EATING THE HEART OF THE BRAHMIN: REPRESENTATIONS OF ALTERITY AND THE FORMATION OF IDENTITY IN TANTRIC BUDDHIST DISCOURSE
by David B. Gray
eating the heart: transgression in tantric buddhist literature
Religious identity, as is now widely recognized, is not monolithic but re-lational, developing and changing through the encounters that continually occur between competing religious traditions.1 In this article I will explore the process by which religious identity was formed in a Tantric Buddhist tradition during the early medieval period, through an exploration of a body of discourse
composed during its period of early development. This tradition, which gave rise to the Buddhist Yoginitantras, is fascinating be-cause it developed in dependence upon a non-Buddhist tradition, and thus faced the challenge of forging a distinctly Buddhist
identity. This challenge was particularly great as this body of scripture, particularly the Cakrasam- vara Tantra, which will be the focus of this essay, exhibited numerous signs of “heretical,” non-Buddhist affiliation, and was also notorious for its transgressive rhetoric.
1 I will argue this below in relation to the early medieval South Asian context. For two re¬cent studies on the formation and change of religious identity in contentious cultural contexts see Chitralekha Zutshi, Languages of Belonging: Islam, Regional Identity, and the Making of Kashmir (New York: Oxford University Press, 2003), and Kathleen Flake, The Politics of American Religious Identity: The Seating of Senator Reed Smoot, Mormon Apostle (Chapel Hill: University of North Carolina Press, 2004).
This article will explore two closely related phenomena. The first is the process by which Buddhists appropriated elements of discourse, both textual and ritual, from a Hindu tradition, focusing on an example notable for its transgressive character. Second, it will examine the process by which the elements of this “charnel ground” (smasana) culture were adapted and transformed within a
monastic Buddhist context. This will be done through, not the examination of a normative instance of these pro¬cesses, but rather, an anomalous instance, one that highlights a limit of Tantric Buddhist discourse, a limit that can be ascertained along the lines outlined by Foucault.
This limit will be highlighted via a comparison of two closely related texts that were composed in the eighth century, one in China and the other in India. The first recounts a myth of the origin of a ritual praxis, narrating the subdual of Saiva dakinis, represented as heretical on account of their engagement in transgressive practices of violent ritual and anthropophagy. The second,
an Indian Buddhist Tantra, describes in some detail the same praxis of anthropophagy, and thus comes close to crossing the line of het- eropraxy established by the former text. These texts are useful not so much because they are representative cases of the processes of appropriation and adaptation, but rather because they are exceptional or extreme cases that represent the limits of the processes. They shed light upon the manner in which Buddhists in eighth-century India struggled to reformulate their identity in response to internal and external pressures.
The first text in question concerns a mantra contained in the Mahavairo- cana-abhisambodhi Tantra, an early and important Tantric Buddhist text likely composed during the mid-seventh century in India. This mantra, hri hah, styled the “dakini mantra,” is listed in the fourth chapter of this text, entitled the “General Mantra Treasury.” This chapter concludes with a long list of mantras
associated with various classes of nonhuman entities, including gods, titans (asura), and a host of nonhuman spirits known for their fondness for human flesh, such as the räksasas, yaksas, pisäcas, in addition to the dakinis. Buddhist mantras are “spells”; that is, they are carefully structured verbal utterances that are recited in conjunction with ritual practices to produce a desired magical effect. In the Mahavairo- cana-abhisambodhi Tantra, these are presented devoid of any contextu- alization, with no explanation of their history, ritual use, or the magical effects of their successful application.
The Indian master Subhakarasimha and his Chinese disciple Yixing addressed this lacuna in their massive Chinese Mahavairocana-abhisam- bodhi Sutra Commentary (MAC), which they composed in the early eighth century in Chang-an. In this work they relate a fascinating myth concern-ing the revelation of the dakini mantra, which occurs as follows:
Next is the dakini-mantra. There are those in the world who are well-versed in this technique, and are practitioners of Isvara's esoteric lore (vidyä, ^W), who are able to know when a person's life is about to end. They know of this six months in advance, and
then knowing it they immediately apply the spell to ex¬tract a person's heart and eat it. It turns out that within the human body there is a concretion, which is thus called human concretion (AÄ). It is like the concre¬tion found in cattle. One who is able to eat it attains the greatest powers (siddhi,
[such as] circling the world in one day, obtaining anything that one de¬sires, and being able to control people in various ways. If they have an enemy, they can use this spell to punish him, causing extreme sickness and suffering. However, this method cannot
kill people. Should they follow this self-devised method, they know when a person is to die six months in advance. Knowing this, they use this spell to extract his heart. Although they take his heart, there is [another] procedure, [whereby] they must replace his heart with something else. [Thereby] this person's life does not [prematurely] end. When he reaches his time of natural death, then [the heart simulacrum] malfunctions.
Their chief was the yaksa Mahesvara, who worldly people say is the ultimate [god]. They were subject to Mahakala, the god called the “Great Black One” (^M)- Vairoc ana, employing the method of Trailokyavijaya and wanting to exterminate them, transformed
himself into Mahakala, exceeding him in an im¬measurable manifestation. His body smeared with ashes in a desolate place, he summoned with his magical art all the dakinis, who had all of the magical powers [such as] flying, walking on water and being completely unhindered. He upbraided them, saying: “Since you alone always devour people, now I will eat you!” Then he swallowed
them, but did not allow them to die. Once they had submitted, he released them, completely forbidding them to [eat] flesh. They spoke to the Buddha saying, “We presently eat flesh to survive. How can we sustain ourselves now?” The Buddha said, “I will permit you to eat the hearts of dead people.” They said, “When a man is about to die, the mahayaksas and so forth know that his life is exhausted, and they race there to eat him, so how can we get [our share]?” The Buddha said, “I will teach you the mantra procedures and mudras.
You will be able to know six months before someone dies, and knowing this, you should protect him with this method, so he will not fear being injured. When his life has expired, then you can seize and eat [his heart].” In this way, they were gradually induced to embark upon the path. Thus there is this mantra, hri hah, which removes the taint of heretical practices.
This myth represents a Buddhist justification of what Phyllis Granoff has termed “ritual eclecticism.” This phenomenon, common in India during the early medieval period, entailed the acknowledgment of the efficacy of re-ligious practices that are openly
recognized as belonging to an outsider group. These are often assimilated into the appropriating group's practice tradition by strategies of subordination, such as via claims that the tradi-tion's own practices are “supermundane” (lokottara), while those of the outsider's are “mundane” (laukika).
Such subordination is often dramatized in Buddhist literature by myths that portray the outsider religious group as dangerous “heretics,” whose misdeeds trigger a cosmic Buddha such as Vairocana to subjugate them, bringing both them and modified forms of their practices into the Buddhist fold. These myths are products of a process in which Tantric Buddhists, having appropriated
elements of Hindu ritual, were seeking to forge an identity through a representation of a radical “other,” in this case Saiva Hindus. This representation does not, naturally, provide us with any re¬liable information about the other group, as distortion, exaggeration, and outright fabrication are common colors in the polemicist's palette. Repre¬sentations of a rival group engaging in radical actions such as cannibal¬ism are relatively common in this genre of religious literature. These constructions of alterity
have the complementary purpose of delimiting the self, “making a total contrast between insiders and outsiders.” As John Henderson has argued, polemical religious discourse inevitably implies “an account of both self and other, of orthodox as well as heretical; for the former positions and defines itself by reference to the latter, even arises and develops historically by constructing an inversion of the heret¬ical other.”
The “heretical other” constructed in this myth are dakinis who are affil-iated with the Saiva deity Mahakala. This passage contains one of the ear-lier occurrences in Buddhist literature of the dakinis, who would become very important in the later Buddhist
Yoginitantras. One of the earliest appearances of the dakini in Buddhist literature occurs in the La“kavatara Sutra, a Buddhist scripture composed in India during the fourth century, where they appear to designate a class of female nonhuman or quasi-human beings, associated both with the anthropophagic Raksasi demons as well as with outcaste groups of human carnivores. The text threatens carnivores with the following fate: “The [carnivore] is born again and again as one who is ill-smelling, contemptuous,
and insane among the families of the Candala, the Pukkasa, and among the Domba. From the womb of a Da- kini he will be born into a carnivorous family, and then into the womb of a Raksasi and a cat; he belongs to the lowest class of men.” The dakinis here are depicted in a negative light, and are particularly associated with meat eating.22 This negative portrayal was evidently still
widespread when the MAC was composed, in which they are portrayed as dangerous and heretical entities in need of reform.23 There is another early reference to a class of texts known as the dakini tantras, which echoes elements of Subhakarasimha's myth. It occurs in the autocommentary to the Pramanavarttika, a text composed by the Buddhist philosopher Dharmakirti, who was active
during the late sixth and early seventh century.24 In this work he mentions a class of texts called dakini tantras, in the context of a passage addressing the issue of whether or not “success” (siddhi) in magical procedures involving mantras is dependent upon adherence to ethical norms (dharma) or not.25 His answer was “No, for it is evident that there are observances in the dakini and
bhagini tantras, etc., which are incompatible with ethical norms and are replete with vio¬lence, theft, sexual intercourse, perverse actions, and so forth, and through which there is distinctive success.”26 Alexis Sanderson has reported that the dakini tantras were texts infamous for their advocacy of ritual killing, which accords well with Subhakarasimha's portrayal of the dakinis.27 A
22 A similar association occurs in Santideva's Bodhicaryavatara, 4.4, in which the dakinis are associated both with (largely) carnivorous beasts and demons. The list occurs as follows: “Tigers, lions, elephants, bears, serpents, all enemies, and likewise
all hell guardians, dakinis, and raksasa demons.” 4.4: vyaghrah simha gaja rksah sarpah sarve ca satravah / sarve nara- kapalas ca dakinyo raksasas tatha, in P. L. Vaidya, ed. Bodhicaryavatara with Commentary (Darbhanga: Mithila Institute, 1960), 51. This text was most likely composed during the eighth century: Santideva has been tentatively dated to 658-763 CE. See Kate Crosby and Andrew Skilton, Santideva: The Bodhicaryavatara (New York: Oxford University Press, 1995), viii.
23 For an excellent comprehensive survey of the dakinis and the lore concerning them see Adelheid Herrmann-Pfandt, Dakinis: Zur Stellung und Symbolik des Weiblichen im Tantri¬schen Buddhismus (Bonn: Indica et Tibetica Verlag, 1992).
24 Toshihiko Kimura has argued that Dharmakirti lived c. 550-620 CE, contra several other attempts to date him either somewhat earlier or somewhat later. See his “A New Chronology of Dharmakirti,” in Dharmakirti's Thought and Its Impact on Indian and Tibetan Philosophy, ed. Katsura Shoryu (Wien: Verlag der Österreichischen Akademie der Wissenschaften, 1999), 209-14.
25 I interpret dharma here in its normative rather than descriptive sense, as described by Richard Gombrich in his How Buddhism Began: The Conditioned Genesis of the Early Teach¬ings (1996; repr., New Delhi: Munshiram Manoharlal, 1997), 34-37. For a similar normative use of the term see the Aganna Sutta, Dïgha Nikaya 27.7-8, translated in Maurice Walshe, The Long Discourses of the Buddha: A Translation of the Digha Nikaya (1987; repr., Boston: Wisdom Publications, 1995), 408-9.
26 My translation of the following text: na / dharmaviruddhanam api krauryasteyamaithuna- hinakarmädibahulänam vratanam dakinibhaginitantradisu darsanat / tais ca siddhivisesat. Text edited in Raniero Gnoli, The Pramanavarttikam of Dharmakïrti: The First Chapter with the Autocommentary (Roma: Istituto Italiana per il Medio ed Estremo Oriente, 1960), 163; cf. the translations in
Ronald Davidson, “The Litany of the Names of Manjusri: Text and Translation of the Manjusrinamasamgïti,” in Strickmann, Tantric and Taoist Studies in Honour of R. A. Stein, 1:8, and Alexis Sanderson, “History through Textual Criticism in the Study of Saivism, the Pancaratra and the Buddhist Yo g i n itantr a s,” in Les Source e11e temps, e d. Fran ç o i s Gri m al (Pondicherry: École française d'Extrême Orient, 2001), 11-12 n. 10.
27 See Sanderson, “History through Textual Criticism in the Study of Saivism, the Panca¬rata and the Buddhist Yoginitantras,”
12 n. 10. commentator on this text, Karnakagomin, identified these texts as follows: “In the Dakinitantras the rule of postinitiatory discipline is that one attains the power of the Mantra if one kills and devours a living creature.” It seems almost
certain that these were not Buddhist texts, despite their similarity to the Yoginitantras, which cannot be dated earlier than the late eighth century. Dharmakirti clearly identifies them as non-Buddhist teachings, as follows: “Violence, sexual union, the doctrine of the Self and so forth are explained as being the causes of bad and good results in mantra ritual texts that are
Buddhist and non-Buddhist, [respectively]. How could both be true if one is designated as being incompatible [with ethical norms]? Since there is no verdict on this matter as a ritual text of anti¬thetical import accords with the other [perspective], there is no certainty [with regard to this issue].” This passage reflects the ritual eclecticism common during this period, which was characterized by an unwillingness or inability to reject the efficacy of the practices of other religious groups, as Granoff has observed.
On the basis of these and other texts, Sanderson has argued that the da- kini tantras were Saiva texts. His claim is supported by Subhakarasimha’s account, which clearly links the dakini to the god Siva, particularly in his terrifying form Mahakala. This deity was particularly favored by extreme Saiva groups such as the Kapalikas, who were infamous for their advocacy of transgressive
practices, including violence and socially disapproved modes of sexuality. Dharmakirti, like Subhakarasimha, thus portrays the dakinis as heretical, largely on account of their propensity toward violence.
This myth bears a striking resemblance to the contents of the Buddhist Yoginitantras, a genre of literature that appeared in India no later than the mid-eighth century, and which was also referred to as dakini tantra.33 These texts were notorious for their transgressive rhetoric, and their apparent ad-vocacy of practices that would normally be prohibited in Buddhist con-texts, such as sexuality and ritual violence.
The Buddhist Yoginitantras, like the Hindu Tantric traditions to which they are closely related, appear to have originated in a distinct subculture that could be termed “the cult of the charnel ground,” consisting of anti- nomian male and female renunciants,
yogins and yoginis, who chose a de-liberately transgressive lifestyle, drawing their garb and, in part, sustenance from the liminal space of the charnel grounds that was the privileged locus for their meditative and ritual activities.34 This was a manifestation of the “siddha movement,” a pan-South Asian religious movement noted for its marginality, and its advocacy of a liminal social identity that was often symbolically associated with marginal social spaces such as the charnel ground.35 The Saiva Kapalikas constituted the best-known group of this subculture, as attested by the numerous references to them in Sanskrit literature.
33 The earliest Yoginitantra is evidently the Sarvabuddhasamayoga-dakinijalasamvara
(To. 366), a text which is described in Amoghavajra's Guidelines to the Eighteen Assemblies of the Vajrasekharasutra-yoga T. 869.18.286. c9-16) , a text composed by him following his return to China from India in 746 CE. See Rolf W. Giebel, “The Chin-kang-
ting ching yu-ch’ieh shih-pa-hui chih-kuei: An Annotated Translation,” Journal of Naritasan Institute for Buddhist Studies 18 (1995): 179-82. To my knowledge, only the term Yoginitantra is attested in Indian sources regarding this genre of Buddhist tantric
texts. How¬ever, the later Tibetan author mKhas-grub-rje (1385-1438 CE) gives dakinitantra/mkha'- gro-ma'i rgyud as an alternative designation. See Ferdinand D. Lessing and Alex Wayman, Introduction to the Buddhist Tantric Systems, 2nd ed. (Delhi: Motilal Banarsidass, 1978), 250¬51. This should not be taken as a reference to the earlier dakini tantras mentioned by Dhar- makirti, but is likely the result of the interchangeability of the terms yogini and dakini in this literature.
34 Dick Hebdige defined the term “subculture” as “cultures of conspicuous consumption. . . . It is through the distinctive rituals of consumption, through style, that the subculture at once reveals its ‘secret' identity and communicates its forbidden meanings. It is basically the way in which commodities are used in subculture which mark the subculture off from more orthodox cultural formations.” Subculture: The Meaning of Style (London: Methuen, 1979), 103. The charnel ground cult did indeed have its
distinctive style, characterized by ornamen¬tation derived from the charnel grounds (skull, bones, ash, etc.), as well as “distinctive rituals of consumption,” such as the extraction and consumption of power substances from human corpses. Hebdige's definition is also cited by Hugh Urban, who relies upon his work as well as Bourdieu's to explicate the development of a distinct
identity by the Kartabhajas in colonial Bengal. See his The Economics of Ecstasy: Tantra, Secrecy, and Power in Colonial Bengal (New York: Oxford University Press, 2001), 60ff., and esp. 241 n. 5. I have adapted Shinichi Tsuda's term “cult of the cemetery,” because “cemetery” is not an appropriate translation for smasana. This expression was first employed in his essay “A Critical Tantrism,” Memoirs of the Research Department of the Toyo Bunko 36 (1978): 167-231. See as well his article “The Cult of smasana, the Realities of Tantra,” in The Sanskrit Tradition and Tantrism, ed. Teun Goudriaan (Leiden: E. J. Brill, 1990), 96-108.
35 For an excellent survey of the siddha movement see Ronald M. Davidson, Indian Eso¬teric Buddhism: A Social History of the Tantric Movement (New York: Columbia University Press, 2002).
The Yoginitantras also focused on female deities, most particularly the yoginis and dakinis, who, like the dakinis of earlier lore,
were particularly associated with meat eating, (potentially violent) sexuality, and outcaste social groups.36 These texts are also notable for their association with “heresy,” namely, Saiva traditions such as the Kapalikas who appear to have been a significant source for these traditions. Alexis Sanderson has argued in several published articles and unpublished papers that the Yo-
ginitantras in general, and the Cakrasamvara Tantra (CST) in particular, were composed in dependence upon earlier Saiva scriptures, and he has pointed out several examples of intertextuality to support this argument.37 The CST, a Yoginitantra composed by the late eighth century,38 contains a fascinating passage that reproduces important elements of Subhakara- simha's myth. It occurs in the eleventh chapter, which reads as follows:
36 These names are interchangeable in the Yoginitantras. Some tantras, such as the Hevajra, primarily use the term yogini for the tradition's female deities, while others, such as the Cakrasamvara, primarily use the term dakini. Parallel passages in the CST and the closely related Abhidhanottara Tantra (AD), however, often attest both terms in otherwise identical passages, suggesting
that they were understood to be interchangeable. For an excellent study of the yoginis and the sexual practices associated with them see David Gordon White, Kiss of the Yogini: “Tantric Sex” in South Asian Contexts (Chicago: University of Chicago Press, 2003). In fact, the very same three groups mentioned in the La“kavatara Sutra reappear as yoginis in the Hevajra Tantra, a Buddhist
Yoginitantra dating to the late eighth century. These are the yoginis Dombi, Candali, and Pukkasi, mentioned in the Hevajra Tantra at 1.3.10. See the translations in David Snellgrove, The Hevajra: A Critical Study (London: Oxford Univer¬sity Press, 1959), 1.58, and Ch. Willemen, The Chinese Hevajratantra (Leuven: Uitgeverij Peeters, 1983), 48. Regarding the dating of the Hevajra Tantra see Snellgrove, The Hevajra, 1.11-14.
37 See his “Vajrayana: Origin and Function,” in Buddhism into the Year 2000: International Conference Proceedings (Los Angeles: Dhammakaya Foundation, 1994), 87-102, and also Sanderson, “History through Textual Criticism in the Study of Saivism, the Pañcarata and the Buddhist Yoginitantras.” There appears to be no doubt that Buddhists did draw upon Saiva sources in the
composition of the Yoginitantras, and my research on the CST generally con¬firms Sanderson's conclusions, as will be noted below. There is, however, a good deal of un¬certainty regarding the exact relationship between Saiva and Buddhist sources, and until more
textual research is completed, some of Sanderson's conclusions remain hypothetical. For a cri¬tique of aspects of Sanderson's argument see Davidson, Indian Esoteric Buddhism, 386 n. 105. Despite this uncertainty, I believe that the preponderance of evidence supports Sanderson's thesis. Among this evidence I would include the CST materials discussed in this paper, which appear to be the product of Buddhist appropriation from sources that earlier Buddhists, such as Subhakarasimha, would have identified as Saiva.
38 A provisional terminus ante quem for the CST is provided by quotations from it in a dat¬able commentary, Vilasavajra's Arya-Namasamgititika Mantrarthavalokini-nama (To. 2533), Vilasavajra having lived in the second half of the eighth century. See Davidson, “The Litany of the Names of Mañjusri” 6-7. Actually, the majority of cases that Davidson notes as cita¬tions from the
Laghusamvara Tantra (an alternate name for the CST), the Sarvabuddhasama- yoga-dakinijalasamvara (To. 366) is actually cited. In Indian texts of this period, unlike later Tibetan texts, citations such as samvaratantre or samvare invariably refer to the latter text. There is one case, however, in which Vilasavajra quotes the CST, chap. 2, and another case in which he refers to a passage in chap.
48 of this text. See my The Discourse of Sri Heruka: A Study and Annotated Translation of the Cakrasamvara Tantra (New York: Amerian Institute of Buddhist Studies, forthcoming) for a more detailed discussion of this evidence.
Now above all I will speak of the power that the adept should attain, through which there is rapid engagement in power by means of eating only. The person who goes perspiring a pleasant fragrance, speaking the truth, blinks after a long time, is not angry, and who has fragrant breath in his mouth, is one who is born as a man for seven lives. Splitting him there is the concretion in his
heart. Taking this makes a drop with one hundred repetitions of Sri Heruka's Essence Mantra. One will fly up and travel tens of millions of leagues. Just through eating [it] one will become one who has knowledge of the three worlds. One will travel five hundred million [leagues] in a day and a night, and will have a divine body. Who¬ever knows Sri Heruka's Essence will be given whatever things he desires.
A systematic comparison of the two texts is in order. Both the MAC and the CST are texts primarily concerned with the ritual uses of mantra, the dakini mantra and Sriheruka's Essence (hrdaya) and Quintessence (upahr- daya) mantras, respectively. Subhakarasimha's MAC claims that there are dakinis who are associated with Siva or Mahesvara in his Mahakala form, and who seek to
extract concretion from the hearts of certain people. Evi-dently due to serious competition over the valuable commodity contained in their hearts, they would extract their hearts six months prior to their death and replace it with a magical simacrulum. The text
does not specify who these people are, but merely tells us that the dakini had a way to identify them. Mahavairocana's forceful intervention, however, led them to renounce this reprehensible practice, and promise that they would only consume the hearts of deceased individuals.
The CST is more specific regarding the apparent victim or bearer of this commodity, the rocana or concretion. It is a person who has been born as a man for seven times in a row. The text lists five characteristics whereby such individuals may be identified, drawing upon the ancient Indian beliefs concerning the marks of divinity, which suggests that the person “born as a man for seven
lives” enjoys semidivine status. The CST also provides a similar ritual means for identifying them. In the MAC myth, the dakini are provided with a mantra, hri hah, which enables the dakini to predict the time of death of the concretion bearers. A close variant of this mantra is contained within Heruka's Quintessence mantra, which is recited over the drop of concretion (rocana) to
effect the miraculous powers. According to CST chapter 12, which lists the various powers attainable through the application of the Quintessence mantra, “Enchanting water with this, if one rinses one's eyes [with it] one will recognize the one born seven times. Those whose life is exhausted will appear as if dead; the long-lived will appear to be full of life.” Thus the CST asserts that its Quintessence mantra will confer the very power promised by Mahavairocana Buddha to the dakinis, namely, the ability to identify the bearers of the precious concretion and accurately predict their life span.
As in the case of the MAC, the concretion is located in the heart, and is to be attained through the action of “splitting open,” presumably, the victim's chest cavity. The text is not specific concerning the conditions under which this violent act is undertaken. Both texts describe the powers that can be attained by consuming this substance. These are quite similar, as both texts promise the power to travel immense distances in one day, as well as the power to acquire whatever one desires.
There is, however, a major difference between these texts. In the MAC, the ritual appropriation of this concretion is represented as the reprehen¬sible behavior of the heretical Saiva dakinis, whose nefarious behavior triggers Mahavairocana's intervention. This myth is a variant of an estab¬lished genre of Buddhist conversion narratives, in which Buddhist deities assume the guise of non-
Buddhist deities, subjugate them and convert them into subordinate deities. The story of Vajrapani's conversion of Mahadeva in the Tattvasamgraha Sutra is a well-known example of this genre. There is likewise a Yoginitantra version of this narrative, which holds that Heruka and his retinue manifested in the world in Saiva guise in order to subjugate the Hindu deity Bhairava and his retinue and put an end to their “evil conduct,” which involved both violence and wanton sexuality.
These myths played an important role in Buddhist discourse, namely, as justification for Buddhist appropriation of elements of non-Buddhist traditions, in this case the Saiva-Kapalika that evidently was an impor¬tant source for Yoginitantras such as the CST. Such appropriation, how¬ever, is typically a process “wherein the borrowed item is transformed through the process of
incorporation, thus fundamentally altering both the appropriated and the appropriator.” Appropriation is not a feature unique to Tantric Buddhism, but rather was typical of Buddhism throughout its history. For example, while Buddhists opposed elements of the Vedic rit¬ual lore, particularly the practice of violent animal sacrifice, they actively appropriated other elements of Vedic
ritual lore, such as the homa fire sacrifice. They also advanced a thoroughly transformed “version” of the Vedic sacrifice, the “bloodless sacrifice” or Mahadana rite of ceremonial gift giving. Refraining from the ritual killing of animals has often been seen as a key marker of Buddhist identity, in contradistinction to one con-struction of alterity, which portrays “heresy” precisely in terms of such behavior. the limit of tantric buddhist discourse
The “ritual eclecticism” exhibited in texts such as the Mahavairocana- abhisambodhi Tantra and the CST was not uncontested; the appropriation of non-Buddhist practice elements could and did lead to the Buddhist iden-tity of these texts being challenged. The author or authors of the former tantra apparently expected such criticism and thus inserted into the text the following prophecy:
In the future there will appear faithless beings with little intelligence, who not believing this teaching will have great misgivings and doubt, and who will just hear it, neither retaining it in their hearts nor accomplishing it. They themselves are unsuitable and they corrupt others. They will say, “This is not what was spoken by the Buddhas, but it belongs to the non-
Buddhists!” But these foolish people do not know that the Bhagavat, the All-knowing One who has attained mastery over all phenomena, who has directly understood what benefits beings, has said: “I shall explain all of these things,” having previously [vowed to] help beings.
There evidently was considerable resistance to ritual eclecticism. A number of strategies were advanced to overcome this resistance, including pro-pounding myths of the conversion of non-Buddhist deities, as well as put-ting forth the claim that Buddhist figures taught ostensibly non-Buddhist ritual elements, such as the dakini mantra, out of a compassionate moti-vation. The
concept of “skillful means” (upaya), a strategy for inducing heretics to enter the Buddhist path, was frequently deployed as a means of neutralizing the tension triggered by signs of heteropraxy, particularly the presence of the transgressive rhetoric advocating ritual violence or sexuality that was quite common in the Buddhist tantras.52 Anandagarba, a ninth-century Buddhist
commentator,53 attempted to justify the violent and erotic rhetoric in the Guhyasamaja Tantra by claiming that it was taught for the purpose of converting to Buddhism54 those “low born ones who are opposed to the other tantras and who are inclined toward malicious deeds, who have the karmic obstruction of the inexorable sins,55 and so forth, who adhere to the teachings in the Visnu,
the dakini and deviant tantras,56 who kill, who do not give but take, who tell lies, and who ‘practice' with their mothers and daughters and who enjoy both suitable and unsuitable foods.”57 Likewise, the Tibetan polymath Bu-ston (1290-1364 CE), claimed that the Yoginitantras or “mother tantras” (ma rgyud ) were taught “for the sake of training women.” Shortly after making this claim, he continued, writing that “father tantras are [taught] so that men of one's own class who have unerring views can attain Awakening; mother tantras
52 The concept of upaya was long invoked by Buddhist authors as means of dismissing as “interpretable” (neyartha) doctrinal teachings that do not accord with their own views. The “interpretable” text is thus explained as an instance of upaya on the part of the Buddha, em¬ployed as a strategy for the conversion of those with “wrong views.” On this issue see Donald S. Lopez, “On the
Interpretation of Mahayana Sutras,” in Buddhist Hermeneutics, ed. Donald Lopez (Honolulu: University of Hawaii Press, 1988), 47-70. Note, however, that the concept of upaya in Buddhist polemic only serves Buddhist apologetic needs and was probably never convincing to the authors' opponents. See Jamie Hubbard, Absolute Delusion, Perfect Buddha- hood: The Rise and Fall of a Chinese Heresy (Honolulu: University of Hawaii Press, 2001), 241-42.
53 Taranatha places Anandagarbha during the reign of King *Mahipala, who died, accord¬ing to him, at the same time as the Tibetan King Ral-pa-can (d. 838 CE). See Lama Chimpa and Alaka Chattopadhyaya, trans., Taranatha's History of Buddhism in India
(1970; repr. Delhi: Motilal Banarsidass, 1990), 284. Taranatha appears to confuse the Pala succession here, as the first Mahipala did not rule until much later, c. 992-1042 CE (Davidson, Indian Esoteric Buddhism, 52). The Pala king who did rule during this period was Devapala (c. 812-50 CE), during whose reign we might tentatively place Anandagarbha. Lessing and Wayman place him during the tenth century, but do not state their reasons for doing so (Introduction to the Buddhist Tantric Systems, 24).
54 That is, cause them to take refuge in the Three Jewels, generate the Spirit of Awakening and enter into the mandala.
55 The pañcanatariya, which are (1) killing one's father, (2) killing one's mother, (3) killing an arhat, (4) drawing the blood of a buddha with ill intent, and (5) causing a schism in the samgha.
56 It is not clear what sort of texts Anandagarba meant when he referred to dakini tantras. It is clear that he is referring to texts that he considered heretical, and it is possible that, like Dharmakirti, he was referring to a now lost genre of Saiva text. But it is also possible that he is referring to the Buddhist Yoginitantras, which were composed by this time, but whose ortho¬doxy was questioned by many Buddhists.
57 Guhyasamajamahatantrarajatika, To. 1917, D rgyud ‘grel vol. bi, fol. 2a. [were taught] in order to train outsiders who delight in killing and so forth, and who adhere to erroneous views and spiritual paths.” He thus portrays the Cakrasamvara and related tantras as representing a Buddhist strategy to facilitate the conversion of heretics, with whom he also associated women.
This strategy was apparently not always sufficient to assuage the doubts of Buddhists. As a result, Buddhists have often attempted to reencode sus-pect entities or practices. As Bernard Faure has observed, Buddhists have repeatedly “felt compelled to convert or subdue the local deities, to erase the memory of the places, to reconvert or desacralize spaces, to decide and re-encode legends.”
In mythic language this complex process of conver-sion, characterized by subdual, erasure, and reencoding, can be symbolized by the trope of consumption and digestion. Even the heretical dakini can be assimilated following their “digestion” by Mahavairocana. Hence the presence of seemingly “raw,” heretical elements in an Esoteric Buddhist text would not ordinarily pose an insurmountable
problem for ingenious commentators. The quasi-heretical female deities of the CST were thus accordingly subjected to a “digestion” or reencoding through their corre-lation to normative Buddhist categories, perhaps because the myth of Heruka's subjugation of Bhairava was insufficient to assuage doubt con-cerning their appearance in Saiva garb.
Such strategies collectively constitute an important component of Tan-tric Buddhist discourse, the development of which is particularly notable during the ninth century. This usually seamless integration of Buddhist and non-Buddhist elements is made
possible by the central elements of Tantric Buddhist discourse. These include the use of radically transgressive rhetoric, as well as claims of the need for secrecy, namely, the symbolic interpretation of this rhetoric, and exhortations to avoid literal interpreta-tion, namely, the naive performance of the acts implied by the rhetoric. The resulting tension is a central element in Tantric traditions and a corner-stone of their political strategy.
Despite the deployment of these strategies, the CST seems to have faced considerable resistance even among those who would be most expected to advocate it, its commentators. Commentators on the CST exhibited typical exegetical ingenuity in their efforts to
reencode it as a bone fide Buddhist text, collectively deploying all of the methods discussed above. Yet the eleventh chapter seems to have presented a significant challenge to this ingenuity. As such, the chapter does not represent a normative example of Tantric Buddhist discourse, but instead represents its limit, the extreme beyond which it could not function. As Charles Orzech pointed out with regard to the development of Esoteric Buddhism in China during this pe-riod, “The negotiation that is part of a living,
complex, and changing tra-dition can often be glimpsed in the disjunctions or seams where divergent meanings are stitched together to respond to the necessities of life. By being attentive to these seams and to the underlying paradigms of a re¬ligious tradition we can deepen our understanding of religion in changing social and cultural contexts.” I believe that texts such as the MAC and
CST highlight one such seam, in which competing and possibly incompatible paradigms collided. They point to a limit of Tantric Buddhist discourse, a limit that can be ascertained along the lines outlined by Foucault, that is, by limits to a discursive formation's forms of expressibility, conservation, memory, reactivation, and appropriation. Of crucial interest here are the points at which erasure and decoding occur during the “conversion” pro¬cess wherein appropriated texts and practices are
accommodated within Buddhist discourse. The limit of this discourse represents the point beyond which it could not safely proceed, the point at which it was vulnerable to being labeled questionable, invalid, foreign, and subject to active censor-ship or passive
erasure from memory. The CST seems to embody this limit precisely because it contains, in a rather raw and unprocessed form, the very elements that, in myths such as those accounting for the origin of both the dakini mantra and the deity Heruka himself, are represented as pertaining to the heretical other and hence are suspect.
The CST as a whole is thoroughly steeped in the charnel ground milieu of the Yoginitantras, a context in which the ritual use of corpse-derived products is de rigueur. Typically, its text is ambiguous enough to permit alternate readings (or creative commentatorial misreadings) of transgres-sive passages. This is also the case in chapter 11, which is unclear re¬garding both the
means of acquiring and also the use of the concretion or rocana. Probably the most scandalous reading is that the rocana is acquired via an act of sacrifice or ritual murder. This is in fact suggested by the earliest and most conservative commentator Jayabhadra, who was active during the mid-ninth century. Jayabhadra characterized the “per¬son” ( purusa ) who bears the concretion as a “sacrificial victim” ( pasu). That “sacrificial victim,” rather than “beast,” is meant here is clear from the fact
that the text identifies him as a “person” ( purusa ). The term purusa itself has a sacrificial connotation deriving from the famous Purusasukta hymn.68 While other denotations of the term pasu might be implied here, such as the Saiva technical sense of an “uninitiated person,” the denotation of “sacrificial victim” is strengthened by a passage elsewhere in the CST, which, echoing the Vedas, lists five suitable candidates for sacrifice ( pan- capasu), one of which includes “man” (manusa).69 Another commenta¬tor,
Bhavabhatta, comments here that from among the human social classes, the Brahmin is the sacrificial victim.70 This is not to say that the CST advocates human sacrifice, a practice that would be completely prohibited in a normative Buddhist context.71 Nor do any of the commentators state this, despite their use of the suggestive but ambiguous term pasu. The full context of this text, however, suggests the possibility of this interpretation; this was, after all, the “heretical practice” that stimulated Heruka's assumption of a Saiva disguise, and his conver¬sion of its non-Buddhist practitioners.
68 Rg Veda 10.90.
69 See Hélène Brunner's “Jnana and Kriya: Relation between Theory and Practice in the Saivagamas,” in Ritual and Speculation in Early Tantrism: Studies in Honor of André Pa¬doux, ed. Teun Goudriaan (Albany: SUNY Press, 1992), 27. The term pasu in the sense
of an uninitiated (and hence ignorant) person also occurs in at least one Buddhist text, chap. 1 of the Dakarnava Tantra (a CST explanatory tantra), as follows: “The Lord Yogin manifests instantaneously in a divine form, assuming a transformed shape through
the yoga of a cast image, etc. [This is] the yoga of form, the self-nature of which is emptiness, the defining mark of self-consecration (svadhisthana). As for the yogin who lacks self-consecration, know him to be like a heap of chaff. This sort of supreme characteristic is not known by men who are beasts (pasu, phyugs).” To. 372, D rgyud-'bum, vol. kha, 139b. Regarding the
“five sacrificial victims,” the Satapatha Brahmana lists include the man, horse, bull, ram, and he-goat; see Brian Smith, Classifying the Universe: The Ancient Indian Varna System and the Origins of Caste (New York: Oxford University Press, 1994), 250-
51. The CST list is anomalous, as it invokes the Vedic category of the pancapasu, but lists in fact six items in its chap. 32, as follows: “The sacrificial victims are thus five, with the donkey, the man, tortoise, camel, jackal and horse, and so forth” (pasavas ca tatha panca kharmanusakUrmostrasrgala- hayadibhih; my edition, cf. Pandey, Sriherukabhidhanam Cakrasamvaratantram,
2.519). Only two of those listed overlap with the Vedic list. It brings to mind another anomalous list, that contained in the Kalika Purana. See Hugh Urban, “The Path of Power: Impurity, King¬ship, and Sacrifice in Assamese Tantra,” Journal of the American Academy of Religion 69, no. 4 (2001): 799-800.
70 Bhavabhatta's name is usually represented in Tibetan as Bhavabhadra. The Sanskrit mss. of his commentary, however, give Bhavabhatta. According to Taranatha (as cited above), he was the fifth Tantric preceptor of Vikramasila, which suggests that he may
have been active c. 900 CE. This occurs in Bhavabhatta's Cakrasamvaravivrtti. In his commentary on a line in CST, chap. 32, varnanam varnatah pasuh, he writes that “among humans, from the class, i.e., from among the [social] classes, the brahmin is the sacrificial victim” (Pandey, Sriherukabhidhanam Cakrasamvaratantram, 520: manusyanam madhye varnato varnebhyo brahmanah pasuh).
71 Indeed, it is not possible for us to know exactly what practices CST, chap. 11 implies, or whether such practices were actually performed in India or elsewhere in the Buddhist world. The commentators give us evidence regarding how it was understood in Buddhist monastic institutions. As will be seen, these sources indicate that the text and its practice tradition were gradually sanitized, such that transgressive practices were neutralized, either via outright erasure or through “sublimation,” in which such
ritual practices were reduced to internalized visualization exercises, devoid of actual transgressive impact. The association of this passage in the CST with the reek of heresy trig-gered various commentatorial responses. Bhavabhatta, in his commen¬tary on chapter 11, provides an interpretation of the text that removes the possibility of ritual violence. He begins by
giving a different reading of the first verse, emending the text's claim that power is attained “by means of eating only” (prasitamatrena). Instead, he reads “through that which is scented only (*ghranitamatrena), namely, scent only (ghranamatrena).” This interpretation seems to accord with the tantra in that exuding a pleasant scent is one of the signs for identifying the bearers of rocana. Nor is it necessary to “split him” (tam bhaksayitva). According to Bhava- bhatta, one need only “smell him” (tam ghranayitva) to achieve the de-sired powers,
which appears to eliminate the need for both ritual violence and transgressive oral consumption of human bodily substances. Bhavabhatta, however, stands alone in advancing this novel commenta- torial solution to the text's troubling advocacy of ritual violence. However, there was another attempt to emend the text's call for anthropophagy. The Tibetan translations, in place of “by means of eating only,” read “through service only” (bsten pa tsam gyis), which likely represents another attempt at emendation or commentatorial misreading. What might the claim that power is achieved “through service only” mean? Again the commentaries provide an answer. *Viravajra wrote in one of his commentaries that
Recognizing these signs of [one born as a man for] seven lives, one should serve him with reverence so long as he lives, and one should pray “May I attain my power (siddhi) when he dies.” As soon as he dies, one should take the concretion that is in his heart, recite one hundred times the appropriate [mantras] such as Sriheruka's Essence and Quintessence. If one forms it into a drop (tilaka) on one's forehead, one will soar into the sky and travel ten million leagues. If one forms it into a drop over one's heart, one will know other's minds, [etc.].
Evidently, this service is due to the one born seven times as a man, once he has been successfully identified. A Tibetan exegete, Sachen Kun-dga' sNying-po (1092-1158 CE), reported that the “pandits” claim that those born seven times a man are bodhisattvas, who, if respectfully asked by an adept, will surrender their bodies. This is a claim that does have a basis in the Indian textual tradition.
Bhavabhatta's creative misreading aside, the Cakrasamvara corpus clearly points to the need to consume the rocana. These readings accord with the myth in the MAC, where the concretion is consumed by the da- kinis. It is also consistent with the alternate reading that power is attained through “service” of the spiritually advanced beings known as “bodhisat-tvas” in Buddhist
literature. This evidence points to the widespread Indian belief that bodies are literally transformed through spiritual practice and that the consumption of the flesh of certain spiritually advanced beings can result in the empowerment of the consumer. Here one might point out the eighth chapter of Santideva's Training Compendium (siksasamuccaya), which begins with a lengthy quote from the Aryatathagataguhyaka Sutra describing certain bodhisattvas who create the aspiration that the beings who consume their flesh upon their death in a charnel ground will attain rebirth as gods in the heavens, or even parinirvana itself.
Are these explanations sufficient to remove the taint of heresy? Appar-ently not, for some Buddhists at least. While Mahavairocana's conversion of the dakinis from the violent practice of heart theft to the (relatively) nonviolent practice of postmortem anthropophagy is a positive one, it is not a conversion to normative Buddhist practice. Viravajra's and Sachen's explanations evoke the passage in Santideva's Siksasamuccaya, although this is hardly the intent of that passage, which concerns bodhisattvas' giving up their bodies to feed and thereby save the carnivorous animals and which derives from well-known Jataka tales. Nor would willfully causing the death of a bodhisattva be condoned in normative Mahayana Buddhism; it would put the
practitioner in the position usually reserved in the legends for the antagonist of the bodhisattva who tests his or her resolve. The presence of “raw” non-Buddhist elements in the CST was an on-going problem with which the commentators were forced to contend, although these were gradually eliminated via emendations that rendered the text more legitimate in the eyes of skeptical Buddhists. It is thus not surprising that another commentator on this text, *Bhavyakirti, seems suspicious of this chapter and reports on what must have been a contro-versy in Indian Buddhist circles regarding
the legitimacy of Tantras such as this one as authentic discourses of a Buddha (buddhavacana). He wrote that With regard to taking the concretion (rocana) it is not merely taking his con¬cretion. This is the explanation of heretics. Is this not suitable to be taught in a yoginitantra? This is not the case, however, since there are instructions to eat medicinal
substances such as cow products in all of the yogatantras such as the Sri-Guhyasamaja, as well as in those of the heretics. Since these involve taking it from a corpse, it is not heretical? How are the yogatantras and yoginitantras different with regard to the explanation that the rocana is taken by oneself from the corpse of one born seven times a man? Someone [says that] the yogatantras
were taught by the Buddha, and the yoginitantras were taught by Mara. I myself am unclear about this. However, it should be understood in accordance with the following explanation: “When two things have the same fault, and when both have the same fault in meaning, in investigating the meaning in such a case, it is not suitable to settle on one of them.”
There is little doubt regarding the non-Buddhist “heretics” to whom he is referring here. They are undoubtedly the Saivas who are represented as the heretical other in the Buddhist origin myths. And it seems likely that he refers to the Saivas particularly focused on fierce deities such as Mahakala and Bhairava, such as the Kapalikas, who were notorious for their prac¬tice of ritual
violence, and not only in Buddhist circles. Indeed, the expla¬nations of *Viravajra and Sachen evoke the stories told about Kapalikas in Hindu polemical literature, such as that related by Madhavacarya in his Samkara Digvijaya concerning the Kapalika Ugrabhairava, who sought the head of an omniscient sage to roast in his sacrificial fire in order to obtain the ultimate siddhi. The Kapalikas have been characterized as having been engaged in a pursuit of power, one that often involved the trans¬gression of social mores and rules of purity, and it seems difficult to characterize this chapter of the CST in any other way than this.
- Bhavyakirti's unwillingness to accept this text as Buddhist is telling and indicates that, for him at least, this text tested the limit of the appro¬priation process. But he was not the only one to harbor such doubts. His invocation of the figure of Mara, the classic Buddhist “evil one” who was believed to zealously strive to lead Buddhists astray, suggests that his doubt reflected a
controversy within the Indian Buddhist community over the orthodoxy of texts such as the CST. There is evidence that the Buddhist tantras in general, and particularly the transgressive Yoginitantras, were re¬sisted by Buddhists adhering to the more conservative Nikaya traditions such as the Theravada. According to the Tibetan polymath Taranatha (1575-1634 CE), “In the temple of Vajrasana there was a large silver image of Heruka and many treatises on mantra. Some of the saindhava and simhala sravakas said that these were composed by Mara. So they burnt these [texts] and smashed the image into pieces and used the pieces as ordinary money.”
Despite such resistance, tantras such as the CST were ultimately accepted in some Indian Buddhist circles, and thence transmitted to Nepal, Tibet, and Central and East Asia. But this acceptance appears to have been contingent upon the transformation of both the
text and its practice tradition, such that the more transgressive and “heretical” aspects of the text were erased or hermeneutically neutralized. Likewise, in the realm of practice, there is clear evidence of a movement away from the actual performance of transgressive practices toward sanitized versions of these practices, in which transgressive elements are either symbolically represented or internally visualized.
Despite the controversy that this myth of anthropophagic dakinis and their concomitant rituals seems to have inspired, these narrative and ritual elements do not seem to have been particularly well remembered or con-served in South Asia, Tibet, or East
Asia. In the latter case, it lived on perhaps only in traces manifesting, for example, in the later dakini cult in Japan. In Nepal and Tibet, where the texts were preserved, the actual practices that may once have been associated with them were not preserved to my knowledge, that is presuming that they were ever even transmitted in the first place, manifesting beyond mere transgressive rhetoric.
This seam and limit reflect, among other things, the fractured and flex-ible nature of Buddhist identity. Religious identities are never monolithic and fixed but are fluid; in the language of James Clifford, they are “con-junctural,” namely, relational, and subject to continual negotiation and re-negotiation. Myth and ritual are simply modes of discourse that are not fixed in their
application, but which in fact can be deployed to serve any number of ideological purposes. As such, they are key elements in the processes by which religious communities construct and reconstruct their identities. Identity, as Ger Duijzing has argued, “represents primarily a so-cial bond between the individual and a collectivity or community, which may vary according to time or place, and may be accepted or contested by both insiders and outsiders.” Buddhism in eighth-century India was not monolithic, and
the attempt to reforge an identity, traces of which lin¬ger in the texts examined here, was contested from within the Buddhist community, and possibly from without as well. The very formation of a distinctly Tantric Buddhist discourse probably arose from such multiple pressures, traces of which are inscribed within the bodies of the texts. It was these multiple pressures that shaped Esoteric Buddhist discourse in India, and that continued to reshape it as it was transmitted and took root in East Asia and Tibet.
Santa Clara University
