The Wrathful Deity Theme

Among the major venerable religious traditions in the world, Esoteric Buddhism is one of the most liable to misunderstanding. From its emergence as a separate tra­dition in sixth-century India, Esoteric Buddhism has valued oral interpretations of texts which employ intentionally elliptical language and metaphors. Initiation into enlightenment-directed rituals requires the instruction of a lineage master.

The secrecy which surrounds its core teachings, the use of a cryptic language for prayers, the centrality of mystic experience and comprehensive many-layered sym­bols contribute to a system of belief and praxis which has proved almost impene­trable to discursive scholarship. These difficulties were so discouraging that as late as 1974 the Dutch Buddhologist J.W. de Jong accurately observed, "Tantrism is still the most neglected branch of Buddhist studies."

Despite the accent on secrecy and orality, material expressions of the enlight­enment experience were produced wherever Esoteric Buddhism spread in Asia. Visual representations were sculpted and painted, molded and cast, by t.li.e hands of initiates or under their supervision. Surviving texts make clear that the produc­tion of these objects was important enough to necessitate codified guidelines gov­erning such matters as the motivation of the maker, the types of

materials used, the forms, colors, attributes and compositional relations of the deities depicted and their symbolic interpretation. The works of visual art made in accordance with rit­ ual requirements were not "merely" decorative, nor were they the passive focus of acts of ritual piety. Art functioned actively for the practitioners of Esoteric Bud­dhism both as visual triggers of integrated states of realization and as expressions of those states, which beggared verbal description.

Most of these objects were not available to the non-initiated public at the time they were made and are not analogous to "art" as conceivd by twentieth century

Euro-American cultures. As they are now considered works of art, however, many are accessible for study in museums, private collections, caves and excavation sites around the world.3 They provide invaluable and sometimes unique information about Esoteric Buddhist priorities, social relations, iconographic developments and symbolic systems. If the tools of formal analysis are sharpened, images without inscriptions may be accurately dated, providing in many cases relatively precise information about doctrines and practices. Moreover, by arranging the images chronologically and analyzing the resulting patterns, a schematic model of the entire development of Esoteric Buddhism may be generated. This is the project I undertake here.

Works of art are reliable documents which contribute to an ongoing historical reconstruction. To date most research has been textually oriented, focusing on edit­ing and studying the original Indian scriptures along with their Tibetan and Chi­nese translations. There have been corollary investigations into the understanding of these texts by later Esoteric Buddhists, through the commentaries of Tibetan and to a lesser extent Japanese Buddhists. In the last few decades field work has expanded to include the observation of contemporary ritual practices.

I believe sculpture and painting have been under-utilized by historians of Eso­teric Buddhism, though the discrepancy may also be attributed to the fact that art historians have not sufficiently related their findings to broader movements. Nev­ertheless, analysis of visual evidence in conjunction with other avenues of study provides valuable insights into the evolution of Esoteric Buddhist doctrines and practices.

The study of artistic remains plays a significant role in the understanding of ideas expressed by the Esoteric Buddhist tradition, as well as the historical situation in which they are found. Historical evidence is of particular interest because the liv­ing tradition of Esoteric Buddhism as it survives today, in the Himalayan regions and in Japan, generally emphasizes the revelatory, a-historical nature of inherited texts and teachings. Learned Tibetan scholastics such as

Buton Rinpoche (1290- 1364), Tsongkhapa (1357-1419) and his disciple Kaygrubjay (1385-1438) composed elaborate explanations of the conflicting priorities of texts and schools within the broader tradition. These schemes have been knit together so as to elide internal con­tradictions stemming from the evolutionary development of Buddhism.

Disjunctures between very different types of texts, all under the rubric of Eso­teric Buddhism, have also been studied by modern scholars using techniques of comparative philology and textual criticism.5 Linguistic methods are-particularly effective in differentiating layers within a single text which were written

at separate times and later joined together. Unfortunately, such studies are often hampered by having to work with redactions of texts which were refashioned centuries after the time they were originally composed, or with translations into Tibetan or Chinese of texts which no longer survive in Middle Indic or Sanskrit languages.6 David Snell­ grove has summed up the difficulties of precisely dating Esoteric-Buddhist texts:

It is also generally known that to date any sutra or tantra with certainty is practically impossible, because most of them continued to change and develop over long periods of time, and the most that can be done is to date them at a, particular stage of development when they happened to be translated into Chinese, or by such references as may be made to them during clearly recorded periods of Tibetan history.

Even a group of texts which can be placed in rough chronological order constitute unevenly spaced points along a complex line of development. Greater or lesser gaps remain unmapped. A corpus of texts cannot represent a continuous progression, but

only stranded markers of quantum leaps of development. Witl:tin the voids, works of art are waiting to be recovered and interrogated. There are hundreds of sculptures and paintings with Esoteric Buddhist themes preserved from South Asia between the seventh and the twelfth centuries. As long

as it is granted that works of art carry intelligible meaning, an art historical analy­sis of the archaeological record can provide a valuable complement to uneven tex­tual evidence. As Edward Conze, the late doyen of Indian Mahayana Buddhist tex­tual studies, comments, "To the historian everything is precious that may help him to trace the often obscure development of Indian religious thought between 600 and 1200 AD.118

When ideas - which may or may not be recorded in texts - have become so well-accepted by one or another part of the Esoteric Buddhist tradition that they have been given visual form, we have unimpeachable evidence that those ideas were held at a specific place and time. This can contribute to the corrective project

Gregory Schopen has initiated by replacing the ready acceptance of what a few peo­ple may have thought with what groups actually did. The replacement forces us to go beyond the preoccupation of historians of religions "with what small, literate, almost exclusively male and certainly atypical professionalized subgroups wrote." By making use of art in the way Schopen uses donative inscriptions, we can also make note of local variations which did not become visually or textually canonical. Most importantly, we can follow the maturation and dissemination of ideas in ways that textual evidence will not permit us to do.

The evidence of art may be useful, but it is discouraging in its very abundance. Some order must be imposed on its disorienting richness before it can shed light on the development of Esoteric Buddhism. Particular Esoteric Buddhist themes may be grouped systematically to organize the mass of material left by the

exigencies of chance survival. The second-level order imposed by thematic grouping, however, should not be merely an artificial system which reveals meaning only in ways which would have been irrelevant or tangential to the aims of the creators. In order to consider the treatment which the theme receives over time, the theme must be continuously present in every period under examination.

Beyond these rather practical constraints, it is essential that the theme selected as a focus be in the nature of an "operative metaphor."1 0 While it need not be the single dominant trope, it should be at least what Mircea Eliade points to when he notes the capacity of symbols to express paradoxical situations and structures of ultimate reality;11 in other words, an integrative symbol as opposed to a subsidiary symbol. A homologous notion is what in art historical

studies has been referred to as "symbolic form," or, necessarily, an interpretable paradigm. The vehicle of meaning must reveal certain fundamental attitudes and developmental pheno­mena or structures ("processes"), while on the local scale it describes the diverse incidences of individual events that embody existential events: the macro-process on a microcosmic scale (i.e., a "pattern").

As embodiments of the "demonic divine," wrathful deities provide just such an appropriate perspective from which to study change in the development of Esoteric Buddhism and to explore its doctrinal commitments. The wrathful deities are present from the sixth to seventh centuries, the very beginning of the

period in which Esoteric Buddhist sculptures appear in South Asia. The theme has formal and doctrinal connections with Mahayana, early Buddhism and non-Buddhist imagery. It endures without hiatus throughout the period of Esoteric Buddhism in India, at least until the end of the twelfth century. The deities were important enough to occur in all of the cultures outside of India which received Esoteric Buddhist transmission of texts, images, doctrines and rituals; that is, in China and Japan, in Southeast Asia and in Mongolia, Tibet and Nepal.

While the wrathful deity theme can be identified in hundreds of works of art, it is not present in all art which may be considered Esoteric Buddhist in inspiration. This absence works in a positive way, serving to limit the scope of inquiry without narrowing it excessively. The "demonic-divine" theme is compatible with surviv­ing texts which can be used to unpack the wrathful symbolic form. Most impor­tantly, the theme allows us to witness distinct changes over

periods in the treatment of the subject matter, in ways which are meaningful and consistent with textual tra­ditions assigned to the same period through other types of evidence. The wrathful deity theme is an extremely sensitive marker of doctrinal change because its signif­icance continues to be reformulated along with Esoteric Buddhism itself. The coher­ence between the evolution reflected in imagery and in surviving texts seems to belie Bharati's observation that there was "no modification of icons, but constant elaboration of philosophical concepts."

The perception that icons were scarcely modified actually sets into relief the usefulness of the wrathful deity compared to other potential themes. The Buddha image, for instance, has proved to be a suitable topic for general surveying of Bud­dhism,16 but is not sufficiently distinctive to Esoteric Buddhism. The special form known as the "adorned Buddha" is a topic which stands in clear need of study with greater reference to extant images. Moreover, the adorned Buddha

does seem to be one of the defining characteristics of the Esoteric Buddhist visual program. The cen­tral issue in the case of the adorned Buddha, however, is essentially one of dating its origin and cataloguing the settings in which it occurs. The image itself under­went little change once it appeared in the late

seventh or eighth century in both India and China. The textual problems and iconographic meanings have been stud­ied in some depth by Paul Mus and others.1 7 Scholars generally seem to have agreed that adornment of crown and jewelry "are emblematic of the Buddha's spir­itual kingship, reflecting the ritualistic elements of royal consecration ceremonies practiced in Vajrayana Buddhism."

The bodhisattva image in general, or individual bodhisattva such as Maii.jusri, Avalokitesvara, Akasagarbha, Maitreya and Vajrapani , would seem to be fruitful topics of study.1 9 Once again, however, the bodhisattva are by no means associated with Esoteric Buddhism in any exclusive sense, and a study of any one of them, however valuable, must cover a great deal of ground outside the realm of Esoteric Buddhism. The wrathful deity theme proves to be much more efficient in profiling the distinct character of evolving Esoteric Buddhism.

One of the primary defining characteristics of Esoteric Buddhism, at least in its later South Asian variety, is its "polarity symbolism." The sexual content of this symbolism has generated a great deal of controversy, which is still not resolved. Scholarship has tended to fit into three principal camps with regard to images or textual descriptions of a deity embracing his consort. The first confines itself to expressions of abhorrence and disgust with a tradition which

would deem such images religious. The second interprets sexual imagery purely as philosophical metaphors, while the third insists that it reflects actual practices of Esoteric Buddhists. 20 As appropriate as sexual imagery may be as an integrating symbol for Esoteric Buddhism, its appearance in both texts and art is extremely limited in South Asia before the emergence of the third stage of Esoteric Bud­dhism. In earlier stages of the tradition, notably those which

were imported into China and Japan before the mid-ninth century, polarity symbolism is only implic­it. Historical data suggest that sexual imagery is a later development and there-

fore this particular aspect is less fruitful for the study of Esoteric Buddhism's overall development.

It has been persuasively argued that the mai;u;lala is the ultimate defining and integrating symbol of Esoteric Buddhism. There is considerable validity in that thesis. It is not a practical theme to take up i an investigation of the visual remains of Esoteric Buddhism, however, because except for a few ninth-

century East Asian painted mai:u;lala and a tiny sample of unusual sculpted ones no ma,;i4ala survive from before the tenth or eleventh century. Therefore a study of ma,;i4ala will either depend on written texts from before the tenth century, or will be confined to the later periods of Esoteric Buddhism.22 The wrathful deity theme by contrast has much to recommend it. Its impor­tance will become fully apparent only as the evidence unfolds.

It is nonetheless not out of place here to demonstrate the centrality of the wrathful deity to the later forms of Esoteric Buddhism. In his discussion of the texts of the later period, known as anuttara yoga tantra, Giuseppe Tucci writes, "Nearly all the Tantras of the Anut­tara class centre round the

symbol of Aksobhya or of his hypostases: Heruka, Heva­jra, Guh yasama ja.2 3 This is significant. Tucci has observed that the important texts and images of the later period put at the very center of their teachings, not the his­torical Buddha, not cosmic Buddha forms, not the popular bodhisattva, but wrath­ful deities such as Hevajra and Heruka (also known as Saqwara). Saqwara, Heva­jra and Cai:,.c,iamaharol?ai:ia along with another wrathful lord, Kalacakra, have elab­orate ritual texts named after them which feature them at the center of complex '

Others have also noted the position of centrality of wrathful deities in the his­tory of Esoteric Buddhism and its art. R.A. Ray observes that "the integration and hierarchical arrangement of [the ma,;i4ala's] terrible deities [indicates] not only their fundamental importance to the Tantric process of transformation, but also the dif­ferent stages of awareness bound up with this process." According to P.H. Pott, one of the two significant and distinctive aspects of Mahayana iconography is "the demonizing of its most effective figures." Stephen Beyer has likewise taken note of the paramount position of the later forms of the wrathful deity class:

[T]hey belong to the type of Heruka - multitudinous arms bearing attributes of power and ferocity, three-eyed faces distorted in scowls of rage These Herukas are perhaps the most potent and symbolically evocative of all the Tibetan deities.

In his treatment of the career of one wrathful deity, Etienne Lamotte states that "Vajrapani in the end rose to the summit of metaphysical reality and acceded to the rank of the supreme being."

It may be argued, then, that the wrathful deity is one of the central paradigms of late Esoteric Buddhism. Yet wrathful deities like Vajrapani did not start out com­peting with the Buddha for the central position. We find in the phase immediately preceding the culminating period that the wrathful deity plays an important but not always central role. During the earliest period his role is definitely subsidiary. Yet even then he is consistently present. Therefore the theme satisfies the requirements of both consistency and volatility and can thus reflect meaningful change over time.

The theme of the wrathful deity in Esoteric Buddhist art has additional advan­tages in the project of clarifying the history of Esoteric Buddhism. A number of studies over the last seventy years have elucidated the evolution of specific wrath­ful deities within the Esoteric Buddhist system. For example, Robert Duquenne has illuminated the career of one of this group, known as Yamantaka. He has made

use of Indian Sanskritic as well as Chinese and Japanese sources. Similar analyses of South Asian, Tibetan and East Asian texts have been produced by Robert van Gulik for the horse-necked wrathful deity Hayagriva, by Lamotte and Lalou for Vajra-pal).i, by Iyanaga Nobumi for Trailokyavijaya, by Frederic A. Bischoff for Ucchusma-Mahabala and Roger Goepper for Aizen-Myoo. In addition, there have been recent translations of a number of texts dealing with various wrathful deities.31 All of these sources, along with the studies of the ritual cycles of Sarp.vara, Car:ic;!amaharo ar:ia and Hevajra mentioned above, provide a solid foundation for the study of this class of deities.

Given the recent strides made in this area, it is perhaps now an overstatement to say with Ray that "Unlike the problem of sexual symbolism, Tantric Buddhist horrific symbolism has received, as Pott has noted, hardly any attention at all in Western interpretations of the tradition." Nevertheless, since few of the previous studies have made more than incidental use of actual images, it is still very much a productive area for research.

Finally, the theme of the wrathful deities, apart from its value for the historical study of Esoteric Buddhism, is of interest in other ways for the history of religion, ideas and art. The forms found in Esoteric Buddhist images seem to share a great deal with other traditions of the "grotesque" in that as a "species of confusion" they violate the expectations of reason and natural order.33 The grotesque is often occu­pied with the transgression of the boundaries between the sacred and the profane and appears in a number of other religious and secular traditions.34 What is rather unique to these wrathful deities is that Esoteric Buddhism has infused forms which are expressly violent and threatening (and, at least to the unfamiliar, troubling and offensive ) with divine rather than demonic significance. They do not represent merely the darker, destructive side of nature. Instead, by fusing an outer expression of malevolence with

an inner valency of compassion, they enfold chaos into the sacred. The are visual expressions of the coincidentia oppositorum. For the adept, the fusion paralyzes or transcends dichotomizing cognition, and dualism is subverted. Thus fundamental questions concerning the nature and function of images of the holy are brought into relief by these works.

In this context there is no question of attempting an analysis of the greater "demonic-divine" category as a pattern of human thought and experience. How­ ever, it is my hope that by disclosing the main structural outlines of one particular set of expressions - that of eastern Indian Esoteric Buddhism - this work will contribute materials which the historian of religion can at some future time inte­ grate into a broader study.

The present research, then, is a systematic study of the images of wrathful deities in the Esoteric Buddhist art of eastern India. It utilizes prior studies, along with a number of primary sources in the form of texts translated into Chinese in the eighth and ninth centuries. The most important primary sources, however, are a large stock of wrathful deity images compiled on three continents over the past fifteen years. From a narrowly aesthetic point of view, only a few might be considered "masterpieces." Many are fragmentary and of mediocre quality, yet we can still readily recover from them information of considerable

historical value . Most are sculptures, which were located, studied and photographed in eastern India. Despite their uneven quality from a technical point of view, the images gathered here represent a nearly continuous record from the seventh through the twelfth centuries. Passages of relevant texts translated into Chinese in the seventh, eighth and ninth centuries bearing on the subject are related to contemporary South Asian images. Early western Tibetan images are also utilized

to a lesser extent to solidify the dates of the predominance of one phase over another and to provide a context which rarely survives elsewhere. A total of more than one hundred and fifty images of wrathful deities dating between the sixth and thirteenth centuries creates a collection of visual data of both local and gen­eral value.

Together these materials will allow us to survey the development of the class of deities as a whole. We can document their rise in importance from subsidiary deities with roots in non-Buddhist imagery to their height of independence, which rivals the Buddha in terms of immediacy and status within praxis. A continuous narrative emerges from the systematic analysis of extant visual images. These are supplemented by inscriptions and texts which allow for the identification and interpretation of individual images.

With this narrative history of the wrathful deities in place, we can propose a model of the historical development of Esoteric Buddhism itself. This model is then superimposed on the model emerging primarily from textual studies. We will see that the two models coincide in broad outlines. The artistic record fills

in gaps in the textual model but fails to reflect certain distinctions made in the scriptural model. There are also a number of interesting temporal disparities between the two models which call attention to the need for a refinement in the periodization of Eso­teric Buddhist movements. The disparities

recall Oleg Grabar's distinction between absolute and relative time, between the time of events and cultural time.39 In our case we must distinguish when an idea is formulated and recorded among a local­ized group of initiates, and when that idea is accepted on a broader scale within the tradition. It is dangerous

to give a date to an Esoteric Buddhist movement when that date derives solely from the earliest layer of its most important texts or com­mentators. Works of art may prove to be better historical markers than texts in such cases.

The primary purpose of this work is to systematically discover patterns with­ in the group of wrathful deity images, identify these patterns along coordinates of time and geography, disclose the relationships between the patterns and reconcile the results with what is already known about the historical and doctrinal develop­ment of Esoteric Buddhism. On this basis, a model is generated and analyzed for its potential for integrating new information of different natures.

In sum, to paraphrase R.A. Ray's prediction by substituting the phrase "wrath­ful deity" for "ma7J-ef,ala," the "central importance of the [[[wrathful deity]]] through the long history of the Vajrayana is such that, at some future time, it may prove possi­ble to write a detailed history of the Diamond Vehicle through detailing the changes and transformations undergone by the [[[wrathful deity]]] image." 40 Notwithstanding an awareness of my own linguistic and conceptual inadequacies, the present work represents an attempt to do just that.

Geographic and Temporal Parameters

At this point it may be appropriate to briefly outline the spread of Esoteric Bud­dhism, at least for the cultural areas which provide relevant images for this study. To begin with India, it is apparent from texts and from concentrations of archaeo­logical remains that Esoteric Buddhism found a home in the monastic institutions of eastern India during the period of the eighth century through the twelfth or thir­teenth century, after which Indian Buddhism declined sharply. These monastic centers sponsored a significant share of the surviving Esoteric Buddhist art, along-

side more conventional Mahayana art. Bihar and Bengal were the most prominent, and Orissa is increasingly recognized as an important site of Esoteric Buddhist activity.

Most of the South Asian objects presented here come from eastern India (i.e., modem Bihar, Bengal, Orissa). The majority of them are stone stele, "the medium most commonly met with in the art of eastern India." 43 They are currently located either in museums or excavation sites. These include the collections of the

major museums in New Delhi, Lucknow, Samath, Patna, Calcutta and Bhubaneswar, as well as many large and small sites in Bihar and Orissa, the larger ones including Bodh Gaya, Nalanda, Samath, Ratnagiri, Udayagiri and Lalitagiri. The concentra­tion of wrathful deity images available for study there justifies eastern India's importance for this subject.

Field work was concentrated in eastern India and in Indian Tibet (Ladakh, Zanskar, Spiti), though it was also conducted in Nepal and in Maharashtra, at Aurangabad, Ellora and Kanheri, not to mention East and Southeast Asia. Howev­er, it is well known that Esoteric Buddhism spread through a wide area of South Asia. This is illustrated in a very concrete way by the fact that dharani inscriptions have been recovered from different sites in Bengal, Bihar and also northwest India, all deriving from the same or similar texts.44 Other areas of India, notably Kashmir and southern and western India, also participated to some degree in the develop­ment of Esoteric Buddhism.

There are a number of indications in the textual record that major centers were located in the south of India and in Sri Lanka. The Tibetan historian of Esoteric Buddhism, Taranatha, mentions Dramil (Tamil), while inferences from the Tamil epic Manimekalai have suggested to some the influence of Esoteric Buddhism. 46 Lilapa, Nagarjuna and perhaps others of the eighty-four Mahasiddhas were origi­nally from southern India.

There are also numerous references to South India and Sri Lanka in the biographies of the three major Esoteric Buddhist masters who came to China in the eighth century. Unfortunately, very little in the way of material arti­facts remains to support this tradition, although such "late" deities as Heruka, Tara and Ugra-Tara were recovered at Amaravati in southeast India. An inscription on an image of Prajnaparamita in a style of the ninth or tenth century found at Nalan­da

which records the gift of a man from Karnataka also underscores the connec­tion between South India and Esoteric Buddhism, albeit indirectly. Von Schroeder has recently gathered together textual and visual data for Mahayana and Esoteric Buddhism in Sri Lanka. Despite the textual and inscriptional evidence for a signif­icant if limited presence of Esoteric Buddhism there, the extant images are precious but few.50 Whatever its role may have been, because of the dearth of visual mater­ial, the south can play very little part in this study.

Western Indian cave sites are now being interpreted as early Esoteric Buddhist sites. In particular, Geri H. Malandra has studied the Buddhist caves at Ellora in Maharashtra. The caves date to the seventh and eighth centuries, and she con­cludes, "Ellora is a remarkably rich resource for the study of early Tantric Buddhist iconography." Caves Six and Seven at Aurangabad, datable to the late sixth or seventh century, have also recently been reinterpreted as

early Esoteric Buddhist sites.52 The western Indian cave sites of Ellora, Aurangabad and Ajanta, however, did not continue to be sites for the production of mature Esoteric Buddhist art. While they are generally fertile fields for the early development of Esoteric Bud­dhist art, one cannot follow a single theme through a long period of time. Thus their value as comparative material in the development of wrathful deities is restricted to the early period.

The present study also includes images from Kashmir, which was an important center of Buddhism, although "documentation is strangely lacking for study of the strictly historic aspect of this problem."53 Surviving images, mostly metal sculp­tures of middle and late period Esoteric Buddhist wrathful deities, are particularly numerous.54 Kashmir was also important because of the regard in which it was held among Buddhists along the land route to China, at such Central Asian centers

as Khotan and Dunhuang. Kashmiri Esoteric Buddhist masters and artists were

also critical for the transmission of Esoteric Buddhism to Tibet.

After a severe persecution of Buddhism in Tibet in the mid-ninth century, Yeshe b, king of western Tibet, sent a mission to Kashmir in the late tenth century. In good measure due to the efforts of Rinchen Zangpo (958-1055 CE), one of the mis­sion's returning members, Esoteric Buddhism was again ascendant in western Tibet. In Ladakh temple sites like Alchi, Sumda and Mangyu house rare ma,:z ala paintings featuring wrathful deities which date to the twelfth or early thirteenth century.56 They can be associated with Rinchen Zangpo's school and bear distinct Kashmiri stylistic and iconographic features.57 The results obtained through field work will play a role in the present study.

In the middle of the eleventh century, Atisa (982-1054) came to western Tibet and became Rinchen Zangpo's teacher.58 He brought with him the doctrinal innovations associated with the legendary Mahasiddhas of eastern India. He was also a highly respected senior monk who had links with the major monasteries of eastern India, including Nalanda and Vikramshila. Moreover, he represents the wider, nearly global, perspective of Esoteric Buddhism. Atisa is recorded as

having left India in 1012, returning in 1025. During the intervening years he studied with an Esoteric Buddhist master in Suvarnadvipa. The exact location of Suvarnadvipa is still a matter for debate, but it is most likely in Sumatra and at any rate was within the large Southeast Asian maritime empire known as Srivijaya.59 Atisa thus also represents the movement of Esoteric Buddhist practice and imagery into West and central Tibet from eastern India.

Although East Asian Esoteric Buddhist art is not included in the present work, I have made extensive use of the texts transmitted from India and translated and preserved in China and Japan. A sketch of the history of Esoteric Buddhism in East Asia may be helpful here. Interest in South Asian proto-Tantric or zomitsu ("diffuse esotericism") texts is attested in China from the fifth century. 1 The introduction of early mature Esoteric Buddhism into China took place

in the late seventh and early eighth centuries. Several Chinese pilgrims went to India to study, bringing back Esoteric Buddhist texts and, we must assume, images. Yijing's Gaoseng zhuan commemorates fifty-six Chinese and Korean Buddhist monks who left China for India in the seventh century, and there is no reason to believe that the flow of Buddhist travelers slackened in the next century. The systematic beginnings of Eso­teric Buddhism in China, however, revolve

around the work of several charismatic Esoteric Buddhist masters who came to China from South Asia, where they had studied at such important centers as Nalanda and Vikramshila.6 5 During the eighth and the first half of the ninth centuries, Esoteric Buddhism received the patronage of the Tang Dynasty emperors and

high officials at cour t6.6 The court financed Eso­teric Buddhist art of great quality, which prominently features wrathful deities.6 7 Following a severe persecution of Buddhism between 842 and 845 CE, Esoteric Buddhist activities in China were extremely restricted. South Asian texts continued to be translated in the tenth and eleventh centuries, though the texts were not pro­mulgated nor their rituals widely practiced.68 In a few pockets, mostly on the fringes of China, Esoteric Buddhist ideas continued to have a certain influence and

crop up in later Chinese art, as at the Mogao caves at Dunhuang. Such is the case in Yunnan into the twelfth century, and a number of wrathful images in painting and sculpture survive from that milieu.7° In the territories of western China ruled by the Tangut Xia (1038-1227 CE) and by the Mongols during the Yuan Dynasty (1260-1367 CE), there was a reintroduction of Esoteric Buddhist themes into China under the auspices of the Tibetans. Not only are there Esoteric Buddhist murals painted during that time at Dunhuang, but at other sites in western China as well, in northern China at the Feilai Feng near Hangzhou, and even in central coastal China.

While still riding a tide of imperial patronage in China during the early ninth century, Esoteric Buddhism was brought to Japan. The most important Japanese monk in this process was Kukai, or Kobo Daishi, as he was titled after his death in 835 c E.72 Esoteric Buddhism found a permanent home in Japan under the name of Shingon ("True Word," i.e., mantra ).73 Its character was shaped by two factors: first, the stage of Indian Esoteric Buddhism which Japan received through China in the ninth century; and second, reformulations in China and Japan. The Indian basis is recognizable, especially in artistic products of the early centuries (i.e., the ninth to twelfth centuries).

There is therefore justification in using East Asian texts to amplify the South Asian textual record for the purpose of interpreting Esoteric Buddhist developments in South Asia. In turn, clarification of the early history of Esoteric Buddhism in India will shed light on the patterns inherited by East Asian practitioners. Both Robert Buswell and RA Stein have pointed out the importance of comprehensive approaches that transcend national traditions.74 It is hoped that the present work will be of use to those interested in both sides of the transmission.

The Schematic Model

[[}n order to draw a general outline of the history of Tantrik ideas in Buddhist literature and life, we must disregard the traditional divisions as embodied in the Tibetan cata­logues or the Western theories on the subject and build a classification of our own. Louis de La Vallee Poussin, "Tantrism"

I define the Esoteric Buddhist group of deities that I label krodha-vighnantaka on the basis of a fundamental expressive characteristic (krodha, wrathfulness) and a pri­mary identifying function (vighnantaka, destroyer of obstacles). The members of this group share the aspect of wrath with other types of deities

within the Buddhist pantheon, such as the dvarapala gate guardians, the Lokapala (guardians of the four directions) and the dharmapala protectors. The krodha-vighnantaka group may also be charged with the same functions as these groups. However, the krodha-vighnantaka group is entrusted with additional functions and attains a higher status. These added functions involve "inner and outer practices," destroying both the inner obstacles to enlightenment (greed, anger, sloth, etc.) and demonic projections which attack the meditator with diseases of mind and body.

The first artists to depict the krodha-vighnantaka in the sixth century in India drew on prior depictions of Yak a, ayudhapurufa (personified attributes, such as Vi 1:m's wheel, Siva's trident) and Siva-Gar:,.a (Siva's auspicious dwarf attendants). The Yak a image as it appears in early Buddhist art was the ultimate source for the image of the krodha-vighnantaka. Two types of Yak a images were created: the dwarfish, big-bellied type (here called vamana-Yak a), and the heroic warrior type

(here known as vira-Yak a ). Following this twin tradition, the krodha-vighnantaka are also found in both forms, though the vamana-Yak a body-type is more common in the early period of Esoteric Buddhist art in India.

Analysis of the images reveals large-scale patterns in the course of develop­ment of the krodha-vighnantaka, suggesting three main phases which I argue corre­spond closely to the historical stages of Esoteric Buddhism. They will be treated separately in Sections One through Three. The origins of the three

stages may be considered diachronically, though after the eighth century Phase One and Two are contemporaneous, and after the tenth century all three are contiguous. Neverthe­less, one may say that Phase One dominates the period between the late sixth to the eighth centuries, Phase Two presides from roughly the eighth to the late tenth cen­tury and Phase Three from the late tenth century through to the twelfth.

2 . Avalokitesvara with Hayagriva, Ratnagiri (Orissa), ca. 9th century

The Phase One krodha-vighnantaka is subordinated to a bodhisattva, who is wor­shipped with offerings and dharani- prayers for specific ends, such as the removal of misfortune . The krodha-vighnantaka performs tasks like gathering together sen­tient beings to whom the bodhisattva preaches, subduing rowdy elements, or per­sonifying attributes of the bodhisattva, such as his adamantine wisdom or the power of his voice. This phase has links to the traditional Tibetan classification of kriya tantra (action tantras; rites of magic) and carya tantra (performance tantras; rites of religious practice) and can be related

to the trikula (three family) system . Phase One grows directly out of orthodox Mahayana belief and practice, making it difficult to distinguish the two. Important texts for this phase which are discussed in relation to the krodha -vighnantaka are the Ma fijusrimulakalpa (MMK) and the Maha Vairocana abhisambodhi sutra ( M VS, transitional between Phases One and Two) as well as a number of minor texts translated into Chinese in the seventh and eighth

centuries. The first is the most enduring phase, continuing into the twelfth century in India , by which time it had been transmitted to Nepal, Tibet, Southeast Asia and to a lesser degree East Asia . Because of the length of time over which Phase One images were produced, the krodha-vighnabtaka image can be followed through sev­eral stages of growth, florescence and finally stylization accompanied by dimin­ished importance. The principal Phase One krodha-vighnantaka deities whose careers are followed here are Hayagriva , Yamantaka , Mahabala, and Vajrapu ru a.

Particular members of the Phase Two krodha-vighn tintaka group become important independent images, often with their own attendants, or are fitted into centrally­ ordered schemes of deities (i.e., ma l;fala ). [3] Throughout this period the krodha­ vighnantaka himself performs the tasks of conversion , aversion of misfortune and destruction of obstacles. The wrathful deities are no longer paired off in a subordi­nate relationship with a bodhisattva; they now correspond

to one or another aspect of the Five Directional Buddhas, either one of the five wisdoms or each Buddha's "commanding power." Some Phase Two krodha -vighntintaka appeared earlier in Phase One imagery, but others are new creations. Phase Two may be compared with the traditional Tibetan classification of yoga tantra (Yoga tantras; rites of yoga), and can be related to the development of the paficakula (five family) system . Impor-

3. Trailokyavijaya , Achutrajpur (Orissa), ca. 10th century

4. Heruka, Ratnagiri (Orissa), ca. 11th century

tant texts for this phase include the MV S , the Sarvatathagata Tattvasamgrha ( STTS ), the Sarvadurgatiparisodhana (SOPS), and sections of the Guhyasamaja Tantra, as well as a number of minor texts translated into Chinese in the eighth century. The prin­cipal Phase Two krodha-vighn ii ntaka deities whose careers are followed here are Yamantaka and Vajrapani-Trailokyavijaya.

Phase Three

The Phase Three krodha-vighniintaka ranks together with the Buddha among the most important deities in Esoteric Buddhism . He is considered a form of the Bud­dha who plays a more prominent role in Esoteric Buddhist art than the historical Buddha. The krodha-vighniintaka appears at the center of the dominant ma,:i ala cycles. As such he supplants both the bodhisattva and the Buddha as the primary yidam, or initiatory deity. The krodha -vighn iintaka

embodies the enlightenment and all the power of the Buddha and, further, the supreme bliss and reality itself . Most of the Phase Three krodha-vighniintaka have new identities which at a deep level are reformulations of Phase Two deities in Phase Three terms. The third phase corre­lates with the traditional Tibetan classification of anuttara yoga tantra (supreme yoga tantras), and its formation had close links with the Hindu tantric smasiina (cemetery) cult. Important texts for this phase are the Hevajra Tantra and the Samvarodaya Tantra. Outside of Tibetan-inspired instances under the Tanguts, Mongols

and Manchus, this phase did not exercise much influence in China or Japan, though some of its texts were translated into Chinese in the late tenth and eleventh cen­turies . The principal Phase Three krodha -vighniintaka deities whose careers are fol­lowed here are Heruka (4), Hevajra, and Sarµva ra.

The tripartite outline demonstrates that the krodha -vighn iintaka deities have undergone a radical development in tandem with Esoteric Buddhism itself, which gradually began to infuse the krodha -vighniintaka deities with greater meaning. This development is reflected graphically in artistic representations. The conceptual model is derived from the study of wrathful deity images, which will proceed as follows . The introduction investigates terminology used for

this class of deities . It also defines the identity of the group in contradistinction to other groups of deities with partially overlapping functions and formal appear­ance. The chapters of Section One focus on the evolution of Phase One krodha­ vighniintaka imagery, following the careers of four of its members .

Two important issues dealt with in the course of this section are a) the origins of the forms used for Phase One wrathful deities in earlier Buddhist, Brahmanical, and cultic art; and b) the religious context of mixed Mahayana-Esoteric Buddhism in which Phase One images are found. The chapters of Section Two deal with Phase Two images, and Section Three treats Phase Three wrathful deities. In the Afterword I compare the historical model based on the krodha-vighniin taka material with other models of Eso­teric Buddhism and summarize the results of the investigation .