There are no truths outside the Gates of Eden.
—BOB DYLAN
Although possession as a religious and sociocultural phenomenon in ancient and classical India has been understudied, in fact rarely mentioned in the secondary literature (to a great extent because it has been regarded as absent in the primary literature),1 the opposite is the case for modern India, for which studies of possession abound. Indeed, it is no exaggeration to say that possession is one of the most studied topics in anthropology and is beginning to find its way into the academic study of religion.2 Its attractiveness as a subject of study is often due to its exotic character (which is to say its spiritual distance from academe), its evident irrational Otherness, and its frequent presence in ritual. Caroline Humphrey and James Laidlaw, who have studied possession and exorcism among largely educated middle-class people at the Padampura Jain temple near Jaipur, assert, “Ethnographers have displayed an almost puzzle-solving pleasure in ‘decoding’ rituals: the more bizarre and serpentine the better. Behind this lies the idea that rituals contain hidden messages about the societies in which they are performed.”3 This, they believe, is the consuming passion of anthropologists; yet they see the scholar’s analysis of ritual (especially that in which possession is contextualized) as usually neglecting the participant’s view. This emic viewpoint is something I attempt to engage in this volume, as a parallel discourse to the etic viewpoints that constitute and regulate contemporary academics (though one may argue that the practice of an etic discourse in thinking about the Other is simultaneously the practice of an emic discourse within the tribe of scholars).4 Conferring at least limited credibility to emic viewpoints is, in my experience of contemporary scholarship (in opposition to Humphrey and Laidlaw’s observation), finding an increasingly prominent place in more recent studies of possession. Furthermore, in defense of those of us who study possession (including Humphrey and Laidlaw), it must be said that public performance of possession is nearly always thrilling and exotic for those who are enculturated. Thus anthropologists and other observers and researchers must not be faulted too severely for yielding to that same odd attractiveness, for succumbing to the spell of possession.5
In order to comprehend possession as presented in Sanskrit texts, it is important to understand the use of the English word and its background in Western thought, especially in anthropological analysis. However, it is neither possible nor desirable to provide a full accounting of the history of experiential states identified as “possession.” Nevertheless, a summary, particularly of possession reports in the Indian ethnographic literature, will prove helpful.6 Nevertheless, I do not intend to write an account of possession in Sanskrit texts according to any particular anthropological paradigm. I am not scrutinizing a broad range of Sanskrit and other classical Indian texts in order to historically validate any single anthropological model or with the specific intention of applying any of them exclusively or strictly to the evidence for possession derived from these texts. But the insights on possession gained by anthropologists and other ethnographers will be useful in evaluating the evidence adduced here—and I shall not hesitate to call on these insights when necessary or convenient.
In this section (Chapters 2–4) I first look at definitions of possession found in anthropological literature, especially in the work of scholars of South Asia. Then I examine a number of intellectual perspectives in which anthropologists and others situate possession. After considering these, I attempt a broad characterization of possession in South Asia according to the anthropological literature (Chapter 2). With this in mind, I consider, as an interlude, the distinction between deity or spirit possession in South Asia and trance channeling in contemporary America (Chapter 3). Then, I enter into a lengthy discussion of the vocabulary and general diction of possession and allied forms of spirit-mediumship in several South Asian languages and linguistic and cultural areas that have been the subject of studies of possession (Chapter 4). I conclude the chapter with a reflection on the relationship between “folk possession” and “classical possession.”
The definition of possession cited earlier by Boddy in her review article on the literature of possession has the advantage of both conceptual precision and elasticity. This definition bears repeating. Possession, Boddy says, is
a broad term referring to an integration of spirit and matter, force, or power and corporeal reality, in a cosmos where the boundaries between an individual and her environment are acknowledged to be permeable, flexibly drawn, or at least negotiable.7
Erika Bourguignon notes a curious statistic: Out of 488 societies examined, 360 of them (74 percent) showed evidence of possession belief and 251 (52 percent) evinced possession trance.8 She defines possession accordingly. After stating that possession “is an idea, a concept, a belief which serves to interpret behavior,” she says:
We shall state that a belief in possession exists, when the people in question hold that a given person is changed in some way through the presence in or on him of a spirit entity or power, other than his own personality, soul, self, or the like. We shall say that possession trance exists in a given society when we find that there is such a belief in possession and that it is used to account for alterations or discontinuity in consciousness, awareness, personality, or other aspects of psychological functioning.9
In addition, we may cite Rex L. Jones, who studied spirit possession in Nepal:
Spirit possession can be defined as an altered state of consciousness on the part of an individual as a result of what is perceived or believed to be the incorporation of an alien form with vital and spiritual attributes, e.g. the spirit of a superhuman form such as a witch, sorcerer, god, goddess, or other religious divinity.10
Carl Becker provides broad leeway in defining possession as:
the phenomenon in which persons suddenly and inexplicably lose their normal set of memories, mental dispositions and skills, and exhibit entirely new and different sets of memories, dispositions, and skills.11
By simply defining possession as a type of behavior, he avoids (perhaps too cautiously) the knotty problem of an individual being taken over by another (presumably discarnate) being or conscious entity. This broad definition should become useful in dealing with the broad category of possession in South Asia. Bearing in mind that these definitions are offered by anthropologists rather than classicists or historians of religion and that they have different emphases—possession as an integrated but fluid psychophysiological state, as an interpretative idea, and as an altered state of consciousness induced by an external agent—they are good starting points.12
Historically, the study of possession has been a Eurocentric concern, mostly by white men studying an experience of mostly nonwhite men and women. Certain fallacies incumbent in this enterprise in early studies of possession, which are apparent to the turn-of-the-millennium eye (but which continue to the present in many cases), were summarized by W. G. Jilek in 1971 in his analysis of the “sick medicine-man” thesis. The fallacies of this thesis, which can be applied to the “possession thesis,” are that the very notion of the “sick medicine-man” is:
(1) Eurocentric, insofar as it tries to explain descriptive data collected in non-Western cultures by Western scientific theories, ignoring the non-Western folk systems of explanation and overlooking Western folk systems analogous to shamanic institutions; (2) Positivistic insofar as it considers behaviour which does not fit into the framework of logico-experimental explanatory theories … as departure from rational norms due to “ignorance and error,” under which is subsumed the “poor reality testing” of psychopathology.13
Although the early British ethnographers who mapped out the “cultures, artifacts, and peoples” of the British Raj were unabashedly Eurocentric and positivistic, these tendencies have been in decline with recent methodological advances in anthropological research. An increasing number of anthropological studies of possession, for example, those by Peter Claus, Edith Turner, Smriti Srinivas, and Daniel Halperin, all discussed below, question the anthropologist’s presumed superior understanding of possession—superior, that is, to the understanding of the people under study.14 A substantial number of anthropologists and folklorists, such as Humphrey and Laidlaw, Claus, Erndl, Caldwell, and Sax, understand the issues at stake here and address them cogently; and, among them, Claus directly addresses the central problematic.
Echoing a theme I note often in this study, Claus contends that the results of anthropological studies of possession have been guided by preselected models, which in most cases means that the method determines the madness at the expense of fair and adequate representation of the cultural context. Although very different from the early British ethnographies, the methodological assumptions and approaches are now much more varied and abstract. An example, discussed more fully below, of the overdetermined nature of many modern studies is revealed in Richard Castillo’s indictment of Gananath Obeyesekere and others whom Castillo believes forced their evidence to fit their theories. In the case of Obeyesekere in particular, Castillo tries to show that his assertion of repressed oedipal desire as the primary causal factor in possession is completely mistaken—that Obeyesekere’s data cannot possibly support this claim. Claus’s point bears on the problematic of the one-way discourse of many studies such as Obeyesekere’s (which, nonetheless, may still be valuable and illuminating). He states, “I doubt that any serious anthropological study gives credence to the native claim that they are possessed by spirits, although we do sometimes acknowledge that we can proceed with the study ‘as if there were spirits.’”15 Fortunately, this attitude of grudging acknowledgment is an increasing trend, which opens the discussion considerably and allows for more nuanced interpretations. Halperin states, along the same lines:
While externally formulated, or “etic,” approaches normally consider possession from the perspective of the human individual or society, the practitioner’s own worldview is nearly always an actively spirit-centered one. Ritualists typically perceive the essence of possession rites not as entering a trance but as being entered by spirit(s), by a force clearly outside themselves.16
As shown below, many recent studies are not exclusively concerned with constructing their own theories of possession or justifying fully formed theories that they have adopted, but also consider the implications of the emic viewpoint for the people under study. This is not to criticize any particular theories or the enterprise of theorizing, but to say that theorizing can be much more fruitful when granting those under study the facticity of their viewpoints.
Studies of possession commonly (and correctly) ask one or more of three questions: What is possession? How does it work? What does it mean? In spite of several cogent definitions, such as those noted above, there is little agreement on what possession actually is. How it works is also not well understood or generally agreed upon. The greatest disagreement, however, is over the last question, which most researchers, curiously, feel qualified to address even as they readily acknowledge ignorance in relation to the first two questions. This is reminiscent of physics, in which the existence and role of black holes can be postulated and surmised, yet “from within which,” declares the science writer Timothy Ferris, “one may rest assured, no reporter will ever file dispatches.”17
Since the earliest serious field studies of possession about a century ago, a considerable amount of work on the subject has been carried out, most of which views possession as “a symbolic expression of other experiences.”18 This characteristic of anthropological interpretations strives to unearth psychological, medical, or social causes of possession. This investigative tendency, Elizabeth Schoembucher maintains, is due to the Western social scientists’ insistence on rationality and “search for a deeper meaning.”19 This has resulted in a series of different interpretations, in which the fact that possession is seen as an ecstatic religious experience by the people themselves is intellectually indefensible, an academic inheritance from the Protestant-dominated cultures of anthropology and the study of religion, which subtly yet pervasively establish hierarchies subordinating sorcery, witchcraft, and “natural” religion to higher “revealed religion.”20 This disregard for indigenous voices and interpretations is noted by Humphrey and Laidlaw, who, not surprisingly, reject the interpretive strategies of most of their contemporaries. They contend that most ethnographers read ritual, a category in which possession most often falls, as “an escape for the ethnographer from the bewildering and dispiriting prospect of thinking about multifarious individuals.”21 One of the features of the present study, I hope, is to permit the texts, our current voices under examination, to speak for themselves, to admit at least some of their own light by allowing their own interpretations to arise more or less unhindered, given the limitations I fully recognize in the oft-stated caveat that every act of reading or translating is an act of interpretation. Another objective is to avoid reifying or “orientalizing” possession—a possibility against which Arjun Appadurai warns in no uncertain terms when he speaks of “the unruly body of the colonial subject (fasting, feasting, hook swinging, abluting, burning, and bleeding)” as a creation of the “exoticizing gaze” of orientalism.22
Possession has usually been examined from the perspective of one of five interpretative frameworks: (1) as demonic, opposed to God and good; (2) as a medically defined psychological state; (3) as a psychological condition engineered for the purpose of gaining social or even political control; (4) as an aspect of shamanism; and (5) as an existential reality. I deal with each of these in turn.
Because possession is deeply embedded in the complexities of local culture and ritual, it is important to exercise caution in formulating general theories about it.23 Although it is not my intention to present a general theory, it is nevertheless necessary to present different viewpoints regarding possession, particularly because most available theories have been applied to possession in India. To begin, the interpretation of possession as the work of demonic, ungodly beings is characteristically the property of certain Western Christian, usually Protestant, denominations, though, as shown below, South Asians themselves often enough accuse tantrikas in the temporary employ of local enemies of setting malevolent spirits on family members afflicted with mental illness. Commenting on studies of possession, though not singling out Christian perspectives, Nicholas Dirks comments: “Most of the literature on possession deals with the nasty kind, when it is the devil rather than the lord who has taken up residence within our mortal coil.”24 This may be stretching it a bit, especially given the explosion of possession studies among anthropologists since the 1960s. But it does constitute a well-represented viewpoint commonly invoked until the mid-twentieth century, one that has not been completely abandoned. It is more accurate to agree with Isabelle Nabokov, who observes that despite Christian missionary accounts of possession as “replete with prejudicial interpretations and factual errors, they still disclose a vibrant, well-established tradition of boundary transgressions between this world and the hereafter.”25
One (1991) proponent of this missionary viewpoint shows, if nothing else, that eighteenth- and nineteenth-century “orientalist” attitudes die slowly, stating that “possession by the evil one means losing control of the self.”26 This is not far from the language used almost three hundred years ago by Bartholomaeus Ziegenbalg (1682–1719), the first Danish Protestant missionary in India, whose pre-enlightenment views were a precursor to imperialist Christianity. Ziegenbalg spoke of “devil dancers” in his “Genealogy of the Malabar Gods” (1713),27 a phrase that has cropped up for centuries. It was not, therefore, a complete surprise when, in 2000, the New York Times quoted the Reverend Bob Larson, a full-time, Vatican-sanctioned Roman Catholic exorcist operating out of Denver, Colorado, on the physical effects of exorcism. According to Larson, these may include violence, vomiting, and cursing on the part of the possessed: “Dealing with the Devil is ugly work. The Devil is ugly. Evil is ugly. When you get to what I call pure extreme evil, it’s not going to be pretty.” The Times article notes that in January 1999, “the Vatican issued a revised Catholic rite of exorcism for the first time since 1614, essentially reaffirming that Satan exists.”28
An article of a much higher caliber of statement than that on Reverend Larson, written by a Protestant missionary in Pakistan, appeared in a1990 issue of the journal Missiology. This article is of interest because it displays an anthropology parallel to that conducted by mainstream scholars—parallel but distinctly detached from it. It is recognizable to conventional scholarship because the data gathering is methodologically rigorous and linguistically sound. Yet, because of its source, it is not referred to in any secondary source that I have examined. The author of the article, Wayne McClintock, was employed by Interserve, formerly the Bible Medical Missionary Fellowship International, and reveals practically no familiarity with current scholarship on possession phenomena in South Asia. In the academic mainstream, this would be regarded as an unacceptable oversight for an article on folk demonology in Pakistani Panjab. However, McClintock’s motivations are different from those of the academic mainstream. After recording a good deal of phenomenological and linguistic data of some importance, McClintock’s conclusion is, at least for this reader, unexpected:
It is incumbent on persons bent on contextualizing the gospel among the rural peasantry of South Asia to grapple seriously with a worldview that perceives the misfortunes of life as attributable to some supernatural agency. Unless serious attempts are made by outsiders to understand this cosmos of demons and ghosts and its relevance to the everyday life of the rural villager, then efforts to incarnate the gospel message in a South Asian context will fail, and new believers will continue to be entangled in the web of syncretism.29
McClintock’s legitimate, if dated, anthropology, not to speak of his conclusions, locates him within a certain tradition which from our present perspective must be regarded as emic, and not peculiarly so, as it deals with faith-based assertions that assume a superior reason (and reasonableness) to what it clearly sees as other local contesting faith-based traditions. It seeks a victory for reason—Christian reason—over what it sees as an irrational, thus faulty, emotion-laden and superstitious belief system.30
This victory of reason over emotion in religious doctrine (if not always in religious practice) is well documented and is rooted as much in institutional prerogatives as in theological ones (though, of course, often there is little difference between them)—reason is regarded as stable, stabilizing, measurable, and the result of conformity with theological writ, all of which is to say that reason is institutionally safe as well as godly. In contrast, emotion is unstable, destabilizing, unpredictable, and attributable to the invasion of evil spirits, which is to say that it is socially unsafe as well as ungodly. Reason, stability, and control are due to possession by God; emotion, instability, and spontaneity threaten theological order and may well be the product of the devil or symptomatic of possession by spirits. It is not difficult here to connect the ascendancy of reason, hence control, expressed in McClintock’s Christian writing to Gombrich’s observation, noted earlier, that “Brahmanism inculcated control.” In both cases, the issue of control serves as a diaphanous cover for a strong undercurrent of behavior that transgresses institutionally imposed limits or, at least, bears the threat of such transgression.
However, acts of faith, which are central to the perpetuation of institutionalized religion, regardless of whether the religion is devotional, monastic, a state-sponsored cult, or a hereditary formalism, often take extreme forms that are permitted by the institution so long as they can be brought under theological, hence institutional, control. In other words, extreme behavior may not be regarded as unreasonably transgressive if it is accorded a peaceable and compliant place at the outer limits of theological or institutional normativity. Thus certain types of extreme behavior can be, and often are, theologically domesticated in order to neutralize any potential threat to institutional control. For example, glossalalia and āveśa are permitted, and sometimes encouraged, by dominant theological elites because they can be assigned an agreeable location in the theology and power structure of the religion. This structural positioning may seem at first glance forced or unnatural—after all, glossalalia and āveśa are uncomfortably close to the more transgressive and threatening specter of “devil worship” or physical violence such as self-impalement. Practitioners are empowered by the personal and social epiphanies that are components of glossalalia and āveśa, while priesthoods and theologians may be equally empowered so long as they are ceded final interpretative power and at least the illusion of overarching social control. In fact, the process is dialogical because it negotiates in both directions. Less privileged nonelite nonprofessionals poach on the discourse of privileged elite professionals by adopting their symbols and images, thus demarcating a realm of social and political safety for their experience, while the elites, the priesthoods, extend their domain by creating discourse mechanisms through which they domesticate popular practice. Indeed, as shown in culture after culture, elite participation in popular practice contributes to this negotiation between professionals and nonprofessionals.31
European Christianity, however, is unlike indigenous Indian religions in that it established more or less clear boundaries with respect to ecstatic experience, then strived to maintain them, at least in the periods immediately before and after the rise of Western colonialism. In order to understand the Protestant Christian, hence orientalist, attitudes toward possession in India (as well as other colonized territories), it is necessary to briefly summarize the Church’s position on possession on its home turf.32 Fortunately, we have Keith Thomas’s compendious Religion and the Decline of Magic to turn to for this.33 The devil was believed to be immanent, lying in wait to spread disease, crime, and havoc. “The belief in the reality of Satan not only stimulated allegations about diabolical compacts; it also made possible the idea of demoniacal possession.”34 Unusual or antisocial behavior, fits or convulsions, unconscious or unwitting blasphemy of the Church, and many other physical and moral ills were attributed to evil spirits sent by witches in league with the devil. “In seventeenth-century England, the epithets ‘possessed’ and ‘bewitched’ came very near to being synonymous.”35
The English Protestants viewed the practice of exorcism with considerable hostility, employing repressive actions to counter possession or bewitchment, including imprisonment or, in extreme cases in which a scapegoat was required, death.36 The Holy Spirit occupied an ambiguous place in the mainstream Church: To be possessed by the Holy Spirit was often intended metaphorically and was often at the outer limits of Protestant discourse.37 Beyond this, all other epiphanies were cordoned off, quite unlike the earlier Church, in which the limits were extremely elastic, at least in practice. Thus the dialogue between theological powerbrokers and representatives of popular movements broke down considerably earlier than was the case in India, and it was the deputies of the official power structure, in the form of missionaries, who came to India and attempted to impose their values on a phenomenological universe, which to them resembled the work of the devil. In view of this precedent for “othering,” it is not surprising that Western advocates of unorthodox intellectual and methodological positions, all to some degree rooted in the Reformation, should look so skeptically at emic assertions of possession that they too end up engaged in one or another type of othering. This othering is also assisted by the relative absence of a Christian framework for experiencing spirits, which “collapses them into a Christian demonic realm, which is much vaguer than spirits’ traditional classification.”38 In this case, as elsewhere in the colonial system, unfamiliarity bred contempt.39
To say that brahmanism or Christianity inculcate control is to say that the human desire and capacity for exotic empowering experience is kept within manageable, recognizable limits. In fact, however, is it actually the case that the elites whose discourse controls the “isms” also control the experience? This begs the complementary question: To what extent does more or less unsanctioned or barely sanctioned experience of the nonelite reshape the discourse of the “isms”? Both the anthropological and the Indological evidence suggest that when institutionally imposed limits are stretched, but not irreparably ruptured, these limits are usually modified, broadened by incumbent priesthoods and other elites in something of a tacit recognition that their control is comparatively thin, and often in name only. This interpretative strategy, which is not the property of anthropologists but of missionaries, is to be distinguished from the following perspectives in that it views possession as a religious problem, rather than as a medical, social, psychological, or existential problem.40
As others have noted before me, spirit (or deity) possession must be viewed from within the cultural systems of which it is a part. In such systems, as Loring Danforth points out, “people who are seriously disturbed are generally not able to express or resolve their idiosyncratic problems through the cultural symbols provided by the idiom of spirit possession.” Spirit possession is, therefore, “an adaptive response through which they are able to resolve social and psychological problems in a widely accepted, and sometimes well-respected, cultural idiom.”41 Thus the question of labeling spirit or deity possession a form of deviant behavior or psychologically definable pathology must be answered within the context of the local culture. This is no less true in South Asia than anywhere else. In spite of (or perhaps because of) a suspected (or suspect) normalcy, and because of the extensive time frame of the topic as I approach it here, it is necessary to engage in a fairly thorough review of the psychological and psychoanalytic interpretations of possession as I see them and their relevance to the topic at hand.
The defining feature of possession to both the possessed individual and the observer is marked psychological change that engenders recognizable modifications in speech and behavior. This is accompanied by an observable shift from one internal focal context to another and from one presumed identity to another. The operative assumption of nearly all scholars who favor psychoanalytic interpretations of possession is that the modifications are generated intrapersonally, strictly from within, the result of the psyche responding to stimuli that may be either internally or externally produced. If the latter, then the stimulus is regarded as secondary; in psychoanalysis the primary concern is to locate and analyze internal reactions and personality complexes that confer on the individual’s experience a unique and idiosyncratic flavor. Regardless, then, of whether an alleged experience of possession is said to result from interpersonal or any other extrapersonal contact (and nearly all of them are), the psychoanalytic project is to address internal psychic and psychological processes, usually released by external stimuli, that induce observable personality modifications.
Possession or external invasion has not been a productive psychoanalytic diagnosis, at least not until recently, because it opposes the assumption of the normalcy of an inviolable and unitary self, on which the Western biomedical system is founded. As much, however, as this philosophical reason for the denial of extrapersonal contact as the etiology of reported experiences of possession is that psychoanalysis (and, we should add, other social and “human” sciences), in the words of J. Moussaieff Masson, is “always skeptical of anagogic or ‘spiritual’ explanations, or of explanations in terms of what we cannot understand.”42 Part of the problem is that these sciences have established the parameters of what is understandable, and this creates an often unbridgeable rift between the science and the object of study. What eventually occurs, then, is that one explanatory model is replaced with another, which hardly solves the problem of communication between models.43
Nonetheless, psychologists and anthropologists who study possession are able to observe and even measure certain of its manifestations, such as altered speech and behavior, following which these investigators construct credible psychological and psychosocial narratives said to be responsible for these modifications.44 What observers, including ethnographers and (we choose to believe) indigenous participants, are unable to perceive are the alien spirits and deities reported to be taking possession of the individual and instigating these psychological and socially interactional changes. Thus most psychoanalytic interpretations either deny the possibility of possession (e.g., Lambek, see below) or leave the question unaddressed. The nature of these interpretations has become more sophisticated and nuanced over the years, though, as always, psychoanalytic interpretations continue to carry more than their fair share of cultural and intellectual biases.45
The three competing theoretical perspectives concerning spirit (and, we might add, deity) possession are Freud’s psychoanalytic theory, which regards possession as a form of hysteria, dissociation theory based on the work of Pierre Janet in the late nineteenth century, and the notion that possession is a culturally mediated altered state of consciousness. We deal with the third of these later; for the moment, it is important to understand that dissociation theory has always opposed Freud’s psychoanalytic theory. In the late nineteenth century most possession experience was regarded as dissociation. After rise of Freud in the early twentieth century, dissociation theory was eclipsed by psychoanalytic theory. A perusal of the different editions of the Diagnostic and Statistical Manual of Mental Disorders (DSM), published by the American Psychiatric Association (in four editions so far: 1952, 1968, 1980, and 1994), illustrates the changes in categorizing psychiatric disorders during the second half of the twentieth century. Initially psychoanalytic theory was ascendant. In the past few decades, however, there has been a shift toward viewing psychopathology “as a set of discrete heterogeneous disorders with distinct biological causations.”46 Along with this, dissociation theory has experienced a revival (for our purposes, note especially the work of Castillo and Littlewood, discussed shortly). At the same time, certain psychoanalytic categories have been completely expunged from the DSM, most recently neurosis and hysteria.47
Until the latter decades of the twentieth century, possession was usually explained by advocates of psychoanalytic theory as little more than a psychological state indicative of “repressed infantile complexes or a reconfirmation of primitive beliefs once surmounted,”48 as repressed oedipal desires in the unconscious, or even (in the case of spontaneous, nonritualized, possession) as compulsion, obsession, epilepsy, or hysteria.49 Although this may be an oversimplification, even a caricature, of most current psychoanalytic interpretations of possession, investigators throughout the twentieth century resorted to explanations that attributed reported possession experiences at least in part to abnormal psychological conditions or, if not abnormal, as culturally sanctioned psychological reactions to social or political oppression—in other words, as compensatory behavior. In either case, the reported “possession” was lightly regarded; rather, the experience was considered a psychic dimension of learned and conditioned behavior, founded on an ultimately fictive belief in possession.50 Stanley Tambiah attempts to interpret Indian notions of spirit possession from a standard psychoanalytic paradigm.
it is to be expected that the Indian view of mental illness would be incompatible with the Western approaches. In the Indian shamanistic or spirit possession beliefs, mental illness may be admitted to be related to undue or unfulfilled desires, but the results of such unfulfilled desires are not internally split-off portions of the self but are externalized entities given phenomenal existence as demons and spirits.51
Later we examine in detail the subject of bhūtavidyā (demonology) in classical Indian medical literature and how a diverse array of invasive spirits are believed to induce disease, especially in children.
Of more immediate concern, however, is to better understand the incompatibility of Indian and Western approaches of which Tambiah speaks. In short, the Western biomedical approach to physical and mental illness excludes the possibility of nonsensory forces influencing the individual’s consciousness. This consciousness, in the Western biomedical model, is a function of the mind, which is equated with and bounded by the brain. The proximate construction in Indian thought to this model of internal functioning is the “inner organ” (Skt. antaḥkartṇa), analyzed as the heart (as the locus of emotional experience), mind, intellect, and ego combined.52 This “inner organ,” however, is not immune to nonsensory influence, from infiltration by external signals. With respect to possession, unarguably an assertion of such nonsensory, external influence, Sudhir Kakar, for example, contends that what Indians regard as spirit possession is, in fact, nothing more than an interpretation. Kakar, perhaps the most literate apostle of Freud in the Indian psychoanalytic community, interprets possession as a self-induced mechanism for psychic release, consistent with the generalization stated above.
Kakar observed reported possession in Mehndipur in eastern Rajasthan, which contains a temple to Hanumān, called Bālājī, as well as several other facilities well known for exorcisms and other treatments for psychological disorders. The classic visible symptoms of possession, very much on display at Bālājī, suggested to Kakar that the evident somatic unburdening was caused by a culturally identifiable and sanctioned mode of psychic release, akin to classically defined hysteria but identified locally as a kind of possession. To Kakar, spirit possession is an act of situated imagination, the classic mechanism of projection.53 Another clinician, Dr. Antti Pakaslahti, has been conducting fieldwork for many years on the processes of psychological healing that occur at Bālājī. He agrees: “Here psychological distress is expressed and treated in terms of spirit affliction, culturally congenial to large segments of the Indian population.”54 However, Pakaslahti avoids the increasingly vague term “hysteria,” consistent with its banishment from the 1994 edition of the DSM, and sees the possession and exorcisms at Bālājī as effective family therapy, regardless of any claims to the facticity of possession.55 The reason for the effectiveness of the therapy, says Pakaslahti, is that at least one member of the family of the afflicted party accompanies him or her to the healing center, and, in the presence of the family, the healer is careful to expunge the afflicted person of guilt by convincingly shifting the blame for the individual’s dysfunctionality from the individual to the afflicting spirit. The anthropologist Michael Lambek, who has studied possession in Mayotte, an island in the Mozambique channel between Madagascar and Tanzania, puts it this way: “To argue that spirits are imagined is not to hypostatize them but to locate the work of imagination in its social context, in history.”56
Lambek states flatly, perhaps more decisively than Kakar or Pakaslahti would find comfortable: “Spirits are products of imagination, partial world constructions that are fictional but not simply fictitious.”57 In other words, whatever condition it may be, it is not possession, even if those experiencing it are not lying or misguided by their cultural preceptors.58 Among such conditions often associated with possession are “dissociation, fugue states, multiple personalities, fainting, functional epileptic seizures and other behavior of an apparently hysterical type,” writes Bourguignon.59 Many of these conditions have been favored diagnoses for possession, particularly dissociation, which falls within the general category of “double consciousness—in which two different human personalities apparently coexist in association with a single physical body.”60 From the point of view of the Western commonsense assumption of a unitary self lurking behind multiple personalities, double or multiple consciousness is an aberration. As Roland Littlewood suggests, however, distancing himself from Lambek’s psychoanalytic rationalism, “this might not be altogether legitimate.”61 Clifford Geertz provides a tidy summary of the Western conception of the person, a description that should be kept in mind throughout this study. This person is a “bounded, unique, more or less integrated motivational and cognitive universe, a dynamic center of awareness, emotion, judgment and action organized into a distinctive whole and set contrastively both against other such wholes and against a social and natural background.” It is no less important that Geertz notes, as an addendum to this definition, that this is “a rather peculiar idea within the context of the world’s cultures.”62
Littlewood, a practicing psychiatrist, suggests that this notion of a “single bounded and volitional self which shares a body which gives rise to it” hangs together as a result of the potential accessibility to awareness of various facets of the personality such as habits, abilities, memories, and idiosyncratic responses to stimuli. However, he adds, “this hanging together becomes unstuck in dreams, or in the usual process of forgetting and inattention, such that chunks of past experience cannot necessarily be recalled simultaneously.” These chunks, says Littlewood, citing Jung, are “split off ‘complexes’” or fragments (of which Tambiah speaks) that “might be so extensive as to actually constitute a parallel secondary self.”63 The concept of double consciousness, of dissociation, popular in the nineteenth century, fell out of favor for much of the twentieth century and has experienced a well-publicized revival since the 1980s under the designation Multiple Personality Disorder (MPD). Most recently, the latter has been rechristened Dissociative Identity Disorder.64 Furthermore, the DSM-IV has dropped the category of hysteria, which served as the bedrock of Freud’s view of possession, and suggested a new category for further research called “dissociative trance disorder.” Among the features of “dissociative trance disorder” are the following:
A. possession trance, a single or episodic alteration in the state of consciousness characterized by the replacement of customary sense of personal identity by a new identity. This is attributed to the influence of a spirit, power, deity, or other person, as evidenced by one (or more) of the following:
(a) stereotyped and culturally determined behaviors or movements that are experienced as being controlled by the possessing agent
(b) full or partial amnesia of the event
B. The trance or possession trance state is not accepted as a normal part of a collective cultural or religious practice.
C. The trance or possession trance state causes clinically significant distress or impairment in social, occupational, or other important areas of functioning.
D. The trance or possession trance state does not occur exclusively during the course of a Psychotic Disorder … or Dissociative Identity Disorder and is not due to the direct physiological effects of a substance or a general medical condition.65
It is important to understand that the DSM-IV is not suggesting that the etiology of such a state is in fact spirit possession, but that this is the symptomatology of a mental state associated with reported possession episodes.
Littlewood does not regard double or multiple consciousness as necessarily an aberration or in general as a kind of pathology. “Humans have the ability to dissociate their mental processes and do so the whole time—through changing moods, selective attention and putting unpleasant issues out of mind, through fantasy and dreaming—for the potential contents of our awareness are hardly accessible simultaneously.”66 Dissociation, he says, is:
the necessary flip side of consciousness: it allows detachment of awareness from the immediate passage of events, part of the evolutionary development of “self-consciousness” as an internal system of representation and self-monitoring where the individual’s awareness can then objectivize their own cognitions (“my senses deceive me,” “I was overcome by emotion”), allowing self-recognition, anticipation, introspection, creative imagination, recognition of another’s motivations and possible identification with them, disbelief, and acting: all requirements for our complex programmes of intersubjective action.67
Such dissociation, he says, can be engendered through hypnosis, sensory deprivation or overload, hyperventilation (or, I might add, other forms of regulated breathing), the ingestion of psychoactive substances, as well as through witchcraft or spirit possession. All of these conditions may influence the individual to bring about concentrated attention, meditative states, a shared sense of community, out-of-body experiences, and so on.68 The fact that beginning in the nineteenth century the Western psychiatric model established these phenomena as intrapersonal in origin rather than as deity or spirit intrusion may be little more than an artifact of a culturally weighted scientific paradigm. There is no reason why the “dissolution of everyday agency”69 and its temporary reassignment cannot occur through a comportment of causation and volition that is differently constituted from that presented by the intrapersonally defined Western psychiatric model. This is not to defend the “reality” of spirit possession, however, but to point out the limitations of the psychoanalytic model in providing a “final” assessment of possession. To state succinctly the distinction between paradigms, I offer the following formulation: Possession as an intrapersonal phenomenon requires (at least) two minds in one body, but as an interpersonal phenomenon requires two minds in one body and one mind in two bodies.70 An immediate objection to this formulation is that the possessor might be a disembodied spirit. However, as we discuss below (Chapter 8), the classical Indian doctrine of embodiment is multilayered and only with certain qualifications allows for true disembodiment.
Robert I. Levy et al. state: “Two conditions are necessary for full possession to flourish: people who are psychologically disposed to dissociation, and a cultural environment that makes conventional use of possession episodes.”71 Few would disagree with the latter, but some would view possession as integration rather than dissociation (leaving aside Littlewood, who would regard dissociation as an alternative mode of integration). In a relatively early article on possession in India, Stephen Fuchs comments on the barwa, male exorcists of central India, and distinguishes their dissociation from more pathological dissociation:
We might be tempted to diagnose the case of the barwa as that of a “dissociated personality.” Whenever he is “possessed,” the barwa assumes a different personality—that of his familiar, who speaks and acts through him with the authority and behavior of a superhuman force. In psychopathic cases, however, this change of personality takes place automatically and cannot be controlled by the individual patient; the barwa, on the other hand, is able deliberately to produce this psychic state.72
In an influential 1963 article, “Spirit Possession and Social Structure,” Edward B. Harper states that the spirit possession he observed among Havik brahman women of coastal Karnataka could be “labeled dissociation, characterized by compartmentalisation of the personality in which one aspect acts more or less independently of another.”73 This dissociation, Harper argues, is comparatively benign and functions as a stress reduction mechanism as well as an outlet for the expression of power by women disadvantaged by the incumbent social structure. As with Lambek, however, it is not spirit possession:
under certain prescribed conditions and in certain prescribed manners, it is relatively commonplace in most parts of the world for some individuals to identify themselves so completely with another as to believe they are the other, and to be considered neither mad nor impostors by their fellow citizens. The particular type of complete-identification role playing I am talking about involves temporarily taking the identity of—not another human being but—a personified supernatural being.74
What Lambek and Harper do is draw a fixed line between psychological disturbances as understood in the Cartesian Western sense, influenced, it appears, by the Protestant perspective outlined above and spiritual disturbances in the Mayotte or Indian sense. I am not arguing that Lambek and Harper are wrong or morally in error by denying or minimizing the emic viewpoint; nor am I attempting to argue that scholars who “admit” the possibility of possessing agents are right. Scholarship, notably in the social and philological sciences, must depart from emic discourses and follow, as their designations suggest, the positivist moorings of science, at least to the extent of searching for mechanisms for distant, difficult, or vague ideas, such as possession. Nevertheless, I contend that both etic and emic categories of interpretation are limited if they are not used to complement, or at least accompany, each other. A case in point is dissociation, a disputed concept that does not necessarily encompass the concept of possession.
The Finnish folklorist Lauri Honko, who has produced a major study of an oral epic named after its heroine, Siri, in the same area in which Harper worked, quizzed his informants on the relationship between domestic problems and possession in the performance of the epic. “Our psychologising questions concerning the identification of possession cult members with their mythical (epic) models began to irritate Gopala Naika [a possession priest, Honko’s principal informant], as they seemed to imply manipulation of the epic plot in relation to prevalent moods and problems in the audience.”75 In other words, Honko’s informants had no use for Westerners psychoanalyzing their possession rituals.76 There is little doubt that South Asians in possession would unanimously reject such psychoanalytic interpretations. This, however, is not to suggest that Honko’s questioning was wrongheaded and culturally insensitive. Indeed, Honko’s thorough work appears, to some extent at least, to bear out Harper’s observations. These observations, in turn, are useful in discussions of bhakti-induced āveśa.
As is so often the case in psychology, these and other psychoanalytical interpretations of possession originate with Sigmund Freud, the revived importance of Janet’s work on dissociation aside. In 1919, while attempting to understand the concept of the “uncanny” (Unheimlich), Freud cited F. W. Schelling (1775–1854), who stated that “‘Unheimlich’ is the name for everything that ought to have remained … secret and hidden but has come to light.”77 Based on this and other definitions, “Freud developed the idea that the uncanny arouses anxiety because the familiar and the unfamiliar appear in inextricable conjunction.”78 The bewilderment other people’s possession induced in Western fieldworkers has often reminded them of Freud’s Unheimlich and his association of it with repeated “doubling, dividing, and interchanging of the self” (Ich-Verdoppelung, Ich-Teilung, Ich-Vertauschung).79 Furthermore, many (if not most) ethnographers tacitly accept Freud’s distinction between the externalizing primitive and the internalizing civilized. Freud bases this distinction on his belief that the premodern subject labored under the delusion that external forces were responsible for neuroses that were, in fact, internally induced. This amounted to a declaration of the superiority of the uncanny by rationality. We revisit this possibility in later discussions of possession in Indian literature (see Chapters 8 and 12). He wrote to W. Fliess, “The medieval theory of possession, held by the ecclesiastical courts, was identical with our theory of a foreign body and a splitting of consciousness.”80 Possession, he concluded, was a mistaken diagnosis for hysteria.81 He was not, it seems, concerned with dissociation and did not address that problematic.
Although hysteria fell out of favor with the American Psychiatric Association, many researchers have chosen to retain and reinterpret it. Clarke Garrett, a historian who has worked on possession among the Shakers of North America, describes hysteria as a legitimate “act of communication, intended for an audience whose shared values and expectations enable it to interpret the behavior in culturally normative terms.”82 In this sense, it shares common ground with possession. Indeed, Garrett asserts, “spirit possession and hysteria become two names for the same thing—a kind of spectacular body language for expressing convictions or emotions too profound, too painful, or too dangerous to be expressed verbally.”83 In this interpretation, the possession is again reduced to an interpretation, rendering it a phenomenon induced solely by internal psychic processes. Thus it ends up asserting the same old truth claim of the psychoanalytic viewpoint and the falsehood of the subject’s assertion. Dissociation theory, however, allows the indigenous or emic voice a role in etic theorizing, as seen above in the construction of “dissociative trance disorder.” Littlewood’s consideration of possession as a nonpathological and even psychologically advantageous state of dissociation—whose causes are both variegated and little understood—not only is thus in keeping with recent advances in psychiatric taxonomy, but is a more prudent route to take in a study of a culture with a long, diverse, and abundant history of reported possession phenomena.
As should be clear by now, possession states, like other states of mind, are psychological states. Interpretations of these states, perhaps most of which are grounded in cultural and religious beliefs, assumptions, and practices, must consider the psychological content of these states within the framework of these cultural and religious contexts. If interpretations of phenomena that are regarded by the experiencer as religious, spiritual, or in some way extraordinary depend too strictly on psychoanalytic (or, for that matter, any other) models, which is to say at the expense of examining what people actually do, then these interpretations risk becoming overdetermined or, more simply, too normative for their own good as they slip into discontinuity with their subject matter. An example of this is Masson’s Freudian analysis of Indian asceticism, mentioned above, which poses many excellent questions but entirely neglects actual ascetic practice and, in the end, disintegrates into puerile oversimplification.84 This is one reason why I am attempting in the present study to temper my textual examination with a survey of the ethnography of possession. Fortunately, most recent studies are a good deal more culturally sensitive and accommodating than their predecessors and are considerably more advanced theoretically because of it.85
Perhaps the best-known studies of possession in South Asia that rely on psychoanalytic interpretative strategies are Stanley and Ruth Freed’s seminal 1964 study of possession in a North Indian village,86 Obeyesekere’s many studies of possession in Sri Lanka,87 and Jeffrey Kripal’s Kālī’s Child.88 I address only the latter because Castillo has effectively deconstructed the work of the Freeds and Obeyesekere, demonstrating that their psychoanalytic interpretations are based on assumptions of “unfulfilled oedipal desires” of those experiencing possession (cf. Tambiah above), assumptions for which, Castillo shows, there is not a shred of evidence.89
In his far-reaching examination of the life of Ramakrishna Paramahamsa, the well-known nineteenth-century Bengali saint, Kripal intrepidly defends his use of psychoanalytic categories in analyzing Ramakrishna’s ecstatic experiences and possession states. In spite of occasional lapses into Freudian overspeculation, Kripal does not wander into the trap into which Masson precipitously plunged of abandoning his primary subject (in this case, Ramakrishna), entranced by Freud’s seductive gaze. Kripal’s often attractive and compelling work is well known for the wrong reason, namely, the criticism it (and he) has sustained at the hands of the Ramakrishna Mission and its allies, who believe that their patron saint has been desecrated by this Freudian upstart.90 Kripal eventually concludes that Ramakrishna’s ecstatic experiences, possession states, and homoerotic visions (to say nothing of his asceticism and misogyny) were to a considerable extent compensatory behavior for the early death of his father as well as for conflicted experiences of sexual seduction.91 Sarah Caldwell, however, presents more solid evidence that in India possession states may be linked to transvestism, sexual conflict, and sexual abuse.92 Among women, the experience of bādhā āveśam (demonic possession) resembles other varieties of dissociative trance that occur as a result of anxiety, depression, withdrawal, low self-esteem, victimization, hallucinations, hearing voices, and so on. These states, she suggests, following Castillo, are not due so much to sexual dysfunction as to sexual abuse at a young age.93 It has proved difficult to ferret out of the earlier Sanskrit literature evidence for such abuse and consequent possession.
At their worst, psychological interpretations of other people’s possession reduce it to sexual dysfunction, illness, or incoherence and erroneous perception arising from dissociation; at their best they consider the indigenous context and explore whether and to what extent that forces modification in their own theories.94 One sociologist who, like Kripal, worked in Bengal was Deborah Bhattacharyya (whose important study Kripal never cites, though it was published more than two decades earlier). She notes that, in Bengal, possession represents “a kind of pseudo-pāgalāmi [insanity], wherein the person only appears to be pāgala.”95 This is distinguished from insanity as a medically defined pathological condition, in Bengali called māthāragolamāla (the disease of head malfunctioning).96 Such recognition of indigenous categories is all too rarely found in psychoanalytic analyses of possession.97 Caldwell, working in Kerala, states: “As in other areas of India, mental imbalance is traditionally believed to be a sign of possession by an evil spirit, either male or female—but, significantly, not by Kāḷi or any other deva (divine being).”98 A psychiatrist in Kerala known to Caldwell pointed out to her that the symptoms of “hysteria and sexual dysfunction” among women were similar to the “feelings of inner heat, trembling, and rage that typify the inner experience of possession reported by actors”99 performing the possession dramas that Caldwell observed. Why this should be so is not clear. It is possible that possession, regardless of the locus of its expression, exhibits external manifestations of feelings of trembling, rage, and so on simply because these are culturally recognized symptoms of possession and because possession is a form of practiced behavior, as Rich Freeman and others point out. In fact, Freeman’s definition of possession is sensitive to this issue: “To be possessed means to perform the possession rituals correctly and manifest possession behaviors.”100 At any rate, a topic not sufficiently discussed by Obeyesekere and others is the relationship between possession drama and compensatory possession brought on by physical, sexual, or mental abuse.
Another advocate of psychoanalytic interpretations of possession is Charles Nuckolls. He supports psychoanalysis as an analytic construct to be used in conjunction with others, however. Notably, he considers the rationality—the inferences and decision-making processes—that contributes to the possession-mediumship that characterizes the ritual healing among the Jalaris of coastal Andhra Pradesh. He notes that despite the “abundance of literature on South Asian possession … answers to the question ‘How does it work?’ are limited, for the most part, to ‘psychological’ and ‘sociological’ perspectives that have been around for years.”101 Jalari possession falls within the realm of what Sax calls “oracular possession”: It is concerned largely with healing and other sorts of problem solving. In this way, it is rather unlike most possession that occurs in festivals or in Sanskrit literature, where oracular possession is unusual.102 In any event, studying possession-mediumship as an inferential process helps to contextualize and modify the usual psychocultural interpretations. Nuckolls recognizes that emotion and thought both contribute to decision-making, and he notes how this contributes to Jalari possession. He cites strong evidence that possession-mediumship among Jalari and non-Jalari informants arose as a resolution to generational crises: ambivalent mother-son relationships that led to the son’s “identification with the mother-goddess as a central aspect of the self” or, in the case of women, “transformation into possession-mediums as a process of recovering deceased sons to whom they felt ambivalently attached.”103 After he establishes the generational and gendered basis of Jalari possession, he then analyzes the ritualized possession process in terms of contextual linguistic strategies and decision-making processes. Thus Nuckolls’s approach represents an advance over the usual psychoanalytic theorizing. Nevertheless, because of the nature of the ancient and classical Indian material, the synthetic approaches of Nuckolls, Kripal, and Caldwell are of limited use to us here.104
Possession is often interpreted as an expression of social control, or at least its attempt, a category that includes political power. Within this broad category of interpretation are almost innumerable variations that have been much pursued by anthropologists in the past several decades. Although control is exercised both actively and passively through political enforcement and cultural ideology, these interpretations have emphasized the active or performative aspects of possession and their roles within implicit psychological, social, and political contexts. This political influence in South Asia, however, is hardly felt because, despite the widespread practice of possession, it has not gained sufficiently broad formal recognition to influence religious or political bodies, as it has, for example, in certain places in Africa where possession argues much more visibly a politics of identity based on a sense of colonialist victimization,105 or in Tibet where the Nechung oracle has had a history of influencing the politics of the Dalai Lamas.106
Interpretations of possession as a product of or reaction to various modes of social control or oppression are usually aspects of broader interpretations that strive to locate more specifically and define more broadly the psychological and sociological dimensions of possession.107 At worst, these broad interpretations reduce possession to rituals of compensation or scripted cultural performance; at best, practice or performance theory, structuralism, and other innovative interpretative frameworks are invoked in conjunction with a sensitive hearing of other people’s ontologies and a solid understanding of the literary or folk-cultural backgrounds of the subjects under study.108 Claus’s comment—“The phenomenon of spirit possession—or rather a tradition of spirit possession—can provide a symbolic medium through which the individual (or group) re-adjusts himself to an appropriate [moral] order”109—provides a thicker accounting than many interpretations. This is because it views possession as a more or less normative and legitimated activity that helps sustain a balance between the possessed (who may or may not be the dispossessed) and the unpossessed. Viewed from this perspective, possession is an integral part of a self-regulating system of social control. It is not merely a one-sided or desperate compensatory response to oppression, nor is it a veiled or exclusive mode of cultural communication, the role of which is to socially or politically strengthen or empower the individual or group.
One of the more impressive and sophisticated studies of South Asian possession is Hancock’s explication of the social dimensions of oracular and ecstatic possession among middle-class Smārta brahman women in Madras. She claims, with Claus, that possession is an aspect of the socially self-regulating, yet ambiguous, mechanism of bhakti within a domestic setting, in which compliance with patriarchal norms is assured through retaining their “patriarchally derived identities as wives and mothers” while also manifesting a resistance to these norms “through their radical appropriation of authority through the avenues of possession and mediumship.”110 In a statement that can as well be true of possession in general, Hancock argues that bhakti “cannot be treated merely as a compensatory safety valve or an interstitial or liminal interlude in an otherwise normatively ordered social existence. Bhakti, especially as pursued by cumaṅkalis [Tamil, Skt. sumaṅgali (fortunate woman)], may be an agency for change in domestic life, whether the bhakta is perceived as succeeding or failing in her efforts.”111 It is this sort of observation that should be valued as a safety valve on the exclusivity of certain psychoanalytic or sociological conclusions about possession.
Although most cases of possession as āveśa or samāveśa attested in Sanskrit and other classical literature are manifestly empowering in a religious sense, I hesitate to overlap this empowerment with externally induced attempts at social control or engineering. While Sax’s observations (predominantly in men) of possession in present-day Garhwal—as a medium for the contestation of prestige through which social control is gained—constitute strong evidence for this interpretation (and this is replicated often enough in the ethnographic literature),112 I am not persuaded that possession as attested in Sanskrit texts is guided by this sort of quest, at least not in all cases. Primarily this is because much of the possession found in these texts is nonritual or at least may be epiphenomenal rather than primary (as in certain tantric texts examined below). Regardless of whether the texts accurately reflect possession outside the texts, in the present project we must evaluate the evidence at hand, which means textual, and only then attempt to infer sociological realities behind the texts. Unfortunately, very few studies have focused on nonceremonial possession (indeed, many scholars would hold that this very notion is mistaken)—and this requires us to go beyond the familiar theories in assessing “Sanskritic” possession. In other words, the evidence from classical texts does not ordain that I base my study on ethnographies from the past hundred years or that I necessarily regard these accounts as normative, as filling the lacunae that are a natural aspect of texts that do not aim at thick description. Rather, it seems, the vast number of texts that speak of possession directly or refer to its social context fragmentarily or obliquely provide evidence for other interpretations or for locating a home for this and other more or less singular and isolated interpretations of possession within other cultural and theoretical explanations, not all of them necessarily contemporary Western models. Part of the reason I feel uncertain about this or any other singular interpretation is the lack of evidence: Quite simply, Indology cannot easily, if at all, locate the political background to first millennium B.C.E. or C.E. events with the precision and depth of events that Caldwell, Claus, Erndl, Hancock, Sax, and others present through the disciplined practice of ethnography.
That said, it is necessary to emphasize that possession is, in virtually all cases including those described in ancient and classical literature, an act of social subversion as well as an act of social confirmation, at least within a small but informed circle, regardless of the visible or conditioned structures of oppression surrounding it. Possession subverts consensual views of consciousness, the nature of the individual, and the orderliness of social and political order. Possession, even when it is nonritual (as in the case of disease-producing possession), behaves like ritual in that it is itself a structure of resistance, a rift in the psychological, social, and political fabric of society. However, in most cases (even in disease-producing possession) it is a contained rift: It is not ultimately counter-hegemonic, as Claus has noted. Thus possession becomes a satisfying, sometimes even a necessary, dissociative experience that ends up presenting very little threat to prevailing orders. It becomes a form of “controlled chaos,” to invoke Heesterman’s depiction of vedic ritual. And, like the latter, what is controlled is (usually) the danger, the violence, the utter alienness, of possession.113 Just as vedic sacrifice cannot be performed outside a sanctified ritual arena, thus limiting its extension into the social and political arenas, possession cannot be manifested outside its own orbit; it is self-contained, limited by the boundaries of the body. Although this observation may seem obvious and tautological, what I mean here is that as a dissociative experience it is segregated from consensual, normatively configured, states of consciousness and communication. As a site of resistance, possession has a limited domain, and when performed it tends to “lapse back into the social order,” as Dirks notes of public ritual.114 Just as both the individual and the social order are reckoned to be permeable and fluid as a result of possession, possession is ultimately circumscribed, penetrated, and permeated by the social and political orders. That this is much more evident in anthropological studies than in the classical South Asian literary record introduces a question that we can attempt to answer only at the end of our study: To what extent can we equate possession as discourse with possession as event? Or, more precisely, to what extent does possession as a discursive event represent possession as a psychophysical enactment?
Because of the prevalence of interpretations that seek to locate (and limit) possession as a product of social and psychological forces, I offer, early in my account, some tentative, if synthetic, conclusions on these issues. Many studies have shown that the splitting of the personality into discrete personae—public and private, outer and inner, authorized and unauthorized, lawful and illicit—is a characteristic of authoritarian societies. This fragmentation, and the consequent struggle to unify the self, or at least to publicly depict a unified self, may be extended to more restricted domains of tyranny: within the family, particularly women under patriarchal oppression; within the school or other educational systems characterized by consensual dominance and inferiority; within prisons or other more modified precincts of legal systems; and within hospitals and other health-care environments in which an afflicted individual becomes psychologically exposed, uncertain, and disadvantaged. In other words, the personality fragments—and becomes habituated to such fragmentation—in encounters of conditioned inequality. Among the time-honored methods of contending with this loss is full or partial dissociation. As we have seen, among the defining features of dissociation is increased intensity of mood, which complements increased objectival focus. This arises as extraneous or impractical aspects of the personality drop away as part of a coping mechanism. Dissociation thus becomes a response to particular varieties of stimuli. However, while these stimuli may not fit into cultural categories sanctioned hierarchically, they may not be destructive and may not necessarily be oppressive in the classical sense. They may simply fall outside the system. But, good, bad, or neutral in any cultural system, they may be said to oppress the openness of the personality by shutting down certain areas of it, areas that are, at least temporarily, deemed inconvenient, unnecessary, and distracting. The difference, then, between dissociation and any ordinary shift in mood or focus is one of intensity and cultural recognition and accountability.
Examples might be dissociation derived from the inspiration of religious or spiritual experience, from an encounter with novelty or from anything that might suddenly push the mind into new and captivating realms of thought. Thus a poor woman suffering patriarchal oppression in an Indian or Nepali village might develop the ability to sequester her mind from the drab familiarity of her world by entering into an intimate and coveted relationship with a deity (or the idea of a deity), then become so engrossed in it that she acts out, manifests, and finally “becomes” the deity. This is possession, regardless of the “reality” of the deity. Similarly, a man, prosperous and seemingly in control, might covet a deep relationship with a deity or other ethereal being (or with the constructed idea of the deity or being) and psychophysically act out the theorized behavior of that being: This may well be surprising and unexpected to the man. Again, this is possession, regardless of the “reality” of the deity or ethereal being. What is cultivated is a heightened awareness and spontaneity that is exciting and attractive to both performer and audience. This increased focus and spontaneity is highly valued as an enhanced level of awareness. What turns this into āveśa in popular culture from the Ṛgveda to the present—or, according to certain Tantras, samāveśa—is that it is a willed and invited, rather than an involuntary, state. Of course, there may be involuntary possession, and this is dealt with in due course.
Possession is often identified or compared to shamanism.115 The most important theorist of this relationship is I. M. Lewis, whose book Ecstatic Religion remains mandatory reading in the field. Lewis asserts that “possession is a culturally normative experience”116 in many places, a statement with which I concur with respect to South Asia. Like most anthropologists, he seeks to determine the social etiology and meaning of possession as well as its culturally determined psychological dimensions. He examines trance, ecstasy, hypnosis, and shamanism, ultimately concluding that the latter shares an important process with possession—a reciprocity between gods and humans. He posits a homogeneity between shamanism and spirit possession, arguing, contra Eliade, that “the Tungus evidence makes nonsense of the assumption that shamanism and spirit possession are totally different phenomena, belonging necessarily to different cosmological systems and to separate historical stages of development.”117 In drawing this equivalence, Lewis, whose model strikes me as excessively deterministic despite many virtues,118 delineates a wholly presumed and uncompromising functionalist sociology of possession states.119 In this he develops two contrasting psychosocial loci of possession, which he labels “central” and “peripheral.” Central possession, generally highly valued, supports prevailing political, moral, and religious beliefs and views spirits as sympathetic to these. Peripheral possession, an invasion of evil, amoral spirits, is undesirable and dangerous. Central possession, understandably, is a preserve of those of high status and power, while peripheral possession is generally associated with those who lack power and status. Although peripheral possession appears to be historically the preserve of women and those of lower social rank, it is definitively less so in recent times, as democratization, modernity, and the disappearance of Christian-dominated colonialism have freed educated middle- and even upper-class individuals and families to embrace the peripheries, bringing them closer to the center. In this way, economics, rather than ideology, caste, or other markers of purity or impurity, defines and situates power and status, at least in this orbit, by collapsing the boundaries between the periphery and the center.120
Gombrich critiques Lewis a decade and a half after Lewis’s first edition.
Possession is sometimes confused with shamanism; but in shamanism one’s spirit travels while one’s body remains unconscious, whereas in possession one’s body is temporarily inhabited by another spirit, while what happens to one’s own spirit at the time is left undetermined. This temporary loss of the sense of self is just like hysteria as clinically defined by Freud and Breuer. In India possession is mainly valued when it is practiced (perhaps one should rather say “undergone”) by specialists; they become possessed by non-human spirits, interact with an audience, and solve problems for clients.121
In order to bring some clarity to the much-abused term “shamanism,” especially as it is used by South Asianists with reference to possession, it is constructive to critique both Lewis and Gombrich. As to Gombrich: while we admit that he offers an adequate (though limited) critique of Lewis, he nevertheless misses the point that in India shamanism cannot be restricted to spirit travel while the body lies inert. In this, Gombrich adheres to the notion of possession offered by Eliade, that soul-journeys, rather than possession, are the definitive markers of the trance states characteristic of shamanism.122 This is also stated succinctly by Bourguignon, who distinguishes between shamanic trance and possession trance: “The [shamanic] trancer sees, hears, feels, perceives, and interacts with another; the possession trancer becomes another.”123 South Asianists, however, do not limit the use of shamanism to “soul-journeys”; indeed, in South Asia there is little evidence for this sort of shamanism.124 Obeyesekere largely agrees with Gombrich’s critique, though, in his study of the Pattini cult of Sri Lanka, he discusses a variety of possession states, not all of which can be described as sharing with shamanism the essential feature of reciprocity.125 These are cases of involuntary possession, usually symptomatic of disease or psychomental instability.
Another criticism of Lewis, at least from the South Asian standpoint, is that the terms “ecstasy” and “ecstatic,” as well as “central possession” and “peripheral possession,” are unproductive and misleading.126 Although Lewis is largely free from cultural biases, his characteristic terms “ecstasy” and “ecstatic” are drawn from Western theological discourse and cannot be applied unhesitatingly to South Asian possession accounts. At the beginning he defines ecstasy as “those transports of mystical exaltation in which man’s whole being seems to fuse in a glorious communion with the divinity.”127 He then equates this with the premodern Christian sense of “enthusiasm.” While this is surely one of the many senses that come under the category of possession states in South Asia, as we see below in reviewing possession in the Mahābhārata, bhakti, Tantra, and medical literature, it is by no means the only one. Through expanded attention to the lexical and linguistic presentation of possession states in South Asia, the present study hopes to set aright the misrepresentation of possession to which even a sensitive researcher, like Lewis, falls victim.
As for “central” and “peripheral” possession, I have no disagreement with Lewis’s distinction between possession states, and even spirits, that support prevailing political, moral, and religious beliefs, as well as other states (and spirits) that are undesirable and dangerous. The problem with his categories is that the terms “central” and “peripheral,” while valid, are of limited application in the South Asian context. Indeed, in South Asia, undesirable and dangerous states and spirits often support prevailing beliefs and moral orders as strongly as do benevolent ones. For example, folk possession in Nepal, Tamilnadu, and elsewhere is often violent, even dangerous, but in many ways it serves as the touchstone of certain deity cults, mediating between the individual and the prevailing social and religious orders.128 Truly malevolent possession is usually associated with physical and mental illness. But this, too, might not fit neatly into the category of peripheral, as it can also support prevailing morality and social systems, garnering participation across the social spectrum. I am not arguing that all possession in South Asia is central, rather that these terms should be used carefully.
Perhaps the most critiqued aspect of Lewis’s work is his association of possession with marginalized people, his “assumption that possession provides the powerless with a means to symbolically express social, economic, or political oppression.”129 Although many of the contexts of possession worldwide confirm this, the preponderance of the evidence in India bears this out only superficially. While no one can deny that power relations are an important factor in South Asian possession, the evidence from both classical and contemporary possession demonstrates that it is by no means the exclusive province of the powerless.
Shamanism broadly indicates manipulation of spirits, the usual process being the intentional, though ritually induced, introduction of a spirit into the shaman, which he or she then employs in a deliberate manner, generally for the benefit of others, often for curative purposes.130 In this seemingly physical entry, little if any distinction of category or function is made between spirit and deity; the shamanistic process is the same in either case. The shaman as spirit medium could as easily be the shaman as deity medium. One can say here that “the medium is the message,” that the shaman’s very presence, the empowering perception of his or her extraordinary gifts, is the most essential component of the shamanistic process. It is this that sets in motion the performative processes and their eventual fructification. Sergei Shirokogoroff lists five essential characteristics of a shaman (at least among the Tungus of Siberia): (1) a shaman is a master of spirits, who has (2) mastered a group of spirits; (3) a shaman commands a recognized array of techniques and paraphernalia that have been transmitted from elders; (4) s/he possesses a theoretical justification for the shamanistic process; (5) the shaman occupies a special social position.131 Whether spirit or deity, these five are repeatedly borne out in studies of South Asian spirit mediumship, which most often assumes the form of oracular possession. In other words, shamanism in South Asia takes the form of spirit mediumship and oracular possession. This is not to argue that all possession in South Asia is shamanistic (disease-producing possession is not, nor is destructive possession analogous to that of Nala by Kali in the Mahābhārata, and other forms of possession emanating from curses that one finds at exorcism sites such as Bālājī), or that the possession states that I describe from Sanskrit literature are shamanistic; it is simply to say that there are points of contact among such phenomena that must be distinguished.132
This technicality is important because shamanism has sometimes been limited to a specific mode of contact, in which the shaman ascends to the province of the gods or to chthonic realms in order to access special knowledge, which is subsequently transmitted to an enculturated and ritually constituted group. The innovation in South Asia, arising in no small measure from a widely distributed and more or less uniform devotional ethos, is that in oracular possession, gods descend into humans as an act of grace.133 Although the shaman may have invited the deity, the prevailing view is that in the final analysis it is often believed to be the deity that instigates the possession. While the shamanistic components of personal empowerment and acquisition of social status are neither lost nor ignored by either ethnographers or their informants, the latter invariably report that the primary value of possession and spirit mediumship is religious, that they are reciprocations of bhakti.
The late G.-D. Sontheimer, the well-known scholar of Maharashtrian folk culture, also understood possession as shamanic in a broad sense, when it was mediumistic, usually involving exorcism or healing.134 In his many essays on Khaṇḍobā, he consistently translated the Marathi word devṛṣi (variants: devṛṣi / devruṣi; Skt. devarṣi) as “shaman,” though he quickly adds that a “devṛṣi is not a shaman in the strict sense. There is no land of the dead into which the devṛṣi accompanies the dead, and no magical shamanic flight. The god ‘comes’ in order to take possession of the medium’s body and to speak to the Dhangars,”135 a shepherd caste whose members are possessed in this manner. Even if it is in a festival context where others may be in ecstatic possession brought on by devotion, the devṛṣi’s experience is qualitatively different. Although his possession (devṛṣis are all male) is brought on by devotional intensity, like that of others in ecstatic possession, it is also oracular or mediumistic and seemingly open to the same rational processes described by Nuckolls. “By entering into the body of the devṛṣī, he [the god śidobā] makes him into a ‘temple’ and enables him to diagnose others’ diseases, adversities, and so on, in the form of bhūta, and to recommend methods for remedying them.”136
Finally, Smriti Srinivas, who studied oracular possession among the Buddhists of Ladakh, recognizes its distant affinities with classical Siberian shamanism. Nevertheless, what she studied was also oracular possession, not journeys of the soul. Srinivas recognizes the terminological problem, the fuzzy understanding and application of different categories: “There is no unity of terminology—they may be called mediums, oracles, or shamans.”137 It is clear that most possession is not shamanic, strictu sensu, even when the term is used.138 In a statement summarizing studies of possession in Chinese religion—apposite for the state of affairs in South Asian studies—Jordan Paper writes: “The general tendency to refer to all ecstatic religious functionaries as shamans blurs functional differences”139 Thus it is more prudent to employ the distinction between spirit possession and spirit mediumship proposed by Peter Claus.
Spirit mediumship—the legitimate expected possession of a specialist by a spirit or a deity, usually for the purpose of the aid of the supernatural for human problems—is perhaps more common in certain parts of South India than in others. Spirit possession—which we may tentatively distinguish from the above as an unexpected, unwarranted intrusion of the supernatural into the lives of humans—is far more common than mediumship and far more widely distributed.140
This distinction is valuable for the current study, though, as seen below, there are many situations in which spirit possession is expected and warranted, but is not mediumistic or oracular.
As for the current study, much of what Lewis refers to is difficult to apply to the material found in Sanskrit texts. It is not possible to align either shamanic trance or ecstasy or social oppression with possession based on the evidence adduced here. Indeed, Lewis’s determinism is unproductive for the present purposes; he overtheorizes or makes his argument too extensive. The problem is the overdetermined nature of the discursive resources at his disposal: His theories all too easily circumscribe his evidence. It may be that the texts are hopelessly inadequate, but I believe that the evidence accumulated below demonstrates that only with great caution is it advisable to link possession (or ecstasy) with oppression and shamanism in premodern India. There is some evidence of mass possession cults in eastern India (i.e., greater Bengal) due to Mughal imperialism before British domination. This is alluded to in some of the Bengali maṅgalkābyas,141 but as far as I can see there is virtually no evidence for this before the Muslim period. It is possible, however, that the evidence I have presented here can reveal a sociology of possession that I have failed to retrieve.142
Possession is understood as ontological or veridical reality by those who undergo the experience, which is to say as imposition or investiture on an individual personality of an independent and unseen external agent. This, of course, is the emic view, the narrative of indigenous tradition, the interpretation given to possession by the groups or individuals studied. This stands opposed to nearly all academic theorizing—to the emic narrative of scholarship. Nevertheless, it is an interpretation not wholly discarded by many ethnographers and others engaged in the issue. For most scholars who recognize, if grudgingly, the possibility of possession, this recognition is tactical, as a contribution to other theoretical agendas. The underlying assumptions of the social constructivist approaches that characterize most academic theorizing about possession—that issues and ambiguities concerning power and authority lie at the basis of much spirit belief and nearly all claims of unmediated possession experience—are being increasingly softened by scholars willing to find a way for their approaches to reconcile with indigenous notions that they find are all too often summarily dismissed. These scholars, most of whom occupy university positions in religion, psychology, and anthropology, are themselves, perhaps, rebelling against the possession–or, to put it more mildly, the enchantment–of their disciplines by modernity (or postmodernity). To transpose Geertz’s observation of cultures as possessing a “curious mixture of borrowed fragments of modernity and exhausted relics of tradition,” we might speak of contemporary academics sorting out for itself exhausted relics of modernity and borrowed fragments of tradition.143
Perhaps the most sympathetic and best-reasoned considerations of this mixed viewpoint are by Felicitas Goodman, whose work has been (perhaps not unsurprisingly) largely overlooked in the past decade or so,144 Edith Turner,145 and Michael Cuneo.146 In 1970, the Indian psychoanalytic literature added the category of “possession disorder” to its list of diagnostic categories, in recognition of the dissimilarities of alternative identity adoption in the possession experience from other types of dissociative disorders.147 This is part of a growing trend (at least in South Asia) to consider possession a viable category in anthropological and psychological theorizing, in which interpretative frameworks supplement one another or overlap.
Boddy notes, “Possession intersects with numerous cultural domains including medicine and religion, but is itself reducible to none.”148 For example, psychological or social factors may be causal in ritually circumscribed, though socially or politically empowering, experiences of possession, experiences that may be characterized by temporary loss of personal control—which under “ordinary” circumstances would be strikingly disempowering. That is, one may take seriously the givenness of spirits in order to understand possession as an idiom of communication, moral discourse, gender, identity formation, and conflictual or power relations.149 A good example of this is Manuel Moreno’s study of possession in Tamilnadu, in which he questions analyses of Hindu gods as “‘disembodied symbols’ of social realities and human relationships.”150 Rather, he argues, gods and humans “are related by complementary, mutually rewarding bodily exchanges,” which, as shown below, is not far from the notion of transfer of essence that characterizes possession in India as early as the Upaniṣads. Notwithstanding a few attempts at multivalent theorizing such as Moreno’s, academic approaches to possession have usually foregrounded the social or psychosocial contexts of possession, or the moral or communicative value of its performance, approaches that generally, it seems, have drained possession of the assumption of existential reality. More to the point, to most academics the existential possibility of possession is epiphenomenal to constructed psychosocial or moral realities, while in the present study the opposite is the case: Because of the relative absence of thick description of possession in the classical literature of India, environmental factors are often forced into the position of epiphenomena. This is why Halperin advises treading a fine line between the usual scholastic etic approaches and an unqualified “emic friendly” approach.
Given the veritable minefield of official discourses, underlying assumptions, deeply held feelings, and occasionally conflicting behaviors of participants, the ethnographer is perhaps best advised to do research equipped with a flexible analytical armamentarium. Preferably, this approach includes a more than superficial appreciation of other emic worldview(s), without being limited to such an orientation.151
The present work should, it is hoped, permit a flexible armamentarium to enter, to possess, its analysis.
We must now discuss some specifications of possession in South Asia, including, first, the possibility of possession as a gendered phenomenon. In South Asia, as elsewhere, more women than men appear to experience possession. Important to understanding why this is the case, and perhaps always has been, is an evaluation of the constitution of personhood and sacred space in South Asia, for these are intimately linked in the physical and metaphysical complex of South Asian women’s possession. It is also necessary to discuss possession as performance, as this is also linked with the constitution of the person and relational space.
Schoembucher notes that the earliest European notice of possession was in 1713 by Bartholomaeus Ziegenbalg, the Danish missionary who lived on the Malabar coast, while the first serious anthropological study was by C. G. Seligmann and Brenda Seligmann, who wrote an account of spirit mediumship among the Veddas of Ceylon in 1911. These and most subsequent studies have remarked on the preponderance of women in possession.152 One reason often attributed is psychosocial: Possession is a refuge and an empowerment for the oppressed and marginalized woman in a patriarchal society. In this sense, possession is a ritually protected means of enabling a woman to interact from a position of authority and status. Many studies of South Asian possession bear this out, though not all are in agreement about how this occurs.153
In fact, the classical feminist model explicates the dynamics of oppression and empowerment and, in possession studies, investigates notions of agency. The most recent general study of this is by Mary Keller, who investigates the interrelationships between women’s bodies, power, and possession. Her principal observations led her to theorize an “instrumental agency,” in which “possessed bodies share the same paradoxical agency in that the body is not speaking, it is spoken through; the body is not hammering, it is being used to hammer; the body is not mounting, it is being mounted.”154 This will come as no surprise to those who study possession in South Asia; the evidence presented in Chapter 4 reveals that language after language in South Asia has terms for riding, mounting, and so on that are characteristic of possession, preponderantly women’s possession. Does this then occlude or nullify any power in the form of cultural rebalancing that women might gain from possession, by shifting the power to a possessing spirit or deity and away from the woman herself as an autonomous agent? My sense is that it does not. Primarily, I believe, this is because the question does not quite work in the context of South Asia. First, much of this power is ritualized,155 which tends to reinforce cultural conservatism, as both the textual and ethnographic evidence reveals. Most women’s possession practice tends to maintain, rather than threaten, the power distribution between men and women by functioning within orbits of culturally sanctioned ritual control. The release of the energies of women’s oppression into possession states does little in the end to rectify the power imbalance between the sexes in the “traditional” cultures of South Asia. However, the reasons for this have less to do with the shifting of power to a spirit or deity or, more broadly, to possession as an ineffective mode for asserting women’s power than with a millennia-old acceptance of possession as, to a great extent, a women’s cultural or religious practice. Notions of agency, including Keller’s of an instrumental agent that functions outside or in opposition to the realm of “selfhood,” simply have not arisen in South Asia, where an autonomous self was rarely regarded as normative, or even particularly sought after, beyond the thin veneer of certain forms of vedāntic theory and practice. The normative self was understood as permeable, hence the distinction was rarely drawn between an instrumental agent and an autonomous self. In this sense, the power women derived from possession was real, and constantly corrective, because they functioned autonomously and powerfully as instrumental agents.
A reason closely embedded in the woman’s marginalization in South Asian culture that is sometimes given for the preponderance of women’s possession is their relatively lower level of education, especially in rural towns and villages, which, some say, contributes to their greater ability to experience “irrational” states.156 Although it is true that positivistic education contributes to the growth of “reason” and a concomitant devaluation of states that run counter to that, the evidence presented in Chapter 3 and verified elsewhere in the world demonstrates that possession states can occur, even abundantly, in literate and educated societies.157 To give this as a reason for possession is to harbor biases, which should have been put to rest by now, against the mental and psychic capacities of the “Other” and to misunderstand the possibilities of our own positivistic education and the challenges it confronts in an ever-changing culture.
The notion that possession is a long-established aspect of “women’s religion,” as opposed to the supposedly more intellectualized and ritually formalized “religion” of men, receives a boost from Margaret Trawick, who contributes to this understanding in her observation that women in South India who are possessed by the goddess Māriyammaṉ are the feminine counterparts to the masculine practitioners of the siddha tradition of Śaiva Siddhānta.158 To whatever extent this may be the case, what appears certain is that women are consistently depicted locally as embodying an inherent power, a pure but malleable active energy that is both dangerous and spiritually potent, called śakti or, in Tamil, aṇaṅku.159 It is this, I might speculate, that is the conductor of women’s permeability or penetrability or, more strongly, that it is a uniquely gendered substance that cultivates women’s power and (among other things) contributes to the capacity for possession. This characterization, especially the assertion of inherence, is largely a male cultural construct, in which śakti (Ta. cakti) or, perhaps less, aṇaṅku, because of its greater ambiguity or multivalence, is believed to derive from the woman’s awesome and incomprehensible menstrual and child-bearing capacities, from her mysterious ability to manifest blood and babies.160 In spite of the culturally conceived purity of the energy that lies behind the phenomenon of manifestation or transformation, whether it is of babies, deities, or spirits, the obverse is equally regarded: a latent impurity that is an aspect of a more delicate female constitution. At any time this can burst forth into dangerous, “negative,” possession, into illness and madness—a topic examined in detail in Chapter 12.
The characterization of śakti as an inherent female power, inherent because it is identified as the essence and energy of the goddess, is noted in many anthropological reports. This śakti may be intensified in any individual through possession, typically goddess possession, which is itself typically an intensified form of bhakti. However, manifestation of śakti, even regularly, neither elevates a woman from poverty nor confers on her equal status in a patriarchal universe, in spite of its temporary empowerment; wealth and status operate independently of śakti or possession. In fact, the situation is quite the opposite: Although exceptionally a person who gains a reputation for oracular possession may benefit financially through developing a consultative practice,161 usually possession and other manifestations of śakti confirm the socioeconomic order as they provide a separate identity comprehensible from within the woman’s world. Hancock notes, based on fieldwork with middle-class Smārta brahman women from Madras who experienced possession of the goddess Karumāriyammaṉ, “Women’s devotionalism, including spirit possession, often involved efforts to renegotiate domestic relations of authority while still retaining their patriarchally derived identities as wives and mothers.”162 Somewhat at odds with Hancock’s sociological observation, Margaret Egnor (Trawick) reports on a life history of a woman regularly possessed by the goddess Māriyammaṉ in Tamilnadu: “Mediums for the goddess are expected to be poor, low caste individuals.… The poverty and low status that Sarasvati endures are considered necessary conditions for possession by the goddess.”163 Nevertheless, Egnor, in agreement with Hancock, adds, “if women have more śakti than men, this is (at least in part) because women stand in the position of servants with respect to men.”164 Thus neither possession nor the śakti on which it both rides and creates upsets the social order. Indeed, śakti legitimates the suffering implicit in woman’s subordination, and it is this, Egnor concludes, that creates “an essential unbreakable unity among females themselves.”165
The manifestation of śakti is therefore self-empowering and at least temporarily subverting in that it confirms and liberates the woman’s suffering and status so that it may be used to her advantage. Yet that universe of empowered suffering is (in general) ritually circumscribed, hence does not confer on her the social or political power to threaten the patriarchal order, at least not in the long term. The male Śiva, the archetypal śaktimān, controls his consort Pārvatī, the śakti. Woman, characterized as śakti, is, in possession, in fact revealed as śaktimati, śakti as controller, in her incarnation as Durgā, in which she controls the male god. This is her act of subversion. Durgā is Mahiṣāsuramardinī, the “slayer of the buffalo demon,” as she mounts a tiger and destroys ignorance incarnated as a man deformed. It is śakti as śaktimatī that is subversive and is therefore feared and overpowered by men. But in possession, especially oracular, the woman—here the goddess (most possessing deities are female)—manifests power that confronts men directly. Erndl has shown that village women in Panjab become Śerāṅvālī, “she who rides the lion,” by becoming possessed, by being “ridden” by the goddess.166 It is this powerful woman who tames the oppressive man. The reversal is empowering to the woman and her community, even if it cannot extend to other sociocultural areas. In other words, women in possession opt out of the dominant culture and create their own path, reversing the socially inscribed hierarchy provisionally and briefly, demonstrating to themselves, at least, its vulnerability and shallowness.167
This function of possession is not unlike that of certain genres of women’s songs, which Gloria Raheja and Ann Gold show empower women while they critique the structure of the dominant male culture, without, however, materially altering it.168 Both women’s possession and women’s songs follow the inherited cultural and religious paradigm in India in which women’s participation and empowerment rarely lead to social dominance. For example, in the Vedic New and Full Moon sacrifice (darśapūrṇamāsa), the wife of the sacrificer undergoes a regular ritual initiation, as both Stephanie Jamison and I have pointed out.169 According to the Vedic Brāhmaṇa texts, this empowers the woman to influence the ritual, hence the workings of the cosmos, in uniquely gendered ways, while not substantially affecting the patriarchal social order.
The nature of śakti as a factor in possession must be contrasted with the classical Tamil concept of aṇaṅku, a term that does not appear in ethnographies, but is nonetheless related to śakti and its modern deployment. Indeed, it is entirely possible that the concept of aṇaṅku, at least certain aspects of it, influenced the discourses of śakti and possession in Sanskrit. The term aṇaṅku was brought to general Indological notice by George Hart in his 1975 book, The Poems of Ancient Tamil. Hart gives this term the broad meaning of “sacred power.” However, upon review of the usages of aṇaṅku in early Tamil texts Kamil Zvelebil amends it to a “fear-provoking divine or demoniac force … [a] hieratic, sacred power, which could be dangerous but not always malevolent.”170 Zvelebil continues:
The functioning of aṇaṅku in human recipients represents often the true power of possession which caused particularly in women a kind of “sickness” (nōy) which had to be dealt with, controlled, removed. However, apart from this possession by aṇaṅku which must have been regarded as dangerous and unwanted, the same sacred force emanating from Murugan caused obviously a sacred possession, a hieratic trance which was welcomed and very probably self-induced.171
Thus aṇaṅku is the substantial and immanent means through which divinity, or any other power, is localized, whether it occurs in objects, places, or persons.172
The phenomenon of “localization” was explicated by David Shulman, who demonstrated that each Tamil shrine “sees itself as the only center of the universe, the one spot that is directly linked to heaven and the nether world.”173 This extensiveness and localization of divinity, though not a concept actively enunciated in the early Tamil texts, was an assumed part of south Indian religion and much more of a force in central and north India than Shulman recognized.174 Shulman’s statement is, after all, a strikingly succinct summary of shamanistic idealism, which has been shown to be distributed throughout the world. Although it is possible that the recognition in Tamil sthalapurāṇas of the multilocality of divinity influenced the acceptance of this notion elsewhere in the subcontinent, it is also possible that the notion was at least partially derived from vedic sources, which are also often quite specific about the localization of divinity. Nevertheless, the acceptance of the multilocality of divinity must surely have contributed to the notion that the divine, or any other less benign force, could penetrate and divinize any receptive individual. In this way, the individual—the person in the form of the body—became a sacred site, confirming the Toda conception of sacred space noted by Murray Emeneau in 1938: “the ‘sacred place’ and the ‘god’ are the same thing.”175 It is remarkable that this notion of the coextensiveness of divinity and place is preserved among the Toda, a nonliterate tribal group that speaks a Dravidian dialect, almost two millennia after the composition of the early Tamil literature. The assumption of localization, immanence, and transportability of divinity, possibly a dispersed relic of early Tamil religious thought, thus became a background metaphysic for the widespread phenomenon of possession in South Asia.176
It is also entirely possible that this phenomenon of localization of divinity was more applicable to women’s lives than to men’s. Most of the religiosity that survives in early Sanskrit literature is men’s religion. Men were the priests, the official mediators between the gods and the rest of the human species; they were largely responsible for the construction of temples, as inscriptions of all ages testify; they constituted the vast majority of renunciates; they were the majority of teachers and yogins. However, it is unreasonable to assume that, in such a culture, women did not also assert the sacredness of their domain, nor is it reasonable to assume that they would cede full control of that sacredness. If I may be permitted to project backward from the modern ethnographies, I would speculate that, for a millennium or two at least, possession has had a presence in women’s religion. Indeed, evidence from the Bṛhadāraṇyaka Upaniṣad, discussed in Chapter 5, permits this. It was an ideal strategy for localizing divinity and the social and religious power incumbent in that divinity, both of which were otherwise in the dominion of more socially and economically privileged men.
This fluidity of role evident in the woman’s assumption of śakti (or aṇaṅku) in possession brings up one of the major issues involved in thinking about possession, one to which Boddy refers in her definition: that possession exposes the fluid and permeable nature of personal identity. I say more about this below, but here I might pose a few of the questions that require attention later on: Is the individual still an individual when possessed? Is possession a state in which both the individual and the ethereal possession agent (deity or spirit) are fully attributed? If so, is the fabric of the individual then destroyed? Is it a mistake to assume an autonomous individuality in the first place?
It is useful here to review Marriott’s notion of “substance-code.” This term indicates a diversity abiding within, and even constituting, a unity or inseparability, a common locus of “codes,” for example, puruṣa, norm, mind, and spirit, and their “substance” counterparts, prakṛti, behavior, body, and matter, respectively. Particles of substance codes mutate and recombine with particles from others, thus manifesting new ones. Not only is this thesis closely related to anthropological theories of the self as a system of symbols and social meanings that is best articulated semiotically,177 but it can be viewed as a restatement of the brahmanical bandhutā (linkage) that will be discussed below. How is it that Viṣṇu is the sacrifice or dawn the head of the horse? “Such ancient and widely held Indian assumptions,” says Marriott, “if abstractly stated, seem compatible with the theories of modern natural scientists, although they conflict with common Western popular beliefs in standard, stable entities and in the normally impermeable, autonomous person.”178 Marriott recognizes that:
persons—single actors—are not thought in South Asia to be “individual,” that is, indivisible, bounded units, as they are in much of Western social and psychological theory as well as in common sense. Instead, it appears that persons are generally thought by South Asians to be “dividual” or divisible. To exist, dividual persons absorb heterogeneous material influences. They also must give out from themselves particles of their own coded substances—essences, residues, or other active influences—that may then reproduce in others something of the nature of the persons in whom they have originated. Persons engage in transfer of bodily substance-codes through parentage, through marriage, and through other kinds of interpersonal contacts.179
Although Marriott has applied this concept to caste transactions, the question “How does possession work?” can be answered in part through application of this concept. Dividual persons, entities, or even concepts transfer parts or essences of themselves, in whole or in part, willfully or by force, to other dividual persons, entities, or concepts. A case of transfer of substance-code in the anthropological record occurs in Obeyesekere’s study of the Pattini cult, in which transfer of the goddess’s essence (Simhala diṣṭi; Skt. dṛṣṭi; glance) is a state called ākarṣaṇa, translated by Obeyesekere as “magnetism,”180 distinguished locally from possession, āveśa or āruḍha (mounted).181 The use of Sanskrit terms here is significant, because this distinction is not found in Sanskrit texts, where ākarṣtṇa, (attraction) is one of the “six acts” (ṣaṭ karmāṇi) of left-handed Tantra.182 At any rate, the point is that ākarṣṇa, in Obeyesekere’s sense, becomes, in Marriott’s, a physiological substance that is transacted during ritual action. The goal of this transaction is to integrate the dividual individual’s activated substance-code with a similar, though more powerful, substance of the divinity. In so doing, rather as a byproduct, incompatible or negatively charged substances are surmounted or removed.
Nabokov critiques Marriott’s view by stating that, in her experience, Tamil ritual does not serve an integrative or reintegrative function, but fragments the participants “to the point of splitting them apart.”183 This is a serious criticism, in spite of Nabokov’s (correct) contention that her “research seems to rather corroborate Marriott and Inden’s ‘dividual’ model of the Hindu person.”184 Her view is that the ritual she observed, which had a major component of possession by deceased relatives, left the individual, in the end, in a state in which she (usually) was forced to dis-integrate psychologically in order to reproduce her received roles and patterns within the dominant culture of oppression. Without the benefit of the thick description that modern ethnography offers, it will become clear that possession in classical Indian literature supports Marriott’s views more readily than Nabokov’s.
A major feature of possession known from ethnographic literature is its frequent occurrence in either religious drama or religious ritual (or both together). One need look no further than the recent work of Alf Hiltebeitel and Richard Frasca on South Indian street theater,185 Kathleen Erndl’s work on goddess worship in Panjab,186 and William Sax’s work on the Pāṇḍav Līlā in Garhwal to recognize the prominence of possession in contemporary religious drama.187 In the Pāṇḍav Līlā, for example, the actors or dancers are regarded as possessed by the characters whom they represent, such as Arjuna, Kṛṣṇa, or Draupadī. Indeed, as in the Kṛṣṇa and Rām līlās in Vrindaban, Varanasi, and elsewhere, the deities are themselves believed to be physically present, temporarily but fully manifested in the performers. This display of power and presence of the sacred permits the deities to be worshipped directly. Also documented is possession judged as higher and lower, positive and negative, not at all unlike the general distinction noted above between ā√viś and pra√viś in “Sanskritic possession.”
Schoembucher recognizes similar categorical pairs in the anthropological record: “controlled and uncontrolled possession, induced and spontaneous possession, … desired and undesired, divine and demonic possession.”188 As an example of this, Hiltebeitel writes, “In Draupadī cult contexts, drunkenness and narcotics are thought of as inducing a kind of impure, or ‘polluted’ possession, one which is essentially demonic and stands in opposition to higher forms of possession which become vehicles for bhakti.”189 A feature of possession common to many of the ethnographies is play. The tendency toward the ludic in South Asian religion is strong,190 and Erndl shows in her book that the goddess “plays” with her devotees in the form of pavan (wind), while Wadley and Sax show that the possessing spirits or deities enjoy “dancing” in the bodies of their temporary hosts. On possession and play, Tom Driver writes,
Possession, in my view, has also to be regarded as a kind of serious game.… [T]he seriousness and the truthfulness of spirit possession does not mean that it involves no role-playing, as is made obvious by the fact that costumes and props are made ready ahead of time, prepared for a panoply of spirits that is just as recognizable as the dramatis personae of any familiar script.… In spirit possession, “playing for keeps” is escalated to a very high level.191
Another feature, noted by Sax, is “oracular possession,” a phenomenon discussed below. Sax writes:
(1) that such possession often, but not always, runs in families, usually from male to male; (2) that according to popular stereotypes, the oracle has no choice in the matter, and also cannot remember any details of his (or very rarely, her) trances; and (3) that discussions I have had with friends who are regularly possessed lead me to suspect that in many cases the oracle experiences an intrusion from outside of another consciousness or being but also adds his own elements of manipulation and self-conscious stagecraft in order to achieve a convincing performance.192
Susan Wadley writes succinctly in a slightly different vein, in order to distinguish between an oracle and an exorcist: “Basically, an oracle is possessed by a spirit and speaks while possessed; the exorcist himself is not possessed and therefore is not an oracle.”193 About the latter there can be no disagreement; but the label “oracle” for anyone who speaks while possessed is much too broad and certainly inexact for the material we consider. A. W. Macdonald also contributes to conceptually locating oracular possession. In describing why he so delicately translates the Nepali word jh
kri (var. jhāṃkri), a magicoreligious specialist often regarded more broadly as a shaman, as “interpreter of the world,” Macdonald provides an excellent description of spirit mediumship in South Asia. “Interpreter of the world,” says Macdonald,
avoids the pejorative connotation “conjurer” or “wizard.” These latter terms are generally associated with irregular activities that either stray slightly from the societal norms or that are blatantly anti-social. The jhkri, however, appears to be the very vehicle of a certain Nepali traditionalism. He is a person who falls into a trance, during which time voices speak through his person, thereby enabling him to diagnose illnesses and sometimes cure them, give advice for the future and clarify present events in terms of their relationship to the past. He is therefore both a privileged intermediary between spirits (who cause and cure illness) and men; between the past, present and future; between life and death, and most importantly between the individual and a certain social mythology.194
Another form of ethnography is the religious or spiritual biography in which possession is explained for the edification of an audience or student as a reproduceable benchmark experience.195 In a contemporary report of a Tantrie practitioner known as the Aghori Vimalananda, written by Robert Svoboda, the Aghori disregards the word “possession” as well as the common Indic terminology discussed here, and speaks instead of avishkara (Skt. āviṣkāra; manifestation, sudden appearance), which he equates with the Hindi baithak (baiṭhak; seat) and the Urdu hazri (hājiri; presence).196 He employs these terms to denote a state in which a deity is invited into the body of the tantric practitioner, to his or her benefit.197 Vimalananda tells Svoboda, “By long worship your subtle body will actually take the form of the deity you are worshipping.”198 In addition to this practice, in which a full manifestation of a deity is produced through spiritual practice, “[t]he deity can enter into someone’s body, and you can worship Him or Her in that way.”199
After reviewing various definitions of possession, methodological approaches including those that study possession (1) as a negatively inspired religious phenomenon, (2) as a psychological event, (3) as a social or sociopolitical event that trades in power relations, (4) as a type of shamanism, and (5) as a “real” incursion of spirits or deities to be taken at face value and addressed with a full cosmological arsenal, and after examining possession as a gendered, situated, and performative experience, we are in a position to move forward in our study. As should be clear from the above remarks, I am not bound to any single methodological approach; indeed, I find all of them problematic, especially for a study that emerges for the most part from the literary record. Most of this record does not satisfy the requirements of thick description, as this was never the intention of the authors. Furthermore, possession as mentioned and described in India is exceptionally diverse in its perspectives and dates. Therefore, we cannot expect a unity of vision to emerge, nor a single methodological approach to work for all the material.
As a classicist and Sanskritist, it is my “natural” mode to resort to text-critical method. This is not, as some readers might suspect, a regressive or default tactic. The methods of Indology, following Lariviere, prove productive for this enterprise. However, I supplement this “natural” inclination with insights from other fields and do not hesitate to employ observations from comparative religion, ritual studies, and a variety of anthropological approaches if they help elucidate the material at hand. This is evident in the next two chapters, which, like the current chapter, deal with modern and contemporary phenomena.
NOTES
1. A recent exception is Padoux 1999.
2. Boddy 1994 cites 221 anthropological and ethnographic studies of possession, and this is far from complete as she cites only English-language sources.
3. Humphrey and Laidlaw 1994:261. The authors add, with respect to the notion that the purpose of ritual is cultural communication: “Our theory suggests that this is a mistake” (ibid.).
4. I am grateful to Scott Schnell, an anthropologist, for this observation.
5. I return to this theme periodically. See also Kapferer 1997:83, on the attraction to the anthropologist of sorcery in general. For a very different reaction, see Nabokov (2000), who admits to being so intrigued by possession and a séance that she witnessed one night at the beginning of her fieldwork in Tamilnadu that she spent the remainder of her research year attempting to understand what she had seen (pp. 3–4). By the end of the year, however, she claims to have “acquired none of the fascination that anthropologists working on possession sometimes develop for states of trance” (pp. 179–80). Her illuminating book provides the reasons for this; indeed, the reasons are embedded in the title, Religion Against the Self.
6. See the work of Biardeau, including her 1989a primer on Hinduism; also Knipe 1991 and Fuller 1992 for introducing anthropological material into introductory books on Hinduism. Fuller presents an especially good summary of deity and spirit possession in popular religion (1992:231–236).
9. Ibid.:7–8. Bourguignon is strongly critiqued in Levy et al. 1996:17ff.
12. Rouget’s definition of possession as a “trance of identification” is too broad and too vague to serve as a starting point, though, as seen in discussions of bhakti, in South Asia identification with a deity is often regarded as a variety of possession; cf. Rouget 1985:25–28, 325.
14. E. Turner 1992a, 1992b, 1993; and Halperin’s (1996) important critique of Turner. I would include Humphrey and Laidlaw in this group except that possession is not their primary object of study.
17. Ferris 1999:35. This might be juxtaposed with a quotation from Aldous Huxley:
No man, however civilized, can listen for very long to African drumming, or Indian chanting, or Welsh hymn singing, and retain intact his critical and self-conscious personality. It would be interesting to take a group of the most eminent philosophers from the best universities, shut them up in a hot room with Moroccan dervishes or Haitian Voodooists, and measure, with a stop-watch, the strength of their psychological resistance to the effects of rhythmic sound. Would the Logical Positivists be able to hold out longer than the Subjective Idealists? Would the Marxists prove tougher than the Thomists or the Vedantins? What a fascinating, what a fruitful field for experiment! Meanwhile, all we can safely predict is that, if exposed long enough to the tom-toms and singing, every one of our philosophers would end up by capering and howling with the savages. (1952:369)
This is no doubt an extreme statement that might be contrasted with Nabokov’s opposing personal view cited in note 5. For an excellent anthropological self-disclosure, see Stoller and Olkes 1987.
18. Schoembucher 1993:242.
19. See Stoller 1998, for a discussion of several approaches: universalist (Horton, Sperber, Lévi-Strauss, Ernest Gellner), relativist (Tambiah, Geertz), and phenomenological (Husserl, Schutz). “Relativists complain that universalists are insensitive, eurocentric, or even racist. Universalists chide the relativists for their scientific naïveté and epistemological imprecisions” (p. 248), while phenomenologists tend to be too ahistorical. Stoller concludes that “the phenomenological approach might lead us to more embodied rationalities that are intellectually and personally transformative” (p. 240).
20. For an excellent account of the “hegemonic” Protestant interpretative ethos in studies of classical India, see Schopen 1991. For a reminder of just how thoroughly these ideas still dominate our thinking, see Tambiah 1990.
21. Humphrey and Laidlaw 1994:261. Elsewhere they write: “anthropologists engaged in symbolic analysis often singularly fail to provide realistic information about what people actually say about such symbols, no doubt because the actors so frequently reply something along the lines, ‘I know this means something, but I don’t know what it is.’ Anthropological interpretations have commonly covered up such unhelpful replies in order to produce a reading from their own deductions” (p. 180). And “we would claim that an across-the-board analysis, with no attention to the cognition of real participants, produces an occluded and ontologically unclear level of analysis” (p. 181).
22. Appadurai 1993:333–334.
23. Bell 1992, esp. ch. 4–6, reminds us of the dangers of positing general theories of ritual: all theorizing, she insists, must be deeply contextualized.
26. Will 1991:9. Wade strongly discourages hypnosis: “Don’t even watch it” (p. 31). He refers to spiritualism as an “ancient monster” (p. 72) and states that, “possession results from submission to the suggestions of an evil spirit” (p. 41). It is not necessary for me to recount here the history of Orientalism or the extensive scholarship on it in the past thirty years.
27. Originally: Genealogie der malabarishen Götter. Aus eigenen schriften und briefen der heiden zusammengetragen und verfasst (Madras, 1867). English translation: Genealogy of the South-Indian Gods; a Manual of the Mythology and Religion of the People of Southern India, Including a Description of Popular Hinduism (Madras, 1869; New Delhi, 1984). A more exact translation of the title would be: Genealogy of the Malabar Gods, Compiled from My Own Writings and Letters of the Heathens. See Sweetman 2003:104-126, for an account of Ziegenbalg’s life and a number of translations from his work, some of which appear shockingly heavy-handed to today’s reader. See also Jeyeraj 2004 for a complete translation of Ziegenbalg’s original manuscript along with a thoughtful analysis.
28. Fountain 2000. See also Cuneo 2001, for an account of the rise in exorcism in the United States in the past several decades, which, he says, is due to the influence of popular culture, e.g., the film The Exorcist, increasing Protestant fundamentalism, and the acceptance of exorcism as an alternative to conventional psychotherapy for those who are wounded, addicted, or sexually abused. This fits with some of the observations of Lewis discussed below, as well as with my own observations in contemporary India (Chapter 12).
30. I am not claiming here that there was an unbroken continuity between Ziegenbalg and McClintock or that the latter was typically “orientalist.” Nor do I intend here to equate Protestantism with orientalism. Indeed, most orientalists (in both senses of the word) were not Protestant missionaries.
31. Cf. Certeau 1984; Ginzburg 1989. The latter is an excellent historical example of this process of negotiating.
32. Catholic and Protestant churches had slightly different formal stances on the issue of possession. Nevertheless, they were alike enough with respect to what they encountered in South Asia to conflate them here without fear of undue essentializing.
34. Ibid.:569. Thomas also explains that the spread of enlightenment rationalism in English Protestantism helped bring witchcraft persecutions to an end, notably with the repeal of the Witchcraft Act in 1736 (p. 535).
37. Nevertheless, many Protestant denominations, including mainstream Methodists, speak of “being filled with the spirit.” Furthermore, Pentacostalists “speak in tongues,” while certain Catholic saints experienced stigmata, all resonant with different manifestations of āveśa.
38. Levy et al. 1996:15. With respect to the Pacific islands, they note that Christianity “radically altered the cultural framework for numinals, selectively overthrowing or demoting the gods, replacing a rich diversity of conceptualizations with an undifferentiated residual and negative category of ‘devils’ or ‘demons’” (pp. 24–25). Missionary attempts to impose this in India met with decidedly less success. For comments on the opposite effects of colonial Christianity and modernity on possession cults, see the articles on possession in Southeast Asia in Mageo and Howard 1996.
39. See Ram 1991:85–105, passim, for comments on possession and healing in a Christian community of Kanyakumari District in southern Kerala. The discourse, language, and actions are little different from those of Hindu possession, reflecting (at least in this case) little political involvement in religion by the ancient (Syrian) Christian Church in Kerala.
40. Cf. Claus 1984:60f; Lewis 1989:18ff.
42. Masson 1976:615 (1999:238).
43. For example, Masson’s statement “I cannot help viewing asceticism … as a defense, not a freely chosen position” (196:620; 1999:243). Masson fails to recognize that the reasons for asceticism in India can be, and always have been, myriad and that freely chosen positions are probably never free of conditioning factors. Procrustean characterizations like this close off the possibility of truly engaging the Other.
44. Ervin et al. (1988) and Simons et al. (1988) take steps toward establishing formal research criteria: Simons et al. show that, in Thaipusam, a possession ritual performed in Southeast Asia, based on South Indian precedents, a training regimen is in general a prelude to the experience of possession; in other words, it is a learned behavior. While seeking to understand the mechanisms of trance physiology, the second study confirms the findings of the first one: that the physiological responses to possession are sharpened through repetition, which is to say through practice, a conclusion discussed further in Chapter 4. The conclusions of Ervin et al. are that muscle tone is increased, that except for the slow pulse the state resembles hypoglycemic shock and that the recovery period of a few minutes is characterized by headache, photophobia, and confusion. “Early in training,” they state, “trances are characterized by poorly organized motor hyperactivity and occasionally the display of intense emotion. With experience, the motor patterns become more coherently organized with few affective displays. In both cases, trances are accompanied by amnesia. This state is suggestive of other situations in which the limbic system dominates the behavioural program” (Ervin et al. 1988:277).
45. An excellent and sympathetic introduction to cross-cultural psychoanalysis, which outlines the theoretical assumptions behind psychotherapy, is Frank and Frank 1991, esp. ch. 5. They state, with respect to their method: “Practitioners of bio-medicine generally refuse to take seriously the evidence that healing can occur through procedures involving the paranormal or supernatural. In seeking to maintain objectivity, we shall try to navigate between the Scylla of scornful scepticism and the Charybdis of gullibility” (p. 88).
46. Castillo 1994a:2. Much of my discussion here is summarized from this excellent article.
47. See Lewis-Fernández 1998, for a critique of what was added to or deleted from the latest edition of the DSM (American Psychiatric Association 1994). Lewis-Fernández was a member of the subcommittee of the American Psychiatric Association that recommended these changes. He notes: “Very prevalent phenomenologies such as ‘possession trance syndromes,’ amok, ‘falling out’ and ataques de nervios, among several others, still did not have an appropriate nosological niche in the manual” (p. 388). He also notes, as do many others, “the inherent normalcy of most dissociative states” (ibid.), adding that specific diagnostic criteria that would enable the clinician to distinguish between normal and pathological dissociation was also lacking.
49. Jonas (1965), a practicing psychoanalyst, links demonic possession with epileptoid phenomena: “In the past, many afflicted individuals thought they had been possessed by the devil or the holy ghost, and it is quite probable that many of the visions and voices described by mystics and prophets were not triggered by hysterical disassociations but epileptic equivalents” (p. 48, cited by Weiss 1977:160). Weiss’s Ph.D. dissertation, now being revised for publication, has analyzed Indian categories of unmāda, “insanity,” and exogenous disease-causing spirits in terms of Freudian categories of psychological pathology (pp. 112–186). Whatever might be said of Weiss’s work (and more is said at length in Chapter 12), it stands on firmer ground than some of the early “pseudo-psychiatric” studies he cites (as bad examples of the psychoanalytic trade), which grandly (and grandiosely) diagnose Indian culture at large, such as Berkeley-Hill 1921 (cf. Weiss 1977:135n5). For more grand theories, see Oesterreich 1930.
50. Significantly, the New York Times reported in 1999 that, “the Vatican again urged bishops and priests not to confuse psychological suffering and possession, and to seek medical help while at the same time offering spiritual consolation” (Tagliabue 1999:4).
51. Tambiah 1990:134–135.
52. Cf. Vallabhācārya, Antaḥkaraṇaprabodha, and commentaries.
53. Cf. Kakar 1982:53–88. Most of the clientele at the temples and facilities in Mehndipur are from first-generation literate families with strong memories of the lifeways of oral cultures (cf. the work of Luria, cited in Ong 1982).
55. Pakaslahti writes (personal communication, October 2001) that he disagrees with Kakar’s view that “possession phenomena in Balaji are manifestations of hysteria or hysteric personality. Such a view—not rare among Indian doctors—is a ‘pathologization’ of the native therapeutic processes and phenomena. It overlooks the cultural construction of possession in Balaji which provides a specific language, the forms and the method of spirit-possession-as-therapy.”
57. Ibid.:238. Like most other researchers on possession, I find Lambek’s positions extreme. Morris states the case with greater subtlety: “For only in the wake of a submission to the fictions that mediums themselves inhabit can one ask what work forgetting performs and, thus, how mediumship works” (2000:125).
58. See Stephens’s 2003 book on the Christian clerical misogynistic preoccupation with stories of sex with demons in the fifteenth and sixteenth centuries and their consequent crises regarding the reality of not just demons, but God as well. This account strengthens my own sense of the emic nature of Western scholarship in the work of Lambek and other anthropological witch hunters. They have, it seems, unwittingly taken up the mantle of their clerical forebears, assigning it a place under the rubric “hermeneutic of suspicion,” then superimposed it onto their anthropological subjects. This hermeneutic has no doubt advanced their (and our) scholarship, but sometimes the results are little different from what they were hundreds of years ago: a totalization of their own ontological worldview and a privileging of it over that of their subjects. This has emerged as part of the dominant and still very Christian scientific paradigm, something that most in the modern Western academy have failed to problematize.
59. Bourguignon 1976:10. Note her inclusion of dissociation as a subset of hysteria.
64. Cf. DSM-IV:487, #300.14. Dissociative Identity Disorder is defined as follows:
A. The presence of two or more distinct identities or personality states (each with its own relatively enduring pattern of perceiving, relating to, and thinking about the environment and self).
B. At least two of these identities or personality states recurrently take control of the person’s behavior.
C. Inability to recall important personal information that is too extensive to be explained by ordinary forgetfulness.
D. The disturbance is not due to the direct psychological effects of substance (e.g., blackouts or chaotic behavior during Alcohol Intoxication) or a general medical condition (e.g., complex partial seizures. Note: In children, the symptoms are not attributable to imaginary playmates or other fantasy play.
65. DSM-IV:729; 490f., 727ff; also Lewis-Fernández 1998:389f. on formally establishing the category “possession trance disorder.”
70. Some of the Sanskrit stories discussed in the following chapters deny the latter. It is not uncommon to find references to a body vacated by its conscious controlling agent after this conscious agent has entered the body of another being. This can lead to ludicrous consequences, cf. the discussion on the Bhagavadajjukāprahasana in Chapter 8.
73. Harper 1963:166. An important article on the value of Western psychoanalytic categories and psychiatric labeling in cultures in which the norms of mental health (and illness) may be different from those in the Western cultures that produce these categories and labels is Murphy 1976. Murphy’s views are accepted uncritically by Weiss (1977:5). This is problematic and is dealt with in Chapter 12.
76. Honko himself admits that the epic, its oral tradition, and the social forces and functions surrounding them are both too complex and too localized to offer such unilateral interpretations (1998:147). Antti Pakaslahti described to me a similar scenario in north India:
I have taped a long conversation between an Indian patient and Indian psychiatrist. More precisely it was not really a dialogue but a composition of two intermittent monologues, quite surrealistic, at times like an Ionesco play. The patient talked about spirits, devotion, and rituals, the psychiatrist about hysteria, depression and therapy. Neither understood what the other was saying. The doctor considered the patient to be an uneducated believer in black magic and supernatural beings. The patient was wondering what on earth the psychiatrist was trying to say with all his complicated, strange words which made no sense to him. (personal communication, October 2001)
77. Freud 1955[1919]:224; 1972:235.
78. Bargen 1997:21. This is one of the few studies of possession in classical Asian literature that I have been able to find. Consistent with the findings of many anthropologists, Bargen states that, in the Tale of Genji, “spirit possession is an individual woman’s response to pervasive discontent rather than the inherited affliction of a single person, family, or clan” (p. 246).
79. Freud 1955[1919]:234; 1972:246.
80. Freud 1962:242; see also Mayes 1998:101.
81. Cf. Freud 1959[1923]:446: “Cases of demoniacal possession correspond to the neuroses of the present day.”
84. See note 41. Masson’s final assessment of asceticism is that it is “unconsciously active sadism that expresses itself in extreme piety or even gets turned back against oneself” (1976:625; 1999:247). That this speaks more about Masson than about asceticism as a vigorous Indian socioreligious phenomenon was revealed almost two decades later in his stinging critique of his father’s relationship with Paul Brunton (1993). Cf. also Kripal 1995:303, 358n71.
85. A recent far-reaching model proposed by Mayes shows promise. Capitalizing on Freud’s notion of the externalizing primitive and internalizing civilized, discussed above, she discusses possession as a “primal fantasy of being moved or contained from without [which] may have either a negative or positive interpretation: at one pole is a desired state of rapturous surrender to external interpenetrating management, while at the other is a feared state of enforced submission to an invasive alien power” (1998:103).
86. Freed and Freed 1964:152–171.
87. See esp. Obeyesekere 1970:97–111; 1977; 1981.
88. Kripal 1995:39, 271, 335, 340n63, 357n46. See also Vaidyanathan and Kripal 1999, which is helpful because many of its essays actively engage the critical question of the validity of Freud’s psychoanalytic enterprise to Indian and Hindu cultures. The most important chapters dealing with this issue in this collection are by Parsons, Ramanujan, Kurtz, and Kakar. All of these essays can be called comparativist in the true sense of looking at cultures that are constituted differently, and they all suggest ways in which Freud must be emended in order for his theories to apply to India.
89. Castillo 1994b: 142–152.
90. Although the review of the book by Svāmī Ātmajñānānanda (1997) is meritorious on many counts, there can be little doubt that the Svāmī’s engagement in the Ramakrishna Mission colored the review. Svāmī Ātmajñānānanda’s suggestions have been taken to heart by Kripal in the second edition of his book (1998). Certain other reviews have been more favorable; see, for example, the reviews by Patton (1996b); Haberman (1997): McLean (1997); and Padoux (1999). The most important and illuminating of the positive reviews, however, is in a review article by Parsons (1997), which illustrates the benefits of Kripal’s comparativist approach, especially when compared to two other seriously flawed psychoanalytically based biographies of Ramakrishna, by Sil (1991) and Kakar (1991). Among the less-positive reviews of this book, the better are by Urban (1998) and Walker (1998). See also the exchange between Larson and Kripal: Larson’s challenge in JAAR (1997), and Kripal’s response (1998). Kripal defends himself against the charge that the book’s final conclusions are monocausally reductive. But Larson was not the only reviewer to level this charge; Walker and Urban would agree with Larson.
91. Echoing what has been stated in certain reviews of the book (cf. Urban 1998), I might also criticize the book, though mildly, for its lack of attention to historical context. Kripal never asks why, in nineteenth-century Bengal, its prosperity punctuated regularly by oppression, floods, epidemics (notably a malaria epidemic in the late 1860s), famine (a major pan-Indian famine in the 1870s, which sent birth rates plummeting), not to speak of the usual “fever” and cholera, most other males (Bengali or otherwise) did not become homosexual or resort to ecstatic behavior. In fact, the average life span in India—Bengal was no exception—was thirty-six years according to the census of 1881, a statistic that was computed after eliminating infant and early childhood mortality. When adjusted for these factors, the average lifespan dropped to twenty-three years. Thus it is a safe guess that most men lost their fathers at a young age. Although this is a minor point given the overall scope of Kripal’s book, it is important to point out, if for no other reason than as an argument for more broadly based and interdisciplinary studies of issues in the history of religions. Cf. Report on the Census of British India (1883) pp. 142ff., 170ff. On Bengal as a historical center of prosperity, contrary to popular impressions, see Greenough 1982, ch. 1.
92. Caldwell 1999, esp. ch. 5.
93. Ibid.:228f.; cf. Castillo 1994b.
94. For perceptive comments on this problem, see Claus 1979.
95. Bhattacharyya 1986:71; see also 59–75.
97. Yet occasionally one finds cases that beg for psychoanalytic interpretations. Vijaya Ramaswamy (1997) had one fall into her lap, but failed to engage psychoanalysis. Ramaswamy writes of a contemporary saint in Tamilnadu, Andavan Pichchi (“Mad”) Amma, who at age fifty was dying when she was possessed by the spirit of Pinnavasal Swamigal, a disciple of the nineteenth-century mystic Sadasiva Brahmendra. This male possession of a female body, says Ramaswamy, “has led to biological problems for her. However it has also given her an in-depth knowledge of Vedanta and the Upanishads which Amma as an illiterate wife had no acquaintance with” (p. 6). I am ill equipped to venture a psychoanalytic interpretation, but any such interpretation would have to balance Amma’s biological problems with the fact that she was also healed of her mortal affliction.
98. Caldwell 1996:222n22.
100. Freeman 1993:134. Humphrey and Laidlaw observe that even “unconsciously,” possession bears ritual patterning: “The phenomenon of spirit possession trance also demonstrates this: people coming out of these trances commonly say they do not know what they were doing and their family does not hold them responsible for their words and actions. Yet during the possession trance people act in specific, patterned ways (the circling of the head and shoulders) and not in the other numberless ways possible. This indicates that even ritualized possession is in some sense ‘directed’ and not totally formless, incoherent activity” (1994:235).
101. Nuckolls 1991a:58. See also Nuckolls 1991b; and for a deeper view of some of the cultural phenomena behind Jalari spirit possession, see Nuckolls 1997.
102. Sax (1991a) reports instances of oracular possession in festivals in Garhwal, where people randomly rush up to a possessed individual to ask for help in solving their problems. This sort of description is absent from Sanskrit literature.
104. Nuckolls replicates the structure that Bell (1992) attributes to anthropologists generally: that the ritual resolves a contradiction.
105. For example, see Sharp 1993, which examines possession as a mode of political consciousness embedded in religious experience. Among the varieties of possession she examines, one is of mass possession in the public schools, largely of adolescent migrant girls (222ff.). In one instance of this, a powerful healer was consulted. He reported that the local ancestors were angry because the French paid no regard to the sacredness of the ancestral ground on which the school was built, moving and destroying tombs. The healer recommended the performance, on the school grounds, of a ceremony honoring the deceased ancestors, including the sacrifice of an ox. After this performance, the possession diminished considerably. It is difficult to imagine a school system in India that would accede to such a request.
106. See Diemberger 2005 for the role of female mediums in Tibetan politics. She writes: “Indeed, from historical sources we know that in ancient Tibet spirit possession had a central position in the political system and that women had an important role in embodying this institution” (p. 144). Diemberger’s excellent study appeared too late for me to take full advantage of it.
107. Claus writes, with respect to the compelling nature of these interpretations to academics, “Spirit possession is almost invariably viewed by anthropologists in psychological or sociological frameworks” (1979:30).
108. In my view, the works of Caldwell, Erndl, Freeman, Hiltebeitel, and Knipe excel here. See, e.g., Knipe 1989; also Sontheimer 1989b.
110. Hancock 1995:62. See Dernbach 2005, for a detailed argument that female possession on the island of Chuuk in Micronesia is both a discourse and a mode of cultural continuity and change.
112. Sax 1991a:41, 184, 194.
113. Heesterman’s grandest statement on this issue in Vedic ritual is his 1993 book, The Broken World of Sacrifice. Heesterman may have overhistoricized what the Vedic poets intended poetically or metaphorically, but his efforts are nevertheless noteworthy. See my review of this book (F. Smith 1995).
115. See Atkinson 1992 for a thoroughly documented account of the decline in the 1960s and subsequent revival of the concept of shamanism in ethnographic discourse. That the word shaman is possibly a derivation from Sanskrit śamaṇa (< śramaṇa), as Altaicists generally believe (though Tocharian is an option), does not help in determining the nature of shamanism as it is thought of today, even in India. See the discussion by Menges (1989), who supports the derivation, citing a letter from Sir Harold Bailey to Carmen Blacker (Bailey, cited in Blacker 1975:321–22n4) that traces the term through Prakrit, Saka, Tokharian, Sogdian “and other forms in Uigur Turkish, Asokan Greek and New Persian until it reached China as sha-men. Thence it made its way to Japan as shamon” (ibid.:23). In addition to the references cited here, see those cited in the articles on South Asian, Himalayan, and Tibetan shamanism in Walter and Fridman 2004:741–798.
116. Lewis 1989[1971]:57.
118. A critique against which he explicitly defends himself in the “Preface to the Second Edition” (Lewis 1989:10).
119. However, for a strong defense of Lewis, at least on this point, see Gellner 1994:27–48.
120. This is being increasingly recognized by scholars; cf. Gellner 1994; Humphrey and Laidlaw 1994; Pakaslahti 1998.
121. Gombrich 1988:36–37.
122. Cf. Eliade: “To be sure, shamans are sometimes found to be ‘possessed,’ but these are exceptional cases for which there is a particular explanation” (1964:6).
123. Bourguignon 1979:261.
124. Ripinsky-Naxon (1993), who has written a wide-ranging and dense distillation of the phenomenology of shamanism, argues for restricting the domain of shamanism to the realm of soul-journeying, while presenting almost unlimited variations on this theme. However, he does state, rightly: “It is easy to confuse folk beliefs and spirit possession with shamanism” (p. 67). Nevertheless, most modern scholars would admit possession into the realm of shamanism, which they regard as a rather open category.
125. Obeyesekere 1984:13–14. See also the excellent (and underutilized) study by Wirz (1954). This study recognizes in twentieth-century Ceylon some of the possessing “seizers” (graha.) that occur in the classical literature, especially the rakṣas, yakṣa, and preta.
126. Pettigrew and Tamu (1994:416–422) demonstrate that ecstasy and possession are uneasy partners in Gurung shamanism in Nepal. Related is Rouget’s criticism of the confusion of “ecstasy” and “trance” (1985:312). Trance, he says, is accompanied by convulsions, music, noise, amnesia, and an audience, while ecstasy is characterized by silence, immobility, and solitude. In the South Asian experience, the characteristics of both may be found alternating in the possessed (see the discussion of Knipe’s work on raudra and śānta in Chapter 8). See also Diemberger’s helpful critique of Lewis, in which she demonstrates that though Lewis’s gendered distinction between central and peripheral possession may be applicable in Tibet (male as central, female as peripheral—a distinction that is largely inapplicable in India), Lewis’s distinction between “lower” and “higher” does not work, largely because such hierarchical positionings shift rapidly in different, even neighboring, communities (2005:145-146).
128. Examples abound. See, e.g., Biardeau 1989a and 1989b; Hiltebeitel 1991; for Nepal see, e.g., Macdonald (1975), who translates the threatening ranting of a gaine (professional beggar/minstrel who is possessed by Bhat Bhateni), “a divinity whose altar is to be found not far from the Snow View Hotel in Kathmandu, a divinity famous for its murders of new-born children” (p. 116).
129. Sharp 1993:15; see also her excellent summary of anthropological perspectives on possession, pp. 14ff.
130. Cf. Shirokogoroff 1935:269.
131. Ibid.:274, cited by Maskarinec 1998:viii. See also Siikala 1982. Also relevant to the discussion is During Caspers (1992), who argues that certain inscriptions, masks, and other items of material culture from the Indus civilization reveal the presence of shamanism in that culture, probably as a survival from earlier, more characteristically shamanistic hunter-gatherer cultures farther north. Although it may be part of During Caspers’s project to argue for an Indus civilization shamanism based on likely regional antecedents, it is not part of my project to argue that possession as an aspect of shamanism is based on models from that earlier civilization. The evidence is simply too fragmentary and random to say any more than that it is possible.
132. Inglis (1985) argues that most of the possession in Tamilnadu that may be considered oracular is not shamanic, strictly speaking.
133. Other instances of gods descending into humans in a manner departing from the shamanistic norm may be cited, most famously the oracle at Delphi.
134. Sontheimer 1997:80, 106, 174, passim; 1989a:142–146. Stablein (1976), who studied Tibetan Buddhism in Nepal, also sees possession as shamanistic. In fact, he sees possession as the main connection between shamanism and Vajrayāna Buddhism, a theme taken up by Samuel (1993), but contested by Gombrich (1997).
135. Sontheimer 1989:142.
137. Srinivas 1998:178. Kakar also recognizes different categories of shaman, which to him, too, is a rather large category of spirit healers (1982:89ff., passim). Also important for shamanism in tribal cultures is Jones 1968. Jones’s primary influence was Eliade; he wrote before Lewis and his critics and was unaware of the article by Rubin discussed in note 138.
138. For an early enthusiastic article on shamanism in India, see Ruben (1940), who regards almost everything dramatic from the Ṛgveda and Avesta through early Buddhism to early twentieth-century ethnographies as manifestations of shamanism and argues for its presence in India. He writes of magical flight, magical dreams, animal sacrifice and totemism as suggestive of shamanism, drawing on a bewildering array of cross-cultural comparisons that would be forbidden in more recent scholarship. His work is a reminder of how difficult it was sixty-five years ago, and still is today, to say exactly what shamanism is.
139. Paper 1995:85. In this cautiously recommended general study of religion “on the ground” in China, Paper has devoted two chapters to “ecstatic functionaries in Chinese religion” (pp. 51–124). For a more thorough critique of this book, see F. Smith 2001a.
140. Claus 1979:29. Nabokov (2000: 187n2) comments lucidly on the use of the terms shaman, oracle, trance-therapist, and possession-medium by various scholars.
141. These texts remain untranslated from Bengali. I am indebted to Frank Korom for this information. The only notable account of the maṅgalkābyas is by Zbavitel 1976:156–169. See also W. Smith 1980.
142. For an interesting critique of Lewis, see Kjaerholm 1982: 185ff. Kjaerholm’s article is seriously flawed from an Indologist’s perspective, though he rightly (and vigorously) calls for an understanding of cultural representations of personhood as a prerequisite for understanding possession. More appropriate for the South Asian context is Jones 1976, which suggests four, rather than two, categories of possession: peripheral, tutelary, reincarnate, and oracular.
143. Geertz 1983:59. I agree with Flood’s methodological suggestions regarding a dialogue between scholarly and indigenous narrativization (1999:139ff.).
144. Goodman 1988. Curiously, this work is not mentioned in Boddy’s review article (1994).
145. See references in note 14.
146. See Cuneo’s discussion (2001:273ff.) on the existence of spirits based on his observation of more than fifty exorcisms in late twentieth-century America.
147. See Castillo 1994a:145, 156, and references therein.
149. See Claus 1979, 1984; Halperin 1996; also Nourse 1996.
152. This is true elsewhere, cf. Lewis 1989[1971], for extended discussions of this; also, e.g., Kendall 1985, 1988.
153. Cf., e.g., Erndl (1993), Harper (1963); and Nabokov (2000), all of whom nuance their studies differently while illustrating the utility of this perspective.
155. Most important here are Bell’s elaborations of Foucault’s notion of “ritualized agency” (1992:197ff.).
156. It need hardly be re-emphasized that “If there can be said to be such a thing as prelogical mentality, it can surely be found in our own midst and not just in Others” (Jensen and Geertz 1991:17).
157. See F. Smith 2001a and the books reviewed therein.
158. Trawick 1992:153–154.
159. As for the danger, see Harper 1969, for an account of the perceived danger of widows and more general fear of women among certain brahmans of coastal Karnataka.
160. See Caldwell 1999:217, for a summary of scholarly views on the relationship between women’s possession and sexual and aggressive emotions. Now see Diemberger 2005, who argues that female oracular possession in Tibet is an aspect of a peculiar “female competence” (p. 141). Women, in part because they are often outside the system of social ranking, in part because they have been largely responsible for child-rearing and in part because they “have a deep knowledge of the kinship-network as patrilocally married wives” (p. 143), are experts at mediating small-scale conflicts and threats to the coherence of the community. As such, it is part of their hereditary religious role to serve as oracles, a role also attested in India (and elsewhere).
161. Cf. Gellner 1994; Hancock 1995.
162. Hancock 1995:62. Although smārtas are followers of the advaitin Śaṅkarācāryas (especially of Kanchi and Sringeri) and are (in theory, at least) guided by the smṛti texts, they engage in popular goddess worship. As Hancock (along with many others) notes, “The Śaṅkarācāryas (all of whom are men) are said to have merged with the goddess” (p. 71). This is an old notion, as the attribution to Śarikara of dozens of hymns to the goddess testifies.
163. Egnor 1980:12–13. This excellent essay discusses, among other things, the relationship between devotion (bhakti), power (śakti) and feeling (Ta. uṇarcci).
166. Erndl 1996; see also Wadley 1976. The equestrian theme is a common and potent metaphor in possession cults throughout the world: a medium is mounted by a spirit who takes the reins of control. Halperin (1996:33) notes that in Afro-Brazilian trance dancing a medium is called a “horse,” upon which a spirit rides. See also Deren 1991; Matory 1993; and Chapter 4, n. 189.
167. The theme of gender role-reversal is well known from Victor Turner’s The Ritual Process (1969). This is also proximate to Foucault’s reflections on the agency and power of subjectivity and on resistant “technologies of the self.” By shaping their own subjectivity, Indian women in possession can potentially thwart, challenge, or question the ways that they have been shaped by men. See Ransom (1997:152) for an explanation of the political impact of this aspect of Foucault’s project.
168. Gold and Raheja 1994.
169. Jamison 1996:42ff. In demonstrating that tying a cord around the waist of the sacrificer’s wife is ititiatory, Jamison has argued against a position that I took in my 1991 essay, that the act of binding the wife was also an act of suppression. While I agree with Jamison’s conclusions on the initiatory aspects of the rite, which I also stated in my article (though not as forcefully as she), I also believe that her conclusions do not necessarily contradict mine.
170. Zvelebil 1981:15–16. Zvelebil examined roughly four dozen occurrences of this word in old Tamil texts.
172. Through examination of a broad sweep of Tamil texts, Rajam 1986 demonstrates that the term aṇaṅku has a much greater semantic range than Hart supposed. Rajam finds instances of the word in the senses of sexuality, undefeatable quality, a celestial woman, a male ascetic, a source of distress, the supernatural capacity of an ocean, and so on. Furthermore, she discovers that, when used for a woman’s force, it is not necessarily inherent or sacred. For a thorough discussion of the word, see also Dubianski 2000:6–19. Burrow, in his 1979 review of Hart 1976 defines aṇaṅku as “the possessing spirit” and “the state of possession” and cites the expression aṇaṅk āṭu (= teyvam āṭ, to dance under possession of a spirit).
173. Shulman 1981:55, esp. ch. 2, “The Phenomenon of Localization” (pp. 40–89).
174. See, for example, Chapter 4, pp. 140–142, on possession during the Debaddhani festival in Assam.
175. Emeneau 1938, cited by Zvelebil 1981:13.
176. This supports Daniel’s observations (1984) on the coextensiveness and “substantial” or “fluid” relationship, or even identity, between place and individual in south India.
177. See Daniel 1984; Singer 1984, esp. chs. 3 and 6.
178. Marriott 1976:110. See now the perspicacious remarks of Sax 2002:9–15.
180. The word ākarṣaṇa is used in Bengali to denote the power of place due to its habitation by a deity; a “magnetic pull.” I thank Frank Korom for this. The “magnetic pull” of a temple or deity is a familiar experience in India; thus it is possible that the Simhala usage reported by Obeyesekere encompasses this.
181. Obeyesekere 1984:14.
182. Cf. Bühnemann 2000; Goudriaan 1978:294ff.; cp. Türstig 1985:101ff. The lists of ṣaṭ karmāṇi in Türstig’s texts do not mention ākarṣaṇa. I am unable to locate āruḍha in the sense of possession in any Sanskrit text.
183. Nabokov 2000:15. Her entire discussion on this issue is worth examining.
184. Ibid. :14. See Marriott and Inden 1977.
185. Blackburn 1988; Frasca 1990; Hiltebeitel 1988, 1991; Upadhyaya and Upadhyaya 1984.
187. Sax 1991b, 1995. Other studies worth mentioning are, on Kerala, the work of Richmond and others in Richmond, Swann, and Zarrilli 1990; on Garhwal pilgrimage, Sax 1991a; on Rajasthan, Gold 1987. Also of interest is Ramanujan’s discussion of the relationship between possession and performance (1986a).
188. Schoembucher 1993:242.
189. Hiltebeitel 1989b.365.
190. Cf. Korom 1999; Norbeck 1974, 1976.
191. Driver 1991:177–78. In tones familiar to some of the issues we are grappling with, Driver notes that in Christianity there is a wide gulf between liturgists/ theologians and advocates of possession:
To “get the spirit”—that is, to become filled with the immediate presence of the deity—is the essence of the sacramental act, although this truth has been obscured, not to say suppressed, in those churches that have come to put high value on their social respectability and thus to shy away from experiences that are strongly antistructural, preferring the shelter of this world’s customs even in ritual. In these circumstances, the experience of possession (being filled with something) during a sacrament has been more or less banished, its place taken by an emphasis on symbolism. (pp. 208-209)
193. Wadley 1976:235; see also Dumont and Pocock (1959:55f.), who note that an exorcist or priest mediates from man to deity, but an oracle mediates from deity to man.
194. Macdonald 1976:310. The reader must compensate for the masculine pronouns here because, as Merz and many others have shown, most of the jhkri are dyah-mā, women. Perhaps the most complete account of exorcism in Tibetan Buddhism is Mumford 1989:140–162.
195. These personal experience narratives have not received the scholarly attention they deserve; cf. Korom 1997, esp. 570ff; 1999.
196. See Bellamy 2004 for an account of different kinds of hāzri at a healing site called Husain Tekdi in Rajasthan. She records the Hindi term vaḥ hājri bhar rahā haiṃ (he is filling that presence).
197. Cf. Nasir 1987. Like much else reported in this chapter, baiṭhak in Peshawar was the province of women. It is also very formal, includes songs, and is used for healing purposes. Vimalananda’s use of the term is nonspecific and does not reflect this context.