tantric siddhas and yogins. Sanderson (1988) discusses how the Saivite versions of this "cremation-ground culture" were in time tidied up and rendered respectable (see also Germano 1994: 232 and n.73). [Return]
 Note however such pairs as "relative bodhicitta" and "absolute bodhicitta". Bodhicitta never loses its association with compassion and altruism, but it acquires a variety of more complex connotations, including internal physiological ones (Samuel 1991). [Return]
 This raises the question of what alternatives local populations might have in relation to landowners. Briefly, these were switching allegiance and/or moving to a marginal area. A continuing labour shortage meant that these options were often realistic. [Return]
 My description marginalizes such stable and institutional aspects as the large teaching monasteries of Central Tibet which tend to dominate many descriptions of Tibetan religion and politics (e.g. Goldstein 1989). For reasons why I feel that this shift of emphasis is appropriate, see Samuel 1993: 142-6. A more extended treatment of the topic would nevertheless need to incorporate more explicitly the effects of the establishment of the dGe-lugs-pa state in Central Tibet in the mid-17th century and the development of state institutions there and elsewhere. [Return]
 As in the formal recognition of heads of the four orders of Tibetan Buddhism (five, with Bon), or the prevalence of the Catholic-derived titles of "His Holiness" and "His Eminence" for high reincarnate lamas. [Return]
Becker, Judith (1993) Gamelan Stories : Tantrism, Islam, and Aesthetics in Central Java. Arizona State University, Program for Southeast Asian Studies (Monographs for Southeast Asian Studies)
Carrithers, Michael (1983) The Forest Monks of Sri Lanka. Delhi.
Germano, David (1994) "Architecture and Absence in the Secret Tantric History of the Great Perfection (rdzogs chen)" J. International Assoc. of Buddhist Studies 17/2, 203-335.
Goldstein, Melvyn C. (1989) A History of Modern Tibet, 1913-1951: The Demise of the Lamaist State. Berkeley, University of California Press.
Lichter, David and Lawrence Epstein (1983), "Irony in Tibetan Notions of the Good Life". In Charles F. Keyes and E. Valentine Daniel (ed), Karma: An Anthropological Inquiry,, pp.223-60. University of California Press.
Lorenzen, David N. (1972) The Kapalikas and Kalamukhas: Two Lost Saivite Sects. New Delhi: Thomson Press.
Mulder, Neils (1992) Inside Thai Society: An Interpretation of Everyday Life . (3rd rev.edn.) Duang Kamol, Bangkok.
Ray, Reginald (1995) Buddhist Saints in India: A Study in Buddhist Values and Orientations. New York, Oxford University Press
Samuel, Geoffrey (1990) Mind, Body and Culture: Anthropology and the Biological Interface. London and New York: Cambridge University Press.
Samuel, Geoffrey (1989) "The Body in Buddhist and Hindu Tantra: Some Notes." Religion, 19, pp.197-210.
Samuel, Geoffrey (1993) Civilized Shamans: Buddhism in Tibetan Societies. Washington, D.C., Smithsonian Institution Press.
Samuel, Geoffrey (1994) Tibet and the Southeast Asian Highlands: Rethinking the Intellectual Context of Tibetan Studies. In P. Kvaerne (ed.) Tibetan Studies: Proceedings of the 6th Seminar of the International Association for Tibetan Studies, Fagernes 1992, pp.696-710. Oslo: Institute for Comparative Research in Human Culture
Samuel, Geoffrey (1995) "The Effectiveness of Goddesses." Typescript.
Sanderson, Alexis (1988) "Saivism and the Tantric Traditions." In Stewart Sutherland et al. (ed.) The World's Religions, pp.660-704. London, Routledge.
Sanderson, Alexis (1991) "Vajrayana: Origin and Function." To appear in Mettanando Bhikkhu et al. (ed.), Buddhism into the Year 2000. Bangkok: Dhammakaya Foundation.
Tambiah, Stanley J.(1976) World Conqueror and World Renouncer: A Study of Buddhism and Polity in Thailand against a Historical Background. Cambridge University Press.
Tambiah, Stanley J. (1984) The Buddhist Saints of the Forest and the Cult of Amulets: A Study in Charisma, Hagiography, Sectarianism and Millennial Buddhism. Cambridge University Press.
Templeman, David (1997) "Buddhaguptanatha: A Late Indian Siddha in Tibet." In Helmut Krasser, Michael Torsten Much, Ernst Steinkellner and Helmut Tauscher (eds) Tibetan Studies. Proceedings of the 7th Seminar of the International Association of Tibetan Studies, Graz 1995, Vol.II pp.955-966.Wien: Verlag der Österreichischen Akademie der Wissenschaften. (Österreichische Akademie der Wissenschaften, Philosophisch-Historisch Klasse. Denkschriften, 256. Band. Beiträge zur Kultur- und Geistesgeschichte Asiens, Nr.21.)