The Self Possessed: Deity and Spirit Possession in South Asian Literature and Civilization

By Frederick M. Smith

The Self Possessed is a multifaceted, diachronic study reconsidering the very nature of religion in South Asia, the culmination of years of intensive research. Frederick M. Smith proposes that positive oracular or ecstatic possession is the most common form of spiritual expression in India, and that it has been linguistically distinguished from negative, disease-producing possession for thousands of years.

In South Asia possession has always been broader and more diverse than in the West, where it has been almost entirely characterized as "demonic." At best, spirit possession has been regarded as a medically treatable psychological ailment and at worst, as a condition that requires exorcism or punishment. In South (and East) Asia, ecstatic or oracular possession has been widely practiced throughout history, occupying a position of respect in early and recent Hinduism and in certain forms of Buddhism.

Smith analyzes Indic literature from all ages-the earliest Vedic texts; the Mahabharata; Buddhist, Jain, Yogic, Ayurvedic, and Tantric texts; Hindu devotional literature; Sanskrit drama and narrative literature; and more than a hundred ethnographies. He identifies several forms of possession, including festival, initiatory, oracular, and devotional, and demonstrates their multivocality within a wide range of sects and religious identities.

Possession is common among both men and women and is practiced by members of all social and caste strata. Smith theorizes on notions of embodiment, disembodiment, selfhood, personal identity, and other key issues through the prism of possession, redefining the relationship between Sanskritic and vernacular culture and between elite and popular religion. Smith's study is also comparative, introducing considerable material from Tibet, classical China, modern America, and elsewhere.

Brilliant and persuasive, The Self Possessed provides careful new translations of rare material and is the most comprehensive study in any language on this subject.

• Language: English
• Publisher: Columbia University Press
• Release date: Aug 14, 2012
• ISBN: 9780231510653

Book preview

THE SELF POSSESSED

the self possessed

DEITY AND SPIRIT POSSESSION IN SOUTH ASIAN LITERATURE AND CIVILIZATION

Frederick M. Smith

COLUMBIA UNIVERSITY PRESS      NEW YORK

COLUMBIA UNIVERSITY PRESS

Publishers Since 1893

NEW YORK   CHICHESTER, WEST SUSSEX

cup.columbia.edu

Copyright © 2006 Columbia University Press

All rights reserved

E-ISBN 978-0-231-51065-3

Library of Congress Cataloging-in-Publication Data

Smith, Frederick M.

   The self possessed : deity and spirit possession in South Asian literature and civilization / Frederick M. Smith.

      p.   cm.

   Includes bibliographical references and index.

   ISBN 0-231-13748-6 (cloth : alk. paper)—ISBN 0-231-51065-9 (electronic)

   1. Spirit possession—South Asia. 2. Spirit possession in literature. 3. Sanskrit literature—History and criticism. 4. Tantrism—South Asia. 5. Spirit possession—Hinduism. I. Title.

BL1055.S63   2006

A Columbia University Press E-book.

CUP would be pleased to hear about your reading experience with this e-book at [e-mail].

As a reflection is to a mirror or other similar surface, as cold and heat are to living beings, as a sun’s ray is to one’s gemstone, and as the one sustaining the body is to the body, in the same way seizers enter an embodied one but are not seen.

SUŚRUTA SAṂHITĀ 6.60.19

This work is dedicated to the memory of the late Pratap Bhagwandas Mody of Bombay, who demonstrated artfully and convincingly that we are never quite alone.

CONTENTS

LIST OF ILLUSTRATIONS

PREFACE

ACKNOWLEDGMENTS

INTRODUCTION

Part I. Orthodoxies, Madness, and Method

1.    Academic and Brahmanical Orthodoxies

Sanskritic Culture and the Culture of Possession

The Sanskritic Vocabulary of Possession

Problematics of Interpretation

Part II. Ethnography, Modernity, and the Languages of Possession

2.    New and Inherited Paradigms: Methodologies for the Study of Possession

Classical Study and Ethnography

Definitions and Typologies

The Devil’s Work

Psychiatric and Psychoanalytic Interpretations

Possession as a Form of Social Control

Possession and Shamanism

Possession as Ontological Reality

Śakti, the Localization of Divinity, and the Possessed

Performative and Biographical Context

Conclusions

3.    Possession, Trance Channeling, and Modernity

4.    Notes on Regional Languages and Models of Possession

Lexicography, Languages, and Themes

Exorcists, Oracles, and Healers

Reflections on Folk and Classical in South Asia

Part III. Classical Literature

5.    The Vedas and Upaniṣads

Embodiment and Disembodiment Among the Ṛṣis

Possession in the Early Vedic Literature

Shape-Shifting and Possession

In the Beginning, God Possessed Heaven and Earth

Transfer of Essence

The Gandharva, the Apsaras, and the Vedic Body

6.    Friendly Acquisitions, Hostile Takeovers: The Panorama of Possession in the Sanskrit Epics

The Mahābhārata, Where Everything Can Be Found

Notes on Possession in the Rāmāyaṇa

7.    Enlightenment and the Classical Culture of Possession

Possession as Yoga Practice

Possession and the Subtle Body in the Yogavāsiṣṭha

Śaṅkara’s Possession of a Dead King

Possession and the Body in the Brahmasūtras

Possession in Indian and Tibetan Buddhism

Conclusions

8.    Vampires, Prostitutes, and Poets: Narrativity and the Aesthetics of Possession

Culture, Fiction, and Possession

Possession in Sanskrit Fiction

Can There Be an Aesthetic of Possession?

9.    Devotion as Possession

Devotional Possession in the Gītā and Ānandavardhana

Vallabhācārya’s Concept of Āveśa

Śrī Caitanya and the Gauḍīya Concepts of Āveśa, Avatāra, and Multiple Bodies

Āveśa and Bhāva

Āveśa, Bhāva, and Alternative Vedāntas

Part IV. Worldly and Otherworldly Ruptures: Possession as a Healing Modality

10.    Possession in Tantra: Constructed Bodies and Empowerment

Samāveśa as Tantric Realization

Discipline and Enlightenment

Divinizing the Body

Possession in Buddhist Tantras

Tantric Possession and Images of a Multiple Self

11.    Tantra and the Diaspora of Childhood Possession

The Śaiva and Buddhist Tantras and the South Indian Texts

Svasthāveśa and the Prasenā

Epigraphical Evidence for the Practice of Svasthāveśa

The Ritual of Svasthāveśa

Possession Across the Himalayas

Aweishe: The Indic Character of Chinese Possession

Svasthāveśa in South India

The Mantramahodadhhi

The Tantrarāja

Indian Āveśa and Chinese Aweishe: A Comparison

Conclusions

12.    The Medicalization of Possession in Āyurveda and Tantra

Disease-Producing Spirit Possession

Bhūtavidyā: Vedic and Āyurvedic Demonologies

Other Indic Demonologies

Piśācas and the Piśācmocan Temple

Childsnatchers and Therapy to Counter Demonic Possession (Piśācagṛhītabhaiṣajyam)

Healing and the Circulation of Knowledge

Possession and Exorcism in Contemporary Āyurveda

Diagnosing Possession

Conclusions: Notes on the Textuality of Āyurveda

13.    Conclusions: Identity Among the Possessed and Dispossessed

Variation and Vocabulary

Possession and Embodiment

Sudden and Gradual

Questions and (a Few) Answers

Bringing It All Back Home: The Mahābhārata and Traditions of Possession

BIBLIOGRAPHY

INDEX

LIST OF ILLUSTRATIONS

PREFACE

THE WRITING OF THIS VOLUME TOOK ME BY SURPRISE. I never envisaged it as part of my research program until it began to form a life of its own. Eventually, it grew from childhood to adolescence and, typical of adolescence, began to exact unreasonable demands on my resources, including time, place, and modes of thought. Having now achieved maturity, it demands to be set free. On the whole, I would have preferred to be in India translating Sanskrit and examining manuscripts, rather than working on a project like this that imposed on me a new and very different set of intellectual, psychic, and even physical demands. As it turned out, I was forced to examine worlds of thought and theory that I had always suspected lay in wait, less quietly than I appreciated, to ensnare me, while the project unceasingly transgressed boundaries I kept setting on it. Its conception and infancy—I thoughtlessly intended to abandon it in childhood—took the form of papers delivered at annual meeting of the American Academy of Religion in 1992, the American Oriental Society in 1993, and the Indic Seminar at Columbia University in 1994. My intention in these papers was to examine a few of the semantic issues surrounding some of the key terms for possession that I saw repeated in Sanskrit texts of many different periods and genres. I foolishly believed that I could accomplish this in ten, twenty, or thirty pages. Soon enough, however, I discovered the vast ethnographic literature of possession in India and became almost hopelessly entangled—and gridlocked—in the theoretical issues surrounding it. This is discussed in due course, but I must confess here that my reading of possession in modernity had a much greater impact on my reading of possession in antiquity than I had expected or desired. Instead of a paper on possession in antiquity—the initial scope of the project—this has become a work more generally on possession in South Asian culture over a long span of time.

After several years, at first intermittently, of data gathering and absorbing stories of possession, then reading and reflecting on theories of possession, and finally engaging it actively, I have arrived, with this book, at a meditation, a perilously intimate one, on personhood, which is sometimes, though not always, contiguous with selfhood. As the title of the book suggests, I find myself attempting to reconcile in this project the self, possessed, with a presentable veneer of self-possession. In this way, the final product has also become a meditation on embodiment and incarnation, gain and loss, transformation and transition, and tradition and imagination, which, my friend Robert Beer reminds me, must become the same thing (1988:9). However it began—and the raison d’être of scholarship is often contested, perhaps especially within the mind and body of the scholar him or herself—it was, upon reflection, inspired by the constant, elusive, and very personal conundrum of embodiment, by a sense of the irreducible strangeness of life, by the shock of an eternally mutating present and presence when we seek only past and future permanence, which is to say by the trauma and bewilderment of continuity when we seek resolution and termination. This was aided by a vision of the simultaneity of multiple selves clamoring for dominance, propriety, order, and voice as they succumb to the inexorable force of entropy, by dreams pushed aside incomplete and irretrievable by the disappointment of awakening, and by awakening to (and within) the disappointedness of dream. In short, the process of creating this book has been a long and complicated exorcism.

If my selection of material appears planned but extravagant, the reason is that the planning came to life as a learning process, like perfecting a rāga: I found a few unique scales and constantly improvised on them. Thus, the extravagance could never be exhaustive. The material turned out to be much more extensive than I initially expected. In many key places, in dealing with the Mahābhārata, Tantra, and bhakti texts, for example, I was forced to be illustrative and selective. As a friend, a veteran of many books, told me (paraphrasing W. H. Auden, as I recall) when I was about three-quarters done, this is the kind of book that cannot be completed but, instead, should be abandoned. The lesson for me was that both data and knowledge can be infinite, especially as they are swept up in an ever-expanding vision with ever-increasing dimensions and vocalities. The evidence of the multidimensionality and multivocality of possession that I have brought to bear on the topic is more than I had ever hoped to find or thought was even possible. Many readers will still say that I left out this or that, especially from ethnographies or modern autobiographies, or could have interpreted something differently, that I should have attended more to feminist perspectives or psychoanalytic theory. I must also mention that our knowledge of Tantra from the mid-first millennium through the first few centuries of the second millennium C.E. is rapidly expanding, in great measure because of the efforts of Alexis Sanderson and his students at Oxford University. Doubtless, there will soon be much more to say about possession in tantric literature that will add considerably to what I have written in Chapter 10, and may force new paradigms on the notion of possession itself as it was configured historically in India. Nevertheless, for me, this exercise—whatever I have adduced on the topic—has turned religion, particularly as observed in South Asia, on its head, as the material ultimately argues against much of what is stated in standard textbooks. If even a tiny amount of that is transmitted to the reader, this project will have been worth the effort.

I should say a few words here about the study of possession. In India and elsewhere, the field has been dominated by compartmentalized ethnographies and, less often, by histories of possession in specific lineages or local cultures. No syncretic history or synoptic account of possession in India has been attempted.¹ While my intention here is to locate and capture such a history, I have tried to keep in mind the problems associated with master narratives and endeavored to avoid them. Even if I were dedicated to a single theoretical model (and it will soon become obvious that I am not), two things would still parry any attempt to create such a master narrative: the sheer variety of the textual and ethnographic source material, and the delicacy with which the layers of their connections must be handled. I have been constantly aware of the pitfalls of both subjectivity and objectification that confront both scholars and participants who think about and live with possession. This inspires in me a certain trepidation, because it sharpens rather than occludes the necessity to define and delimit, to construct and deconstruct, to know when to intervene and when to leave alone, to know how strongly to invoke situated histories, to know when to allow tradition and imagination to merge, and to feel comfortable if all my data and conclusions are not scrubbed clean of contradiction. Nevertheless, I take full responsibility for lapses in clarity, errors in judgment, and oversights in the use of material.

After this volume went into production, two films dealing with spirit possession were released that deserve comment here because they illustrate the point of the title of this book. One was an American film, The Exorcism of Emily Rose; the other was an Indian film, Paheli (The Riddle).² What is striking about these films is their representation of the self possessed. Emily Rose was founded on the assumption of a naturalized, identifiable, and unitary self that was breached by demonic possession. This replicates the standard view within Western culture, even if it is belied by the very fact of, for example, acting (as in a movie) or role playing, which presupposes a fluidity, multiplicity, or even nonexistence of a fundamental identity on which personalities, or even persons, are constructed or superimposed. Another standard Western view, also casually supported by this film, is that possession can only be sinister, demonic, and evil. All of these assumptions are called into question in the present study.

In the Hindi film Paheli, a wife is seduced by a bhūt (spirit) who falls in love with her and takes the form of her husband after he is sent by his father, a successful businessman, to a distant city for an extended period. Soon, the bhūt is drawn into his own self-construction, and at the end of the film he appears to merge with the character that he has replicated, making the possession complete and, we believe, satisfying and permanent. Thus, the bhūt possesses, first, his own shape-shifted construction and, eventually, its more substantial prototype. Almost as casually as Emily Rose replicates Western assumptions of selfhood and fixed identity, Paheli illustrates the Indian (and other Asian) recognition of selfhood as mutable, multidimensional, nonlinear, and (at least in Buddhism) fabricated, a moving part among other moving parts. The bhūt maintains his self-possession, his own identity as a bhūt (the wife also knows this—it is their great secret), aware that he has instigated his own construction. In certain important respects, this is similar to some of the cases I discuss here; indeed, it emerges from the same religious and cultural tradition. These include, for example, the Upaniṣadic case of Brahmā possessing the inert world that he has created and the eighth-century philosopher Śaṅkara possessing the body of a dead king. We also see in Paheli a striking sympathy for the character of the bhūt, a portrayal of possession that would not be possible in Western cinema, except perhaps as comedy.

As startling as the film Emily Rose may be, equally startling to this viewer is the cinematic figure of Dr. Adani, an anthropologist. Dr. Adani is perhaps not entirely unlike this book’s author or some of its readers. About this character, and the film in general, A. O. Scott writes, in his review of Emily Rose in the New York Times, that the anthropologist studies demonic possession and is studiously noncommittal as to whether it really exists. The movie pretends to take the same tolerant, anything’s-possible position.… Its point of view suggests an improbable alliance of postmodern relativism and absolute religious faith against the supposed tyranny of scientific empiricism, which is depicted as narrow and dogmatic (Scott 2005). This volume addresses the issues brought up by Scott, all of which are implicit in Emily Rose and Paheli, and tries to present them from a variety of perspectives.

Finally, I have established a Web site that will include some of the material in this book, including the plates, which will be in color, and the bibliography. In due course I will also put on the site audio and video clips of possession phenomena as well as photographs and accompanying explanations. I invite interested readers to contact me about adding entries to the bibliography and placing other possession material on the site, or links to theirs. I envision this as a clearinghouse for the topic of possession in South Asia. The URL is www.possession-southasia.org, and I can be contacted at [e-mail].

NOTES

1. The nearest attempt so far has been in the collection of articles edited by Assayag and Tarabout (1999). As good as this collection is, it lacks a general historical context and the syncretism that only a single-authored study can provide. The same is true for possession studies elsewhere in the world. For Africa, see Behrend and Luig 1999; for Indonesia and Oceania see Mageo and Howard 1996.

2. Both films were released in 2005. Emily Rose was directed by Scott Derrickson, Paheli by Amol Palekar. See Philip Lutgendorf’s discussion of Paheli at www.uiowa.edu/~incinema/Paheli.htm.

ACKNOWLEDGMENTS

BECAUSE OF THE MULTIDIMENSIONAL NATURE OF THIS STUDY, I have sought and received help in the form of information, discussion, and, eventually, critical reading of different parts of the book from many friends, colleagues, and acquaintances. This particularly long gestation period of a decade and a half has, therefore, placed me in their debt, a debt that I am delighted to acknowledge. First, I extend special thanks to Frank Korom, William (Bo) Sax, and Lindsey Harlan for reading early drafts of the chapters on theory and ethnography. Thankfully, Frank, Bo, and Lindsey think very differently about both anthropology and religion and offered correspondingly different critiques, all of which helped me immensely. I am also heavily indebted to Chris Minkowski for comments on the Mahābhārata chapter, to George Thompson and Steven Lindquist for comments and moral support on the chapter on possession in the Vedas, to Alexis Sanderson for time-consuming and selfless assistance on the second of the two Tantra chapters, and to Claudia Welch for insight and extensive comments on the Āyurveda chapter. I am grateful for the University of Iowa International Programs office for grants that allowed me to travel to Kerala for fieldwork on Āyurvedic mental health care in 2002 and 2004. I can confidently say that I could not have accomplished any of my work in Kerala were it not for the ready kindness of N. V. Ramachandran of Palakkad, to whom I remain heavily indebted. His ethnographic skills as well as knowledge of local medical traditions and, conveniently, all the back roads of Kerala were decisive in that work.

I deeply thank my colleagues at the University of Iowa: Philip Lutgendorf, Susan Lutgendorf, and Janine Sawada, for their many insights and comments on various incarnations of this work. Any work touched by Philip Lutgendorf emerges better for his contact. Indeed, one of the great boons of my life has been his presence in the next office, not to speak of his enormous generosity in countless areas of life, for a substantial number of years. Susan Lutgendorf, professor of psychology at Iowa, guided me in readings on psychology and constantly challenged me in thinking about the inner dimensions of possession. Her engagement with the topic has vastly advanced this work, especially Chapter 2. Wendi Adamek has been an indefatigable supporter of this project throughout. Like Philip Lutgendorf, she is a brilliant and exacting reader and a luminous thinker, whose advice on matters of theory, consistency, and style is always to be heeded. Many other friends and colleagues have offered productive comments and substantive help along the way, including, most formidably, Rich Freeman, Robert Svoboda, Antti Pakaslahti, Laurie Patton, Jeffrey Kripal, John Dunne, David Gray, Kathleen Taylor, Stephanie Jamison, David White, David Knipe, Anne Feldhaus, and E. Muralidhara Rao. I also express my appreciation to the exceptional staff of Columbia University Press—Wendy Lochner, Christine Mortlock, Leslie Kriesel, and Debra Soled. Their kindness and enthusiasm helped enormously to validate the years of time and effort I put into this project. All errors are, of course, my own responsibility, and in a work of this size and omniformity I am sure there will be more than a few, for which I beg the reader’s indulgence.

INTRODUCTION

THE EVALUATION OF THE EVIDENCE FOR DEITY AND spirit possession in South Asia, in both classical texts and modern practice, is framed by prevailing characterizations of such gravity that only with great effort is it possible to escape from beneath their weight—and in doing so it is still impossible to escape their shadow. These characterizations are literary, marked by vocabulary, images, and themes distinguished from an array of related phenomena. They are recognized because of a rich history of scholarship on the general subject of possession. Most such studies, whether Indological or anthropological—or even brahmanical, if we consider the long history of Sanskritic and other indigenous Indic scholarship—have commenced with the necessary chore of phenomenological categorizing and entertaining a potpourri of theoretical approaches. Just as brahmanical and Indological scholarship have set certain limits and challenges on the subject, thereby creating orthodoxies, so have the shifting grounds of anthropological studies, thereby creating mini-orthodoxies of their own. Although the word orthodoxy may sound harsh and unjust to some, scattered in this study is an accounting of methodological approaches to and models of possession employed by Indic authors and scholars throughout the millennia, by anthropologists and other ethnographers, as well as psychologists, in the latter cases reflecting Western academic deployment and establishing for them new dimensions. The approaches, addressed in Chapter 2, range from outright condemnation (by apologists for religious orthodoxies) to sophisticated psychoanalytic interpretations, sociopolitical views, notions of possession as a facet of shamanism, and last (and probably least) an increased acceptance, with certain qualifications, by ethnographers, of emic viewpoints regarding the existential reality of spirit and deity possession.

Both separable and inseparable from these approaches are other general issues, including those of gender and illness. Why, in popular culture, are women subject to possession more regularly than men? Do we find the same gender configuration in possession described in Indic literature? To what extent or under what circumstances does possession signal disease, particularly mental illness? These questions are addressed in due course, and answers to them must consider the purport and environment of the Sanskrit texts as well as their contents. In order to more clearly envision the Sanskritic and classical contexts, it will also be necessary to view the subject from the top down, that is, to discuss the vocabulary and linguistic characterizations of possession in postclassical and modern South Asia. To this end, I explore the semantics of possession and their significations in several South Asian languages (Chapter 4). This discussion begs a consideration of the relationship between folk possession and classical possession, a topic that is frequently revisited in this study.

After explorations of the ethnography and linguistics of possession in (mostly) postclassical South Asia, we turn to the Sanskrit texts. First, we consider the extensive evidence for possession and models of thought based on possession in the vedic literature (Chapter 5). References to possession begin in the Ṛgveda itself, where we first encounter the word āveśa (and other forms derived from the root ā√viś) in the ninth maṇḍala, the book that includes most of the soma hymns. It is significant that āveśa is the most commonly attested word for possession in the Ṛgveda through Classical Sanskrit and Middle-Indic languages down to modern times, where it has found its way into most Modern Indo-Aryan and Dravidian languages. I argue that the vedic soma drinkers experienced (at least in part) a type of divine possession. The word āveśa occurs in nearly all vedic texts, with a preponderance in texts of the early and middle vedic periods. By the late vedic period, derivatives from another Sanskrit verbal root, √gṛh (to seize), especially the substantive grahaṇa, begin to appear, indicating a rather less divine form of possession. In attempting to address the vedic evidence, I highlight four features in particular: first, the production of soma and the sages’ experience of its consumption; second, āveśa as a paradigm for cosmic creation and personal and divine incarnation; third, a phenomenon called transfer of essence (medhas), found in many vedic texts beginning with the Brāhmaṇas; and, fourth, the relationship between women (mostly) and gandharvas, celestial musicians with a predilection for possession. It is this latter notion that has perhaps the greatest impact on later classical literature and folk possession. In fact, it is in the vedic texts that the Indian notion of the self is first expressed as permeable and multivalent. Some readers may find this notion overly complex because it runs counter to the received Vedāntic idea of the self as an autonomous ātman identified with a universal brahman. However, the evidence here demonstrates that the Vedāntic notion of the self is primarily a normative, albeit a popular and attractive, idea that, because of the force and elegance of much of the Sanskrit philosophical literature, has overshadowed more complex and, I believe, more fundamental notions of self and personhood in classical India.

Possession comes into its own as a literary motif with accompanying procedural specificity in the Mahābhārata, a subject addressed in Chapter 6. Among the possession stories in this great epic are the well-known tale of Nala’s possession by Kali; another in which Vipula Bhārgava protected Ruci, the wife of his guru, Devaśarman, against the sexual advances of the notorious womanizer Indra by entering (anupravisya) the body of Ruci through his yoga-power; the account in the Sauptikaparvan of Asvatthaman’s destructive possession by Rudra/öiva; and the account near the end of the epic in which Vidura, an incarnation of Dharma, employs his yogic abilities to leave his own body, permanently, and enter that of Yudhiṣṭhira. Although the Mahābhārata contains many stories in which possession is central, the other Sanskrit epic, the Rāmāyaṇa, frequently refers to possession as a dimension of intense emotion but has few stories in which it is central. Thus the nexus of possession and emotion appears fully developed in the epics, a feature that characterizes Indic possession from that point on. Countless narrative passages in the epics and later literature describe extreme anger, love, anxiety, concern, or faith in terms of possession. Taking rhetorical leads from the epics, many of the Purāṇas frame possession in the same way; indeed, possession in the same key appears in the Purāṇas with remarkable regularity. It is no exaggeration to say that the Mahābhārata is the most influential text in Indian history, and the tone and timbre of possession, in both its positive and negative manifestations, is established in the Sanskrit epics. Because of certain important conceptual links between the epics and the Purāṇas, specifically the way devotion frames possession, it is essential to have a brief look at possession in the Purāṇas as well.

Among the most visible manifestations of early Indian thought are Vedānta, classical Yoga, Buddhism, and Jainism, all of which are nuanced reactions to the perceived excesses of vedic ritualism. Chapter 7 discusses possession in these important branches of knowledge. Although possession is not a primary concern of the literature of these schools, which, on the whole, avoids discussion of personal experience except in the vaguest, most normative, and most fragmentary ways, it is nevertheless mentioned and occasionally discussed at some length in their foundational philosophical texts. It occupies a small place in the Yogasūtras of Patañjali and a rather larger one in the Jain Yogaśāstra of Hemacandra. Possession is also discussed in the Brahmasūtras of Bādarāyaṇa and in the commentaries by Śaṇkara, Rāmānuja, and Vallabhācārya. The Buddhist Pali canon, where demonologies appear for the first time (in the Petthavatthu), mentions it, as do the Jātaka stories and the more scholastic Buddhist Sanskrit literature. This is an appropriate place to begin to examine possession in Buddhist practice, where it is, in the best-known instance, a highly visible performative phenomenon by the oracle who has assisted the Dalai Lamas in matters ranging from statecraft to personal practice. In all the relevant texts, as in observable practice, possession is not condemned, but considered accessible for individuals who have achieved particular stages of mental and physical discipline. In addition, it is considered useful for gaining certain kinds of knowledge. An example of translating yogic theorizing about possession into practice appears in the hagiographies of the philosopher Śaṅkara, who, it is said, took possession of the body of a dead king named Amaruka in order to gain carnal knowledge without sullying his pure brahmanical and ascetically trained body. This incident must be discussed at some length.

Chapter 8 explores possession in Sanskrit and other classical Indic fiction, drama, and aesthetic theory. In the same way that the 1973 film The Exorcist and Henry James’s 1898 novella The Turn of the Screw influenced (and were influenced by) the Western—particularly the Anglophone—imaginaire in its views and experience of possession as a nasty and terrifying phenomenon (discussed in Chapters 2 and 3), stories of possession in India influenced its experience and occupied a major place in the Indian imaginaire. The evidence from Chapters 4 and 12, discussing the language of possession and extensive demonologies, testifies to this. But the influence and experience was not all of a negative sort, as tales of possession—especially from the Mahābhārata, the text with the greatest influence of all—did not simply create a psychic locale for dumping one’s worst fears and (one hopes) imprisoning one’s most terrifying ghosts. In addition, these tales opened the psychic doors to possibilities of oracular, divine possession, which is subsequently taken up by different genres of Sanskrit, Middle Indic and modern Indian literature; the tales are a major reason why possession is the most common and, at the grass roots, most valued form of spiritual expression in India.

Possession as a motif in Sanskrit fiction was initially addressed by the American Sanskritist Maurice Bloomfield in 1917. However, he covered only a few of the important texts and examined possession more as a curiosity than as a cultural phenomenon or the reflection of a widespread experience. The most important texts in this category are the well-known vampire (vetāla) stories and the massive Kathāsaritsāgara. Possession is also referred to in Sanskrit and other classical dramas, including the Bhagavadajjukāprahasana, a comedy by the seventh-century Pallava king Mahendravarman in which it is the primary theme; indeed, that possession is presented as social satire is itself telling. Perhaps more than elsewhere, in fiction and drama possession is revealed as a well-known (though not particularly well-respected) feature of popular religion (at least among the literati). It is also appropriate in Chapter 8 to attempt to derive a formal aesthetics of possession by examining theories of mood and emotional construction, as well as audience response, in the works of Bharata (Nāṭyaśāstra) and Abhinavagupta (Dhvanyālokalocana), then to locate resonances of this in modern popular culture.

Chapter 9 places possession in the context of devotional practice. This is perhaps the most important aspect of possession, as the intensity of personal engagement is revealed in both text and practice to be the signal determining feature in personality identification. And in India bhakti it is nearly always linked to an intensity driven by a single-minded love for the deity. Thus, in the context of devotional practice, personal identification with a deity is often interpreted as possession. The primary textual locale for this is late first-millennium Sanskrit and Tamil bhakti literature, eventually extending to Hindi, Marathi, and other regional literatures. Although this narrow, interiorized, single-minded (ekāgratā) focus is ostensibly resolute and impenetrable, it is, in South Asia, indicative of the porousness of the self and the fluidity of personhood.

Chapters 10–13 address possession in Tantra and Āyurveda, discourses that must be examined together to some extent because they increasingly share the same intellectual context. It is here, for the most part, that the notion of possession as grasping or seizing (grahaṇa) is highlighted, negative possession brought on by a malevolent spirit (the influence of the Mahābhārata is also felt here). But this is only a part of the story, as it shares conceptual territory with the positive possession most often associated with deities, as well as with interpersonal possession found in texts ranging from the Ṛgveda to the Upaniṣads, the Mahābhārata (again), the Purāṇas, the texts of classical Yoga, and most other texts encountered in the earlier chapters. It is important that the Tantras feature oracular possession as an active practice. This is dealt with in Chapters 10 and 11, which cover the ground of normative brahmanical possession and possession found in Buddhist Tantras. Chapter 11 addresses a single topic, the practice of using children as agents of oracular possession and the dissemination of this practice from North India to Tibet, China, and South India.

The usual way in which the Sanskrit texts dealt with negative possession was to identify and classify invasive spirits in demonologies, then to medicalize their possession. The most important and influential texts undertaking this are the foundational works of Āyurveda, as well as a number of Tantras. I offer new translations of the relevant sections of the three primary Āyurvedic texts—the Caraka, Suśruta, and Aṣṭāṅgahṛdaya Saṃhitās—which provide specific demonologies in the context of discussions of insanity (unmāda). I also provide extracts from some of the relevant tantric texts, the most important of which is the Īśānaśivagurudevapaddhati, and a hybrid text of Āyurveda, ritual, and dharmaśāstra called Madanamahārṇava.

Because of the importance of astrology in Indian culture for at least two millennia, and because of the perception of planets as graspers or possessors (indeed, the Sanskrit word for planet is graha), I must also discuss passages from astrological texts, notably the Praśnamārga, dealing with possession. In summarizing the ritual and pharmacological aspects of exorcism recommended by various āyurvedic, tantric, and astrological texts, I bring into the discussion my own fieldwork on possession among these traditions as they are constituted in present-day Kerala.

The final chapter offers a few conclusions based on my years of living with this topic. The first identifies typologies of possession that are correlated with language use in order to elucidate the semantics, constituted within the sociolinguistics, of possession. Then follows a discussion of the influence of possession on notions of corporality and personal identity. Essentially, I conclude that possession was both a cause and a result of the generally accepted South Asian notion of a permeable embodiment. This notion, derived from a study of practice, is somewhat at odds with notions of selfhood developed in philosophical texts that have been adopted as canonical from the time of the Upaniṣads to postcolonial India. Perhaps not incidentally it is exhibited more consistently among women or at least applied more generally to women, in accordance with anthropological findings. However, our evidence demonstrates that such permeability is equally applicable to men, though in rather different configurations. Finally, I return to the Mahābhārata for an examination of the diachronic aspects of possession. We see in this paramount epic prototypes and patterns of possession that recur throughout Indian history, not just in performance and philosophical construction but in basic problems of self-definition that lie at the root of human multivocality.

Finally, I should say something about Sanskrit text passages in the book, which may appear to be irregular. Because of the length of the book I had to be selective in my choices of transliterated Sanskrit text passages. I include text of the Ṛgveda in Chapter 5 and of passages that are of particular lexical or philological importance scattered elsewhere throughout this work. I have usually omitted the text of passages I have translated of Sanskrit works that are generally available—such as the Upaniṣads, the Mahābhārata, and the published āyurvedic texts—and those that have been translated by others. I hope these decisions do not prove to inconvenience the reader excessively. All translations are my own unless otherwise noted.

I

Orthodoxies, Madness, and Method

CHAPTER 1

Academic and Brahmanical Orthodoxies

I am large, I contain multitudes.

—WALT WHITMAN, Song of Myself, v. 51

Sanskritic Culture and the Culture of Possession

Ethnographic work at the beginning of the twentieth century, which has been supplemented and revitalized during the past forty years, shows that spirit or deity possession is a widespread epstēmē—a historically situated discourse, phenomenon, and practice—in Indian thought, culture, religion, and medicine.¹ However, if our knowledge of the subject were limited to the accounts of classical Indologists and others who have privileged the high intellectual and religious traditions while eschewing the history of actual religious practice, we would scarcely know of the existence of this phenomenon, much less its pervasiveness. One might say that it is the ism of Hinduism, the Hinduism essentialized by its most literate propagators and translators that has nullified and delegitimized its breathtakingly broad spectrum of popular practice, including possession.² More specifically, non- or scant recognition of possession in premodern India has been due to a longstanding aversion among educated Judeo-Christians as well as educated Hindus, for whom possession has fallen outside the realm of both reason and social accountability.³ Perhaps the most common perception—and consequent stigmatization—of possession among both academic and indigenous orthodoxies is that it is a nettlesome aberration, a blemish on the face of epistemological order, a phenomenon subject to benign neglect, or, at most, sanitized into nonrecognition. To whatever degree this perception is borne out by taxonomies in both Sanskrit and regional languages that distinguish various loci and states of possession (a topic dealt with extensively below), the more important point is that possession and, indeed, emotion itself, the kinder, gentler ancestor of possession, have been considered to bear the stigma of primitivism:⁴ they are associated with people of lower social rank, including low castes, tribals, and women,⁵ or more generally with those lacking literacy, the great (and, before modern times, extremely rare) tool generally believed capable of bestowing introspection, self-knowledge, and control.⁶ Therefore, discussion of it has been avoided or denounced by the self-conceived (and, by no accident, highly literate) orthodox among both scholars and indigenous practitioners. And among scholars, even when this stigma has not been inflicted, possession as the exclusive property of lower-ranking individuals has been assumed so casually as to preclude a search for it among other groups and individuals. Even the eminent sociologist Louis Dumont declares possession a mystic ecstasy … which so far we have not encountered at a learned level.⁷ The present study aims to rectify these shortcomings.

In this study, I press two overriding contentions, both of which not only are paramount for the present study but have implications for the methodology of future scholarship. The first is that the force of the ethnographic accounts should elicit a re-examination of classical texts for evidence of possession. The second is that possession as described by anthropologists and other ethnographers, and as understood by native and scholastic orthodoxies, does not represent the full spectrum of possession as revealed in Sanskrit texts. No doubt, such possession is described in or assumed by certain Sanskrit texts, and the bulk of the present project is dedicated to the exploration of Sanskritic possession. However, as shown below, the category of possession as understood in Sanskrit literature must be expanded beyond the parameters assigned to it by previous scholarship, which is largely drawn from prevailing Western notions. As an indigenous category in ancient and classical India, possession is not a single, simple, reducible category that describes a single, simple, reducible experience or practice, but is distinguished by extreme multivocality, involving fundamental issues of emotion, aesthetics, language, and personal identity.

Both of these contentions contribute to my principal project here, which is to retrieve, as far as possible from texts, an understudied aspect of ancient and classical Indian cultural practice, religious experience,⁸ and disease production. I become suspicious of the apparent nonexistence of possession in antiquity when we read present accounts, such as that of David Knipe, who states, based on fieldwork in East Godavari District, Andhra Pradesh, it appears that the number of householders subject to possession states is astonishingly high, and the phenomenon occurs within families of all communities as a central component of religious life.⁹ This suspicion is verified, at least for non-Sanskritic classical culture, when we read David Shulman’s reports on Tamil Hinduism. Beginning in the Pallava-Pāṇṭiya period in the sixth century C.E., writes Shulman, The focus is entirely on the interaction between the devotee and the god who has entered him, mastered or ‘possessed’ him without destroying his empirical, sensually motivated, autarchic being.¹⁰

Despite vivid descriptions of possession in Tamil texts of the first millennium C.E., which on the whole bear out the contemporary ethnographies, as lucidly discussed by Glenn Yocum,¹¹ most Sanskrit texts, conspicuously canonical religious texts, entirely neglect—or, perhaps, avoid—discussing what people actually do or experience. The reason for this is well known: The major commentaries on the philosophical darśanas and primary texts of aesthetic theory set the epistemological agendas for the remainder of religious discourse in Sanskrit. The principal epistemological feature of these primary texts and commentaries is a quest for programmatic perfection, with unwritten rules that discourage discussion of personal experience except as a mythical, fictional, or paradigmatic figure might have a paradigmatic experience. These texts clearly are mindful of the requirements of canon, which, as mentioned, privilege theory over experience.¹² Thus Sanskrit texts are almost always oriented uncompromisingly toward concepts, prescriptions, and mythology—what their authors understood to be true—but rarely toward human concern (except insofar as theorizing is a human concern)—what we (and probably they) deemed to be real.¹³ However, with persistent searching—a task doomed to incompleteness—perhaps we can begin to retrieve descriptions of this deemed reality from texts that both approximate those found in ethnographic literature and shed light on the parallel issues of emotional construction and personal identity in India.

The problem of retrieval, however, is not just one of oppositional texts; equally problematic is scholarly re-engineering of possession states, a problem that continues to the present day, albeit rather reduced from that of previous decades or centuries. It is well known by now that the academic agendas of respectable Indological scholarship were set in the middle of the nineteenth century by Europeans who cast a long shadow of doubt on the authenticity, hence validity, of a number of subjects.¹⁴ The important subjects, which quickly became canonical, which is to say orthodox, in Indology, following to some extent a sanitized brahmanical presentation, were the Vedas, Sanskrit grammar (vyākaraṇa), the six systems of Indian philosophy (ṣaḍdarśana), Buddhism and Jainism, law (dharmaśāstra), epic poetry (kāvya), and, to a lesser extent, poetic theory (alaṃkāraśāstra). Among the subjects neglected by respectable scholars were Tantra, bhakti, and astrology (jyotiṣa), the texts of which, interestingly, are included in disproportionately high numbers in manuscript collections throughout India, but were cast into disfavor by the orientalism of the day.¹⁵ Factors weighing against these latter texts were their relative modernity, their popularity, and their intellectual movement away from the more classical texts, which, following the orientalist prerogative (consonant with their nostalgia for classical Greece and the Renaissance), expressed in general the ascendancy of reason. Thus certain topics and texts were destined for neglect because they participated in and typified a perceived intellectual and cultural malaise that rendered the defeated, superstitious, caste-ridden, economically straitened, and technologically underdeveloped India the natural subjects of the European Enlightenment. In other words, not only were most of these texts composed after the close of the first millennium C.E., which is to say on the wrong side of the tracks, temporally, to count as true or ideal representatives of ancient India, but their very subject matter was beneath the intellectual and cultural prestige or dignity of contemporary European scholarship.¹⁶

If the brahmanical orthodoxies engineered popular and experiential content out of their texts, if they cavalierly shredded context and imposed an artificial and hegemonic discursive order, as standard Indology suspects—and this is a big if—it becomes immediately necessary to pose a question that will inevitably bear on everything dealt with here: Are Sanskrit (and other classical Indic) texts appropriate sources for our inquiry? More precisely, is Sanskrit literature or Sanskritic culture capable of representing the interests of a clientele more broadly based than a brahmanical elite that is usually characterized (and caricatured) as unregenerate and epistemologically prejudiced? At this point, which is to say until we have recounted and taken stock of the evidence—or its absence—we must be content with the unsatisfying observation that this question is fraught with ideological considerations and is thus unanswerable.

Nevertheless, a preliminary inquiry into this question is possible through a consideration of two opposing viewpoints on the capabilities of Sanskrit literature. In one corner is the eminent historian and Sanskritist D. D. Kosambi, who does not mince his words in the introduction to his critical edition of a collection of Sanskrit court poetry and miscellanea, the Subhāṣitaratnakoṣa, compiled in the eleventh or twelfth century C.E. in Bengal by one Vidyāpati. Kosambi observes that the life depicted in such poetry was not shared by most Indians of the poet’s or of any other time; in essence, not even by the poet himself.¹⁷ According to Kosambi, Sanskrit poetry, as well as classical Sanskritic culture, was the shared property—as well as the shared fantasy (Kosambi would likely regard it as a shared conceit)—of a brahman/ ruler class collusion.¹⁸ The end product was the empty bombast of a Rājaśekhara, thoroughly detached from the lives and livelihoods of the majority.¹⁹ In the other corner is the Sanskritist Richard Lariviere, who parries criticisms of Sanskrit philology in a wonderfully contentious article entitled Protestants, Orientalists, and Brāhmaṇas: Reconstructing Indian Social History.²⁰ It is Lariviere’s view that if we return to the philological techniques and values that have been exhibited with such consistency in the study of Greek and Latin classics, and that were once an important part of Sanskrit philology, but seem, in recent years, to have fallen out of favor, … [the result will be] to give the fullest possible voice to the views of classical Indians—of all social classes. To achieve this, he states, we must carefully and thoroughly edit their texts. In Lariviere’s view, [t]he better the data we have the better our answers will be, and philology provides the data. Although Lariviere sanctimoniously calls for more and better critical editions of Sanskrit texts as the central requirement in the noble mission of expanding our knowledge of the historical South Asian underclasses, it must be remembered that in the Subhāṣitaratnakoṣa Kosambi has also taken on the task of constructing a critical edition, always with a view toward understanding Indian social history—and has arrived at radically different conclusions. Regardless of how this debate eventually plays out—and I am not particularly optimistic that Lariviere’s hypothesis will be upheld in any great measure²¹—the present work is, I hope, an example of data-gathering from a large number of Sanskrit texts and of how such data gathering can help us to retrieve and identify some of the religious views and practices of a broad spectrum of ancient and classical Indian society.²² In the process I hope to demonstrate implicitly some of the areas in which both Kosambi and Lariviere might be right.

Several scholars have declared that possession is conspicuously absent in Sanskrit texts, at least possession as classically understood from ethnographies. In addition to Dumont, cited above, Richard Gombrich states, Later brahmanism … denied all value to possession states and they were screened out of brahmanical religion. To be possessed is to lose one’s self-awareness and self-control. Brahmānism inculcated control.²³ Gombrich’s remarks were to some extent responsible for alerting me to the possibility of possession in Sanskrit texts. While it is true that brahmanism inculcates control, to what extent, I wondered, did brahmans strictly observe brahmanism, and to what extent was this aspect of brahmanism an instigating force in a broad sampling of Sanskrit texts? I wondered, in other words, was the relationship between brahmans, brahmanism, and brahmanical texts very different a thousand or two thousand or three thousand years ago from what it is today? It is important for our purposes to note that brahmans, perhaps as much as others, participate in rituals of possession (and enter into states of possession), as is evident in the bhūta cults of South Kanara District,²⁴ the pilgrimage to Nandadevi documented by William Sax,²⁵ bhakti derived possession by smārta women in Madras by Mary Hancock,²⁶ and rites of exorcism in Maharashtra, Rajasthan, Varanasi, and doubtless elsewhere.²⁷ This is analogous to the participation of Buddhists, unswervingly loyal to their textual traditions and text-based practices, in an array of possession rituals in Tibet, Nepal, Sri Lanka, and Thailand, if not elsewhere.²⁸ As is the case with Buddhists and Buddhism, participation by brahmans in ritual must be distinguished from brahmanism as Gombrich employs the term. Indeed, many if not most brahmans demur in practice, at least part of the time, from the ism of brahmanism, and probably always have. We examine in Chapter 5 a passage from the Bṛhadāraṇyaka Upaniṣad that confirms this in part.²⁹ The tidy conflation of brahmans and brahmanical practice with the ism is all too hastily associated with both and leads to conclusions such as Dumont’s, which are little more than weary stereotyping. To be more generous, however, what Dumont and Gombrich are doubtless referring to are the official brahmanical doctrines espoused by the literary elite for the past two or three millennia, as they appear in the normative dharmaśāstra texts and philosophical darśanas. In this sense, Gombrich is correct in noting the importance of brahmanism in delimiting the parameters of control, self-awareness, and the self as a single discrete independent entity. Indeed, these are some of the issues I must address in unpacking the category possession.

Restated in the form of fundamental questions, these issues may be expressed as follows: Is the self or person revealed and explicated in possession states genuine? Is there a true self (assuming for the moment that this is a viable concept and can be located) contaminated or camouflaged by possession states, spirits, or deities? Is the self a singular stable undiluted entity? Or is it a referential, explanatory convenience, an individualized composite construction, and thus unrealizable beneath its layerings—in agreement with the Buddhists? If the former, is deity or spirit possession a superimposition on a pure and realizable self? If so, then possession should be distinguished from that insular self, which is to say that the self and the personality are distinct. If the latter, is such possession a random or occasional component in a mutating multiform self, and if so can we really call this a self rather than a person? Or—another possibility—is possession merely a trope, a significator for an essentially plural or composite self? These are all questions intended to explore the experiential and soteriological trajectories of possession as well as its historical and sociopolitical contexts.

With respect to Buddhists, to whom control and mindfulness were paramount, Gombrich observes that they set themselves against such vulgar lack of self-control in favour of what they considered more ‘civilized’ standards.³⁰ Thus, possession skirted the various crossroads of native scholasticism, civilized standards, and control, self or otherwise. The assumed—and unaddressed—questions were: Is the deity or the true self revealed more accurately when one is in control or out of control? Is the self revealed out of control different from or, as some would have it, inferior to the one revealed while under control? Or, is there not much difference; which is to ask, does being out of control lead to states of knowledge, surrender, and spontaneity that are similar or identical to those manifest while under control? Or, is spontaneity in religious expression possible only under controlled (which is to say ritualized) circumstances?³¹

With respect to the integrity of the self, which cannot be easily separated from the body, the early and middle Vedic texts are clear. For example, the Śatapatha Brāhmaṇa (9.5.1.11) insists on a self that is constitutive of an intimate interplay between human, divine, and sacrificial bodies, while the Atharvaveda (11.3.1–2,7–8) holds that of this porridge Bṛhaspati is the head, Bráhman is the mouth, heaven and earth the ears, sun and moon the eyes, the seven seers the in- and out-breaths … dark metal its flesh, red metal its blood, tin its ash, gold its complexion.³² Whether porridge or human being (and often there is little difference), this theme of a composite body and self recurs in Vedic and Purāṇic literature (the figure of Yajñavarāha, the sacrificial boar, and the kāmadhenu or wish-fulfilling cow is employed in Chapter 10 to help explicate this) and, to a great extent, sets the tone for widespread textual and extratextual acceptance of possession as a natural phenomenon and of power sharing across apparent individual boundaries.

As suggested, one of the primary goals of this study is to examine the notion, generally recognized by anthropologists, though not by religious orthodoxies, that possession is not just one thing and apply this insight to the evidence from Sanskrit and other classical Indian texts. What I hope to show is that the category of possession as it has been commonly understood in religious studies and Indology, where it has been addressed at all, does not work well in the context of the classical Indian view of the self. As I demonstrate, it is a self with permeable layerings and boundaries, both of which constantly shift and mutate, and this can be known in part through a study of possession, as suggested by recent anthropological studies. Janice Boddy, an anthropologist who has studied possession in Sudan, writes that possession, within a wide spectrum of anthropological situations, 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.³³ Furthermore, I hope to show that the questions asked above can be answered, at least in part, through a better understanding of the experience and construction of possession in premodern India.

I retain the term possession, though it may not work very well in some of the circumstances that I discuss.³⁴ But I preserve it because the broad semantic boundaries of the terms under study here provide no attractive alternative. Consistent with this, I also suggest, in agreement with Gombrich, that the disregard or oversight of possession is a product of its construction (or lack thereof) by orthodoxies. In spite of concerted efforts at ideological exorcism, possession is frequently found in Sanskrit texts, but locating it, thinking about it, and understanding it cannot be undertaken by applying our usual category of possession, even if Gombrich is quite correct given his assumptions of what possession actually is.

Close scrutiny leads me to conclude that forms of religious experience and expression since the time of the Ṛgveda disclose much more possession than hitherto believed. After reviewing notions of possession derived from ethnographic literature, important because these notions define the commonly understood terms of our investigation, I discuss examples from a broad range of Sanskrit texts, including the Ṛgveda and Upaniṣads, the two great epics, the Mahābhārata and Rāmāyaṇa, the primary classical medical texts, classical poetry and fiction, classical yoga and Vedānta texts, tantric texts in both their philosophical and ritual incarnations, sectarian devotional texts, and of utmost importance for negative possession an array of āyurvedic texts. Throughout I attempt to summon a few conclusions about possession that might augment our appreciation of the phenomenon.

In spite of the generally accessible chronology of the selected texts and genres (and presumably of the development of the concepts contained therein), I must emphasize that, in view of the thematic organization of the subject matter presented here, I cannot rigorously adhere to the chronological order of text composition in assembling my account of possession in Sanskrit texts. Although text chronology does not necessarily correspond to the logical order of my presentation, neither is my assemblage of evidence, often lexical in nature, merely associative. Not unexpectedly, many of the text passages are ritual in nature, though a notable feature is that the objects and intentions of the rituals as expressed in the early texts are often not the same as in later texts. In the Ṛgveda, for example, possession was associated with ritual consumption of soma, while in the later Tantras and other explicitly ritual and devotional texts it was associated with initiation, ritual lapses, hostile takeovers by spirits, immersion in the regulated (hence ritualized) life of a deity, and the very equivalent of enlightenment. This shift is significant because the relevant passages from most of the later texts as well as occurrences of possession found in modern religious practice, including religious drama, appear to be predicated more explicitly on an ethos of possession that in time became more characteristic, recognizable, and predictable. It is not that the early texts suggest spontaneous states of possession that were later suppressed or domesticated through ritual confinement, but that the earliest accounts may be skewed by frustratingly incomplete descriptions of both ritual and states of possession.³⁵

Regardless of the extent to which early texts may be incomplete or later texts and ethnographic accounts complete, an unchanging fact was that spiritual and social forms of control were to a great extent mediated through intricate ritualizing, in which the domestication of possession was undoubtedly a factor. But this probably came about not because of more possession, or ritualization of possession, but because of increasing recognition of it through a more developed vocabulary of possession. After possession was recognized and named—and this appears to have already occurred by the time of the Ṛgveda—the category or categories denoting this activity became multivocal; they expanded to include varieties or shadings of itself. As is often the case with an action or phenomenon that is inherently uncertain or difficult to identify, possession as experienced and thought about in India suffered the discomfort of certainty and confidence arising from linguistic identification. In other words, one of the methods of controlling possession was to identify it, however vague or variant it might be, as something that was already known. Thus the forms of experience reflected in this vocabulary appeared to require control, given the social and ritual fabric of control in ancient India and the increased tendency to codify such control through the careful use of language and the composition of texts.

Perhaps as important as any other feature inhibiting orthodox recognition of possession was the very fact of literacy itself, which presupposes the composition of Sanskrit texts (not to speak of scholarly exegesis). One might say that in general textuality itself was, or at least epitomized, the antithesis of possession. Walter Ong notes, By separating the knower from the known, writing makes possible increasingly articulate introspectivity, opening the psyche as never before not only to the external objective world quite distinct from itself but also to the interior self against whom the objective world is set.³⁶ In other words, literacy—or, more simply, text—potentially releases the knower from the necessity of intense bodily engagement, from interaction with other beings, human or nonhuman, real or imagined, and, at the same time, establishes self-sufficiency beyond the pale of relational intrusion, a realm in which possession naturally abides. Indeed, to the culture of the situated word, possession dictated a relational intrusion beyond the realm of self-sufficiency; it was, rather, a technology or mode of representation, an improbable and intractable self-effacement and a wholesale replacement of one set of memories for another.³⁷ The response of most of the orthodox among Sanskrit paṇḍitas to the problem of possession, or more generally of human experience itself, was a civilizing and, not incidentally, a benignly controlling neglect, a response not far from that of their more modern Indological counterparts.

Regardless of whether writing was a universal fact of life among the educated (and in India education did not always require the written word), the brahmanical culture was one in which certain behavioral norms were propagated and largely observed. The very propagation and observance of these norms, regardless of how closely tied they were to written texts (and eventually their link was very close), constituted at least a cultural or brahmanical literacy that became extratextual, though emphatically not universal, as the discipline of anthropology constantly confirms. Indeed, I try to demonstrate, inspired to some extent by Ong, that the ideological literacy of Sanskritic culture was not universal, that it was, at various points, strongly under the sway of nonliterate and vernacular modes of thought.³⁸ While it is probably true that literacy, wittingly or unwittingly, inculcates control, what we find in India is that certain cultures, highly refined (saṃskṛta) yet, at the same time, comfortably vernacularized, share certain