Anda di halaman 1dari 673

When the Goddess was a Woman

Numen Book Series

Studies in the History of Religions

Texts and Sources in the

History of Religions

Series Editors
Steven Engler (Mount Royal University, Calgary, Canada)
Richard King (University of Glasgow, Scotland)
Kocku von Stuckrad (University of Groningen,
The Netherlands)
Gerard Wiegers (University of Amsterdam,
The Netherlands)

When the Goddess was
a Woman

Mahbhrata Ethnographies
Essays by Alf Hiltebeitel, Volume 2

Edited by

Vishwa Adluri and Joydeep Bagchee

On the cover: A processional image of Draupad Temple in T. Kuliyanur,
Dharmapuri District, Tamilnadu. Photo by Alf Hiltebeitel.

This book is printed on acid-free paper.

Library of Congress Cataloging-in-Publication Data

Hiltebeitel, Alf.
When the goddess was a woman : mahabharata ethnographies : essays / by Alf
Hiltebeitel ; edited by Vishwa Adluri and Joydeep Bagchee.
p. cm. (Numen book series ; v. 132)
Volume 2.
Includes bibliographical references and index.
ISBN 978-90-04-19380-2 (hardback : alk. paper) 1. MahabharataCriticism,
interpretation, etc. I. Adluri, Vishwa. II. Bagchee, Joydeep. III. Title.

BL1138.26.H46 2010

ISSN 0169-8834
ISBN 978 90 04 19380 2

Copyright 2011 by Koninklijke Brill NV, Leiden, The Netherlands.

Koninklijke Brill NV incorporates the imprints Brill, Global Oriental, Hotei Publishing,
IDC Publishers, Martinus Nijhoff Publishers and VSP.

All rights reserved. No part of this publication may be reproduced, translated, stored in
a retrieval system, or transmitted in any form or by any means, electronic, mechanical,
photocopying, recording or otherwise, without prior written permission from the publisher.

Authorization to photocopy items for internal or personal use is granted by Koninklijke Brill NV
provided that the appropriate fees are paid directly to The Copyright Clearance Center,
222 Rosewood Drive, Suite 910, Danvers, MA 01923, USA.
Fees are subject to change.
Madeleine Biardeau

These two volumes would not have been possible without the support
of many. We would also like to acknowledge the encouragement we
received from many scholars, including Alf, of course, and Greg Bailey,
Ashok and Vidyut Aklujkar, Jan Houben, Saraju Rath, T. P. Mahadevan,
Graham Schweig and Satish Karandikar. Special thanks are due to Jeny
Ruelo and to Roman Palitsky for their technical assistance. We thank
Maarten Frieswijk for his tremendous support of this project and Saskia
van der Knaap for her work in laying out the volumes.
Thanks are also due to our dear parents Dr. and Mrs. Adluri, and
Sandeep and Dr. Aruna Bagchee. We also thank Dr. Madhava Agusala
for his great support. A special thanks to all those who sustained
us with their love: Joachim Eichner, Thomas Komarek, and Elena
Garcs. Finally, we would like to thank colleagues who supported our
work: Barbara Sproul, Arbogast Schmitt, Danielle Feller, and Simon

Acknowledgements ............................................................................ vii

Introduction ........................................................................................ xi
Chronology of Works ....................................................................... xxxv

i. millenial draupads

Chapter One Draupads Hair ....................................................... 3

Chapter Two Draupads Garments ............................................. 33
Chapter Three iva, the Goddess, and the Disguises of the
Pndavas and Draupad ................................................................ 53
Chapter Four Purity and Auspiciousness in the Sanskrit
Epics ................................................................................................. 83
Chapter Five The Folklore of Draupad: Srs and Hair .......... 101
Chapter Six Orders of Diffusion in Indian Folk Religion ........ 125
Chapter Seven Draupad Cult Lls ............................................. 147
Chapter Eight Colonialist Lenses on the South Indian
Draupad Cult ................................................................................ 167
Chapter Nine Review of Landscapes of Urban Memory ......... 191
Chapter Ten Draupads Question ................................................ 195

ii. the sacrificial death of a co-wifes son

Chapter Eleven Dying Before the Mahbhrata War: Martial

and Transsexual Body-Building for Aravn .............................. 207
Chapter Twelve Hair Like Snakes and Mustached Brides:
Crossed Gender in an Indian Folk Cult .................................... 243
Chapter Thirteen Kttnta var: The Divine Lives of a Severed
Head ................................................................................................. 275
Chapter Fourteen Kttnta vars Cross: Making That Young
Bride, Whoever She Is, a Widow ................................................ 315
x contents

iii. companion studies

Chapter Fifteen The Indus Valley Proto-iva: Reexamined

through Reflections on the Goddess, the Buffalo,
and the Symbolism of Vhanas ................................................... 399
Chapter Sixteen Fathers of the Bride, Fathers of Sat: Myths,
Rites, and Scholarly Practices ...................................................... 433
Chapter Seventeen Two Ways to Tell a Story: lh in the
Bhavisya Purn a ............................................................................. 463
Chapter Eighteen Boar and Twins: Comparing the Tulu Kt i-
Cennaya Pddana and the Tamil Elder Brothers Story ........... 487
Chapter Nineteen On the Handling of the Meat and Related
Matters: Two South Indian Buffalo Sacrifices ........................... 517
Chapter Twenty Transmitting Mahbhratas: Another Look
at Peter Brook ................................................................................ 547

iv. apparatus

Bibliography ........................................................................................ 583

Index .................................................................................................... 607

The second in a two-volume edition of Alf Hiltebeitels collected essays,

When the Goddess Was a Woman presents twenty articles drawing mainly
on his fieldwork on the cult of Draupad as south Indian goddess and
the related cult of Kttnt avar/Aravn. Published originally between 1978
and 2005, these articles provide a comprehensive overview of Hiltebeitels
development from his earliest textual studies on the symbolism associated
with the goddess to his most recent pieces that discuss the reception and
interpretation of the Sanskrit epics in folk and other genres.
However, rather than follow strict chronological order, we have chosen
to arrange the essays thematically: Part I under the heading Millennial
Draupads, moves across ten chapters from treatments of Draupad in
the Sanskrit epic Mahbhrata to portrayals of other Draupads in folk-
lores and literature; Part II, under the title The Sacrificial Death of a Co-
Wifes Son, contains four articles, the last newly written, about regional
variations in the Tamil cult of Aravn/Kttnt avar; Part III, under the title
Companion Studies, closes with six articles that compare Draupad with
goddesses and heroines from other contexts in Indian mythology, oral
epic, and local ritual, and one that discusses the ethnography behind Peter
Brooks relatively recent dramatization of the Mahbhrata.
In contrast to Reading the Fifth Veda, the first of this two-volume
edition of Hiltebeitels selected works, which focuses mainly on the
Western academic reception of the Sanskrit epics, this volume mainly
seeks to address the relation between the classical Mahbhrata text
and lived Mahbhratas illuminated through ethnography. The vol-
ume is organized around two complementary aspects of Hiltebeitels
work: ethnographic and textual.

From Indogermanisches Urepos to a Str-dra-Veda

In his article Draupads Hair1 from 1980, included here as the first
chapter in the volume,2 one can see a preliminary culmination point

Alf Hiltebeitel, Draupads Hair, Autour de la desse hindoue, ed. Madeleine
Biardeau, Purusrtha 5 (1981): 179214.
Chapter 1 thus sets the stage for what might be called a bhakti re-reading of the
xii introduction

of this process in the formulation of four methodological working

assumptions. As these assumptions are formative for the remaining
articles included in this volume, it is well worth considering them
here. Hiltebeitel writes:
These [assumptions] are, first, that certain themes connected with a number
of the heroines of the Hindu epics can be illumined by looking to materials
on the Hindu Goddess that often appear only later in post-epic literature
and folk traditions; second, that the epics themselves present a complex
theological vision that gives pride of place not only to heroes and heroines
who represent aspects of Visnu, but also of iva and the Goddess; third, that
the epics thus provide narrative reflections on (and of ) the cult and mythol-
ogy of the Goddess as it emerges into the literary light of day; and fourth,
that the epic poets made selective use of oral traditions that probably had
some affinities with oral and vernacular epic traditions still popular today.3
If one looks carefully at these four assumptions, one will see how they
map the future course of Hiltebeitels research. While the first assumption
leads Hiltebeitel to understand aspects of the epic retrospectively out of
ritual traditions,4 the second assumptions influence can be clearly seen in
the two articles from 1980Draupads Garments (1980)5 and iva, the
Goddess, and the Disguises of the Pnda vas and Draupad (1980).6 Next,
a set of articles from the period 1985 to 1995 (reprinted as chapters 47 of
this volume) are motivated in the attempt to understand something of the
reciprocal relation between the textual and the ritual traditions; and Two
Ways to Tell a Story: lh in the Bhavisya Purna (1999)7 in particular
undertakes a mapping of the classical epic onto oral and vernacular epics.

epic, whose influence is not just limited to Hiltebeitels ethnographic studies, but can
also be seen in the article The Two Krsn as on One Chariot: Upanisadic Imagery and
Epic Mythology from 1984; an article in which Hiltebeitel adopts as his point of depar-
ture the assumption that the Mahbhrata in its classical form is a work of bhakti
through and through. The Two Krsn as on One Chariot: Upanisadic Imagery and Epic
Mythology, History of Religions 24 (1984): 1; reprinted in volume 1, pp. 485512.
Hiltebeitel, Draupads Hair, 181.
See especially the articles Dying before the Mahbhrata War: Martial and
Transsexual Body-Building for Aravn (1995), Hair Like Snakes and Mustached
Brides: Crossed Gender in an Indian Folk Cult (1998), and Kttnta var: The Divine
Lives of a Severed Head (1999) published as chapters 1113 of this volume and the
new article, unique to this volume, Kttn t avars Cross: Making That Young Bride,
Whoever She Is, a Widow (chapter 14).
See n. 1 for the full citation.
Alf Hiltebeitel, iva, the Goddess, and the Disguises of the Pndavas and
Draupad, History of Religions, Twentieth Anniversary Issue 20 (1980): 14774.
Alf Hiltebeitel, Two Ways to Tell a Story: lh in the Bhavisya Purn a, in
The Resources of History: Tradition, Narration, and Nation in South Asia, ed. Jackie
Assayag, Institut Franaise de Pondichry. tudes thmatiques, 8 (Pondicherry: cole
Franaise dExtrme Orient, 1999), 96112.
introduction xiii

In 1974, after finishing his first book on the Mahbhrata, The Rit-
ual of Battle: Krishna in the Mahbhrata,8 drawn from his doctoral
dissertation, Hiltebeitel made his first trip to India with a sense that
the Indo-European connections he had argued for in that book were
getting less and less compelling, and with a hunch that he might find
the Mahbhrata known better on the ground there than in Western
or, for that matter, Indian halls of academe. In particular, he had
begun to notice while proposing his project, on a suggestion from
Madeleine Biardeau that he study the Tamil Draupad cult, that there
was total bifurcation between knowledge of the Sanskrit Mahbhrata
and knowledge of Mahbhrata vernaculars.9 Colonialist Lenses on
the South Indian Draupad Cult (1992)10 documents the total indif-
ference to the Draupad cult Mahbhrata during a formative period
of Western Indology. If Sanskritists owned the text, vernacular
Mahbhratas, not to mention a Tamil Mahbhrata of dra farm-
ers, were only to be belittled.11
But what was it that Hiltebeitel hoped to learn from a Tamil
Mahbhrata of dra farmers? Since the institutionalization of modern
Epenforschung in German universities in the late 18th/early 19th cen-
turies, it had been an article of faith among Mahbhrata scholars that
research into the epics was primarily concerned with textual sources.12

Alf Hiltebeitel, The Ritual of Battle: Krishna in the Mahbhrata (Ithaca: Cornell
University Press, 1976; 2nd ed.: Albany: State University of New York Press, 1990).
This section repeats material also found in the Introduction to volume 1.
Alf Hiltebeitel, Colonialist Lenses on the South Indian Draupad Cult, in Rit-
ual, State and History: Festschrift for Jan C. Heesterman, ed. A. W. van den Hoek,
D. H. A. Kolff, and L. M. S. Oort (Leiden: E. J. Brill, 1992), 50731.
This question of who owns the Mahbhrata crops up in a different guise in the chap-
ters that end Parts I and III. Each is about modern literary treatments the Mahbhrata:
one by the Bengali short story writer Mahasweta Devi, the other from the British drama
director Peter Brook. Each of these essays can be said to reinforce the hunch mentioned
above: that it is a mistake to overlook Mahbhratas known on the ground.
A prejudice that continues to be reflected in the comments of contemporary Sans-
krit philologists. Ehe unser Text vom philosophischen oder vom religionskundlichen
Standpunkt studiert werden kann, mu er verllich herausgegeben werden. Daran
ist nicht zu rtteln, wenngleich diese Binsenwahrheit gern und immer wieder in jenen
Kreisen ignoriert wird, die sich lieber an Upanisad-bersetzungen erbauen, als da sie
sich die Mhe machten, jene Texte philologisch-kritisch im Original zu lesen und zu
verstehen. Wilhelm Rau, Bemerkungen zu ankaras Brhadranyakopanisadbhsya,
Paideuma VII (1959/61), 299; It is nothing short of a scandal that still, after some 200
years of study, instead of preparing reliable texts and translations, a lot of ink keeps
being spilled in work with inadequate materials. M. Witzel, Introduction, in Inside
the Texts, Beyond the Texts, ed. Michael Witzel (Cambridge, MA: Harvard University
Press, 1997), VIVII. Philology . . . is the fundament of our science. It is the founda-
tion on which we must build. The texts are our best source of testimony about classical
xiv introduction

In contrast, the reception and interpretation of the Mahbhrata

among audiences in India could only be of secondary concernand
that too, only insofar as it could illuminate the texts history and
transmission. In other words, if there was anything to learn from the
dra farmers, it could only be, paradoxically, what Western scholars
could tell them about where they stood in the hierarchy of the texts
Moreover, conventional scholarly wisdom held that they stood quite
low in this hierarchy. Present day dra farmers of Tamilnadu, so ran
the implicit reasoning, were not only geographically and historically
remote from the epics hypothesized origins among ryan tribes or
kingdoms in north India in the early first millennium BCE.13 Rather,
they were also cut off from the epic through distinctions of caste and
religious and social perspective. The epic, it was held, was the product
of a heroic ryan lineage and reflected its ethical and social norms.
dra farmers could not be further removed from the concerns of the
Ksatriya kings and heroes immortalized in this supposedly original
oral bardic epic.14 Nor could they claim, as the Brahmins couldwho,
it was asserted, had appropriated and refashioned the old Ksatriya-
centric poem to their own socio-political ends15to find their theolog-

India. . . . in many ways the only window that we have on classical Indian society. . . .
only through the creation of edited texts . . . can [we] begin to place these texts in their
proper context. Richard Lariviere, Protestants, Orientalists, and Brhman as: Recon-
structing Indian Social History (Amsterdam: Royal Netherlands Academy of Arts and
Sciences, 1995), 1617.
Cf. n. 25 of the Introduction to volume 1.
Cf. Hermann Oldenberg, who makes a distinction between the Hindu and the
Aryan and asserts that it is given to the Germans philologists to know the
Aryan of old India better than the British colleagues who live in his [the Hindus]
country and breathe his air. Indische und klassiche Philologie, Kleine Schriften, Teil
II, ed. Klaus L. Janert (Wiesbaden: Fritz Steiner Verlag, 1967), 15171518; originally
published in Neue Jahrbcher fr das klassische Altertum, Geschichte und deutsche
Literatur und fr Pdagogik 17 (1906): 19.
The thesis that the epic was originally the possession of the Ksatriya- or warrior-
class and only later taken over by Brahmins is a fundament of German epic scholarship.
One finds the claim throughout 19th century literature, especially in the form that there
was an older epic (the so-called Urepos), to which the Brahmins later added masses
of didactic, theological, and devotional material, ruining the simple noble lines
of the former. Although the earliest references to a Bhrata as opposed to a Mah-
or Great Bhrata may be found in Lassen (1837), it is Goldstcker who gives the
thesis its classic form: The groundwork of the poem, as mentioned before, is the great
war between two rival families of the same kin; it occupies the contents of about 24,000
verses. This, however, was overlaid with episodical matter of the most heterogeneous
kind. Nor was this merely matter of accident in the sense in which such a term
introduction xv

ical or philosophical doctrines reflected in it. If the remaining Indian

castes saw anything at all of value in the text, this could only be due to
the fact that they were the dupes of Brahmin ideologyand it was up
to a Western scholarly elite to enlighten them about their miscon-
ceptions.16 In other words, they needed to be taught how to read their
texts critically, i.e., to be attentive both to historical differences in the
text (to the layers, strata, rings, nodes, repetitions, interpo-
lations, and various other terms Western scholars had come up with)
and to the vested interests (primarily religious, but also political and
social) behind these historical changes to the text.17

might vaguely be used. A record of the greatest martial event of ancient India would
have emphatically been claimed as the property of the second or military caste, the
Kshattriyas. But such an exaltation of kingly splendour and of the importance of the
military caste, would as naturally threaten to depress that of the first or Brahmanical
caste. Brahmans, therefore, would endeavour to become the arrangers of the national
epos; and as the keepers of the ancestral lore, as the spiritual teachers and guides, as
priestly diplomatists, too, they would easily succeed in subjecting it to their censor-
ship. It became thus the aim of the Brhmanas to transform the original legend of the
great war into a testimony to the superiority of their caste over that of the Kshattriyas.
And this aim was effected not only by the manner in which the chief story was told,
but also by adding to the narrative all such matter as would show that the position
and might of a Kshattriya depends on the divine nature and favour of the Brhmana
caste. Here and there an old legend or myth might be found in the epos, apparently
not betraying such a set purpose. Theodore Goldstcker, Literary Remains of the Late
Professor Goldstcker, vol. 2 (London: W. H. Allen, 1879), 97-99. For an analysis of the
deep Lutheran ressentiment this recurring Gestalt of German epic scholarship reveals,
see our forthcoming The Nay Science: A History of German Indology.
The 19th century German scholar Albrecht Webers remark to the Prussian Min-
ister of Culture (Kultusminister) Karl Otto von Raumer is paradigmatic: The study
of Indian antiquity has, in the last fifteen years, with the availability of the oldest holy
scriptures of the Indians, the Vedas, gained unimaginably and increasingly in both
practical and academic significance. The practical significance has affected England in
particular and has been acknowledged both there and in India, by the Christian mis-
sions as well. The entire weight of the religious and cultural structure of contemporary
India appears to rest on the Vedas. As soon as they are unveiled from the mysterious
darkness surrounding them till now [sobald nun diese nicht mehr in ihr bisheriges
mysterises Dunkel gehllt sind], and made accessible to all, all the untruths shall be
automatically revealed, and this shall, in time, put an end to the sorry plight of reli-
gious decadence [dem traurigen Zustande der religiser Versumpfung] of India. The
critical analysis and publication of Vedic texts shall assume a role among the Indians,
similar to Luthers translation of the Bible. A. Weber, Letter to Karl Otto von Raumer,
12.10.1855 (Humboldt University Archives, P. F. 1433); translated and cited in Indra
Sengupta, State, University, and Indology: The Politics of the Chair of Indology at
German Universities in the Nineteenth Century, in Sanskrit and Orientalism: Indol-
ogy and Comparative Linguistics in Germany, 17501958, ed. Douglas T. McGetchin,
Peter K. J. Park, Damodar Sardesai (Delhi: Manohar, 2004), 278279.
To cite but two recent instances, Malinar, in her recent book on the Bhagavad
Gt, dismisses classical Indian philosophical commentaries on the Bhagavad Gt as
xvi introduction

Against the dominant prejudice, in an essay on the symbolism of the

Pndavas disguises written in 1980,18 Hiltebeitel had already rejected
the assumption that epic scholarships gaze had to be permanently
turned backward, seeking to find traces of an Urepos beneath the
detritus of what later ages had added to the Mahbhrata. Rather than
distinguishing a Bhrata (original, oral, heroic, bardic, and histori-
cal) as opposed to the Mahbhrata (derivative, written, Brahmanic,
and mythic), one could, he suggested, observe instance[s] of epic
themes[s] that could only be understood from such later sources.
One would thus have to, he writes,
. . . interpret the Mahhrata not only retrospectively but, in a sense, pro-
spectively. Possibly the epic simply anticipates later forms. More likely,
however, it evokes forms which we know of only (or largely) from later
sources, but which are earlier than is commonly thought. One would thus
need to recognize the pliancy and selectivity of an oral tradition in its
symbolic articulation of some of the fundamental continuities of Hindu
culture, for which the epic is not only the first great effort at synthesis
but a means to transmit this synthesis through the centuries, in India
and abroad. It is thus impossible to study the epic as a story frozen in
its Sanskrit textual forms. For one thing, there are good grounds to sus-
pect that certain features of the story descend from an Indo-Iranian and
Indo-European past. But more than this, one must assume that the epic
poets made selective use of oral traditions and popular cultural themes.
Preposterous as it sounds, considering the immensity of the text, one

well as modern Hindu interpretations of the text on the grounds that each author
establishes his own hermeneutics on the basis of the religious or philosophical tra-
dition he adheres to. Angelika Malinar, The Bhagavadgt: Doctrines and Contexts
(Cambridge: Cambridge University Press, 2007), 17. Are we to understand by this
statement that German scholars have achieved perfect Standpunktfreiheit? Or are we to
understand by it that their hermeneutics are acceptable, while Indian hermeneutics
are not? Or that every single Indian author does so, ignoring the important differences
between different schools, traditions, periods, philosophies, and standpoints? Surely,
one of the contributions of the text-historical school has been its greater sensitivity to
historical and textual variation, not to mention the subtle differences between differ-
ent schools and traditions? Yet von Stietencron, a scholar known for his contribution
to the dialogue between India and Germany, voices near-identical sentiments in his
foreword to Malinars 1996 book: The analytic thinking of Western interpreters who
were schooled in historico-philological methods stands in contrast to the traditional
Indian commentators, who not only harmonized and freely downplayed all breaks in
the text [i.e. the Bhagavad Gt], but, above all, sought to read their own philosophi-
cal-theological concepts out of individual textual passages, in order to secure Krsn as
divine authority for them. Angelika Malinar, Rjavidy: Das knigliche Wissen um
Herrschaft und Verzicht (Wiesbaden: Otto Harrassowitz, 1996), 1.
See n. 6 for the full citation.
introduction xvii

can pretty safely assume that the bards knew more about the main story,
both in terms of variants and underlying symbolism, than they told. It is
thus worth investigating whether what they left untold but implicit, or
what they alluded to through symbols, is not still echoed in the vast oral
and vernacular epic and epic-related traditions that perpetuate the story
to Indian culture to this day. I have come to suspect that living traditions
of and about the Mahhrata are often in close touch with traditional
epic meanings that have escaped the classically based literary scholars.19
In rejecting the old Indological prejudice of the older the more
authentic (Pollock),20 Hiltebeitel was also calling attention to the way
the text had been transformed, interpreted, and disseminated within
the Indian tradition: in Sanskrit dramas such as Bhatt a Nryan as
Ven isam hra (discussed in Draupads Hair)21 or in Villiputtr
lvrs Tamil rendition of the Mahbhrata in Villi Pratam (dis-
cussed in both Orders of Diffusion in Indian Folk Religion [1997]22
and in Draupad Cult Lls [1995])23 or the terukkttu tradition
of street-theatre (discussed in The Folklore of Draupadi: Srs and
Hair [1991]24 and in Transmitting Mahbhratas: Another Look at
Peter Brook [1992]25). In addition, Hiltebeitel proposed examining
the breadth of the South Asian epic tradition as it had been transmit-
ted in regional martial oral epicsthe Pbj epic from Rajasthan, the
Tamil Elder Brothers Story, the Telugu Epic of Palndu, and the lh
of the Hindi-speaking heartland of north Indiafor what it could tell
us about the classical Sanskrit epics.
But Hiltebeitel was not just drawing attention to oral and literary
re-creations or re-interpretations of the epic, but also to its place in the
ritual and sacrificial traditions of south India. In three ethnographic
studies written beteween 1995 and 1999, Hiltebeitel discusses a cult

Hiltebeitel, iva, the Goddess, and the Disguises of the Pndavas, 151152.
See Sheldon Pollock, Is There an Indian Intellectual History? Introduction to
Theory and Method in Indian Intellectual History, Journal of Indian Philosophy
36,5 (2008): 541.
See n. 1 for the full citation.
Alf Hiltebeitel, Orders of Diffusion in Two Tamil Mahbhrata Folk Cults,
South Asian Folklorist 1 (1997): 936.
Alf Hiltebeitel, Draupad Cult Lls, The Gods at Play: Ll in South Asia, ed.
William Sax (New York: Oxford University Press, 1995), 20424.
Alf Hiltebeitel, The Folklore of Draupadi: Srs and Hair, in Gender, Genre, and
Power in South Asian Expressive Traditions, ed. Arjun Appadurai, Frank J. Korom,
and Margaret A. Mills (Philadelphia: University of Pennsylvania, 1991), 395427.
Alf Hiltebeitel, Transmitting Mahbhratas: Another Look at Peter Brook, The
Drama Review 36,3 (1992): 13159.
xviii introduction

that exercises a remarkable fascination on him for its wealth of clues

to the interpretation of the Mahbhratas sacrificial, votive, and tragic
meanings: that of Kttn t avar/Aravn, the son of one of Draupads co-
wives, the snake-princess Ulp. Aravn is celebrated both at Draupad
festivals and in a Tamil cult of his own for offering his body in sacrifice
to the goddess Kl and thus enabling Draupads five husbands to win
the Mahbhrata war. In particular, the resonance that the cult enjoys
among the lower status Vanniyr (a dra community whose members
regard themselves as Ksatriyas, as it were in a disguise forced upon
them by history)26 and the Alis (Tamil transsexuals or eunuchs who
show up for the main festivities at Kvkkam, in large numbers
perhaps up to a thousand not only from throughout Tamilnadu but
from Tamil Ali communities in cities all over India, and also from
Singapore)27 puts paid to the notion of the Mahbhrata as a Brah-
min imposition upon an unwitting lower-caste audience. The Alis are
enthusiastic participants in Kttnta vars main procession, whom they
marry in a ritual ceremony prior to his sacrificial death and mourn,
as Hiltebeitel notes, in scenes reminiscent of the Strparvans heart-
rending depictions of the grieving Kaurava widows. More than their
unfolding of these hypotheses, however, Hiltebeitels ethnographic
studies are remarkable for illuminating an aspect of the Mahbhrata
that has been all too rarely taken into consideration in Western schol-
ars thematization of the epic: its analysis of war and sacrifice as gen-
era of becoming with the intent of providing a soteriological response
to the problem of becoming.28 Here Hiltebeitels description of the
climax of the Kvkkam festivities is worth citing for its expressive
qualities, which succeed in communicating something of the ecstatic
rapture the audience experiences through participating in this ritual-
ized cycle of loss and renewal:
The climax of the festival begins around 5:30 a.m. Awaiting the return of
the ketyam procession, a vast crowd fills the open area among the vil-
lage temples. On top of the Kttn t avar temple, cocks are offered: since

Hiltebeitel, iva, the Goddess, and the Disguises of the Pn davas, 169.
Alf Hiltebeitel, Dying before the Mahbhrata War: Martial and Transsexual
Body-Building for Arvn, Journal of Asian Studies 54,2 (1995): 453454.
For a discussion of the four genera of becoming (sacrifice, cosmology, genealogy,
and war or agn) and on what is meant by becoming, see Vishwa Adluri, Sacrificial
Ontology and Human Destiny in the Mahbhrata (unpublished manuscript). The four
genera are crucial to understanding the Mahbhrata as the entire epic is articulated
in terms of these four genera.
introduction xix

the recent prohibition on animal sacrifice, no longer cut but thrown up

to the temple roof. Finally, the post or kampam on Kttn t avars tr
stands ready for the building of the gods body. Already in place are the
horses, forearms, feet, bells, and two flags (koticclai) from Nttan ham-
let, and straw to fill out the frame. The tiruvcci and umbrella (kutai),
removed from the ketyam, are danced toward the tr and raised to a
support atop the kampam, followed in similar fashion by the umbrella,
from Cevaliyankulam village, and the winged half-tubular epaulets,
raised to shoulder position. While the head is danced around the tr,
the gods shoulders, beneath the epaulets, are filled with flowers: strings
of jasmine thrown up to the devotees assembling the deity on the body-
scaffold. Penultimately, the mr patakkam is raised to the chest. Last,
the head is raised and set on the poles top. A great boom of firecrackers
goes off; 200 kilograms of smoky camphor is lit in front (east) of the
deity; the entire concourse flings strings of jasmine, spectacularly filling
the air before the flowers land on the gods body as two lakhs worth
of garlands (200,000 rupees). The Alis fling garlands removed from their
hair (Nrull 1990: 40). Devotees on the gods frame gather the flowers
onto and into his body.
Aravn thus goes forth in heroic pose, ready to fight, his frame-and-
straw body complete, decorated, draped and filled with flowers. The
giant image is drawn forth as a vran to repeat the march around the
village streets. Now, however, he is said to begin looking sad: happy as
a head, he is sad as embodied. Exhibiting tears and sweat, he is like
Nala, as one informant put it. As he prepares for the rites that recall his
kalappali and eighth-day fight, his embodied state registers his readiness
for sacrificial suffering and death. While his tr goes around the village,
he takes on an increasingly dead look, and when he reaches the south-
ern street where the Alis rent, the Alis have a married man with them
who provides them with a white saree for their collective widowhood.
Here they begin to ritualize Aravns death, undeterred by the additional
deaths to follow.29
Certainly, the scholars who had posited Indo-European epic as the pri-
mary framework for understanding the Mahbhrata were barely, if at
all, aware of the existence of this parallel tradition of transmission and
ongoing interpretation. But even making room for some awareness of
early reports, how was one to decide in favor of one tradition or the
other? Aside from circular argumentation, where everything that was
thought not to belong to the original Indo-European or ryan epic was
excised as late, what evidence was there for an original Bhrata?
In particular, could one account for the Mahbhrata in terms of an

Hiltebeitel, Dying before the Mahbhrata War, 459460.
xx introduction

epic core plus accretions, or were such attempts hopelessly reduc-

tive? And, more immediately, what was the evidence for considering
the south Indian folk and vernacular traditions as non-epic and as
deriving, at least in part, from an alternative source than the Vedic
and ryan sources posited for the Mahbhrata? In the third of his
three ethnographic studies of the Kttnta var cult, Hiltebeitel offers
some conclusions that not only demonstrate how problematic this
hypothesis is, but also how limited scholars interpretive frameworks
were in attempting to account for the Mahabharata tradition in all its
Kttn t avars recurrent lives, deaths, and reanimations, in fact, seem to
spill into and out of all conventional boundaries, including those Hindu
ones that normatively define death, reanimation (the law of karma),
and the rites of establishment (pratisth ) and dismissal (visarjana)
whereby deities are invited to enter into icons, enliven them with their
breath (prn a) or Self (tman), and depart from them. But there is little
if anything in the Kttnt avar cult that one can ascribe to tribal or other
un-Hinduized influences, and muchbeginning with the Vedic conun-
drum of the severed head and the dismemberment of Purusathat taps
deep Hindu sources.30
Hiltebeitels ethnographic studies of the Draupad and Aravn cults
were also to lead to a broadening of the definition of the epic tradi-
tion. V. S. Sukthankar, the editor and initiator of the Critical Edi-
tion of the Mahbhrata, had already remarked in 1933 that the
Mahbhrata is the whole of the epic tradition: the entire Critical
Apparatus.31 Hiltebeitel, however, was now proposing to broaden
the definition of epic tradition to include ritual and performance
in addition to the manuscript tradition. One would thus, he sug-
gested, have to posit an underground Mahbhrata32 in addition

Alf Hiltebeitel, Kttnt avar: The Divine Lives of a Severed Head, in Ways of
Dying: Death and its Meaning in South Asia, ed. Elizabeth Schmbucher and Claus Peter
Zoller (Delhi: Manohar, 1999), 309310.
Vishnu S. Sukthankar, The Mahbhrata for the First Time Critically Edited,
vol. 1: diparvan (Pune: Bhandarkar Oriental Research Institute, 1933 [1997]), cii.
Cf. Hiltebeitel, The Folklore of Draupadi: Srs and Hair, 395: In this chapter
I would like to address some folkloric material bearing on the same twin subjects
[folklore of sarees and hair] and discuss it toward some additional ends, taking up
the wider issue of pan-Indian Mahbhrata folklores, and raising the question of the
relation between the distinctly Tamil folklore about Draupad that is found in her
cult and wider pan-Indian themes. Are Tamil and other south Indian Mahbhrata
folklores (some of which are almost certainly older than the Draupad cult) a source
of diffusion for similar themes found elsewhere in India? Or does the classical epic just
suggest common folk responses? Is there a sort of underground Mahbhrata, one
introduction xxi

to and alongside the text recovered in the Critical Edition, and this
underground tradition was of no less significance in determining the
epics meaning than scholarly theories on the literary Mahbhrata.33
Indeed, one could not, ultimately, separate the two, for, as Hiltebeitel
remarks in the conclusion to the article where he first introduced the
idea of an underground Mahbhrata,
. . . there may well be an underground folk Mahbhrata. But it cannot
be monolithic. It has no prototype outside the Sanskrit text (which can
never be assumed to have fallen out of the folk epic frame of refer-
ence). If such a folk Mahbhrata exists, however, it would seem to
be centered on images of the goddess and the control of the land. Its
lines of transmission and adaptation are too vast to ever trace fully. But
those lines that do emerge suggest the crossing of many geographical
and linguistic boundaries, and symbols and motifs that recur in a wide
spectrum of reflexive and interpenetrating genres: from Mahbhrata
vernaculars to folk dramas, from folk dramas to ritual idioms, from
ritual idioms to temple tales, from temple tales to sisters tales, from sis-
ters tales to regional folk epics, from regional folk epics to Mahbhrata
In the present volume itself, one can see how the cross-fertilization
of ethnographic and textual approaches leads to complementary out-
Indeed, not all of Hiltebeitels essays in the volume are, as the
subtitle might lead one to expect, examples of ethnography, strictly
speaking. Only those at the center of the book, in the section titled
The Sacrificial Death of a Co-wifes Son, are fully ethnographies in
the everyday anthropological sense of writing about culture, based
immediately on the authors fieldwork.35 Yet, all the other essays in this

that is perhaps even reflected in the Sanskrit epic itself but also different from it in cer-
tain basic accentuations concerning the goddess? What are some of the features that
distinguish Mahbhrata folklores from other Indian folklores? And how are folkloric
themes concerning the Mahbhrata related to distinctive modes of transmission and
performance? How and why is the Mahbhrata linked with certain regional folk
epic traditions and not with others? And in such regional folk epic traditions where
there is a connection, how do we understand their portrayals of virgin heroines at the
center of conflicts over land? Draupads sarees and hair provide a fitting entre into
the problematics of such questions. (This article has been reprinted as chapter 5 of
the present volume.)
Literary here being meant in the limited sense of the narrative and not, as Hilte-
beitel has been arguing since the early 90s, in the broader sense that the Mahbhrata
is a work of conscious literary and artistic design.
Hiltebeitel, The Folklore of Draupadi, 421.
Section 2 includes three previously published articles and one new one on
Kttnt avar. These four articles together form a little book within this larger one and
xxii introduction

book are ethnographic in wider senses. Minimally, they draw on Hilte-

beitels fieldwork on the Draupad cult, with which the Kttn t avar
cult, as he demonstrates, shares a history and terrain. In that regard,
those in Parts I and III are essays that surround his strictly ethno-
graphic work, and open paths to and from Hiltebeitels Draupad cult
and Kttn ta var cult ethnographies into mainly (but not exclusively)
wider South Asian terrains with which they can be profitably com-
pared. In The Cult of Draupad, vol. 1, which first established his field-
work project, Hiltebeitel wrote of convergences that further widen
the scope of our inquiry into other topics with which the Draupad
cult intertwines, including the mythologies of local village god-
desses, the Hindu mythology of the goddess in its greatest extent, and
the Mahbhrata. There he went on to mention a hypothesis that he
had begun to investigate in the essays that now provide chapters 1 to 4
of this book: despite the fact that the classical Mahbhrata makes
little direct reference to the goddess, the epic narrative would seem to
be informed by the goddesss mythology. If this is so, it means that a
folk interpretation of the Mahbhrata that places the goddess at its
center has every chance of revealing much to us about the classical epic
itself. And that is what the South Indian Draupad cult is: an adroit
and compelling multileveled interpretation of a living Mahbhrata.36
We may now say, with regard to the subtitle of this book, that Hilte-
beitel was beginning to envision the project of writing an ethnography
not only of the Draupad cult, but of the classical epic itself and the
culture (or cultures) that can be hypothesized to have produced and
disseminated it. We see some of the outcomes of such intentions in
this volume and in volume 1. In this volume, moreover, we can also
see inklings of ethnographies of more modern communities of inter-
pretation: from those of British colonial administrators to schools of
interpretation found in Western and Indian academia, from those who
stage Mahbhrata pageants in downtown Bangalore to those who put
the epic on the international stage.

round off Hiltebeitels longstanding wish to produce a balanced interregional study of

the cult of the god Kttnt avar.
Alf Hiltebeitel, The Cult of Draupad, vol. 1: Mythologies: From Gingee to
Kuruksetra (Chicago: University of Chicago Press, 1988), 135.
introduction xxiii

Millennial Draupads: Part I of volume 2

Like volume 1, volume 2 opens with chaptersthe first fivethat for-

mulated and set the stage for a program of research: in this case, one that
carries through (not without some reconsiderations and reservations)
through all the chapters that make up this book. The first three chap-
ters explore working assumptions37 that opened up new positions for
Hiltebeitels research in the late 1970s to early 80s.
Unlike the literary turn that shaped a new direction for his work
on the Mahbhrata in circa 199238 and provides the backbone for vol-
ume 1, the essays in volume 2 developed these working assumptions
around two main topics. One was the study of The Goddess in India
and Beyond (the title of a course Hiltebeitel created, and often gives
at George Washington University).39 The other topic was the study of
symbolism, particularly that of womens hair, srs, and jewels, along
with the theme of disguises, as seen in chapters 15. Through these
studies and convergences, Hiltebeitel was attempting to work out
the relationship between ethnographic findings, principally in Tamil
culture (and more extensively, south Indian cultures with Dravidian
languages), and the classical Sanskrit Mahbhrata text. That rela-
tionship began to be well formulated in the three studies of chapters
13. These formulations sought to focus on the goddess through the
epic, and the question of what light Mahbhrata ethnography and
the Sanskrit epic can mutually shed on each other: a matter in which
the goddess turned out to be pivotal. In effect, it could be suggested
that it was the study of the goddess that would force open the lit-
erary turn that followed, by calling attention to the literary artistry
through which Draupad and other heroines of the Sanskrit epics were

See pp. viiviii above.
See now the Introduction to volume 1 of Collected Essays by Alf Hiltebeitel; cf.
especially pp. viixxxii.
Hiltebeitels The Indus Valley Proto-iva: Reexamined through Reflections on
the Goddess, the Buffalo, and the Symbolism of vhanas (chapter 15) looks to the
cult of the goddess Durg and her combat with the Buffalo Demon Mahissura to see
if it might shed some light on a well-known figure on an Indus Valley seal. Other god-
desses, including especially Sat (chapter 16) and Kl (chapter 1114, 17), also figure
prominently in this book, as does a close look at two south Indian buffalo sacrifices
in chapter 19.
xxiv introduction

portrayed.40 Whereas chapters 1 (1981)41 and 2 (1980)42 are pieces that

began a search to look at Draupad in the epic more or less on her
own, chapter 3 (1980)43 is an investigation into the Virtaparvan and
takes up the idea that not only her disguise shows her relation to the
goddess, but likewise her husbands disguises show how each one of
them is capable of reminding us of iva. All three of these articles
had behind them the beginning years of Hiltebeitels ethnographic
research on the south Indian Draupad cult. Chapter 4, Purity and
Auspiciousness in the Sanskrit Epics (1985),44 begins to register that
the same questions and formulations can be productive in studying
the Rm yan a. Chapter 5 (1991)45 then marks a rounding off of these
working assumptions by arriving at some tentative conclusions formed
by the late 1980s. The conclusion of chapter 5 then points further
toward Hiltebeitels literary turn46 where it says, there may well be an
underground folk Mahbhrata. . . . [But] It has no prototype outside
the Sanskrit text. . . .47
Chapters in this book from the early 90s on begin to reflect Hilte-
beitels growing involvement in the George Washington Univer-
sity Human Sciences Program, which he was to direct from 1997
to 2002. Many subsequent chapters reflect the widening of interdis-
ciplinary questions that interested Hiltebeitel in that period, includ-
ing post-colonial studies (chapter 8; 1992),48 gender studies (chapter

For a demonstration of the validity of this suggestion now in his most recent
work, see Alf Hiltebeitel, Dharma: Its Early History in Law, Religion, and Literature
(New York: Oxford University Press, 2011), chapter 8.
See n. 1 for the full citation.
Alf Hiltebeitel, Draupads Garments, Indo-Iranian Journal 22 (1980): 97112.
See n. 6 for the full citation.
Alf Hiltebeitel, Purity and Auspiciousness in the Sanskrit Epics, Purity and
Auspiciousness in India, ed. Frdrique Apffel Marglin and John Carman, Journal of
Developing Societies 1 (1985): 4154.
See n. 24 for the full citation.
Although my own work on the Mahbhrata has taken various turns and
grounded itself in changing scholarly approachesIndo-European studies, history of
religions, anthropology, historyI believe that the largest inadequacy in Mahbhrata
scholarship, including my own up to 1991, is simply the failure to appreciate the epic
as a work of literature. Alf Hiltebeitel, Reconsidering Bhrguization, in Composing
a Tradition: Concepts, Techniques, and Relationships, ed. Mary Brockington and Peter
Schreiner (Zagreb: Croatian Academy of Sciences and the Arts, 1999), 156.
Chapter 5, 124. These concerns were further and more extensively explored in
Hiltebeitels Rethinking Indias Oral and Classical Epics (Chicago: University of Chi-
cago Press, 1999), of which chapters 17 and 18 can be called spin-offs, and in the
essays that make up chapters 1114, all focused on the Kttnt avar cult.
See n. 10 above for the full citation.
introduction xxv

10; 2000),49 and psychoanalysis (chapters 12 [1998]50 and 16 [1999]).51

Running through several chapters is also a basic determination to
contextualize both ethnographic and textual scholarship historically,
and to resist ahistorical formulations about either the goddess or
the Mahbhrata, whether in local, regional, or pan-Indian forms.52
Orders of Diffusion in Indian Folk Religion (chapter 6; 1997)53
raises wider questions about the problem of diffusion in thinking
about the goddess, both as played out on the local level with different
aspects of the Draupad cult and investigating likewise the problem
of thinking about wider questions of diffusion from sources outside
India (ancient Near East, etc.). Draupad Cult Lls (chapter 7;
1995)54 is a comparison of drama performances at Draupad festivals
and at Rm- and Ras Lls in north India, which are focused on Rma
and Krsn a. Colonialist Lenses on the South Indian Draupad Cult
(chapter 8) is an examination of how early reports on the Draupad
cult and other south Indian goddesses set an early agenda that later
scholars have had to consider in looking at the cult. Review of Land-
scapes of Urban Memory (chapter 9; 2003)55 is a discussion of how a
modern interpreter of a very different kind of Draupad festival found
in Bangalore, but still part of the same tradition, has raised questions
that still call for further discussion of some of the deeper traditions
of Draupad worship. Finally, Draupads Question (chapter 10),56
which is addressed on a textual level in Hiltebeitels Rethinking the
Mahbhrata: A Readers Guide to the Education of the Dharma King,57

Alf Hiltebeitel, Draupads Question, in Is the Goddess a Feminist? ed. Kathleen
Erndl and Alf Hiltebeitel (London: Sheffield, 2000), 11322.
Alf Hiltebeitel, Hair Like Snakes and Mustached Brides: Crossed Gender in an
Indian Folk Cult, in Hair: Its Meaning and Power in Asian Cultures, ed. Alf Hiltebeitel
and Barbara D. Miller (Albany: State University of New York Press, 1998), 14376.
Alf Hiltebeitel, Fathers of the Bride, Fathers of Sat: Myths, Rites, and Scholarly
Practices, Thamyris: Mythmaking from Past to Present 6,1 (1999): 6594.
Chapters 11 and 14 on eastern and western wings of the Kuttn t avar cult, and
chapter 16 on sat, in particular, are insistent on formulating historical hypotheses on
the background of cults and myths.
See n. 22 for the full citation.
See n. 23 above for the full citation.
Alf Hiltebeitel, Review of Landscapes of Urban Memory; the Sacred and the Civic
in Indias High-Tech City, by Smriti Srinivas, Journal of Asian Studies 62,1 (2003):
See n. 49 above for the full citation.
Alf Hiltebeitel, Rethinking the Mahbhrata: A Readers Guide to the Education
of the Dharma King (Chicago: University of Chicago Press, 2001).
xxvi introduction

takes up not only a brief discussion of Draupads humiliation in the

Kuru court, but a modern fictional relocation of that story into tribal
and Marxist confrontations with the government in Bihar.
These four articles form a kind of subset within this first section and
give support to Hiltebeitels sense that Draupad is actually central to
the Sanskrit epic; in his words, one of the epics primary foci is the
question of who Draupad is as a figurea rebel, a figure who is inde-
pendent, vigorous, challenging, a principled woman, a very different
kind of woman, intellectually shrewd, on top of things to the extent it
is possible to be on top of such things.58

The Sacrificial Death of a Co-Wifes Son and Companion Studies:

Parts II and III

At the center of this book are four chapters on a cult that cross-sections
with the Draupad cult: that of Kttnta var/Aravn. These are ethno-
graphic pieces, so the literary turn is not as major an influence here as
studying the distinctive and much less accessible cult and temples of
Kttnta var, which Hiltebeitel sought to map in the mid to late 1990s.
Although it is doubtful that he found them all, he was able to find 44
through mainly oral sources. Hiltebeitels main interest in these pieces
is in figuring out how Aravn/Kttnta var becomes central to a cult of
his own outside the Draupad cult, and why this cult is so very differ-
ent in its stories and rituals in different parts of northern Tamilnadu.
In this section, one of the questions that is important is the relation
between these two cultsDraupads and Kttn t avars. The first three
essays, written in the 1990s,59 were all concerned mainly with different
facets of the most publicized and exuberant festival for Kttn ta var
found in the village of Kvkkam. There and at some other nearby
villages, it is a cult which features very prominently the role of trans-
vestites, and this raises major gender-related questions across the cult
in other areas as well, especially around the question of who it is that
marries Aravn, since in Kvkkam it is none other than Krsn a who

Personal communication.
Hiltebeitel, Dying before the Mahbhrata War; Hair Like Snakes and Mus-
tached Brides; The Divine Lives of a Severed Head.
introduction xxvii

takes the form of Mohin. Chapters 11 to 13 have undergone the most

revision of any chapters for this book so as to avoid overlap and build
up three separate arguments centered on the spectacular worship of
Kttn t avar at Kvkkam village, located at the eastern end of the belt
of Kttnta var cult diffusion.
The last chapter of this section, Kttn t avars Cross: Making That
Young Bride, Whoever She Is, a Widow, is an entirely new article,
written just for this book, which filled an obligation to round this
research out. It focuses on the festival at a temple and stories of Aravn
at the other end of the map as far as Kttn t avar worship goes: in
Coimbatore in southwestern Tamilnadu, where Kttn t avars body is
made of a cross. Here, what we learn about Kttn t avar from other
temples, including Kvkkam, is very germane as is the case in reverse.
There seems to be history here and a way in which we can identify this
history around the different brides who marry him. In Coimbatore, he
marries a totally different woman from Mohin. This essay begins with
the question of what it means for Draupad herself to be relying upon
the death of a co-wifes son to bring about the death she seeks, and
concludes with a consideration of the possibilities of familiarity with
Jesuit missions in the 16th century. The Kttn t avar cult findings offer
a virtual laboratory on the relation between Mahbhrata cults, and
on the goddess and the epic in local and regional variations.
Finally, Part III of this book features six essays on broader themes
relating to the epic under the title Companion Studies. These are
essays which were all written in conjunction with larger questions
raised by Hiltebeitels research on the Draupad cult and the goddess.
The Draupad cult is in the background of these essays, but it is not at
the center of what they are about. The section begins with a 1970s arti-
cle that focused on the as-yet-unresolved question of the figure on the
Indus Valley seal, that many have called the Indus Valley Proto-iva
(chapter 15; 1978).60 This is a landmine area for anyone, but this article
does state a position that still has attractions for scholars interested in
this discussion.61 Fathers of the Bride, Fathers of Sat: Myths, Rites,

Alf Hiltebeitel, The Indus Valley Proto-iva: Reexamined through Reflections on
the Goddess, the Buffalo, and the Symbolism of vhanas, Anthropos 73 (1978): 76797.
An interesting and provocative has been published by A. Hiltebeitel, who drew
heavily on the work of B. Volchok, one of the Russian scholars who worked on their
xxviii introduction

and Scholarly Practices (chapter 16; 1999)62 brings Draupads incar-

nation as Bel in the Hindi lh epic into a discussion of women who
become Sats and what their relations are with their fathersone of the
problematic tensions in Sat mythology, if not also in practice. Two
Ways to Tell a Story: lh in the Bhavisya Purn a (chapter 17; 1999)63
is in the best tradition of the way that an oral epic is, as it were, liter-
ally Sanskritized by being retold in a Purn ic setting. Draupad figures
in this story, reborn as Bel. Boar and Twins: Comparing the Tulu
Kti-Cennaya Pddana and the Tamil Elder Brothers Story (chapter
18; 2005)64 takes up the question of whether south Indian Mahbhrata
folklores enter into a Tulu (South Kanara) version of a story also found
in Tamilnadu, about twin brothers combat with a giant boar. On the
Handling of the Meat, and Related Matters: Two South Indian Buffalo
Sacrifices (chapter 19; 1985)65 compares a buffalo sacrifice in the fort at
the south Indian capital where Draupad has her original temple with
vivid archival documentation of a 19th century buffalo sacrifice from
Karnataka. Finally, in chapter 20 (1992),66 Peter Brook is the subject of a
discussion of the way Indian theatrical representations of Mahbhrata
were looked at by Brook, what he got from them, and what he missed
in particular, by snubbing the performances of the terukkttu in Tamil-
nadu, which Hiltebeitel arranged for him to see.
The title of this book is meant to guide the reader through these three
parts: first through two millennia of varied representations of Draupad;
then into a companion cult of the Tamil Draupad cult, on which Hilte-
beitel has centered much of his Mahbhrata ethnography; and then on
to companion pieces that follow up trails opened up by this ethnogra-
phy from the Indus Valley civilization to the 20th century international

attempted decipherment of the Indus script. Hiltebeitels critique is much like Srini-
vasans, but he makes much of the fact that the horns on the central figure are those
of a buffalo . . . but not proved. Gregory L. Possehl, The Indus Civilization: A Contem-
porary Perspective (New Delhi: Vistaar Publications, fifth printing 2009; first pub. 2002
by AltaMira Press), 142.
See n. 51 above for the full citation.
See n. 7 above for the citation.
Alf Hiltebeitel, Boar and Twins: Comparing the Tulu Kti-Cennaya Pddana
and the Tamil Elder Brothers Story, in In the Company of Gods: Essays in Memory of
Gunther Dietz Sontheimer, ed. A. Malik, A. Feldhaus, and H. Brckner (Delhi: IGNCA
and Manohar, 2005), 14175.
Alf Hiltebeitel, On the Handling of the Meat, and Related Matters: Two South
Indian Buffalo Sacrifices, Divisione delle Carne: Dinamica Sociale e Organizzazione
del Cosmo, ed. Christiano Grottanelli LUomo 9 (1985): 17199.
See n. 25 above for the full citation.
introduction xxix

stage of Peter Brook. The title points to an especially salient feature of

Draupad: that she is an incarnation of a goddess named r. It would be
a bit incautious to say that When the Goddess Was a Woman is a title
that could only apply to Draupad. The Sanskrit Rmyan a of Vlmki
gives some hints that it could also apply to that epics heroine St, and
in some later versions of the Rmyan a it certainly does. But in the
Mahbhrata, the goddess having become a woman is a more salient
and complex matter than it ever is for St. And as Hiltebeitel found,
it is also one that is more open to theoretical, historical, ethnographic,
and literary questions. Together with the unfolding implications of the
title, the three-part structure is one through which theoretical, historical,
ethnographic, and literary threads can be traced through the book, tying
together themes from the three different sections.

Exegesis: Mythic and Ritual

As we have laid out the two volumes in this series, we have made a dis-
tinction between the more theoretical studies of the Sanskrit epics in
volume 1 and the ethnographic and topical treatments of the goddess
and her related cults in volume 2. In spite of their differing thematic
foci, however, we see the two volumes as complementary. One can-
not fully grasp Alf Hiltebeitels contribution to the study of the epic
without also reading volume 2. Nor can one hope to understand the
significance of Hiltebeitels turn to the reception of the Mahbhrata
among dra communities in south India in the late 70s and early
80s without first having some understanding of the prejudices of epic
scholarship against such a move until that point in time. Hiltebeitel
can rightly be called the first scholar to have taken the Mahbhratas
claim to being a Veda for all classes and beingsa str-dra-veda
seriously.67 As Hiltebeitel shows in his studies of the cult of Aravn,

On the epics popular title as a str-dra-veda or a Veda for women and
dras, Blacks discussion in a forthcoming article is especially useful. Black notes
that Although the Critical Edition does not contain the well known description of
the epic as a text for women and dras, the Mahbhrata does seem to regard itself
as delivering a universal message. In addition to the numerous phalarutis through-
out the text that address audiences beyond those who are male and of the twiceborn
classes, Vysa himself, in the ntiparvan, instructs his disciples to teach his story
to members of all four varnas (12.314.45). In light of the authors own instruction
to his students, what better way to reach a diverse and inclusive audience than to
have Brahmanical knowledge communicated by someone of lower birth. Indeed, with-
out making any claims about the real history of the text, this scenario seems to be
xxx introduction

it becomes increasingly difficult to sustain the fiction of the epic as

religious rhetoric or as an ideological vehicle in light of the con-
crete evidence of the reception of the text among non-Brahmanical
communities in India.68
Such an approach need not imply a refusal to take the epic seriously
as a work of history. Rather, it requires a willingness to suspend total-
izing constructions of history such as the ryan hypothesis (Holtz-
mann, Oldenberg, Jakob Wilhelm Hauer)69 or the racial hypothesis of
white Aryans [die weissen Arier] versus black natives [schwarzen

the one that the Mahbhrata tells about its own transmission: originating among
brahmins, but learned by stas such as Ugraravas who, implicitly, share such tales
and legends with a wide audience, particularly when they frequent popular pilgrim-
age sites, such as the ones Ugraravas visited before arriving in the Naimisa For-
est. Brian Black, Transitions and Transmissions in the Mahbhrata: Revisiting the
Ugraravas/aunaka Frame Dialogue, in Revisiting Transitions in Indian History, ed.
Ranabir Chakravarti and Kumkum Roy (Delhi: Oxford University Press, forthcom-
ing), 11. On the term str-dra-veda itself, Black notes that the description appears
in the Bhgavata Purna (1.4.25), which says that Vysa composed his story out of
compassion for women, dras, and uneducated twice-borns. Nonetheless, there are
a number of individual phalarutis throughout the text that offer rewards for dras
and women. Ibid., n. 16. (This note repeats a discussion found in Vishwa Adluri,
Frame Narratives and Forked Beginnings: Or, How to Read the Adiparvan, Journal
of Vaishnava Studies 19,2 [Spring 2011]: 192, n. 7.)
See especially Raf Gelders and Willem Derde, Mantras of Anti-Brahmanism:
Colonial Experience of Indian Intellectuals, Economic and Political Weekly, vol. 38,
no. 43 (Oct. 2531, 2003): 46114617. As the authors, both social scientists, demon-
strate, there is little to suggest the kind of hierarchical top-down imposition of Brah-
manic ideology that has been a central trope of German scholarship on the epic until
now. As Hiltebeitels field studies of the Aravn cult demonstrate, the epics textual
porosity has allowed a number of communities either to find themselves reflected in
some aspect of it or to positively appropriate and retell the narrative in a way more
germane to their specific context.
Jakob Wilhelm Hauer was a founder the Aryan Seminar (das Arische Seminar)
at the University of Tbingen and a member of the SS and SA. Interned after the
war and found guilty of collaboration with the Nazis, Hauer was banned from teach-
ing until 1950. On Hauers life and work, see l Bauman, Die Deutsche Glaubens-
bewegung und ihr Grnder Jakob Wilhelm Hauer (18811962), trans. Alma Lessing
(Marburg: Diagonal Verlag, 2005). The classic Hauer biography is Margarete Dierks
Jakob Wilhelm Hauer: 18811962. Leben, Werk, Wirkung, mit einer Personalbibliogra-
phie (Heidelberg: Schneider, 1986), but it is almost entirely inaccurate; Dierks herself
was a member and sympathizer of the NSDAP and her Hauer biography a patent
attempt at rehabilitating him; see her dissertation Die preuischen Altkonservativen
und die Judenfrage 1810/1847 (Ph.D. Diss., University of Rostock, 1939). On the
Aryan Seminar, see Horst Junginger, Das Arische Seminar der Universitt Tbin-
gen 19401945, in Indienforschung im Zeitenwandel: Analysen und Dokumente zur
Indologie und Religionswissenschaft in Tbingen, ed. Heidrun Brckner and Angelika
Malinar (Tbingen: Attempto Verlag, 2003), 177207.
introduction xxxi

Urbewohner]70 (Lassen) or the theory of a Brahmanic Counter-

Reformation71 (Holtzmann) against the reformatory impulse of
Indian Buddhism. As a passage from Hiltebeitels 1980 article on the
disguises of the Pn davas demonstrates, the question of the place of
the Mahbhrata in the history of Hinduism is a crucial one, but one
that can only be answered through a sensitive and nuanced appre-
ciation of how this text evolves in concert with this history. Such an
inquiry requires us to take seriously all the available evidence: histori-
cal, textual, ritual, social, and mythic.
First, I regard the Mahbhrata as a text which attempts a great syn-
thesis of Indian civilization in the name of Hinduism. By synthesis,
however, something different is meant from the confluence of Epos
and Rechtsbuch stressed in the last century by Joseph Dahlmann. The
recent work leading to an understanding of this synthesis has been car-
ried out by scholars who have stressed the transpositions or connec-
tions worked out by the epic poets in relating the story to para-Vedic (in
some cases Indo-European), Vedic, Brhman ical, and Upanisadic sym-
bols, myths, and rituals, and also to the mythic material fully developed
for the first time in the background myths told in the course of the
narrative itself. But it becomes increasingly clear that a full understand-
ing of this synthesisand thus of the place of the Mahbhrata in the
history of Hinduismrequires a recognition that the epic also evokes,
through its symbolism, certain cultural themes, myths, ritual practices,

Since the Pnkla assuredly belonged to the Aryan races, we may not interpret
the relationship between them and the Pndava as though these ought to be described
as belonging to the black aborigines [Urbewohnern] of India on the basis of the black
color that is attributed to Krishn, these [others] as the white Aryans [weisse Arier].
Nonetheless, the distinction according to color must have a significance and this
can only be that the Pnkla, like the Jdava, who are represented by Krishn a, both
belonged to the Aryan races who entered [India] earlier, [and] became darker through
the influence of the climate as the youngest entrants from the north, and in contrast to
these are called the black. C. Lassen, Indische Alterthumskunde (Leipzig: Verlag von
L. A. Kittler, 1867), 791; editors translation.
Adolf Holtzmann, Zur Geschichte und Kritik des Mahbhrata (Kiel: C. F. Hae-
seler, 1892), 98. The entire passage is of interest for its evocation of the horrors of this
supposed conservative reaction: Around the time of the beginning of our calendar,
the hard and bloody battle [hart und blutige Kampf ] seems to have gotten underway,
in whose course the resurrected Brahmanism finally became master over its danger-
ous foe. Incidentally, the course of the gigantic battle [Riesenkampfes] can be a
matter of indifference to us here, its result at any rate remains firm. Buddhism was
violently exterminated in those areas where it had blossomed for centuries. The causes
are clear: . . . a violent effort, an energetic Counter-Reformation [Gegenreformation] of
the Brahmins, which, through its adoption of the peoples gods especially the beloved
Vishn u-Krshn a, and of the entire folk-superstition [Volksaberglauben], had won over
the masses for itself. . . . (editors translation).
xxxii introduction

and social norms that are not fully attested historically until post-epic
times, sometimes in later texts, sometimes even in contemporary folk
cults and practices.72
In other words, one cannot simply examine texts as though they were
mere marks upon paper. A text grows out of and reciprocally influ-
ences a culture: hence, if our aim is to truly appreciate that text, we
must evolve an analysis along both vectors: textual and ethnographic,
or, as Hiltebeitel terms it in his earliest reflections upon the subject,
mythic and ritual. Yet, as Hiltebeitel argues in a critique of Dumzil
in his 1976 book, to do so requires the ability to suspend our massive
historicist prejudice against other levels of meaning, such as meton-
ymy, myth, religious practice, and philosophical and cosmological
understanding.73 As Hiltebeitel demonstrates, a reductive system of
transpositions, however ingenious, fails to account for the com-
plexity, not just of the Mahbhrata, but any narrative and any cul-
tural phenomenon.74 Hence, transpositions cannot be regarded as the

Hiltebeitel, iva, the Goddess, and the Disguises of the Pn davas, 151.
Goldstckers 1879 review of the Mahbhrata provides us what is perhaps the
locus classicus for this historicizing view and, in particular, provides an understanding
of the interpretive framework of much early Orientalist writing on the Indian epics.
When, by priestcraft and ignorance, a nation has lost itself so far as to look upon
writings like these as divinely inspired, there is but one conclusion to be drawn: it
has arrived at the turning-point of its destinies. Hinduism stands at this point, and
we anxiously pause to see which way it will direct its steps. . . . All barriers to religious
imposition having been broken down since the Purnas were received by the masses as
the source of their faith, sects have sprung up which not merely endanger religion, but
society itself; tenets have been propounded, which are an insult to the human mind;
practices have been introduced, which must fill every true Hindu with confusion and
shame. There is no necessity for examining them in detail . . . nor need we be at pains
of convincing the intelligent portion of the Hindu community; for, the excellent works
which it sent forth from Calcutta, Benares, and Bombay, and the enlightened views
which it propagates through its periodical press, fully prove that, equal in mental
accomplishments to the European mind, it requires no evidence of the gulf which
separates the present state of the nation from its remote past. . . . The cause of the
gradual degeneracy of Hinduism, are, indeed, not different from those to which other
religions are subject, when allowed to grow in the dark. In Europe, religious depravity
received its check when the art of printing allowed the light of publicity to enter into
the book whence her nations derive their faith; and no other means will check it in
India than the admission of the masses to that original book which is always on their
lips, but which now is the monopoly of the infinitesimal fraction of the Brahminical
caste able to understand its sense; and admission, also, to that other and important
literature which has at all periods of Hinduism striven to prove to the people that
their real faith is neither founded on the Brhmana portion of the Vedas, nor on the
Purnas, but on the Rigveda hymns. Theodore Goldstcker, Literary Remains of the
Late Professor Goldstcker, vol. 2 (London: W. H. Allen, 1879), 7778.
Yet there is a point where I would disagree with Dumzil over the nature of
introduction xxxiii

only key to the mythical exegesis. More fundamental is what I call a

method of correlation or correspondence . . . These correspondences,
as Hiltebeitel demonstrates, cannot be limited to those between the
bare historical event and its mythic encrustation, nor can the scholarly
task be restricted to mapping equivalences between historical facts
and mythic retellings. Rather, we must look at these correspondences
as occurring between two levels of continually changing and growing
tradition: myth and epic. In that case . . .
The epic poets . . . emerge not so much as programmers, transposing one
set of information into another form, but as rsi s, in this case the rsi s of
the Fifth Veda whose school is covered by the name of the elusive but
ever-available rsi Vysa. By calling attention to this term for visionaries
and poets, I refer in particular to the rsi s faculty of seeing connections,
equivalences, homologies, and correspondences discussed by Jan
Gonda. This faculty of seeing connections would have involved the
epic poets not only with correlations between myth and epic, but also
between epic and ritualespecially that of the Brhman ic sacrifice. Thus
the mythic exegesis must coexist with a ritual exegesis. Moreover, if
this was the procedure and orientation of the poets, it helps to explain
why they have told certain myths at key points in the epic narrative.
In some cases, they seem to have perceived correlations between myths
and adjacent portions of the epic plot, correlations which were meant to
deepen ones awareness of the meanings of the both the myth and the epic
planes, and ultimately, perhaps, to afford a glimpse of broader unities.75

the relation between myth and epic in the Mahbhrata. First, with regard to the
question of individual transpositions, Dumzil speaks of a veritable pantheon
which has been transposed into human personages by an operation as meticu-
lous as it was ingenious. These transpositions attest to an authors will, and the
poets themselves emerge as erudite, skillful, loyal to a design, perhaps an academy
of priests or several generations of a single school who would have composed the
work before writing, at the time of the four Vedas and the fifth. All this makes
it possible for Dumzil to speak of the heroes as inflexible copies of their mythi-
cal prototypes. To me, this seems too mechanical and too short-term a process.
Hiltebeitel, The Ritual of Battle, 356357.
Ibid., 359360.

1. Books

1976 The Ritual of Battle: Krishna in the Mahbhrata, Symbol

Myth Ritual Series, Victor Turner, ed. Ithaca: Cornell Uni-
versity Press. (Reprinted Albany: State University of New York
Press, 1990; Delhi: Sri Satguru Publications, 1991.)
1988 The Cult of Draupad, vol. 1. Mythologies: From Gingee to
Kuruksetra. Chicago: University of Chicago Press. (Reprinted
Delhi: Motilal Banarsidass, 1992.)
1991 The Cult of Draupad, vol. 2. On Hindu Ritual and the Goddess.
Chicago: University of Chicago Press.
1999 Rethinking Indias Oral and Classical Epics: Draupad among
Rajputs, Muslims, and Dalits. Chicago: University of Chicago
Press. (Reprinted Delhi: Oxford, 2001.)
2001 Rethinking the Mahbhrata: A Readers Guide to the Educa-
tion of the Dharma King. Chicago: University of Chicago Press.
(Reprinted Delhi: Oxford, 2002.)
2010 Dharma. South Asian Spirituality Series, Henry Rosemont ed.
Honolulu: University of Hawaii Press.
2011 Dharma: Its Early History in Law, Religion, and Narrative.
South Asia Research. Series ed. Patrick Olivelle. New York:
Oxford University Press.

2. Edited books

1989 Criminal Gods and Demon Devotees: Essays on the Guardians

of Popular Hinduism. Albany: State University of New York

* Except for a few articles that were newly written, the articles in these two volumes
represent material previously published elsewhere. The editors would like to take the
opportunity here to thank the many publishers & journals for granting us permission
to reuse this material. Below we also explicitly acknowledge the original source of each
of these contributions
xxxvi chronology of works

1998 Hair: Its Meaning and Power in Asian Cultures. Co-edited

with Barbara D. Miller. Albany: State University of New York
2000 Is the Goddess a Feminist? The Politics of South Asian God-
desses. Co-edited with Kathleen M. Erndl. New York and
London: New York University Press and Sheffield.

3. Translations ( from French)

1969 Georges Dumzil, The Destiny of the Warrior. Chicago: Univer-

sity of Chicago Press.
1973 Georges Dumzil, The Destiny of a King. Chicago: University of
Chicago Press. (Paperback reprint 1988.)
1985 Mircea Eliade, A History of Religious Ideas, vol. 3: From
Muhammad to the Age of Reforms. Co-translated with Diane
Apostolos-Cappadona. Chicago: University of Chicago
2004 Madeleine Biardeau, Stories about Posts: Vedic Variations on the
Hindu Goddess. Co-translator with James Walker and Marie-
Louise Reiniche; co-edited with Reiniche. Chicago: University
Chicago Press.

4. Selected Articles, Including Long Review Articles

[Articles included in this volume are denoted either with an asterisk

(if in volume 1) or a double asterisk (if in volume 2)]

1972 The Mahbhrata and Hindu Eschatology. History of Religions

12: 93115.
1974 Dumzil and Indian Studies. Journal of Asian Studies 34:
1975 Comparing Indo-European Epics, review of Mythe et
epopee, vols. 2 and 3, by Georges Dumzil. History of Reli-
gions 15: 90100.
1976 The Burning of the Forest Myth. Hinduism: New Essays in
the History of Religions. Ed. Bardwell L. Smith. Leiden: Brill.
1977* Nahusa in the Skies: A Human King of Heaven. History of
Religions 16: 32950.
chronology of works xxxvii

1977 Review Mahbhrata: Das Geschehen und seine Bedeutung,

by Heino Gehrts. Erasmus 29, columns 8692.
1978** The Indus Valley Proto-iva: Reexamined through Reflec-
tions on the Goddess, the Buffalo, and the Symbolism of
vhanas. Anthropos 73: 76797.
1979 Krsna in the Mahbhrata: A Bibliographical Essay. Annals
of the Bhandarkar Oriental Research Institute 60: 66110.
1979 Hindu Mythology and its Evils, review of The Origins of
Evil in Hindu Mythology, by Wendy Doniger OFlaherty.
History of Religions 19: 26975.
1980 Rma and Gilgamesh: The Sacrifices of the Water Buffalo
and the Bull of Heaven. History of Religions 19: 197223.
1980** Draupadis Garments. Indo-Iranian Journal 22: 97112.
1980** iva, the Goddess, and the Disguises of the Pndavas and
Draupad. History of Religions, Twentieth Anniversary
Issue 20: 14774.
198081 St vibhst: The Jewels for Her Journey. Indological
Taurinensia, Ludwik Sternbach Commemoration Volume
89: 193200.
1981** Draupads Hair. Autour de la desse hindoue. Ed. Mad-
eleine Biardeau. Purusrtha 5: 179214.
1982 Sexuality and Sacrifice: Convergent Subcurrents in the
Firewalking Cult of Draupadi. Images of Man: Religion
and Historical Process in South Asia. Ed. Fred W. Clothey.
Madras: New Era Publications, 1982. 72111.
1982* Brothers, Friends, and Charioteers: Parallel Episodes in
the Irish and Indian Epics. Homage to Georges Dumzil.
Ed. Edgar C. Polom. Journal of Indo-European Studies
Monograph, no. 13: 85112.
1982 Firewalking through the Mahbhrata: The Cult of
Draupad and the Great Indian Epic. The India Magazine
2,4: 1827.
1983 Toward a Coherent Study of Hinduism, review article on
the studies of Madeleine Biardeau. Religious Studies Review
9: 20612.
1983 Die Glhende Axt: Symbolik, Struktur und Dynamik in
Chndogya Upanisad 6. Trans. M. K. Ramaswamy. Sehn-
sucht nach dem Ursprung: Zu Mircea Eliade. Ed. Hans Peter
Duerr. Frankfurt am Main: Syndikat, 1983. 394405.
xxxviii chronology of works

1984 Two South Indian Oral Epics. History of Religions 24:

1984* The Two Krsnas on One Chariot: Upanisadic Imagery and
Epic Mythology. History of Religions 24: 126.
1985 Two Krsnas, Three Krsnas, Four Krsnas, More Krsnas:
Dark Interactions in the Mahbhrata. Essays in the
Mahbhrata. Ed. Arvind Sharma. Journal of South
Asian Literature 20: 7177. (Reprinted in Essays on the
Mahbhrata. Ed. Arvind Sharma. Brills Indological
Library, vol. 1. Leiden: E. J. Brill, 1991. 1019.)
1985** Purity and Auspiciousness in the Sanskrit Epics. Papers
from 1980 Conference on Religion in South India on Purity
and Auspiciousness. Ed. Frdrique Apffel Marglin and John
Carman. Journal of Developing Societies 1: 4154.
1985** On the Handling of the Meat, and Related Matters, in
Two South Indian Buffalo Sacrifices. Divisione delle Carne:
Dinamica Sociale e Organizzazione del Cosmo. Ed. Chri-
stiano Grottanelli. LUomo 9: 17199.
1988* Krsna at Mathura. Mathura: A Cultural Heritage. Ed.
Doris Srinivasan. New Delhi: Manohar and American Insti-
tute of Indian Studies. 93102.
1988 South Indian Gardens of Adonis Revisited. Essais sur le
rituel, vol. 1. Ed. Kristofer Schipper and A. M. Blondeau.
Colloque du Centenaire de la Section des Sciences Religieu-
ses de lEcole Pratique des Hautes Etudes, Bibliothque de
lEPHE Sciences Religieuses, vol. 92. Louvain and Paris:
Peeters. 6591.
1988 The South Indian Draupadi Cult. Journal of Asian Studies
(Madras) 6,1: 3137.
1989 Introduction, Draupads Two Guardians: the Buffalo
King and the Muslim Devotee, and Selected Bibliography.
Criminal Gods and Demon Devotees: Essays on the Guard-
ians of Popular Hinduism. Ed. Alf Hiltebeitel. Albany: State
Unniversity of New York Press. 118, 33971, 46372.
1991 Of Camphor and Coconuts. The Wilson Quarterly 15,3:
1991** The Folklore of Draupadi: Srs and Hair. Gender, Genre,
and Power in South Asian Expressive Traditions. 395427.
Ed. Arjun Appadurai, Frank J. Korom, and Margaret A.
Mills. South Asia Seminar Series. Philadelphia: University
of Pennsylvania.
chronology of works xxxix

199192 South Indian Mahbhrata Folk Cults. Confrence de

M. Alfred Hiltebeitel, Directeur dtudes associ. Annuaire:
Rsum des confrences et travaux. Paris: cole Pratique des
Hautes tudes, Section des Sciences Religieuses. 13539.
1992** Colonialist Lenses on the South Indian Draupad Cult.
Ritual, State and History. Festschrift for Jan C. Heesterman.
Ed. A. W. van den Hoek, D. H. A. Kolff, and L. M. S. Oort.
Leiden: E. J. Brill. 50731.
1992** Transmitting Mahbhratas: Another Look at Peter
Brook. The Drama Review 36,3: 13159.
1994 Epic Studies: Classical Hinduism in the Mahbhrata and
the Rmyana. Annals of the Bhandarkar Oriental Research
Institute 74: 162.
1994 Are Mantras Meaningful?, review of Understanding Man-
tras, edited by Harvey P. Alper. Journal of Ritual Studies 8:
1995** Draupad Cult Lls. The Gods at Play: Lila in South Asia.
Ed. William Sax. New York: Oxford University Press. 20424.
1995 Religious Studies and Indian Epic Texts. Religious Studies
Review 21,1: 2632.
1995** Dying Before the Mahbhrata War: Martial and Trans-
sexual Body-Building for Arvn. Journal of Asian Studies
54,2: 44773.
1995 Folk Religion and the Human Sciences in South India: A
Summer Workshop. Human Sciences Review 2: 25.
1997** Orders of Diffusion in Two Tamil Mahbhrata Folk
Cults. South Asian Folklorist 1: 936.
1998 Introduction: Hair Tropes, and **Hair Like Snakes and
Mustached Brides: Crossed Gender in an Indian Folk Cult.
Hair: Its Meaning and Power in Asian Cultures. Ed. Alf
Hiltebeitel and Barbara D. Miller. Albany: State University
of New York Press. 19, 14376.
1998* Empire, Invasion, and Indias National Epics. Interna-
tional Journal of Hindu Studies 2,3: 387421.
1998 Conventions of the Naimisa Forest. Journal of Indian Phi-
losophy 23: 6979.
1999** Kttntavar: The Divine Lives of a Severed Head. Ways
of Dying: Death and its Meaning in South Asia. Ed. Eliza-
beth Schmbucher and Claus Peter Zoller. Delhi: Manohar.
276310. + Plates 914.2
xl chronology of works

1999** Fathers of the Bride, Fathers of Sat: Myths, Rites, and Schol-
arly Practices. Thamyris: Mythmaking from Past to Present
6,1: 6594.
1999 Reconsidering Bhrguization. Composing a Tradition: Con-
cepts, Techniques and Relationships. Proceedings of the 1st
Dubrovnik International Conference on the Sanskrit Epics
and Purnas (DISCEP). Ed. Mary Brockington and Peter
Schreiner. Zagreb: Croatian Academy of Sciences and Arts.
1999** Two Ways to Tell a Story: lh in the Bhavisya Purna.
The Resources of History: Tradition, Narration, and Nation in
South Asia. Ed. Jackie Assayag. Institut Franaise de Pondi-
cherry. Etudes thmatiques, 8. Pondicherry: Ecole Franaise
dExtrme Orient. 96112.
2000 Introduction (co-authored), and **Draupads Question.
Is the Goddess a Feminist? Ed. Kathleen Erndl and Alf Hilte-
beitel. London: Sheffield. 1123, 11322.
2000 Review of The Sanskrit Epics, by John Brockington. Indo-
Iranian Journal 43: 16169.
2001* Bhsmas Sources. Vidyrnavavandanam: Essays in Honor
of Asko Parpola. Ed. Klaus Karttunen and Petteri Koskikallio.
Studia Orientalia. Finnish Oriental Society, No. 94. Helsinki.
2002 Allusion and Explication in the Gts Epic World. Journal
of Vaishnava Studies. Issue in honor of Dennis Hudson. 11,1:
2000* The Primary Process of the Hindu Epics. International
Journal of Hindu Studies 4,3: 26988.
2003 Indias Epics: Writing, Orality, and Divinity. The Study of
Hinduism. Ed. Arvind Sharma Columbia, SC: University of
South Carolina Press. 11438.
2003** Review of Landscapes of Urban Memory; the Sacred and the
Civic in Indias High-Tech City, by Smriti Srinivas. Journal
of Asian Studies 62,1: 31922.
2004* More Rethinking the Mahbhrata: Toward a Politics of
Bhakti. Indo-Iranian Journal 47: 20327.
2004* Role, Role Model, and Function: The Sanskrit Epic Warrior in
Comparison and Theory. Playing for Real: Hindu Role Models,
Religion, and Gender. Ed. Jacqueline Suthren Hirst and Lynn
Thomas. New Delhi: Oxford University Press. 2750.
chronology of works xli

2004 Kla [= Time]. Co-authored with Randy Kloetzli. The

Hindu World. Ed. Sushil Mittal and Gene Thursby. London:
Routledge. 55386.
2004 Some Preliminary Notes on Women Worshiping Draupad.
Festschrift for Dr. S. D. Lourdu, South Asian Folklorist 7: 145
2005* Buddhism and the Mahbhrata. Boundaries, Dynamics,
and the Construction of Traditions in South Asia. Ed. Federico
Squarcini. Florence: University of Florence Press. 10731.
2005* On Reading Fitzgeralds Vysa, review of The Mahbhrata,
vol. 7, trans. by James L. Fitzgerald. Journal of the American
Oriental Society 125,2: 24161.
2005* Not Without Subtales: Telling Laws and Truths in the Sans-
krit Epics. Journal of Indian Philosophy 33: 455511.
2005 Review of Dancing the Self, by William S. Sax. Journal of Ritual
Studies 19,1: 13840.
2005* Weighting Orality and Writing in the Sanskrit Epics. Epics,
Khilas, and Purnas: Continuities and Ruptures. Proceedings
of the 3rd Dubrovnik International Conference on the Sans-
krit Epics and Purnas, Sept. 2002. Ed. Petteri Koskikallio.
Zagreb: Croatian Acedemy of Sciences and Arts. 81111.
2005** Boar and Twins; Comparing the Tulu Kti-Cennaya Pddana
and the Tamil Elder Brothers Story. In the Company of Gods:
Essays in Memory of Gunther Dietz Sontheimer. Ed. A. Malik,
A Feldhaus, and H. Brckner. Delhi; IGNCA and Manohar.
2006 Avaghosas Buddhacarita: The First Known Close and Criti-
cal Reading of the Brahmanical Sanskrit Epics. Journal of
Indian Philosophy 34: 22986.
2006* The Nryanya and the Early Reading Communities of the
Mahbhrata. Between the Empires: Society in India 300 BCE
to 400 CE. 22755. Ed. Patrick Olivelle. New York: Oxford
University Press.
2006 Looking for the Kaurava Widows. Incompatible Visions:
South Asian Religions in History and Culture. Essays in Honor
of David M. Knipe. Ed. James Blumenthal. Madison, WI: Uni-
versity of Wisconsin-Madison, South Asia Center. 20516.
2007* Krsna in the Mahbhrata: The Death of Karna. Krishna: A
Sourcebook. Ed. Edwin F. Bryant. New York: Oxford Univer-
sity Press. 2376.
xlii chronology of works

2007* Among Friends: Marriage, Women, and Some Little Birds.

Gender and Narrative in the Mahbhrata. Ed. Simon Brod-
beck and Brian Black. London: Routledge. 11043.
2007 Review of The Self-Possessed: Deity and Spirit Possession in
South Asian Literatures and Civilization, by Frederick M. Smith.
Journal of the American Oriental Society 127,3: 63134.
2009 Recontextualizing Satire of Dharmastra in the Aggaa
Sutta. Religions of South Asia 3.1: 7792.
2010* Authorial Paths through the Two Sanskrit Epics: Via the
Rmopkhyna. Epic Undertakings: Papers of the 12th World
Sanskrit Conference, vol. 2. Delhi: Motilal Banarsidass. 169214.
2010 Mapping Bhakti through Friendship in the Sanskrit Epics. Epic
and Argument in Sanskrit Literary History. Festschrift for Robert
P. Goldman. Ed. Sheldon I. Pollock. Delhi: Manohar. 91116.
2011 On Sukthankars S and Some Shortsighted Assessments and
Uses of the Pune Critical Edition (CE). Journal of Vaishnava
Studies 19,2 (Spring): 87126.
2011 Just My Imagination? Puzzling through a Duryodhana Fes-
tival near Dharmapuri, Tamilnadu. In Hinduism in Practice.
Ed. Hilary Rodriguez. 87103. London: Routledge.
2011 You Have to Read the Whole Thing: Some Reflections on
Madeleine Biardeaus Mahabharata. Du texte au terrain,
du terrain au texte: dialogues disciplinaires autour de louvre
de Madeleine Biardeau, Journe du Centre dtude de lInde
et de lAsie du Sud. To appear on the website of the Centre
dtudes de lInde et de lAsie du sud;

5. Interviews and Profiles

1992 Mahbhrata, interview by Steven J. Rosen. Vaishnavism.

Contemporary Scholars Discuss the Gaudiya Tradition. Ed.
Steven J. Rosen. New York: Folk Books. 4959.
2001 What I Wanted to Avoid Was an Easy Equation . . ., inter-
view by Venu. Folklife: A Quarterly Newsletter from National
Folklore Support Centre (Chennai), vol. 1, issue 1 (July): 58.
2007 Professor Profile. The Asian Connection. Issue 2, Fall 2007, 4.
chronology of works xliii

6. In Press

N.d. Dialogue and Apostrophe: A Move by Valmk? In Dialogue

in South Asian Religious Texts. Ed. Brian Black and Laurie Pat-
ton. London: Ashgate.
N.d.* Mapping Bhakti through Hospitality and Friendship in the
Sanskrit Epics. Battles, Bards, Brhmans: Papers from the
Epics Section of the 13th World Sanskrit Conference, vol. 2. Ed.
John Brockington. Delhi: Motilal Banarsidass. 11338.
N.d.* Epic Avamedhas. Papers from the Fourth International
Vedic Workshop: The Vedas in Culture and History. Ed. Joel
Brereton. Delhi: Motilal Banarsidass.
N.d. Ritual as Dharma: The Narrowing and Widening of a Key
Term. Festschrift for Fred Clothey. Ed. Linda Penkower and
Tracy Pintchman.
N.d.* The Archetypal Design of the Two Sanskrit Epics. Proceedings
of the 5th Dubrovnik International Conference on the Sanskrit
Epics and Purnas. Ed. Petteri Koskikallio. Zagreb: Croatian
Academy of Sciences and Arts.
N.d. Why Itihsa? New Possibilities and Limits in Considering
the Mahbhrata as History. Ways and Reasons for Thinking
about the Mahbhrata as a Whole. Ed. Vishwa Adluri. Pune:
Bhandarkar Oriental Research Institute.
N.d. The Southern Recension Reading of the akuntal Story as Its
First Reading: What It Can Tell Us about the Original and the
Second Reading by Klidsa. Revisiting Klidsas Abhij-
nakuntalam: Land, Love, Languages: Forms of Exchange in
Ancient India. Ed. Deepika Tandon and Saswati Sengupta. New
Delhi: Orient Blackswan.
N.d. Just My Imagination? Puzzling through a Duryodhana Festi-
val near Dharmapuri, Tamilnadu. Hinduism in Practice. Ed.
Hillary Rodrigues. London: Routledge.
N.d. Moksa and Dharma in the Moksadharma. Paper presented
at the Brown University Conference on Early Indian Philos-
ophy in the Mahbhrata, Providence RI, April 9, 2010. In
preparation for a conference volume. Ed. James L. Fitzgerald.
N.d. Between History and Divine Plan: The Mahbhratas Royal
Patriline in Context. Paper presented at the Cardiff Univer-
sity Workshop: Genealogy in South Asia, May 2628, 2010. In
preparation for a conference volume. Ed. Simon Brodbeck and
James Hegarty.
xliv chronology of works

7. Book Reviews

1969 Dumzil: Epic in the Balance, review of Mythe et pope,

vol. 1, by Georges Dumzil. History of Religions: 9093.
1975 On Reading van Buitenens Vysa, review of The
Mahbhrata, vol. 1, trans. by J. A. B. van Buitenen. His-
tory of Religions 15: 230232.
1978 Review of The Bhrgus and the Mahbhrata: Gods, Priests,
and Warriors, by Robert P. Goldman. American Academy
of Religion Journal 46: 7678.
1980 Indo-European Marriages, review of Mariages indo-
europens, by Georges Dumzil. History of Religions 19,4:
1981 Review of Les dieux et les hommes: Etude des cultes dun vil-
lage du Tirunelveli Inde du Sud, by Marie-Louise Reiniche.
Journal of Asian Studies 40,4: 18485.
1984 Review of Les Ftes dans le monde hindoue, ed. by Grard Tof-
fin. Special issue of LHomme: Revue franaise danthropolo-
gie 22,3; reprinted in History of Religions 24 (1984): 89.
1986 Review of Discourses on iva, ed. by Michael W. Meister.
Religious Studies Review 12,2: 18485.
198687 Review of Wives of the God-King, by Frdrique Apffel
Marglin and Devi and the Spouse Goddess, by Lynn Gatt-
wood. Pacific Affairs 59,4: 71315.
1987 Folklore and Hinduism, review of Another Harmony:
Essays on the Folklore of India, ed. by Stuart H. Blackburn
and A. K. Ramanujan. History of Religions 27,2: 21618.
1987 Review of LEspace du temple II: Les Sanctuaires dans le
royaume, ed. by Jean-Claude Galey. Religious Studies Review
13,4: 364.
1987 Review of Hindu Goddesses: Visions of the Divine Feminine
in the Hindu Religious Tradition, by David Kinsley. Pacific
Affairs 60,2: 350-51.
1990 Review of Mitra-Varuna: An Essay on Two Indo-European
Representations of Sovereignty, by Georges Dumzil, trans.
by Derek Coltman. Journal of Religion 70,2: 29596.
1990 Review of Oral Epics of India, ed. by Peter J. Claus, Stu-
art H. Blackburn, Joyce B. Flueckinger, and Susan Wadley.
Folklore Newsletter 7,2: 34.
chronology of works xlv

1990 Review of The Theatre of the Mahbhrata: Terukkuttu Perfor-

mances in South India, by Richard A. Frasca. Religious Studies
Review 17,3: 280.
1990 Review of Visions of God: Narratives of Theophany in the
Mahbhrata, by James W. Laine. Religious Studies Review
17,4: 374.
1990 Review of The Mahbhrata: A Play Based upon the Classical
Indian Epic, trans. by Jean-Claude Carrire with Peter Brook.
Religious Studies Review 17,4: 280.
1991 Review of Encountering the Goddess: A Translation of the
Dev-Mhtmya and a Study of its Interpretation, by Thomas
B. Coburn. Religious Studies Review 18,2: 164.
1991 Review of The Rmyana of Vlmki: An Epic of Ancient India,
Volume III, Aranyakanda, trans. by Sheldon I. Pollock. in Reli-
gious Studies Review 18,2: 164.
1991 Review of Mountain Goddess: Gender and Politics in a Hima-
layan Pilgrimage, by William S. Sax. Committee on Women in
Asian Studies (CWAS) Newsletter: 11,1.
1991 Review of Many Rmyanas: The Diversity of a Narrative Tra-
dition in South Asia, ed. by Paula Richman. Religious Studies
Review 18,4: 355.
1992 Review of Rmyana and Rmyanas, by Monika Thiel-Horst-
mann. Religious Studies Review 19,4: 373.
1993 Review of Purna Perennis: Reciprocity and Transformation
in Hindu and Jaina Texts, ed. by Wendy Doniger. Journal of
Asian Studies 53,2: 58789.
1994 Review of Mythologies du XXe sicle (Dumzil, Lvi-Strauss, Eli-
ade), by Daniel Dubuisson. Journal of Asian Studies 54,2: 49496.
1995 Review of From the Margins of Hindu Marriage, ed. by Lind-
sey Harlan and Paul Courtright. Religious Studies Review 22,3
(1995): 26566.
1998 Review of Cooking the World: Ritual and Thought in Ancient
India, by Charles Malamoud. Religious Studies Review 24,4: 445.
1998 Review of Conversions and Shifting Identities: Ramdev Pir and
the Ismailis in Rajasthan, by Dominque-Sila Khan. Religious
Studies Review 24,4: 446.
1998 Review of Gods and Temples in South India, by Winand M.
Callewaert. Religious Studies Review 24,4: 44647.
2003 Review of Religious Doctrines in the Mahbhrata, by Nicholas
Sutton. in Journal of Asian Studies 63,2: 24345.
xlvi chronology of works

2003 Review of Questioning Rmyanas: A South Asian Tradition,

ed. by Paula Richmond. Religion 34: 8892.
2004 Review of Destiny and Human Initiative in the Mahbhrata,
by Julian Woods. Journal of the American Oriental Society
124,1: 18486.

8. Encyclopedia and Dictionary Articles and Editing

1982 Mahbhrata and ten short entries on Mahbhrata charac-

ters. In Abingdon Dictionary of World Religions. Ed. Larry D.
Shinn et al. Nashville: Abingdon.
1987 Hinduism, Gambling, Mahbhrata, Kurukshetra,
Arjuna, and Indus Valley Religion (co-authored with
Thomas Hopkins). In The Encyclopedia of Religion. Ed. Mircea
Eliade et al. New York: Macmillan Free Press.
1995 Area Editor. Religions of India. In The Harper Collins Diction-
ary of Religion. Ed. Jonathan Z. Smith. San Francisco: Harper-
San Francisco.
1995 Hinduism, Sacrifice, Hinduism, and about ten other short
entries. In The Harper Collins Dictionary of Religion. Ed. Jona-
than Z. Smith. San Francisco: Harper-San Francisco. 42440,
2001 Board of Editorial Consultants. South Asian Folklore: An Ency-
clopedia. Ed. Margaret A. Mills, Peter J. Claus, and Sarah Dia-
mond. New York and London: Routledge.
2001 Aravn, Draupad, Gods and Goddesses, Sacrifice,
am (tree). In South Asian Folklore: An Encyclopedia. Ed.
Margaret A. Mills, Peter J. Claus, and Sarah Diamond. New
York and London: Routledge. 2122, 16566, 26265, 52930,
2006 Mahbhrata and Rmyana. In Encyclopedia of India,
vol. 3. Ed. Stanley Wolpert. Detroit: Thomson Gale. 8293;
2009 Draupad and St. Brills Encyclopedia of Hinduism. Ed. Knut
A. Jacobsen, Angelika Malinar, Helene Basu, and Vasudha
Narayanan. Leiden: Brill. 51733.
chronology of works xlvii

9. Work in Progress

N.d. Conversations with Murugan: The Mahbhrata World of a

Priest of Draupad and Duryodhana. University of Chicago
Press. Co-author with Perundevi Srinivasan.
N.d. Translator, alya and Sauptika Parvans. The University of Chi-
cago Press Translation of The Mahbhrata. Gen. ed. James L.
Fitzgerald. Chicago: University of Chicago Press.


According to well known Indian popular traditions, Draupadthe

heroine of the Mahbhratamade a vow that she would not bind
up her hair until she could comb into it the blood of either Duryo-
dhana or Duhsana. She uttered this terrible oath after being insulted
and molested by these two Kauravas at the epics dice match, and
it required not only her own thirteen years of dishevelment during
her exile with her five husbands, the Pndavas, but also the eventual
killing of these two foes by them in battle. The theme has circulated
long and widely. It was first clearly expressed by a north Indian dra-
matist, Bhatta Nryana, who took it as the main theme of his play
Ven sam hra, The Binding-Up of the Braid. Bhatt a, who appar-
ently hailed from Kanauj and settled in Bengal, probably lived around
675725 A.D.1 And it was reworked most recently by Subramanya
Bharati in the second decade of the twentieth century when he por-
trayed Draupad as the image of Mother India defiled in his Tamil play
Panchali Sapatham, The vow of Pcl (or Draupad).2 It would
seem, in fact, that through the centuries the themes popularity has
been most consistent in the south. It is found in the Tamil rendition
of the Mahbhrata by Villiputtr l vr (ca. 1400), in the Kannada
version of Kumra Vysa (fifteenth century), in the Malayalam rendi-
tions of the story by the sixteenth century Ezhuthachan and others,
and in the dramatic styles of the kathakali from Kerala, the yaksagna
of Karnataka, and terukkttu of Tamilnadu.3 In the latter case, the

See discussion in A. B. Gajendragadkar (ed. and trans.), The Ven sam hra of
Bhatta Nryana (1933; 3rd. ed. Bombay, New and Second Hand Book Stall, n.d.),
part I, Ven sam hra. A Critical Study; part II, Text with Translation; part III,
Critical and Explanatory Notes, pp. 116; Francine Bourgeois (ed. and trans.),
Ven sam hra. Drame Sanskrit (Paris, Institut de Civilisation Indienne, 1975), p. 1,
proposes the eighth century.
See Amarakavi Subramanya Bharati, The Vow of Panchali, trans. H. K. Valam
(1957; rev. ed. Mylapore, Mohana Trust Pathippagam, 1972), p.124 (here it is Duryo-
dhana and Duhsanas blood mixed).
For summaries of this episode in Villiputtr (hair tied at death of Duryodhana)
and Kumra Vysa (blood of Duhsana), see M. V. Subramanian, The Mahbhrata
4 chapter one

terukkttu drama Turiytana-vatam (Death of Duryodhana) or

Patukalam (Battlefield, or Place of Lying Down) completes the
cycle of Mahbhrata-based plays that enact the story as background
myth for the cult of the goddess Draupad (Tamil: Tiraupatiyamman).
The study of Draupads cult demands investigation of Draupad in the
epic; and the epic heroine can be illumined by the study of her cult.4
According to A. B. Gajendragadkar, in whose edition of the Ven -
sam hra one finds what seems to be the only major discussion of
our theme, the vow is Bhatt a Nryanas innovation and the
Mahbhrata knows nothing of it.5 As we shall see, these are extreme
or incautious statements. But before examining the epic for its treat-
ment of Draupads hair, we must see how the Ven sam hrawhich
makes it its central themediffers in its handling from the epic. Let it
simply be noted that certain methodological working assumptions are
involved in this study, which itself should be regarded as an explor-
atory test of their explanatory value rather than a demonstration of
their validity. These are, first, that certain themes connected with a
number of the heroines of the Hindu epics can be illumined by look-

Story. Vysa and Variations (Madras, Higginbothams 1967), p. 90; on Ezhuthachans

treatment, see Krishna Chaitanya, A History of Malayalam Literature (New Delhi,
Orient Longmans, 1971), pp. 85, 139140 (Duhsanas blood). Martha Ashton
has described the yaksagna episode to me orally. The kathakali play Duryodhana
Vadha (in which it is Duhsanas blood) and the terukkttu drama mentioned below
(Duryodhanas blood) I have seen myself.
As subsequent research brought out, the terukkttu drama on Duryodhanas
death is usually called Eighteenth-Day War. For an initial synthesis of my research
on the Draupad cult, and for a companion piece to this article also exploring relations
between cult and epic, see Hiltebeitel 1980a and 1982. Fieldwork (1975 and 1977) on
the Draupad cult was made possible by grants from the American Institute of Indian
Studies and the National Endowment for the Humanities, for which I express my
Gajendragadkar, part I: 21; in agreement see G. V. Devasthali (ed. and trans.)
Bhatta Nryan as Ven sam hra (Bombay, D. M. Tilak, 1953), p.214. In addition to
Bourgeois (see n. 1), I have also consulted M. R. Kale (ed. and trans.), Ven sam hra
of Bhatta Nryan a (Delhi, Motilal Banarsidass, 1977); both are silent on this issue.
After completing this article I learned of Daniel Dubuissons, La desse chevelue et
la reine coiffeuse. Recherches sur un thme pique de lInde ancienne, Journal Asia-
tique, 166 (1978), 291310. His attempt to interpret Draupads hairdresser disguise
through Indo-European comparisonsstressing links between hair and vegetationis
intended as a completion of the Wikander-Dumzil interpretation of the Pndavas
disguises. I am unconvinced by this article, which overlooks Dumzils equally signifi-
cant insights into the relation between the Pndavas disguises and the caste structure,
which confuses hair-pulling and baldness (see p. 298), and which fails to take into
account the epic narrative (and wider religious and cultural) context in which the
theme is expressed.
draupads hair 5

ing to materials on the Hindu Goddess that often appear only later in
post-epic literature and folk traditions; second, that the epics them-
selves present a complex theological vision that gives pride of place
not only to heroes and heroines who represent aspects of Visnu, but
also of iva and the goddess; third, that the epics thus provide narra-
tive reflections on (and of) the cult and mythology of the goddess as
it emerges into the literary light of day; and fourth, that the epic poets
made selective use of oral traditions that probably had some affinities
with oral and vernacular epic traditions still popular today.6

A. Draupads Hair in the Ven sam hra

While Gajendragadkar and others have noticed in the Ven sam hra
numerous narrative and characteriological departures from the epic,
the following have all been overlooked.
Taking the departures in their narrative order, first the Ven sam hra
gives the impression that Draupads hair was made loose initially
by Duhsana. Thus Bhma, covered with Duryodhanas blood
and unrecognized by Draupad who thinks he is dead, says to her:
Pcl, not indeed while I am alive should your braid (ven ) dishev-
elled by Duhsana, be tied up by your own hand.7 This implies that
Duhsana dishevelled some kind of ven , braid, a conclusion that
other Ven sam hra passages reinforce.8 This is different from the
Mahbhrata where, as we shall see, Draupads hair is already dishev-
elled when Duhsana seizes it because she is menstruating. One may

These assumptions, and particularly the second, owe much to Madeleine
Biardeau, whose work will be frequently cited below. I continue to explore them, but
especially the first and third, in Hiltebeitel forthcoming, chapter 8, which focuses not
on Draupad but on the women who precede her in the Bhrata dynastic line, from
Gang and Satyavat to Kunt, Mdr, and Gndhr.
pcl, na khalu mayi jvati sam hartavy duh sanavilulit ven rtmapn in
(6.38). All Vensam hra citations are from Gajendragadkar, part II. All Mahbhrata
citations will be from the Poona Critical Edition. Translations, unless otherwise indi-
cated are my own.
Thus see 6.42: Bhma says, Let this braid, dishevelled by Duhsana, . . . be tied up
(sam yamyatm duh sanaviluliteyam ven ); 6.35: she whose braid was dishevelled
from the pulling of the hair by Duhsana (duh sanena kacakarsan abhinnamaulih );
2.25: she whose garments and hair were [both] dishevelled, pulled by the hand of
Duhsana (duh sanena hastkrst a vilolakeavasan pcl)translating similarly,
see Gajendragadkar, part III: 126127, with explanation.
6 chapter one

only guess that Bhatta Nryana has either overlooked or suppressed

this indication of Draupads impurity.
Secondly, Bhatt a Nryana has Duryodhana order both Duhsanas
pulling of Draupads hair and Duhsanas thwarted attempt to
pull off her garments. Duryodhana recalls this himself, referring to
Draupad as she whose garments and hair were both dishevelled,
pulled by the hand of Duhsana at my command.9 And at several
other points Duryodhana is described as the one who was the cause
of both pullings.10 Here Bhatt a Nryanawho is joined on this point
by Villiputtr l vr11departs from an important detail of the epic,
where, as we shall see, it is Karna who orders the disrobing. But in
deepening the parallelism between the pulling of the hair and the pull-
ing of the garments by attributing both commands to Duryodhana as
well as both acts to Duhsana, Bhatta Nryana has evoked an impor-
tant complementarity between the two violatory acts, and driven it
home by reinforcing it over and over with a variety of compounds, the
most memorable and basic being the kembarkarsan a, the pulling
of the hair and the garments.12 I will return to this complementarity
and its significance in the conclusion of this chapter.
The third departure is not so clearly one from the epic as from
the popular tradition. Nonetheless, as we shall see, there are certain
indications that the epic is closer to the popular tradition than to the
Ven sam hra. It is a question of whose vow it is: Bhmas or Draupads.
The popular traditions are clear that it is essentially Draupads vow,

mama jay duh sanena hastkrst a vilolakeavasan pcl; 2.25 (cf. n. 8
Most explicitly at 6.4, where Duryodhana is the base cause of the great sin(s) of
pulling the hair and garments of the queen (devkembarkarsan amahptakapra-
dhnahetur); similarly he is the wind for removing the garments and hair of Krsn
(krsn keottaryavyapanayanamarut; 5.26); a great sinner (for causing) the pulling
of the hair and garments of the Pcla kings daughter (pclarjatanaykem-
barkarsan amahptakin; 6.7); a great sinner (for causing) the pulling of the hair
and garments of Draupad (draupadkembarkarsan amahptakin; 3.47). Noting
without comment Bhatt a Nryanas attribution of both commands to Duryodhana,
see Gajendragadkar, part III: 163.
See Subramanian 1967:90.
See above, nn. 810, citing 2.25, 3.47, 5.26, 6.4, and 6.7. Cf. also 1.8: Bhma
lamenting that the Kauravas still live having dragged the hair and garments of the
Pndavas wife (krsy a pn davavadhparidhnaken); 4.1: Duhasana as he who
dragged the hair and garments of the Pndavas wife (pn davavadhkembarkar-
sin ah ). It is worth noting that one or another modification of this basic compound
is found in every act of the play, with two appearances in the last act. It is thus an
important leitmotif.
draupads hair 7

with Bhmas role in killing Duhsana and Duryodhana one of ser-

vice to her. According to Gajendragadkar, however, the vow is primar-
ily Bhmas: with his hands smeared with the blood of Duryodhana
[he vowed to] rearrange the dishevelled hair of Draupad, who was
therefore to allow her hair to remain in that disordered condition till
he fulfilled his vow.13 Gajendragadkar notes that Bhma makes no
such vow in the epic,14 and thus proposes that Bhatt a has invented
the whole thing. He is certainly right that the Ven sam hra attributes
the vow in the main to Bhma.15 But there are also points where the
Ven sam hra recognizes Draupad as making a complementary vow.
In a passage that we shall return to later, Bhma tells Draupad that
while he and his brother will officiate, she will be the wife whose
vow is maintained (grhtavrat) for the sacrifice of battle (1.25).
Again, evoking Draupads own motivations, her maid recalls ask-
ing Duryodhanas wife: How will the hair of our queen be bound
up when your [plural: i.e., the Kaurava wives] masses of hair are not
loosed ( yusmkam amuktesu keahastesu devyh keh sam yamyanta
iti) (1.21)? And similarly, when Bhma readies himself to kill Duryo-
dhana, he says: Pcls fire of wrath is evidently well-nigh extin-
guished now that the masses of hair have been confusedly loosened
among the women of the Kaurava court, their husbands forcibly slain
by me.16 These passages suggest that Bhatt a Nryana is well aware
of a tradition that Draupad too is undergoing some sort of vow, sus-
tained by her anger and focused on the reversal that will take place
when it will be the Kaurava wives who must dishevel their hair. In this
case, as we will see, the theme that Bhatt a subordinates can be traced
clearly to the Mahbhrata. It would also seem to have echoes in the
popular traditions.
The fourth divergence concerns the manner in which Draupad
wears her hair. According to Gajendragadkar and G. V. Devasthali,
in the Ven sam hra Draupad keeps her mass of hair tied in a single
unornamented long braid known as an ekaven , literally a single

Gajendragadkar, part I: 20; part III: 2.
Kale 1977: 9, mistakenly says he does.
Act I is called Pratijbhma, the Vow of Bhma, and in it Bhma makes his
gory vow (1.21). Cf. also Ven sam hra 6.6 and 6.39, and, on the primacy of Bhmas
role, Bourgeois 1971: 56.
pncly manyuvahnih sphutam upaamitaprya eva prasahya pronmuktaih
keapair hatapatisu may kauravntah puresu; 6.8.
8 chapter one

braid.17 Unfortunately, this terminology can be misleading, for an

ekaven is not actually a braid but a style of tying up or clasping the
hair once in back of the head and leaving the rest to flow loose as a
mass.18 Unlike the ordinary threeplait braid called triven and a five-
plait style called pacaven ,19 which are suitable for auspiciously married
women, the ekaven is worn by virahin s, women separated from their
husbands. When the husband returned, the women untied the single
mass and rearranged the hair properly with appropriate decorations.20
Sanskrit literature knows numerous heroines of this type. Separated
from Rma, St wears an ekaven (5.18.8; 55.27; 57.12) in Rvanas
aoka grove.21 Klidsas akuntal bears an ekaven (ekaven dhar)
when she finally confronts her husband Dusyanta after years of rejec-
tion.22 And the wife of Klidsas exiled Yaksa hero in the Meghadta
also awaits her husbands return with an ekaven .23
Although the term ekaven is never used in the Ven sam hra, Gajen-
dragadkar is probably right that Bhatt a Nryana views Draupad
as a virahin . Once Bhma is specifically referred to as shaking her
ven (ven m avadhya; 6.42), which is certainly implies some form of
ven ; and other references to her hair can also be taken in this sense,
although a different interpretation is often possible.24 Wearing the hair

Gajendragadkar, part III: 3; cf. pp. 5257; see also Devasthali 1953: 236237.
See Vaman Shivaram Apte, The Practical Sanskrit-English Dictionary, 3 vols.
(Poona, Prasad Prakashan, 19571959), s.v. ekaven ; Gajendragadkar also makes this
clear: hair tied in a single mass, [allowed] to hang loosely on their back, somewhat
analogous to what they now do when bathed overhead, to allow their hair to get dried
(part 3:3); in English terms, more a pony tail than a braid.
This style is known in the Cilappatikram. See Adigal, Prince Ilang, Le Roman
de lAnneau. Trad. Du tamoul par Alain Danilou et R. S. Desikan (Paris, Gallimard,
1961; London 1965), pp. 36, 41, 57, of the English edition.
Gajendragadkar, part III: 3.
Rmyan a, Baroda Critical Edition. See also 5.17.18; 5.13.24 and notes; 5.23.9;
5.26.17. Rvana has also seized her by the hair (3.47.16). However, see Mahbhrata
3.265.25 (Rmopkhaynam): her ven is susam yat, either well tied up or well
braided. In an interesting context, Kaikey is an ekaven with a soiled garment
(malinmbar; Rm, 2, App. I, no. 8, lines 1718) when she enters the chamber of
wrath (krodhgra) to demand Bharatas enthronement and Rmas exile. She thus
takes on a symbolic estrangement from her husband.
A. B. Gajendragadkar (ed. and trans.), The Abhijna-kuntala of Klidsa
(Surat, The Popular Books Store, 1951), p. 185; she also wears dusky garments (vasane
paridhsare; idem), on which see idem, pp. 519520, Gajendragadkars comments.
Franklin and Eleanor Edgerton (trans.), Klidsa, The Cloud Messenger (Ann
Arbor, University of Michigan Press, 1964), p. 67 (stanza 90).
She is frequently said to be wearing a mass of hair (keahasta: 1.21, 6.41;
keapa: 6.6, 6.42). On the sense of these terms and keapaksa as mass of hair, see
draupads hair 9

in the style of a virahin would thus symbolize for Draupad that her
husbands are absent, or even dead. As Gajendragadkar puts it: the fact
that she suffered that egregious insult at the hands of Duhsana even
in the presence of the Pndavas meant that to Draupad they were as
good as absent and continued to be so till the insult was avenged.25
If this is the way it is represented in the Ven sam hra, then the
play has at least on this point considerably narrowed the symbolism
of Draupads dishevelment. In popular traditions, Draupads hair is
worn completely loose and free; it is not tied in back.26 And, as we shall
see, the same seems to hold for the Mahbhrata. Taken altogether,
the discrepancies between the Ven sam hra and the epic thus indicate
some consistent patterns. The action is primarily in the hands of the
men. It is Duhsana who dishevels her hair; it is Bhma who makes
her vow; and it is her husbands symbolic absence that dictates her
hairstyle. But Bhatta Nryana is alive to much more, as his reiteration
of the hair and garments theme already suggests. In discussing the
Mahbhratas treatment of Draupads hair and the fuller symbolic
meanings that it evokes, his Ven sam hra will not be forgotten.

B. Draupads Hair in the Mahbhrata

The Mahbhrata seems to know more about Draupads hair than

it ever makes explicit. And what it does tell us seems to rely on an
understanding of Draupadsand Bhmasbehaviour that is in cer-
tain respects different from the Ven sam hras. The pertinent epic pas-
sages, discussed in sequence, are the following.

1. Pulling the hair

When Duhsana, at Duryodhanas command, drags Draupad into
the Kaurava mens hall (sabh) by the hair, two things are most

Gajendragadkar part III: 5758. Problematic verses include Draupad being described
as muktake, loose haired (6.33, 6.42) and a description of Duhsana touching
Krsn whose ven was loosed (muktaven m; 1.19). On the latter, Gajendragadkar, part
III: 34 (cf. pp. 5254), takes ven as tri- or pacaven . Would this not suggest that
Duhsana pulled hair that Draupad had already loosened?
Gajendragadkar, part III: 3.
E.g., in the kathakali and terukkttu. Cf. the artists representations in the Poona
Critical Edition of the Mahbhrata, Sabhparvan facing p. 304, and Virtaparvan,
facing pp. 67 and 272.
10 chapter one

significant. First, as the text indicates here several times and reiter-
ates elsewhere, she is menstruating (rajasvalsmi; 2.60.25b; cf. 27a;
32b). Secondly, when Duhsana seizes her, the hair is described as
long blue and flowing (drghesu nlesvatha cormimatsu [. . .] keesu;
22cd), and further as strewn about or dishevelled (prakrn ake;
28a). These two facts are not unrelated. Draupads hair is dishevelled
because she is menstruating. The Mahbhrata draws here on a well
known prohibition on wearing the hair braided during menstruation,
and not binding it up until the ritual bath that ends the period of
impurity. More particularly, during her menses a woman should not
arrange her hair with a comb,27 a prohibition which provides back-
ground to a further detail of certain popular versions of Draupads
vow that she would wait to braid her hair not only with Duryodhanas
blood but with a comb made from his ribs.28 Today the hair is worn
loose during menstruation by Hindu and Sikh women in the Punjab,
as reported by P. Hershman: [. . .] times of pollution occur [at the
death of an unweaned child;] at the death of a husband; following
childbirth; during menstruation; and following intercourse when she
has not washed. It is highly significant that in the ritual bath which
brings to an end these periods of pollution, it is absolutely crucial that
she washes her hair, grooms it and binds it in a proper fashion.29

This detail, and an ample discussion of Rajasval-dharmh , rules of menstrua-
tion, is found in Pandurang Viman Kane, History of Dharmastra, 5 vols. (Poona,
Bhandarkar Oriental Research Institute, 1974), vol. II, part 2, pp. 802805: quote from
p. 803, citing the earliest text, Taittirya Sam hit 2.5.1: see Arthur Berriedale Keith
(trans.), The Veda of the Black Yajus School Entitled Taittirya Sam hit, Harvard
Oriental Series, vols. 18 and 19 (1924; repr. Delhi, Motilal Banarsidass, 1967), vol. I,
p. 189. The hair-braiding prohibition is obviously related to similar prohibitions in the
texts against spinning and making ropes.
Lawrence Babb, Walking on Flowers: A Hindu Festival Cycle (Singapore, Depart-
ment of Sociology, 1974), p. 19; Subramanya Bharati 1972: 125126.
P. Hershman, Hair, Sex and Dirt, Man, IX (1974), 282283; cf. Gabriella Eich-
inger Ferro-Luzzi, Womens Pollution Periods in Tamilnad (India), Anthropos,
LXIX (1974), 128, and Mrs. Sinclair Stevenson, The Rites of the Twice Born (Delhi,
Munshiram Manoharlal, 1971), p. 247. On hair symbolism more generally, see Charles
Berg, The Unconscious Symbolism of Hair (London, George Allen and Unwin, 1951)
Edmund R. Leach, Magical hair, Journal of the Royal Anthropological Institute of
Great Britain and Ireland, LXXXVIII (1958), 147164; J. Duncan M. Derrett, Reli-
gious Hair. Man, VIII (1973), 100103; Raymond Firth, Symbols Public and Private
(Ithaca, Cornell University Press, 1973), pp. 262298.
draupads hair 11

2. Departure for the forest

In the interim between the preceding and present scenes, much of
importance has happened. Draupad has raised the moot and never
resolved question of whether Yudhisth ira could have bet her after he
bet himself; she has withstood Duhsanas attempts to disrobe her;
she has accepted the offer of boons from Dhrtarst ra and chosen
her husbands freedom and the return of their weapons; and, having
set out with her husbands toward their kingdom at Indraprastha,
she has come back with them to the Kaurava sabh at Hstinapura.
There Yudhisth ira loses the second invitation of the dice match,
the winner-take-all throw that sends the Pndavas and Draupad into
exile, and Draupad and the Pndavas prepare for their departure. Our
first verse finds her bidding adieu to Kunt, promising to look after
the youngest Pndava Sahadeva: So be it, said that lady. Stained
with flowing tears, her single garment smeared with blood, her hair
loose, she went out.30 Then a second passage finds Vidura describing
to Dhrtarstr a Draupads departure for the forest:
Wearing a single garment, weeping, her hair loose, menstruating, her
wet garment stained with blood, Draupad spoke this word: Those on
account of whom I have reached this condition, their wives, in the four-
teenth year, with their husbands slain, their sons slain, their relatives and
dear ones slain, their limbs smeared with the blood of their relatives, hair
loose, menstruating [? rajasvalh ], having thus offered water to the dead,
will enter the City of the Elephant [Hstinapura].31
One noteworthy feature of this second passage is that it presents one
of the two instances in the epic where Draupad says something that
resembles a vow. But unlike the vows (pratijh ; 2.68.46 and preced-
ing passages) of her husbands, which take the form of commitments to
acts they will themselves carry out against their foes, Draupads state-
ment has more the look of a prediction or a curse. It is a word that
commits her husbands and their allies to a course of action, but with

tathetyukt tu s dev sravannetrajalvil
on itktaikavasan muktakeyabhiniryayau (2.70.9; see above, n. 7).
ekavastr tu rudat muktake rajasval
on itktrdravasan draupad vkyam abravt
yatkrte ham imm prpt tesm varse caturdae
hatapatyo hatasut hatabandhujanapriyh
ban dhuon itadigdhngyo muktakeyo rajasvalh
evam krtodak nryah praveksyanti gajhvayam (2.71.1820).
12 chapter one

no stated commitment to any acts of her own. Yet if it does not men-
tion any such commitment, she does predict that among the reversals
that will take place, the Kaurava women will have to wear their hair
dishevelled. As we have seen, Bhatt a Nryana also plays upon this
reversal in a context where it has clear reference to Draupads (i.e.,
Bhmas) vow.
The most striking fact indicated by these two epic passages, how-
ever, is that Draupad has not changed her garment or her hair since
she was dragged into the sabh by Duhsana. It is difficult to know
just how long the poets took this interval to be. On the one hand,
it is twice indicated that the Pndavas had gone far on their way
(vyadhvagata; 2.66.24; 67.1) back home, from Hstinapura to Indra-
prastha. How long it would take a royal procession to cover what
would thus seem to be at least half the roughly seventy crow-flown
miles between these two cities, and then cover the same ground in
return for the rematch, is perhaps best left to the poetic imagina-
tion. But it would not be amiss to suggest that it would have been
long enough for Draupad to change her garments and her hair if she
had wanted to do so. On the other hand, both passages seem to affirm
that upon returning to Hstinapura, Draupad is still menstruating.
But one may wonder at this. In at least one of its contexts, the term
rajasvalh suggests a more general condition of impurity, as in the
meaning covered with dust. Otherwise we have the implausibility of
Draupad predicting that when the Kaurava widows return from per-
forming their husbands and kinsmens funerary rites, they will all be
menstruating at the same time. The fact that Draupads prediction is
uttered in the form of a parallelism to her own and her husbands con-
dition after the dice rematch suggests that Draupad now may also be
rajasvalh in this more general dust-covered sense. Rajasvalh may
thus in these instances simply mean impure, in a state of defilement.
As the Kaurava women after the war will enter the city covered with
dust or in an impure condition, so Draupad and the Pndavas leave
the cityDraupad being described just after Nakula, who is literally
covered with dust (pm spacita; 71.17a)into a realm of death
symbolized by the gestures of their chaplain Dhaumya, who strews
darbha grass to the death goddess Nirrti and sings Sman verses to
Yama (71.21).
We cannot, of course, be any more precise than this on the physi-
ological and temporal specifics of Draupads menstrual cycle. But cer-
tain symbolic connotations now seem unmistakable. Whether she is
draupads hair 13

still having her period or not, Draupad enters the forest as if she is.
That is, she enters the forest in an extended condition of defilement.
And though, as we shall see, she will change her garments during the
thirteen years of exile, if she does not change her hair it is because it is
the primary symbol of that defilement which cannot be cleansed until
her condition has been reversed by the deaths of her tormentors. And
that reversal will occur, as she says, when the Kaurava women mourn
their dead as rajasvalh s, with dishevelled hair.

3. In exile
On just their third night (3.12.3) in the forest, the Rksasa Kirmra
bars Draupad and the Pndavas way. Draupad faints and is caught
by her five husbands: With her hair widely dishevelled and ruffled
by Duhsanas hand, she (looked like) a river run wild amidst five
mountains.32 This verse thus provides the first indication that Draupad
continues to wear her hair dishevelled in the exile. And it would seem
that she keeps it so throughout the twelve years of forest wanderings.
Nowhere else is it so described, but the pulling of her hair is sev-
eral times alluded to as if it were an ever-present source of outrage
(3.13.108; 48.3335), and there are also reminders that Draupad is
engaged with her husbands, during the exile, in various forms of tapas
(3.141.2122), wandering as very lean and good ascetics (sukrh
sutapasvinah ; 3.125.19), she herself wearing bark skirt and deerskin
(3.226.20b) or a brownish-red garment (ksyavasan; 227.10d).33 It
is not unlikely that such allusions to forms of tapas are directly related
to Draupads tacit (in the epic) vow, especially in view of the use made
of such themes in the period of concealment.

duh sanakarotsrst a viprakrn airoruh
pacaparvatamadhyasth nadvkulatm gat (Mbh. 3.12.17).
Draupads tapas lasts through the war and is said by Duryodhana to have
sustained the Pndavas victory (9.4.18; 58.10). Ksyavasana are brownish-red or
yellowish-red robes used by monks, and also in sam nysa-entry and tantric dks
rites; see Jan Gonda, Change and Continuity in Indian Religion (The Hague, Mou-
ton, 1965), pp. 379 and 436437; cf. J. A. B. Van Buitenen (ed. and trans.), The
Mahbhrata. 3 vols. (Chicago, University of Chicago Press, 1973, 1975, 1978),
vol. II, p. 676 (her ochre robe).
14 chapter one

4. Concealment
The thirteenth year of exile, which the Pndavas and Draupad must
spend incognito (ajta), is rich in its surprising twists and complex
symbolism. For the present, we must limit ourselves to a discussion of
Draupads disguise alone, though it should be emphasized that its full
eludication would require a discussion of her husbands disguises as
well, and of the Virtaparvan as a whole.
As J.A.B. Van Buitenen has remarked, there is a Hol-like inver-
sion of roles in the disguises which the three oldest Pndavas assume
before entering Virtas capital in the kingdom of Matsya. The same
principle, he notes, also applies to Draupad, but he mentions only
the lesser of two inversions: In Draupads case the skewness lies
not so much in the choice of occupation, but in the self-imposed
isolation from her husbands [. . .] with whom she had gone through
everything.34 Rather, the essential and significant inversion with
Draupad lies in the paradox that she whose hair is hidden, and, I
would argue, still dishevelled,35 will take on the occupation of chief
hairdresser for Virtas wife, Queen Sudesn. Of the many crucial pas-
sages in this Parvan concerning Draupad, two put this paradox into
the clearest relief. First, donning her disguise, Draupad prepares to
enter the capital: Then, having tossed back her curly-ended faultless
soft hair, that dark-eyed one concealed it on her right side. And hav-
ing wrapped herself in a single large black garment that was very dirty,
having assumed the dress of a Sairandhr, Krsn [Draupad] wandered
about distressfully.36 Next, having applied for the job of Sairandhr,
Draupad hears Sudesns first words of reply to her: I would have
you dwell on my head (mrdhni tvam vsayeyam vai; 4.3.20). This
is a transparent pun. As a Sairandhr Draupads primary task will
be to serve Sudesn as hairdresser.37 Thus Draupad, incarnation of

Van Buitenen , vol. III: 510 (quote from p. 8).
So she is depicted by the Poona Critical Edition artist; see above, n. 26.
tatah ken samutksipya vellitgrn aninditn
jugha daksin e prve mrdn asitalocan
vsaca paridhyaikam krsn am sumalinam mahat
krtv vesam ca sairandhryah krsn vyacarad rtavat (4.8.12).
Thus when she tells her husbands about her planned disguise, her identifying
trait is to be skilled in hairdressing (kual keakarman i; 4.3.17: one must note the
possible pun here, in that Draupad is, in a much wider sense, skilled in work that
concerns hair). In her job interview with Sudesn, she mentions hairdressing first,
followed by skills in pounding unguents and weaving garlands (4.8.18).
draupads hair 15

rthe goddess of the sovereignty that is symbolized in the royal

diadem (kirta)is reduced to dressing the royal hair.38 Indeed, her
disguisewith her long hair concealed under a large black very dirty
garmentis the very antithesis of what one would expect from the
incarnation of r, whose other primary associations include good for-
tune and auspiciousness.39
The skewness in Draupads disguise thus involves not only an
inversion of her identification with r-Laksm, but a deepening accen-
tuation of the symbolism of her tacit vow of dishevelment. Actually,
the Southern Recension of the Mahbhrata leaves little doubt of a
connection between Draupads vow and her disguise. There, when she
tells Yudhisth ira how she will conceal her identity, she says: I will
be a Sairandhr, belonging to that caste, by the name of Vratacrin
(sairandhr jtisam pann namnham vratacrin ; 4.94* line 6; before
4.3.16). Vratacrin , meaning she who is undergoing a vow, can
hardly be anything but a paradoxical reference to her vow of dishev-
elment.40 Here we should recall one of Bhatt a Nryanas allusions
to Draupad vowing. At the outset of battle in the Ven sam hra, she
asks about the sound of drums and is told by Bhma that a sacrifice
is beginning. What is this sacrifice? she asks. And Bhma answers:
The Sacrifice of war (ran ayaja). So indeed. We four [the younger
Pndavas] are the officiating priests; Lord Hari is the director of rites;
the king [Yudhist hira] is the one consecrated (dksitah ) for this sacri-
fice of war; [our] wife is the one whose vow is maintained.41

Cf. also 2.60.23. One of Arjunas names is Kirtin, evoking his connections not
only with Draupad but with kingship; see Biardeau 1978: 17778. On Draupad and
r, see Hiltebeitel [1976] 1990: 6268, 9699, 144147, 166178, 223224.
Cf. the disguise and inverted appearance of r weeping and having an
impoverished share (mandabhgy; 1.189.13) discussed in Hiltebeitel [1976] 1990:
17075. On r and auspiciousness, see Frdrique Marglin, Devadasis as Specialists
in Auspiciousness. Paper presented to the Conference on Religion in South India
(Chambersburg, Pa., May 1978).
There is also a possible allusion in the Critical Edition when Yudhist hira responds
approvingly to Draupads plan for her disguise with the words Good, you are estab-
lished in a vow of rectitude (sdhu sdhvvrate sthit; 4.3.19)although the words
could simply refer to Yudhist hiras confidence in Draupads chastity during this
period when she will appear wanton and accessible to others. The name Vratacrin is
exceptional. Draupad tells Sudesn her name as a Sairandhr, when formerly employed
by Draupad, was Mlin (garland girl: 4.8.19). She is never called this, however, but
only Sairandhr.
ran ayajah / tath iti / catvro vayam rtvijah sa bhagavn karmopadest harih
sam grmdhvaradksito narapatih patn grhtavrat (1.25).
16 chapter one

As Gajendragadkar indicates the wife of the sacrificer [. . .] had to

observe certain vows. In the present case the vow refers to the vow
observed by Draupad of wearing her hair loose.42 This is surely right,
and also most illuminating. Although she is relying on a textual vari-
ant, Madeleine Biardeau has called attention to an epic recognition of
the Pndavas as dksitas. As they embark on their period in the forest,
during which, as we have seen, they and Draupad will undertake vari-
ous forms of tapas, they are prophetically described in some manu-
scripts as dksitas.43 The thirteenth year, spent like creatures dwelling
in the womb (garbhavsa iva prajh ; 4.66.10d),44 would thus mark
the completion of their dks, the consecration through which the
sacrificer is reborn in the dks hut and thereby consecrated to per-
form sacrifice.45 The Ven sam hra passage just cited thus relies on a
suggestive epic theme when it presents Yudhist hira as a dksita being
consecratedas he is in the epicfor the sacrifice of battle.46
Now one aspect of the dks, particularly in connection with the
theme of rebirth in the dks hut, is the taking on of ritual danger,
death, and impurity.47 This takes us near to the heart of the symbol-
ism of the Pndavas disguises. Indeed, impurity taken in the sense
of tasks inappropriate to ones caste, is a common denominator of
all the disguises. Says Bhma in the Ven sam hra, the Pndavas

This recognition is surprising, for, as we have seen, the Ven sam hra gives central
place to Bhmas role in the vow concerning Draupads hair.
Thus a few Northern manuscripts (including Nlakanth as) and one Southern.
The Critical Edition supplies the more regular parjith , defeated.
Cf. also the Northern interpolation after 4.12.11ab they went about in the city of
Virt a as if again born in the womb (virtanagare ceruh punar garbhadhrt iva).
For the source of these ideas and further elaborations, see Biardeau 1976: 2078;
1978: 14957, 18788; idem, 1978: 149557, 187188 (n. 3); Biardeau, Religions de
lInde, Annuaire de lcole Pratique des Hautes tudes, Ve section, t. LXXXII (1974),
fasc. III, p. 94; Biardeau and Malamoud, Le sacrifice dans lInde ancienne (Paris, Presses
Universitaires de France, 1976), p. 133. On correlations between dks, tapas, and vrata
(vow), see Walter O. Kaelber, Tapas, Birth, and Spiritual Rebirth in the Veda, His-
tory of Religions, XV (1976), 257277; Gonda 1965 359, 326, 340345, 381.
There is a growing scholarship on this sacrifice of battle theme. In addit-
ion to Hiltebeitel 1976 [1990], and Biardeau 1976, 1978, see also Heino Gehrts,
Mahbhrata. Des Geschehen und seine Bedeutung (Bonn, Bouvier Verlag Herbert
Grundmann, 1975).
On precautions and themes concerning danger, death, impurity and purification
during the dks, see Gonda 3965 339, 35860, 371372; W. Caland and V. Henry,
LAgnisto ma: Description complte de la forme normale du sacrifice du soma dans le
culte vdique (Paris, Ernest Leroux, 1906), p. 22; Walter O. Kaelber, The Dramatic
Element in Brhmanic Initiation: Symbols of Death, Danger, and Difficult Passage,
History of Religions, XVIII (1978), 5765.
draupads hair 17

and Draupad were concealed by means of improper occupations

(anucitrambhanibhrtam ; 1.11). But it is Draupad whose occupation
is most deeply bound up with impurity.48 This is obvious in connec-
tion with her very dirty garment. But it is also to be perceived in her
tasks as a Sairandhr.
The term Sairandhr, following van Buitenens insightful discus-
sion, is the feminine form of the name of a low caste people, vari-
ously sanskritized.49 Thus when Kcaka declares his passion for her,
Draupad can try to discourage him by telling him she is a low caste
Sairandhr, an abhorrent hairdresser (vihnavarn m sairandhrm
bbhatsm keakrikm; 4.13.13). The Southern Recension amends
this verse to specify that the low caste is the fourth (caturtha),
that is, the dra caste (notes to 4.13.13).50 But Manu and other clas-
sificatory texts (see Mahbhrata 13.48.1920) hint at more complex
and still lower caste associations.51 In this light van Buitenen suggests
deriving the name from ilindhra, silindhra, mushroom, also a bor-
rowed word in Sanskrit, and proposes that the Sairandhras might
be a tribe of mushroom gatherers or eaters.52 As ilndhra, silandha,
slandhra also refers to a kind of fish, however, it might be preferable
to suggest a tribe of fisherfolk. This would present a deep irony, as
the Pndavas are of course hidden in the kingdom of Matsya, Fish.53
But if such etymologies are guesswork, it is worth noting that in south
India the term Sairandhr, with the meaning prostitute, is cited as
the etymological explanation for the namegiven variously as Sin-
dollu, Chindiwandlu, Sindi, Sindhuvalluof a group of courtesans
and actors among the outcaste Mdigas.54

I reserve discussion of the five brothers on this and other points, and of the
ajta period as a whole, for treatment elsewhere.
Van Buitenen 1978, vol. III: 8: other forms include sairam dhra, sarim dha,
sairidya, sairitya, sairidya, sairim dhya, and sarim dhra; van Buitenen sensibly rejects
the etymology siram-dhra.
The Southern Recension also refers to a Sairandhr jti, or sub-caste (4.94* line
6, cited above).
Manu 10.32 identifies the Sairandhra as the offspring of a dasyu father and an
yogav mother: a dasyu is born outside the four varn as (see Manu 10.45), and yogavas
are carpenters (10.48) born from dra fathers and Vaiy mothers (10.12).
Van Buitenen 1978, vol. III: 9.
On the symbolism of fish in connection with the kingdom of Matsya, see some
very subtle insights of Biardeau 1978: 9699.
P. K. Nambiar (ed.), Census of India, 1971, vol. II, part V-B, series 19. Tamilnadu.
Ethnographic Notes. Scheduled Tribes of Tamilnadu (Delhi, Controller of Publica-
tions, 1975), p. 284 [quoting an extract on Mdigas from Syed Siraj al Hassan. Castes
18 chapter one

Cutting and dressing hair are, of course, low caste occupations. To

take a modern instance, the hairdresser for women at Punjabi mar-
riages and funerals is the barbers wife.55 Such handling of anothers
hair naturally involves the handling of impurities. As already indicated,
ritual washing and dressing of womens hair occurs after times of pol-
lution, including sexual intercourse and menstruation. According to
Hershman, in such womens rituals female hair becomes in abstract
terms a symbol of female sexuality or more concretely of vaginal men-
strual blood, a correlation made explicit when Punjabi women refer
to the regular pollution of menstruation [. . .] as syr nn i (having to
wash the head).56 It is thus not surprising that Draupads reduced
status calls forth the variety of associations with low caste that were
discussed above.57
It would thus appear that if, as Bhma says in the Ven sam hra,
the Pndavas and Draupad conceal themselves through improper
occupations, it is Draupads occupation that is the most improper
and impure. It is surely the only one with dra and even outcaste
connotations.58 One might thus suggest that it is through her that cer-
tain aspectsparticularly those involving impurityof the Pndavas
dks-like rebirth are given shape and direction. For example, while
Arjuna (in this same parvan) sees to the marriage of Abhimanyu and

and Tribes of H. E. H. The Nizams Dominions, vol. I (Bombay igo), pp. 409420,
which is unavailable to me]; Edgar Thurston, Castes and Tribes of Southern India,
7 vols. (Madras, Government Press, 1909), vol. IV, p. 315.
Hershman 1974: 281, 285; on the widespread impurity of barbers in India, and
on their low (but not untouchable) caste ranking, see Louis Dumont, Homo Hier-
archicus: The Caste System and its Implications (Paris, 1966: London, Weidenfeld and
Nicholson, 1970), pp. 4849, 5758.
Hershman 1974: 282286.
Also Duryodhana, upon winning her in the dice match, says let her sweep the
house (sam mrjatam vema; 2.59.1). Choosing to be a Sairandhr, however, gives
her certain options beyond those of a sweeper, for, as she says, Sairandhrs, though
unprotected in the world (araksit loke; 4.3.16), can move about freely and set limits
to the tasks they will accept: see, for Draupad, 4.8.2732, and the similar conditions
Damayant imposes when she too disguises herself as a Sairandhr (3.62.3740).
Biardeau has recognized this in a brief note, where she speaks of Draupad
going through une priode dimpuret in which elle est comme intouchable. Cf.
Biardeau 1976: 207, n. 1; 1978: 187, n. 2. Stig Wikander and Georges Dumzil have
perceived that the brothers disguises are differentiated according to the three twice-
born castes; see Wikander, La lgende des Pndava et la substructure mythique du
Mahbhrata (trans. Georges Dumzil ), in Dumzil, Jupiter, Mars, Quirinus, IV:
Explication de textes indiens et latins (Paris, Presses Universitaires de France, 1948),
pp. 4849; Dumzil, Mythe et pope, I: LIdologie des trois fonctions dans les popes
des peuples indo-europens (Paris, Gallimard, 1968) pp. 7073, 91.
draupads hair 19

Uttar, which will eventually lead to the rebirth of the Pndava line,
Draupad concerns herself with impurities that affect the rebirth of
the Pndavas themselves. One will thus note that whereas Arjuna
brings Uttar multicolored garments with which to dress her dolls
(4.35.2223)an image of the baby, Pariksit, whom she will eventually
bear to Arjunas son, Draupads garment is a very dirty black. If
the former garments represent the amnion and chorion as auspicious
dks-related symbols of rebirth,59 Draupads black garment would
represent the necessary taking-on of impurity and death as the pre-
condition for that rebirth. Similarly, when Virta strikes the disguised
Yudhisth ira, and Draupad catches the blood from his nosebleed in a
golden bowl (4.63.4647), she prevents the Pndavas from using their
power destructively against the kingdom of Virta, the very womb
of their rebirth. Says Yudhisth ira to Virta: Surely, if that blood from
my nose had fallen on the earth, you and your kingdom, O king,
would undoubtedly have perished.60 This scene evokes myths of the
Raktabja type, where blood touching the ground leads to the release
of destructive forces. Like Kl or other representatives of the Goddess
who prevent disaster by drinking up drops of blood before they touch
the ground,61 Draupad prevents disaster by catching them in a bowl.
But also, by implication, in preventing the destruction of Matsya, she
prevents the Pndavas from revealing their identities too soon, which
would terminate the possibility of their rebirth.
Both as hairdresser and blood-catcher, Draupad thus turns her
contact with impurities to her own and her husbands advantage. But
the scene which gives such themes their fullest play is the killing of
Kcaka. This tormentor, although Queen Sudesns brother, is a sta
by caste, and thus there is a resonance of association with Karna, a
figure whom Draupad already holds in high enmity for his role in
ordering her disrobing, and whom, at least according to the Northern
Recension, she rejected as a suitor on the very grounds that he was a
low caste sta (see 1.178.17 and 1827*).62 Let us look at the episode
primarily from the standpoint of Draupads hair: its association with

See Gehrts 1975: 206207, 224225; Hiltebeitel 1980a: n. 38; Biardeau 1978: 197.
yadi hyetad pated bhmau rudhiram mama nastatah
sarstr as tvam mahrja vinayeth na sam ayah (4.64.68).
See Dev Mhtmyam 8.4063, and Hiltebeitel 1980c: 193, 20011.
On Kcaka and Karna, see Biardeau 19751976: 576; idem 1978: 17.
20 chapter one

the themes of purity and impurity, and its evocation of further themes
connected with the goddess.
Frustrated by his efforts to seduce Sairandhr-Draupad in his own
quarters, Kcaka chases her into King Virtas mens hall. There, in
front of Bhma and Yudhisth ira, who are incapacitated by their dis-
guises, he seizes her by her mass of hair (keapaksa; 4.15.7) and
kicks her. Her eyes blood red with wrath, having loosed her hair
(ken muktv; 15.36), she rushes off to remonstrate with Sudesn,
and then goes back to bathe. Ken muktv, having loosed her hair,
is problematic.63 I take it Draupad spreads or shakes loose her already
unbound hair (her keapaksa; earlier hidden beneath her black gar-
ment, perhaps now revealed for the first time?) before her husbands
and Virta as a sign of outrage and additional defilement. The South-
ern Recension supports this inference with an alternative line: having
held together her loose hair which was sprinkled with blood (ken
pramuktn sam yam ya rudhiren a samuksitn; 4.338*, line 2, apud
4.16.37)Kcakas kick, according to the Southern Recension, having
made her mouth bleed (mukhd rudhiram sravat; 4.309*, line 3, apud
4.15.7). One is reminded here of the sexual connotations of blood in
the hair discussed by Hershman: foiled rape in this instance turns into
hair sprinkled with oral blood. But if there is not precise and evident
consistency in every epic reference to Draupads hair, one can at least
find it in her actions. When she goes back to her quarters, eagerly
wishing for Kcakas death, she purifies herself by washing her limbs
and garments (gtrn i vsas caiva prakslya salilena s; 4.16.2). And
again, after Bhma has slain Kcaka, Draupadfreed (moksit) by
Bhma (4.23.11)goes back to the city and washes her limbs and gar-
ments (gtrn i vsas caiva prakslya salilena s; 23.12). The striking
thing in these precisely repeated verses is that her two purificatory
baths do not involve any reference to the washing of her hair.
In Bhma and Draupads collaboration to bring about the death of
Kcaka, there are also further echoes of the mythology of the Hindu
Goddess.64 In all this, one must understand that Bhma and Draupad
are acting as one. As Biardeau puts it, when he slays Kcaka, Bhma is

Thus Van Buitenen (1978: 50): undid her hair.
In addition to what is discussed below, cf. the effortswhich the Goddess turns
against their messengersof umbha and Niumbha to have Durg dragged to them
by the hair to fulfill their lust (Dev Mhtmyam 5.1217.20).
draupads hair 21

le bras de la desse.65 This is not the place to underscore the many

affinities between Bhma and the goddess throughout the epic.66 It
must suffice to note that when Kcaka goes back to his apartments to
doll up for his expected nighttime tryst with Draupad, the fool was
exceedingly delighted, not realizing that it was Death (mrtyu) in the
form of a Sairandhr.67 And when he finally arrives at the trysting
place to fulfill his desires, finding Bhma lying there in the bed, the
sta [Kcaka] touched Death (mrtyu) [who was] burning with wrath
that was born from the insult to Krsn .68 Since they meet in the dark,
it is not clear whether Kcakas humorous failure to realize that he
has not met Draupad but Bhma is due to his lust, or to Bhma being
dressed in womens clothes, as he is in south Indian dramatizations
of the scene.69 But in any case Bhma is acting as and for Draupad,
and both are identified with Death, Mrtyu, as she is envisioned and
touched by Kcaka. In the epic, Mrtyu is a goddess, a form evoking
Kl and Durg as they break into the Hindu literary tradition.70 For
Death to appear to Kcaka in the form of a Sairandhr would seem
an unmistakable evocation of Kl, the wild goddess with the dishev-
elled hair.

5. Draupads call for war

Having emerged from exile and concealment and found the Kauravas
unwilling to return to them their share of the kingdom, the Pndavas
about to send Krsn a to negotiate with Duryodhanameet with their
allies and debate strategy. A surprisingly conciliatory speech by Bhma,
urging Krsn a to appease Duryodhana, sets off a chain of reactions that
concludes and culminates with Draupads call for war and revenge.
Rebuking her husbands, she urges Krsn a to turn his full wrath on the
Kauravas, and continues with the following words and gestures:

Biardeau 197374: 96.
See Biardeau as just cited, and also, on his birth, 1976: 23136, 250.
. . . bhram harsapariplutah
sairandhrrpin am mdho mrtyum tam nvabuddhavn (4.21.19).
aynam ayane tatra mrtyum stah parmrat
jajvalyamnam kopena krsn dharsan ajena ha (4.21.42).
See Martha Bush Ashton and Bruce Christie, Yaksagna: A Dance Drama of
India (New Delhi, Abhinav Publications, 1977), p. 43. Bhma also dresses as Draupad
in the terukkttu of Tamilnadu.
See Hiltebeitel 1978: 78385.
22 chapter one

Having said this, the dark-eyed, large-hipped one took with her left hand
her mass of hairsoftly gathered back and curled at the ends, very beau-
tiful, deep black, perfumed with fine scents, endowed with every good
mark, glossy like a great cobra. And the lotus-eyed one of elephant gait,
having approached the lotus-eyed one, her eyes filled with tears, Krsn
to Krsn a, said this word: This [hair] pulled by the hand of Duhsana
is to be remembered by you at all times, O lotus-eyed one, when seeking
peace with the foe . . . [She adds that if her husbands wont avenge her,
her father, brothers, and sons will.] Unless I see Duhsanas dark arm
cut off and covered with dust, what peace will there be in my heart?71
In its interplay between eyes and names, the passage is rich in evoking
the mythological and theological dimensions of the relationship
between Krsn -Draupad and Krsn a-Vsudeva.72 Draupad, r
incarnate, would seem to show an inauspicious and destructive side
of the Goddess when she left-handedly brandishes her cobra-like
mass of black hair and demands the black arms that seized it. But
most interesting, this is the closest the epic comes to attributing to
Draupad one of her popular vows. As with the vow demanding the
blood from Duryodhanas thigh to anoint her hair, the condition of
her hair is directly connected with a call for revenge concerning one

ityuktv mrdusam hram (*) vrjingram sudaranam
sunlam asitpng pun yagandhdhivsitam
sarvelaksan asam pannam mahbhujagavarcasam
keapaksam varroh grhya savyena pn in
padmks pun darkksam upetya gajagmin
aruprn eksan krsn am vacanam abravt
ayam te pun darkksa duh sanakaroddhrtah
smartavyah sarvaklesu paresm sam dhim icchat
/. . ./
duh sanabhujam ymam sam chinnam pm sugun th itam
yadyaham tam na paymi k ntir hrdayasya me (5.80.3336, 39).
* On mrdusam hram, the Ganguli-Roy and Dutt translations both follow Nlakanth as
commentary: mrdusam hram ven rpen a samhrtam api mrdum (Ramachandra
Kinjawadekar, ed., Shriman Mahbhratam, with Bharata Bhawadeepa by Nlakan th a,
6 vols. [Poona, Chitrashala Press, 19291933], Vol. III, p. 159). Manmatha Nath
Dutt (trans.), A Prose English Translation of the Mahbhrata, I 3 vols. (Calcutta,
H. C. Dass, 18951905), Udyoga Parva, p. 121, thus for example renders it soft
though bound up in a braid; cf. Kisari Mohan Ganguli (trans.) and Pratap Chandra
Roy (publisher), The Mahabharata (18841896; repr. in 12 vols., New Delhi, Mun-
shiram Manoharlal, 1970), Vol. IV, Udyoga Parva, p. 177. It is noteworthy, however,
that the Chitrashala Press artist (like his Critical Edition counterpart; see n. 26) depicts
Draupad with dishevelled hair on the page directly facing Nlakanth as comment!
Van Buitenen 1978: 357, translates it simply as soft. Softly gathered back seems
reasonably neutral; sam hra does not require Nlakant has interpretation of a ven .
See Hiltebeitel [1976] 1990: 8190, 14546, 22628; Biardeau 1978: 130, n. 1.
draupads hair 23

of her tormentors (here Duhsanas) limbs. It is unlikely that the

well known popular vows concerning the blood from Duryodhanas
thigh or from Duhsanas chest could derive from this speech. But the
epic poets give expression to much the same tone, and to a parallel,
probably sexual symbolism.73

6. The rebraiding
The war over, Yudhisth iras self-recriminations begun, Bhma consoles
him: By good luck, the sinful Duryodhana has been slain with all his
followers in battle. By good luck, you have gone the way of Draupads
mass of hair.74
This passage establishes beyond any reasonable doubt that Draupad
has worn her hair loose since the dice match. The Ganguli-Roy and Dutt
translations of the epic both seem to recognize this, with Ganguli mak-
ing it explicit in a footnote. Yet these two translators have missed some
of the verses simple precision. Ganguli takes keapaksasya [. . .] tvam
padavm gatah as you have attained to the condition of Draupads
locks, and notes: i.e., thou hast been restored to the normal condi-
tion. Draupad had kept her locks dishevelled since the day they had
been seized by Duhsana. After the slaughter of the Kurus, those
locks were bound up as before, or restored to their normal condition.75
But padavm is more simply a way, as when Nakula says that the
Pndavas enter the forest following the way of Draupad (draupadyh
padavm caran; 2.68.45b). And Draupads keapaksa is not simply her
locks but, as we have seen, the mass of hair that has been seized by
Duhsana and Kcaka and brandished at Krsn a. Thus, while the pas-
sage certainly alludes to Draupads vow, it does so more pointedly
than Ganguli has realized. In following the way of Draupads mass
of hair, the Pndavas have moved from the depths of defilement to
rebirth, from rebirth to revenge, and from revenge to coronation. For
in the rebinding and anointing of Draupads hair the latter is implicit:
she is Royal Prosperity or SovereigntyRjarincarnate, and

On overtones of castration in the breaking of Duryodhanas thigh in the Draupad
firewalking cult, see Hiltebeitel 1982. See also below, n. 88.
disty duryodhanah ppo nihatah snugo yudhi
draupadyh keapaksasya disty tvam padavm gatah (12.16.25).
Ganguli-Roy 1970, vol. VIII: 31, n. 1; Dutt (1905 Shanti Parva: 20), translates with
still more contrivancy: By good luck, you too, like Draupads locks, have regained
your normal condition. Nlakanth a is silent.
24 chapter one

her anointing represents the restoration of Yudhist hiras access to

sovereignty.76 In this connection it is all the more interesting that the
good luck attendant upon following the way of Draupads mass of
hair is mentioned along with the good luck of Duryodhanas death.
Again, one can only wonder how deeply, and with what other themes
in mind, the poets mean to connect the death of Duryodhana with the
binding up of Draupads hair.

C. Binding and Loosing

It is plain that Draupads hair, as a powerful symbol, is not exhausted

by either of the two literary contexts so far discussed. One must reach
deeper into Indian culture to appreciate what the Mahbhrata and the
Ven sam hra are themselves evoking through the use of this theme.
The closing scene of the Ven sam hra opens some surprising vis-
tas. Having just recognized Bhma, begored after killing Duryodhana,
Yudhisth ira asks him, What else remains? A very great remainder,
says Bhma; I will just bind together Pcls mass of hair, dragged
down by Duhsana, with this hand moistened with the blood of
Duryodhana. May you go, replies Yudhithira; let the tapasvin
experience the great festival of the tying of the hair.77 The reminder of
Draupads tapas is impressive. But why a great festival (mahotsava)?
One does well to recall the south Indian village ceremonies in which
the festive tying up of the goddess Draupads hair not only marks the
end of the war (with Duryodhanas sacrificial death), but is a necessary
precondition for the firewalk. In both cases, the festive aspect is related
ritually to the consecration and coronation of Yudhisth ira.78 And one
is reminded further of the three queens unbinding and rebinding their

On Draupads connection with r, see above, nn. 38 and 39; the restoration
of sovereignty may also be evoked when Draupad bestows Avatthmans jewel on
Yudhist hira for the latter to wear on his head (10.16.3334).
Yudhisth irah kim aparam avaista m / Bhmasenah sumahad avaista m /
sam yacchmi tvad anena suyodhanaon itoksitena pn in pcly duh sanvakrst a m
keahastam / Yudhisth irah gacchatu bhavn / anubhavatu tapasvin ven sam hrama-
hotsavam / (Ven sam hra 6.41).
Thus Ven sam hra 6.12 has a messenger report to Yudhisth ira that Duryo-
dhana has been found: Let jewelled pitchers be filled with water for your royal
coronation; and let Krsn hold festivity at [the near prospect of] the tying of her
braid, discontinued for so very long (pryatm salilena ratnakala rjybhisekya
tyantacirojjhite ca kabarbandhe karotu ksan am); on ksan am as festivity,
te / krsn
see Gajendragadkar, part III: 193. At south Indian Draupad festivals, the braiding is
draupads hair 25

hair, slapping their thighs, and circumambulating the horse in the

Avamedha just prior to the chief queens (mahiss) under-cover sex-
ual union with the victim, an act essential to the ritual establishment
of the kings sovereignty.79 One may thus posit that the unbinding and
eventual rebinding of Draupads hair resonates with themes of royal
ritual. But why is it so especially festive to her? And why the blood?
One must go back to the matter of impurity. As we have already
seen, it is the Mahbhrata itself which indicates that Draupads thir-
teen year dishevelment represents a condition of extended menstrual
defilement. Ethnographic evidence has also been cited to the effect that
washing of the hair is an absolutely crucial part of the ritual bath that
ends the period of menstrual confinement. One must, however, under-
stand here that conventional notions of menstrual blood as defiling
are somewhat one-sided. To the woman, at least, menstrual blood is
also purifying. There are numerous statements in the Dharmstra
Sam hits to the effect that women are cleansed by the menstrual flow.80
And the same point is made by Hershman: Punjabi women are not
happy unless they have a strong menstrual flow and they eat hot
foods in order to speed this. As Punjabi women say: the dirty must
be bled out so that we can become clean again. 81 This purifying func-
tion of menstrual blood can be perceived in numerous myths where
the sins of various male divinities are deflected onto women (in com-
bination with such other entities as Waters, Trees, and Earth), thence
to be absolved through the womens menstruation.82 One should also
keep in mind that female blood is thought to combine with male
seed in the act of conception, and thus to be indispensable to the

linked not only with the firewalk but with Yudhist hiras patt piskam, coronationa
ceremony that follows directly alter the firewalk.
pastamba rautastra 20.18.13; Baudhyana rautastra 15.29; Vdhla
rautastra Fragment no. 90. I thank David Knipe for bringing this scenario to my
attention. These gestures are omitted in the texts from the White Yajur Veda school,
which Paul mile Dumont treats synoptically in the main body of his LAvamedha:
Description du sacrifice solennel du cheval dans le culte vdique daprs les textes du
Yajurveda blanc Vjasaneyi-sam hit (Paris, Paul Geuthner, 1927); see pp. ix, 175182,
and, for the texts cited above. pp. 277, 337339, 370371.
E.g. Parara Sam hit 10.25; Atri Sam hit 19099.
Hershman 1974: 286.
See Wendy Doniger O Flaherty, The Origins of Evil in Hindu Mythology (Berke-
ley, University of California Press, 1976), pp. 15759. A case could be made that all
these entities are feminine, or at least connected with femininity.
26 chapter one

life process.83 It is thus essential for a womans period to go through

its natural course, and especially, while it is doing so, for her to be
Now, when Duhsana drags Draupad by the hair into the sabh,
this natural course is interrupted, and Duhsanas act is a molestation
in probably more senses than one. As Hershman points out, there are
close symbolic connections between womens head hair and female
genitals.84 The pulling of the hair, with all the sexual insults that attend
it in the Mahbhrata (including Karnas order to disrobe Draupad
and Duryodhanas bared thigh), is thus a form of sexual assault, and
a symbolic prelude to rape. In this connection, it is instructive to see
how the Atri Sam hit covers the case of a woman raped by a Mleccha
(barbarian): A woman, once she has been possessed by a Mleccha
or an evil-doer, may purify herself by the Prjaptya and the menstrual
flow. If she is taken by force, or else is herself imposed on by another,
a woman, once she has been possessed, purifies by the Prjaptya.85
Duhsana, an incarnate Rksasa like all the hundred Kauravas, fills
quite ably the Mleccha, evildoer role. It is most curious that the puri-
fication for such an assault is not only the completion of a menstrual
cycle, but the Prjaptya vow, a most difficult penance which lasts
for twelve days.86 Draupads twelve years of exile as a tapasvin are
probably an extension of this or some such vow. But this puts her
dishevelment into a new light. If her vow responds to a rape-like
defilement, then it represents not simply an extension of menstrual
impurity, but an extended condition of interrupted menstrual purifi-

See Wendy Doniger O Flaherty, Sexual Fluids in Vedic and Post-Vedic India,
chapter 2 of Women, Androgynes and other Mythical Beasts (Chicago, University of
Chicago Press, 1980).
Hershman 1974: 282, 286.
sakrdbhukt tu y nr mlecchair v ppakarmabhih
prjpatyena udhyeta rtuprasravan ena tu
balddhrt svayam vpi parapratrit yadi
sakrd bhukt tu y nr prjpatyena udhyati
(Atri Sam hit 19798; text from Manmatha Nath Dutt (ed.), The Dharma stra, Text,
vol. I (Calcutta, Elysium Press, 1908), pp. 19798. See Kane 1974, vol. II, part I: 575,
taking the passage in the sense of rape.
Kane 1974, vol. IV, pp. 132, 145146; vol. II. part I, p. 377, n. 916: it consists in
only one morning meal for three days, only one meal in the evening for three days,
subsisting on alms for three days and total fast for three days. Cf. Manu 11.211 and
Yjavalkya Sam hit 3.320. This is not to say that there arent other correspondences,
e.g. the twelve night dks of the Avamedha; see Dumont 1927: 54.
draupads hair 27

cation.87 And the blood combed into her hair, at least when it comes
from Duryodhanas thigh, must be at least in part a reversed image of
sexual revenge. The interruption of female menstruation is purified
by a ritual hair bath not with water, but with the strong flow of
male blood; blood, moreover, which comes from her chief tormentors
right thigh, a seat of virility and of phallic associations.88 Yudhisth iras
words in the Ven sam hraLet the tapasvin experience the great
festival of the tying of the hairthus call for the appropriate comple-
tion of Draupads purification.
If the twelve years of Draupads exile thus disclose the purifica-
tory, tapasic dimension of her vow, it is, however, still the thirteenth
yearspent incognitothat reveals her most hidden nature. Though
she is r incarnate, and thus directly affiliated with Visnu (normally
regarded as rs husband), she is also Krsn (the Dark Woman), a
name which in the Mahbhrata certainly connects her with Krsn a-
Vsudeva, but also with Mrtyu, Death, whom we have recognized in
the epic as a prefiguration of Durg and Kl. The name Krsn , in fact,
comes to be a common epithet for these latter two goddesses, as in the
Mahbhratas two apparently interpolated Durgstavas (4, App. I, 4,
D, line II; 6, App. I, I, line 17) and in the Purnas.89 Its usage for
Draupad is instructive. When she emerges, dark (ym; 1.155.42a),
from the sacrificial altar, a disembodied voice announces Best of all
women, this Dark Woman (krsn ) will lead the ksatriya caste to its,
destruction (sarvayosidvar krsn ksayam ksatram ninsati; 155.44cd).
Thus, while ostensibly it is because of her complexion (varn atah ), it is
also implicitly because of her destructive role that she given at birth the
name Krsn (155.50). This dark and destructive side of Draupad is
further compounded by certain affiliations with iva.90 She is thus not
simply an image of r-Laksm, but of the full vision of the Hindu

Cf. Biardeau 1976: 207, n. 1, connecting Draupads impurity with the non-
completion of her husbands Rjasya. More than one motive is likely.
See Hiltebeitel 1982: conclusion. On Indian traditions concerning the male thigh,
see O Flaherty 1976: 322324, 334, and especially the notions about Duryodhanas
thigh being his vital spot in south Indian vernacular traditions of the Mahbhrata,
on which see Subramaniam 1967: 283 (in Villiputtr), and T. V. Subba Rao, Telugu
Folk Additions to Mahabharatha, Folklore (Calcutta) 89 (1976), p. 272.
E.g. Dev Mhtmyam 5.22; 5.88 (Durg); Matsya Purn a 155.2833 (Prvat as
Kl, when iva calls her black).
She was fashioned by the holder of the trident (nirmit lapn in; 18.4.10);
and it is iva who sets events in motion that lead to her polyandry; see Hiltebeitel
[1976] 1990: 7985, 16667, 17788.
28 chapter one

goddess. To Kcaka, as we have seen, she is Death (Mrtyu) in the form

of a Sairandhr, and thus an epic intimationor to be on the safe
side, prefigurationof Kl herself.
Kl is, of course, the exemplary goddess of the dishevelled hair.91 The
Mahbhrata provides rather equivocal evidence as to recognizing her,
but the term kl, black, is used adjectivally to describe the goddess
Klartri in a crucial passage.92 While Avatthman, possessed by iva,
slays the sleeping remnant of the Pndava army, Klartrithe Night
of Time or Night of Deathis the form taken by the goddess as the
very embodiment of this night of destruction. The dying warriors thus
envision her: Black, with bloody mouth and eyes, with red garlands
and unguents, wearing a single red garment, noose in hand, [in the
form of] ikhandin, standing there smiling.93 If the passage does not
explicitly identify Klartri as Kl, it describes a goddess sufficiently
like her to eliminate any other contestants, and Klartri is a common
form of Durg and Kl in later texts.94 But the important thing is that
Klartri-Kl is connected with Draupad. Nowhere is this more
explicit than in the Draupad firewalking cult. It is understood that on
the night of her vengeance against Duryodhana, Draupad takes on
Klrp, the form of Kl, appearing not only with dishevelled hair
but in some cases with eighteen arms.95 The Mahbhrata itself does
not explicitly identify the black (krsn ) Draupad who will lead all

See the many references in David R. Kinsley, The Sword and the Flute: Kl and
a, Dark Visions of the Terrible and the Sublime in Hindu Mythology (Berkeley,
University of California Press, 1975), pp. 81159, especially the remarkable quotation
from Arthur Avalon, Principles of Tantra: The Tantratattva of ryukta iva Candra
Vidyrn ava Bhatta crya Mahodaya (Madras, Ganesh, 1960), p. 600: I see my Mother,
the mad, disordered girl, dancing with gentle movements of Her body, now taking up
the flute instead of the sword, or again seizing the sword instead of the flute, or yet
again at times making both the sword and the flute into one in Her hand; mingling
Her laughter with Her dancing; now loosening and now binding up Her hair.
See Kinsley 1975, taking kl as a proper name and mistakenly attributing dishev-
elled hair to her (a mistake easily produced by a misreading of the somewhat confusing
Ganguli-Roy translation). I must correct an earlier minimization of the importance of
this passage in Hiltebeitel [1976] 1990: 326.
klm raktsyanayanm raktamlynulepanm
raktmbaradharm ekm pahastm ikhan dinm
dadruh klartrim te smayamnm avasthitm (10.8.6465ab).
See Dev Mhtmyam 1.78cd, addressing the Goddess as Klartri, Mahrtri,
and the terrible Night of Delusion (klartrir mahrtrir mohartrica drun )
the first two referring to the nights of the occasional and great pralayas, on which
see below. Cf. iva Purn a, Um Sam hit 45.5859.
See M. S. Gopalakrishnan, Mother Goddess: A Regional Study (Madras and
Malabar) (University of Madras, M. Litt. Thesis, 1953), p. 107.
draupads hair 29

the Ksatriyas to destruction with the black (kl) Klartri who actu-
ally does so. But there are connections. Klartri appears directly after
Avatthman has killed Draupads five sons and four brothers, one of
whomikhandinis then immediately evoked in Klartris epithet
ikhandin.96 One should also recall that when Kcaka goes to meet
Death, Mrtyu, in the form of a Sairandhr, it is in the dead of night. But
it is the Ven sam hra that ties all such themes together, through the
words of Bhma, at the very point where he rebraids Draupads hair
Honored lady, dishevelled by Duhsana, let this braidthe Night
of Time of the family of Dhrtarstr anow be tied up.97 Draupads
dishevelled hair is thus itself an image of Klartri, the Night of Time,
the night of the dissolution (pralaya) of the universe.
This theme of dissolution brings our discussion to a close. Dishev-
elled hair connotes dissolution in several metaphoric senses. In
poems to Krsn a, the unbraiding of Rdhs hair connotes dissolution
in love.98 As has already been seen, dishevelled hair is connected with
the temporary dissolution of marriage ties in seclusion during men-
struation. And it also connotes the dissolution of such ties even more
explicitly at a husbands death, at widowhood. A most vivid expression
of this latter theme occurs in the Cilappatikram, the Tamil epic, when
the heroine Kannaki goes before the king of Madurai to confront him
with the injustice he has committed in ordering the execution of her

The name ikhandin (see above, n. 93, line 2) is interesting, although it is a
variant along with kutumbin, elderly lady, female householder; see Biardeau 1976
a: 210, n. 2, remarking that ikhandin is also a name for r (12.218.3a, 5a). This is
paradoxical, contrasting the destructive with the auspicious. ikhandin is of course
the true identity of ikhandin, born a girl from her previous life as Amb, and turned
male to defeat Bhsma. This identification of Klartri with ikhandin takes place
immediately after ikhandins death, and thus evokes the image of Klartri as the
Goddess in her full aspect, manifesting herself in a form which includes associations
with r and Amb.
bhavati sam yam yatm idnm dhrtarstr akulaklartriduh sanaviluliteyam
ven (Ven sam hram 6.40).
See Edward C. Dimock, Jr., and Denise Levertov (trans.), In Praise of Krishna:
Songs from the Bengali (Garden City, Anchor Press, 1966), pp. 7*, 24, 56*, 65; Deben
Bhattacharya (trans.), The Songs of Chandids (New York, Grove Press, 1970), pp. 64,
100*, 108; idem, W. G. Archer, ed., Love Songs of Vidypati (New York, Grove Press,
1963), pp. 55, 61, 81, 84*, 116; on the theme of love as pralaya, mentioning eveything
but the hair, see p. 44. Asterisks indicate especially forceful treatments of the theme.
30 chapter one

Her lily eyes streamed with water;

in her hand, a single anklet [proving her husbands innocence]; her form
untenanted by life.
What evil have I done that I saw her,
her black hair spread like a forest?
The king of Maturai beheld her
and terrified he died.99
This is of course only the beginning of the destructions wrought by
this dishevelled goddess of Chastity. As to Draupad, both the dis-
solutions of menstrual separation and widowhood apply during her
and her husbands forest exile and concealment. Throughout the thir-
teen years, she retains the impurity of menstruation compounded by
the violatory touch of Duhsana; and having seen her kingly hus-
bands treated as slaves, she, a princess, Royal Prosperity incarnate, is
symbolically a widow until they are reenthroned.
But beyond, or intertwined, with all these themes are what we may
call the ultimate cosmological and theological themes of dissolution. We
are back to the compound so persistantly invoked by Bhatta Nryana:
the kembarkarsan am, the pulling of the garments and the hair.
These two form a significant diptych, one which might be called the
two intangiblesthe garments and the hairintangible, that is, in
both the senses not to be touched and theologically subtle.
One will recall that in the Mahbhrata it is not Duryodhana (as
it is in the Ven sam hra) who orders the disrobing of Draupad, but
Karna. As I have argued elsewhere,100 the disrobing scene suggests
most basically, and probably most originally, a mythology connecting
Draupad with the Earth.101 Karnas identification with his father Srya,
the Sun god, is a theme that the epic poets never tire of reiterating.
Their choice of Karna to order Draupads disrobing is not arbitrary.
The relationship between Draupad aud Karna is, as we have seen,
one of continued opposition. Their incompatibility is in fact dictated
by a cosmological impossibility: a union of the Sun and the Earth.
Paradoxically, speaking cosmologically, one could say that the Earth
is for all to share (as Draupad is by the five Pndavas) except the Sun.

Cilappatikram 20, ven p 2, as translated by George L. Hart, III, The Poems of
Ancient Tamil: Their Milieu and their Sanskrit Counterparts (Berkeley, University of
California Press, 1975), p. 105. Cf. also idem, p. 93, loose hair as dangerous.
See Hiltebeitel 1980a; Karna orders the disrobing at 2.61.38.
See Hiltebeitel [1976] 1990: 6768, 240.
draupads hair 31

Karnas command to strip Draupad bare (2.61.38) is thus an image

of the occasional (naimittika) pralaya, the dissolution of the three
worlds. Were Duhsana to succeed in fulfilling Karnas command,
Draupad would be denuded like the Earth prior to its combustion,
dessicated by the pralayas seven suns (sprung from the single Srya)
and, as several texts put it, bare like the back of a tortoise.102 It is, of
course, significant that in ordering Draupads disrobing, Karna does
not touch Draupad or her garments himself. The miracle of the sarees
thus symbolizes that between the yugas, where the Mahbhrata story
is set, the Earth holds the power of her own renewal. Or, from another
angle, except at the time of the naimittika pralaya, the Earths power
to restore her garments is inexhaustible.
If the disrobing connects Draupad with a mythology of the Earth,
the dishevelment of her hair evokes still deeper identities of the
goddess. A triven is not only a type of three-plaited braid, but a
confluence of riversin actuality of two rivers but mythologically
of three. The model triven is at Prayga, confluence of the Gang,
Yamun, and the underground Sarasvat. Here at the very point
where the three rivers are braided together, one finds the great-
est purity on earth. This trtha is said to outlast the pralaya (Matsya
Purn a 111.410), which would seem to imply that it will never be
unbraided. Moreover, Prayga is the exemplary trtha connected
with ritual hair cutting for men and widows, and even in some con-
troversial instances for married women who offer their own ven s
into the triven with the mantra: May all my sins be destroyed by
this offering of my braided hair into the ven [. . .].103 The three rivers
are of course goddesses. One will recall Draupads dishevelled hair
also being compared to a river run wild (3.12.17), and think of the
Gangs wildness being tamed and channelled by its passage through
the dishevelled hair of iva.104 Connections between the goddess and
the waters are also drawn in festivals of the goddesss menstruation,
a condition that probably further evokes her dishevellment. In the

This image appears at Visn u Purn a 6.3.23 and Krma Purn a 45.23; see
Biardeau, EMH (III), part I. Cosmogonies purniques (suite), BEFEO, t. LVIII
(1971), p. 69.
Kane 1974, vol. IV, 602603.
Compare also Dimock and Levertov 1966: 56: Her cloud of hair eclipses the
luster of her face, like Rhu greedy for the moon; the garland glitters in her unbound
hair, a wave of the Ganges in the waters of the Yamun.
32 chapter one

Kongu region of Tamilnadu, such a festival occurs on the eighteenth

of the month of ti (JulyAugust), when the overflow of the Kvr
river is identified with the goddesss period, and correlated with the
victory of the Pndavas at the end of the eighteen day Mahbhrata
war.105 If the goddess in such festivals is Prvat or Kl,106 it is clear
how her regulated overflow is associated with themes of bounty and
victory. In Draupads case, however, it is precisely the violation of
this regularity that calls forth images of unbounded dissolution. The
untying of Draupads braid represents the potential untying of the
universe. For the universe is itself a braid, composed of the three gun as
or strands of prakrti, nature. Here, however, it is not the naimittika
pralaya that is evoked, but the prkrta pralaya, the very dissolution of
nature (prakrti) that unravels all seven worlds and occurs after a thou-
sand naimittika pralayas.107 Mythologically, this pralaya is the Night
of Time, Klartri, with whom Draupads hair is directly compared
in the Ven sam hra. Or, in more familiar terms, one thinks of iva
whose own dishevelled hair not only channels the Gang but runs
wild during his dance of dissolution, and finally of Kl herself as the
dishevelled Goddess of death and destruction.
By regaining, her garments, Draupad thus shows that the Earth is
inexhaustible, and the Kauravas cannot truly seize her or it. But by
wearing her hair dishevelled for thirteen years, Draupad also shows
that the full dissolution of the universe is at least metaphorically pend-
ing should her husbands, with the help of Lord Krsn a, not regain the
sovereignty and restore dharma on earth.

Personal communication from Brenda E. F. Beck.
The festival of Rajah sam krnti in Orissa celebrates the menstruation of Kl;
anticipating the rainy season, the festival marks a three day period during which
the earth must not be ploughed, unmarried girls ride swings, and boys take part
in strenuous games: personal communication from Frdrique Marglin, and see
Mayadhar Mansinha, The Saga of the Land of Jagannatha (Cuttack, J. Mohapatra,
1971), pp. 2527. Neither Beck (see n. 105) nor Marglin recalled any association of the
festival with dishevelment, but it is more than likely, since menstruation itself evokes
dishevelment (see above, nn. 2729).
On the distinction between pralayas, see Biardeau 1971: passim, and above,
n. 102.


One of the Mahbhratas most famous scenes is the multiplication

of the sarees at Draupads disrobing. But the passage has never been
closely examined either textually or symbolically. Recent developments
now make such an examination singularly inviting. For one thing, the
Poona Critical Edition provides an excellent guide to the transforma-
tions of the text. Secondly, recent scholarship into Mahbhrata sym-
bolism has reached a point where certain insights are beginning to
converge and distinctions between interpretations are becoming clear.
A discussion of the miracle of the sarees provides an occasion both to
reap the benefit of others insights and to clarify the lines of difference
between scholarly views. The approach here is stimulated by fieldwork in
the cult of Draupad (Tiraupatiyamman), found mainly in Tamilnadu,
in which Draupad is regarded as a form of Dev or akti and the
Mahbhrata is celebrated as her cult myth.1 One premise must thus
be stated at the outset. The notion that Draupad is a form of the Goddess
is not alien to the Sanskrit Mahbhrata. Indeed, her relation to early
concepts of the Hindu Goddess requires careful and detailed study. A
second premise will not bear so directly on this study of the epic text,
but is worth stating to avoid misunderstanding. The Draupad cult can-
not be traced back farther than about 500 years.2 Nonetheless, study
of the cult will frequently enliven and expand our knowledge of the
epic, particularly its symbolism. Garments figure very prominently in
Draupads cult. As in the epic, they are used ritually to mark transitions

Fieldwork was carried out in 1975 under an American Institute of Indian Studies
grant, and in 1977 on a National Endowment for the Humanities Summer Stipend.
For an initial statement, see Hiltebeitel 1982; see also Hiltebeitel 1981.
The oldest reference I know of is to a Draupad temple built and dedicated at the
north of Gingee town by Tubaki Krishnappa, founder of the Nyak line of Gingee
kings (rule ca. 14901520); see C. S. Srinivasachari, A History of Gingee and its Rulers,
Annamalai University Historical Series, No. 2 (Annamalainagar: The University,
1943), p. 88, n.
34 chapter two

or dramatic intervals in the story. This is true in particular of the dis-

robing scene.3
Draupads reluctant entrance into the Kaurava assembly-and-
gambling hall (sabh) is preceded by a question which she insists be
asked before she will appear: did Yudhist hira bet her before or after he
wagered himself? The question remains moot through the entire
episode. To the wisest counsellors it is irresolvable, and it drives
Yudhisth ira to silence. For, as J.A.B. van Buitenen says: Yudhist hira
cannot very well confirm that she was either won or not, for in either
case he would have to confirm a lie: if she was won, he lied about his
own stake, for he would still have been free to stake her; if she was
not won, because he was no longer free, his staking her was a lie.4
Duryodhana, delighting in the chance to catch Yudhist hira in this pre-
dicament, orders Draupad dragged into the sabh. Here she raises her
moot question on her own, and provokes the only responses pro and
con that the epic offers. They are found as a carefully constructed pair
in the form of a debate between two figures whose names, in this con-
text, have the look of a contrived opposition: Vikarna (Dhrtarstras
youngest and noblest son) and Karna. Whatever the significance of
the two names, Karn as position here is of the greatest significance.
Vikarna takes the compassionate view, arguing that the throw is null
for three reasons: (1) Draupad was staked after Yudhisth ira bet himself;
(2) it was only due to the prodding of akuni that Yudhisth ira bet her;5
and (3) the blameless or faultless (anindit) Draupad is common
to all the Pn avas (sdhran ca sarvesm pn davnm; 2:61.2324;
on citations, see n. 4). Karn a outraged at such assertions by a mere
youth, offers a point for point rebuttal: (1) it is irrelevant when she
was wagered: Yudhist hira could bet her because Draupad is included
within his total property (abhyantar ca sarvasve draupad; 61.31);
(2) akuni may have prodded Yudhisth ira, but he did it audibly and
her wager was allowed by the Pn avas (krtit draupad vc anujt

See Lawrence Babb, Walking on Flowers: A Hindu Festival Cycle (Singapore:
Department of Sociology, University of Singapore Working Papers, No. 27), 1974,
pp. 89.
J. A. B. van Buitenen, trans. and ed, The Mahbhrata, Vol. II: 2. The Book of the
Assembly; The Book of the Forest (Chicago: University of Chicago Press, 1975), p. 817,
note to 2.62.25. All Mahbhrata citations are from the Poona Critical Edition (see n. 7),
unless otherwise indicated.
Yudhist hira makes all his bets voluntarily except for the last, his wager of Draupad;
in this instance akuni prods him out of his uneasy silence.
draupads garments 35

ca pn davaih ; 61.33). But everything turns on the twist he puts on

Vikarnas third point: Draupads blamelessness and her polyandry.
Karna begins by introducing a point concerning Draupads condi-
tion that Vikarn a had not mentioned: Or if you think it was through
adharma that she was led into the sabh in a single garment, hear my
final word. One husband per wife is ordained by the gods, O scion of
Kuru; but she, whose submission is to many, is for certain a whore.
For her to be led into the sabh is not strange to my thinking, whether
clad in one garment or even unclad (61.3436). Karn as mention of
Draupads garments is thus gratuitous in the context of the argument
with Vikarn a. But his speech does relate the garments to the subject
of Draupads chastity, or fidelity to a single husband as ordained by
dharma. He thus dispels Vikarnas third point that, being common to
all the Pn avas, the blameless Draupad could not bet by any single
one of them. Hardly blameless, she is a whore (bandhak), com-
mon to all and protected by none.6 It is thus on this pretext, having
overturned Vikarn as assertions about Draupads blameless polyan-
dry, that Karna calls out: O Duhsana . . . Strip the Pnavas and
Draupads clothes (61.38).

I. Textual Transformations

Among the many passages excised from the Mahbhrata story in

the reconstituted text of the Poona Critical Edition, none is so popu-
lar as Draupads prayer to Krsn a in this moment of distress. But an
examination of the Northern and Southern variants of this story fully
justifies Franklin Edgertons judgment: the evidence of the manu-
scripts is entirely conclusive7 that the gesture is interpolated. Edg-
ertons comments focus mainly on questions of style and continuity,
but he also cannot avoid some of the theological implications of the
alteration. No prayer by Draupad; no explanation of the miraculous
replacement of one garment by another; no mention of Krsn a or any
superhuman agency. It is apparently implied (though not stated) that

It is actually Bhma who introduces this term in the episode when he angrily
denounces Yudhist hira for treating Draupad worse than a common gambler would
treat a whore (2.61.1).
Vishnu S. Sukthankar and S. K. Belvalkar, general eds., The Mahbhrata for the
First Time Critically Edited, Vol. II, The Sabhparvan, edited by Franklin Edgerton
(Poona: Bhandarkar Oriental Research Institute, 1944), p. xxviii.
36 chapter two

cosmic justice automatically, or magically if you like, prevented the

chaste and noble Draupad from being stripped in public. It is perhaps
not strange that later redactors felt it necessary to embroider the story.
Yet to me, at least, the original form, in its brevity, simplicity, and
rapid movement, appeals very forcefully.8 The evidence accumulated
in the Critical Edition leaves no grounds to refute Edgertons conclu-
sions: the passage has continuity without Krsn as interference, and the
tendency of later redactors (both Northern and Southern) to embroi-
der the story is evident. The varied interpolations, however, must have
a history, and one which may inform us as to how the miracle was
interpreted in terms of its theology and symbolism.
Rarely does the Mahbhrata offer better materials for historical
reconstruction of the textual process. To begin with, the Southern and
Northern variants of Draupads plea to Krsn a differ from each other,
and show many variations within themselves. Moreover, both North
and South provide prime and excellent manuscripts9 that entirely
omit the plea.
The reconstituted Critical text now reads (2.61):
40. Then Duhsana, O king, forcibly tore off Draupads garment in the
middle of the sabh, and began to undress her.
41. But whenever one of Draupads garments was removed, O king,
another garment like it repeatedly appeared.
42. Then there was a shout of approval there, a terrible roar from all the
kings, having watched that greatest wonder in the world.
It is after verse 40 that most of the interpolations were made. The main
Southern variant, to which shorter additions were frequently made,
has the distressed Draupad repeatedly call out Govinda and Krsn a,
and then recite a one loka prayer which several versions refer to as a
song (gt) taught to her by the sage Vasist ha (548*, apud 2.61.40):
Holder of the conch, wheel, and mace, whose residence is Dvrak, Go-
vinda, Lotus-Eyed, protect me who have come for refuge. (raksa m
aragatm 547*; cf. 548*)

Ibid. p. xxix.
The terms are Edgertons, referring to the Northern rad codex and two Southern
manuscripts, one in Grantha script and one in Malayalam; ibid., pp. xi and xxix.
draupads garments 37

One sees here a fully developed and intentionally highlighted bhakti

theology and iconography, perhaps even colored by ri Vaisnava
The chief Northern variant (543*), familiar from the Roy and Dutt
translations of the Mahbhrata from the Vulgate, and commented
on by Nlakan t ha,11 is actually found in only a few Devanagari
manuscripts.12 It is, however, much longer than the main Southern
variant, and has many more contextual references to other facets of
the Mahbhrata. Draupad also refers to Krsn a as Beloved of the
Gop Folk (gopjanapriya; 543*, line 2; also 542*, line 1), one of the
very few references to Krsn as childhood among the cowherds in
any recension of the epic13 and a likely indication that the passage was
interpolated by bards familiar with the work of Northern Paurnikas.
The passage goes to bizarre lengths, however, to bring Krsn a to
Draupads aid.14 At his residence in Dvrak, having heard the words
of Yjasen [Draupad], Krsn a was deeply moved. And having aban-
doned the couch where he slept, the benevolent one came there on foot
out of compassion (543*, lines 1011). From Dvrak to Hstinapura,
Krsn a had to cover some 800 miles as the crow flies. To cover that
distance on foot (padbhym) in time to rescue Draupad is more than
would normally be required even of a deity.
Most likely the demands on Krsn a were as much textual as devo-
tional. Immediately following the statement that he came on foot is a

On aran gati in South Indian Vaisnavism, see Robert C. Lester, Rmnuja and
r Vaisnavism: The Concept of Prapatti or arangati, History of Religions 5 (1966),
Kisari Mohan Ganguli, trans., and Pratap Chandra Roy, Publisher, The Mahbhrata
(18841896; repr. most recently in 12 vols., New Delhi: Munshiram Manoharlal,
1970); Manmatha Nath Dutt, trans., A Prose English Translation of the Mahbhrata
(Calcutta: H. C. Dass, 18951905). See also Ramachandra Kinjawadekar (ed.),
Shriman Mahbhratam, with Bhrata Bhawadeepa by Nlakaha (6 vols.; Poona: Chi-
trashala Press, 19291933).
Edgerton, Sabhparvan, p. xxix.
See Sadashiva L. Katre, Krsn a, Gopas, Gops, and Rdh, in H. L. Hariyappa and
M. M. Patkai (eds.), Professor P. K. Gode Commemoration Volume (Poona: Oriental
Book Agency, 1960), Part 3, pp. 8385.
Here it is like one other passage which the Poona Critical Edition shows is also an
interpolation; see my The Burning of the Forest Myth, Bardwell L. Smith (ed.), Hin-
duism: New Essays in the History of Religions (Leiden: E. J. Brill, 1976), pp. 221222.
That article presents a statement on the use of the Critical Edition which holds also
for this article.
38 chapter two

curious verse (544*). It is found extensively in the Northern recension,15

sometimes after the long interpolation just cited, more often directly
after verse 40, with the long interpolation omitted. It shows its anoma-
lous character by being the only verse among those either interpolated
or accepted into the Critical text that is in the tristu bh rather than the
loka meter. Theologically, however, it is consonant with a loka verse
(553*, apud 2.61.41) also interpolated widely in the Northern recen-
sion, and found almost uniformly in the same manuscripts as 544*. It
would thus appear that the interpolated verses 544* and 553* together
constitute the original interpolation in the Northern recension.16 Set
into the reconstituted narrative, the altered passage would then read
as follows:
40. Then Duhssana, O king, forcibly tore off Draupads garment in the
middle of the sabh, and began to undress her.
544* Yjasen invoked Krsn a, Visn u, Hari, and Nara for protection. Then
Dharma, concealed, the magnanimous, having a multitude of garments,
covered her.
41. Whenever one of Draupads garments was removed, O king, another
garment like it repeatedly appeared.
553* Thereupon garments of many colors and whites appeared, O lord, by
hundreds, due to the protection of Dharma . . .
In verse 544* Draupad does invoke Krsn a under the first three names
mentioned, but the fourthNarausually refers to Arjuna. But the help
seems to come more directly from Dharma. In verse 553*, her rescue
is more clearly due to Dharmas protection (dharmasya pariplant)
alone, with no mention of any other figure. To be sure, Dharma, like
Nara, could refer to Krsn a as Visnu. But in the Mahbhrata Dharma
is also a deity on his own. He is incarnate in Yudhist hira and Vidura,
and on two occasions acts in disguise. The concealed (antarhita)
Dharma who rescues Draupad here is certainly reminiscent of the
figure by that name who disguises himself as a Yaksa and a dog to test

As Edgerton observes (Sabhparvan, p. xxix), the rad Codex alone omits it,
although one Kamr manuscript has it written on the margin (Edgertons italics).
They appear in all the same Northern Maithil, Devangar, and Bengali mss.
544* appears in one more Kamr ms than 553*. Both appear together in the only
Southern ms in which either appears, G5 (Grantha script), with the apparent exception
of 553* being present in a Telegu ms which Edgerton claims not to have seen (?);
see Sabhparvan, p. xvi. On the basically Northern character of these verses, see ibid.,
p. xxix.
draupads garments 39

the Pndavas at other moments of mortal peril.17 But dharmas pro-

tection would seem also to have a more impersonal connotation. This
apparently initial double interpolation was perhaps what Edgerton had
in mind when he spoke of cosmic justice automatically rescuing the
chaste Draupad. Justice (dharma) may be set in motion by a prayer
to Krsn a, but it is available to Draupad precisely because she herself
is just, virtuous.
Most likely these interpolated verses proved devotionally uninspiring
to later Northern redactors, who must have sought uniformity in the
ambiguous references to Krsn a, Nara, and Dharma by prefacing verse
544* with a long and explicit plea to Krsn a culminating in the totally
unambiguous yet ridiculous assertion that he came to Draupads rescue
on foot. I would further suggest that Krsn as intercession, once
accepted and standardized in the Northern recension, became part of
an oral tradition and was then interpolated into the Southern recension
in the form of a simple devotional song.

II. Inexhaustible Garments

The two Northern verses concerning Dharma thus probably provide

the oldest intra-textual interpretation of the disrobing scene. Now
what does it mean here that Dharma has a multitude of various gar-
ments (dharmo . . . vividhavastrapgah ; 544*, lines 34)?18 The likely
answer is the most obvious. Dharma is often associated with garments,
as in the compound dharmakacuka, having the garb (or armor) of
virtue. The English idiom cloak of virtue is identical in purpose. In
the Rmyan as most pivotal ethical scene, the wounded Vlin uses

On the Yaksa disguise, see Hiltebeitel [1976] 1990, 18690, 24748. The dog dis-
guise is taken to test Yudhist hira on his way to heaven in the epics final book. There
is a significant controversy over who the deity Dharma is: a rejuvenated Vedic Mitra
(Georges Dumzil, Mythe et pope, vol. I: L idologie des trois fonctions dans les popes
des peuples indo-europens (Paris: Gallimard, 1968), p. 172; Yama, known in the epic as
Dharmarja (Madeleine Biardeau, tudes de mythologie Hindoue (IV), Part II: Bhakti
et avatra, Bulletin de lcole Franaise dExtrme Orient 63 (1976), p. 172); or an
independent deity (Jan Gonda, The Vedic Mitra and the Epic Dharma, Journal of the
Royal Asiatic Society (1971), pp. 120133, throughout in disagreement with Dumzil,
but also, p. 125, asserting that the epic scarcely knows Dharma as Yama).
Several Northern scribes were evidently also puzzled by this question, and
amended the text to read Dharma covered her with various beautiful garments (thus
the Roy and Dutt translations), . . . pairs of garments, or . . . meritorious garments;
see Critical Edition, notes.
40 chapter two

similar terms in excoriating Rma for his treachery in shooting at him

from ambush: Rma, says the dying monkey king, is a sinner wearing
the garment of the good, covered entirely with the deceptive dress of
virtue (sat vedhara ppam . . . dharmacchadmbhisavtam;
Baroda Critical Edition, Rm. 4.17.19). Similar terms are used to
describe Yudhisth iras major perfidy, his lie to Dron a leading the latter
to drop his weapons thinking his son was slain. According to the son,
Avatthman, Yudhisth ira is one possessed of the banner [or more
generally, garb] of virtue (dharmadhvajavat; 7.166.19); and Arjuna
admits that Yudhisth ira spoke untruth in the garb of truth (satyaka-
cukam . . . anrtam; 167.35). Dharma thus has many garments, many
disguises, as in the god Dharmas appearances as a Yaksa and a dog.
The metaphor corresponds to the oft reiterated epic statement that
dharma is subtle (sksmo dharmah ). In the disrobing scene, however,
the metaphor becomes reality. Dharmas garments are not just dis-
guises; they are actual apparel. And they appear in order to confirm
that one of the Mahbhratas most surprising disguises of virtue
Draupads polyandryis indeed virtuous. Vikarna is proven right:
Draupad, though wed to many, is faultless. Karn a is proven wrong.
Virtue clothes its own, and is inexhaustible in doing so.
The kernel verses interpolated into the Northern recension may thus
introduce the theme of praying to Krsn a, but their main burden seems
to be a rather profound ethical interpretation of Draupads reclothing.
Indeed, as far as questions of dharma go, these interpolations cannot
be far from the spirit of the passage as a whole, even without the inter-
polated verses. But what about the reconstructed passage as it stands;
with no plea to Krsn a and no reference to Dharma?
Here again the answer may be quite simple, although the symbolism,
being a rich one, has many ramifications. The basic insight was pro-
vided quite early in the history of Mahbhrata scholarship by an
advocate of a long since discredited school of interpretation: nature
mythology. According to Alfred Ludwig, the Mahbhrata conflict
truly begins with Duryodhanas success in the dice match: his victory
signifies the triumph of winter, and the attempt to disrobe Draupad
symbolizes winters baring of the Earth.19 Ludwigs instinct is certainly

Alfred Ludwig, Uber die mythische Grundlage des Mahbhrata, Sitzungsberichte
der Knigl. Bhmischen Gesellschaft der Wissenschaften. Classe fr Philosophie, Geschichte,
und Philologie (1895), p. 12.
draupads garments 41

right. He has just superimposed a Germanic climatology on India. It

is not winter that threatens to disrobe the Indian earth, but the over-
bearing sun of the hot season (usn a) that precedes the rains. It is not
Duryodhana who has ordered Draupad disrobed. It is precisely Karna,
son of the Sun. And it is Arjuna, son of Indragod of rainswho will
eventually kill Karna.
These points of convergence provide strong evidence that nature
mythology is a decisive factor in the scene of Draupads disrobing.
Karnas identification with his father Srya is a theme which the epic
poets reiterate at every possible chance.20 They did not pick Karn as
name out of a hat for his role at the disrobing.21 As to Draupad,
she is at many points identified with the Earth. For instance, when
Yudhist hira laments his sins at the end of the war, he blames himself
for the deaths of Draupads five sons, and compares her to the Earth
bereft of her five mountains (12.27.2122).22 And in the long Northern
interpolation at the disrobing, Draupad compares herself to the Earth,
or more precisely to the proverbial sinking earth of epic and Purn ic
avatra myths:
Sunk in the Kaurava ocean, lift me up (mm uddharasva), O Janrdana,
O Krsn a, Krsn a, Great Yogi, Soul of All, Creator of All. Suppliant to you,
protect me, O Govinda, sinking down (avasdtm) amidst the Kurus
(543*, lines 57).
In fact, Draupads true given name at birth, Krsn , probably has as one
a) being the Earths color.23
of its connotations the Earth, black (krsn

For discussion of examples, see Dumzil, Mythe et pope, I, pp. 125144; Biar-
deau, Bhakti et avatra, pp. 229230; Hiltebeitel [1976] 1990, 25466, 28185.
Later renditions of the scene sometimes attribute the fundamental blame for the
disrobing to Duryodhana rather than to Karna (see G. V. Devasthali, ed, and tr., Bhatta
Nryan as Ven samhram (Bombay: D. M. Tilak,1953), pp. 156 and 160), or even
specify that Duryodhana gave the command (Villiputtr lvrs Villi-Pratam, in
Tamil: see M. V. Subramanian, Vysa and Variations. The Mahbhrata Story (Madras:
Higginbothams, 1967), p. 90. By such changes in emphasis and alterations, Karna is
of course absolved and symmetry results in both the disrobing and the hair-pulling of
Draupad being blamed on Duryodhana as cause and on Duhsana as executor, thus
motivating Bhmas double revenge against these two.
See Hiltebeitel [1976] 1990, 240.
See Biardeau, Brhman es et potiers, Article liminaire, Annuaire de l cole Pratique
des Hautes tudes 84 (19711972), pp. 4041; Hiltebeitel [1976] 1990, 66.
42 chapter two

The relationship between Draupad and Karn a is presented through-

out the Mahbhrata as one of continuous opposition.24 When seen
in conjunction with the disrobing scene, other passages which oppose
them gain in significance. In a passage found in some Northern manu-
scripts, Draupad refuses Karn a an attempt at winning her in marriage.25
Similarly, Karn a rejects Krsn as pre-war proposal that Karn a assume
his rightful place as oldest brother of the Pndavas, with Draupad
coming to him as her sixth husband (5.138.15).26 Such a startling prop-
osition from Krsn as mouth only serves to underline the forces which
prevent such a resolution of hostilities. Not only is Karna fiercely loyal
to Duryodhana and resentful of Kunt. By this time Draupad and
Karn a hate each other, for he has insulted her in the sabh. But most
important, their incompatibility is dictated in all these passages by a
cosmological impossibility: a union between the Sun and the Earth.
Karn as command to strip Draupad bare is an image of the pralaya,
the dissolution of the worlds. Were he and Duhsana to have their
way, Draupad would be denuded like the earth prior to its combus-
tion, dessicated by the pralayas seven suns (sprung from the single
Srya) and bare like the back of a tortoise.27
In offering this interpretation, I must abandon my earlier skeptical
tone toward Madeleine Biardeaus remark that Draupads refusal of
Karna-Srya represents a refusal of the solar hero who symbolizes the
conflagration of the world.28 But precision is required. The disrobing
scene offers a confirmation that, as I have argued elsewhere,29 the epic

This is in contrast with a popular tradition that they are attracted, as is told in a
well known Tamil short story: T. M. C. Ragunathan, Venriyan Enraptum, in Ragu-
nathan Kathaigal (1953). I wish to thank K. Kailasapathy for this reference. See also
Hiltebeitel [1976] 1990, 226 n. 63.
1827*, apud 1.17817; for Southern variants, see ibid., 1830*1834*.
See Hiltebeitel [1976] 1990, 226.
This image appears at Visn u Purn a 6.3.23 and Krma Purn a 45.23; see
Madeleine Biasdeau, tudes de Mythologie Hindoue, Part 3, Bulletin de lcole Fran-
aise dExtrme Orient 58 (1971), p. 69.
See Biardeau, Brhmanes et potiers, p. 42, n. 2, and Hiltebeitel [1976] 1990,
311, n. 38.
The argument here extends others already made in Hiltebeitel [1976] 1990, 309
12 (no model in the pralaya myth for the epics dualistic battle; the allegoric character
of pralaya symbolism in the epic), 4849, 109113 and 278 (Biardeau entirely dis-
poses with the epic and Purnic tradition that the Mahbhrata occurs at the juncture
between the Dvpara and Kali Yugas; she thus misses the significance of that transi-
tion as constituting Indias heroic age. For her, the pralaya cataclysm that she sees in
the Mahbhrata myth implies a full destruction and renovation: thus the end of Kali
draupads garments 43

story has different contours from the pralaya myth. At the time of
the pralaya, the Earth is passive to its dissolution. On the contrary,
the primary epic heroine to represent the Earth is not passive; and
precisely because of her non-passivity, she is able to rescue herself.
If anything, Draupads actions are more reminiscent of the earths
role in myths where, as the goddess Earth, she pleads with the gods
(by epic times, usually headed by Visnu) to prevent her from sinking
(usually from overpopulation) into the great ocean.30 In these myths,
the Earth is active, though fully reliant on the male deities for her
rescue. Yet even this theme does not provide a satisfactory model for
Draupads role at the disrobing. Rather than plead to the gods, headed
by Visn u, to rescue her as the Earth does in such myths, Draupad
saves herselfand saves the Pndavas, sons of gods, to boot! In fact,
the disrobing scene explicitly inverts the symbolism of the myth of the
uplifting of the earth. Thus Karna, after the disrobing has failed, taunts
the Pn davas for being saved by a woman:
Draupad, Krsn , has become the salvation (nti) of the sons of Pndu;
when they were sinking, drowning, in the boatless and fathomless ocean,
Pcl, going to the other shore, became the Pndavas boat (2.64.23).
Similarly inverted is a later passage which recapitulates the scene:
Krsn lifted up (ujjahra) the Pn davas, as also herself, as with a
ship from a swell of the ocean (5.29.35). Such passages, particularly

yuga and the beginning of a Krta yuga, an Age of Perfection: see Biardeau, Bhakti et
avatra, pp. 135161, 170174, 203207).
See F. B. J. Kuiper, Cosmogony and Conception: A Query, History of Religions 10
(1970), pp. 101102; Wendy Doniger OFlaherty, The Origins of Evil in Hindu Mythology
(Berkeley: University of California Press, 1976), pp. 28, 3637, 25860; Madeleine
Biardeau, tudes de mythologie hindoue, Part 2, BEFEO 55 (1969), pp. 6068;
Michel Defourny, Note sur le symbolisme de la come dans le Mahbhrata et la
mythologie brhman ique classique, Indo-Iranian Journal 18 (1976), pp. 1922. These
myths as a type are pre-epic, and have been adjusted to many contexts: the primal
creation by a cosmogonic boar (see Kuiper); creation between kalpas by the same
(see Biardeau); a myth of overpopulation due to suspension of activities by Yama in
the Krta yuga (Mbh. 3, Appendix I, No, 16, 11. 58124, after 3.142); the Boar avatra
form taken by Visnu to rescue the Earth from the demon Hiran yka (see Biardeau,
tudes, Part 2, pp. 164165, n. 3, citing Hiran ykas opposition as the chief distinc-
tion between the cosmogonic and avatric boars), plus other instances connected with
avatras, including the Mahbhratas myth of the sinking earth which accounts for
the origins of the epic heroes and heroines, born to relieve the Earth by ousting the
demons who have infiltrated the lineages of the Earths kings (1.58; 11.8.2026). See also
Paul Hacker, Zur Entwicklung der Avatralehre, Wiener Zeitschrift fr die Kunde Sd-
und Ostasiens, Vienna (1960), pp. 4760, who sees the burden of the Earth myth as
the kernel of the avatra mythology.
44 chapter two

the taunt from Karna, imply that the Pn davas are inadvertent ktas,
carried to the other shore and lifted up from the ocean (the verb
ud-hr is especially evocative of the Boar avatra lifting up the Earth31)
by a woman. It is only the passages interpolated into the disrobing
scene that sought to impose the pattern of the usual avatra scheme
by having Draupad pray to Krsn a for her rescue, and, in one instance,
complain about sinking in the Kaurava ocean (Mbh. 2, 543*, 11. 57,
cited above).
The epic story thus follows contours which are distinct from pralaya
themes, and explicitly opposite to themes found in classical myths
about the uplifting of the earth. One final point confirms the distinc-
tiveness of the epic tale. As already noted, at the pralaya the Sun, or
Suns, does bare the Earth; Draupads robes, however, are restored.
The poets have shown their hand in establishing this distinction.
Karna orders Duhsana to disrobe Draupad. He does not touch her
or her garments himself. Between the yugas, where the Mahbhrata
story is set, the miracle of the sarees thus symbolizes that the Earth
holds the power of her own renewal. And Karn a, the figure of the
Sun and potential agent of destruction, remains explicitly distanced
from Draupad. In simplest terms, the restoration of Draupads sarees
shows that, except at the time of the pralaya, the Earths power to
restore her garments is inexhaustible.
On this point, there is a strikingly similar tradition from the Ancient
Near East. In their descents to the nether world, the Sumerian goddess
Inanna and the Akkadian Ishtar must pass through seven gates, at each
one losing garments and jewels, until finally they stand naked before
their sister Ereshkigal, goddess of the Underworld. Inanna and Ishtar
undergo a death and resurrection before they are able to return, their
garments and jewels restored.32 In the Akkadian account, the absence
and disrobing of the goddess are clearly connected with infertility on
Since Ishtar has gone down to the Land of No Return, the bull springs not
upon the cow, the ass impregnates not the jenny. In the street the man

Michel Defourny, Le mythe de Yayti dans Ia littrature pique et purnique
(doctoral thesis, University of Lige, 1973), pp. 165171, 194195.
S. N. Kramer, trans., Inannas Descent to the Nether World, in James B. Pritchard
(ed.) Ancient Near Eastern Texts Relating to the Old Testament (2nd. ed.,; Princeton:
Princeton University Press, 1955), pp. 5257; E. A. Speiser, trans., Descent of Ishtar
to the Nether World, ibid., pp. 106109.
draupads garments 45

impregnates not the maiden. The man lies down in his (own) chamber,
the maiden lies down on her side.33
The loss and return of garments and their connections with the exhaust-
ibility/inexhaustibility of nature are themes found in the disrobing of
Draupad. Draupad has not descended to the nether world; rather, the
nether world in the form of incarnate Rkasas and Asuras has infil-
trated this world in the person of the kings who provoke and watch
her disrobing in the sabh (see above, n. 30). This is but one of many
thematic connections between Indian and Ancient Near Eastern cults
and mythologies of the goddess, a complex matter which has gener-
ated some significant study but has yet to gain the attention and careful
study which it deserves.34

III. Nature: Dharma and the Earth

Having revived a nature mythology interpretation of the Mahbhrata

disrobing scene, some further refinement is in order on Indian concepts
of nature. For one thing, the bounty and devastation of the Earth are
forever bound up with the principle of dharma. When a king rules by
the dharma, the land is fertile.35 In fact, according to the Mahbharata,
a just king need not even rule for the land to be fertile where he lives.
Thus Bhsma says that wherever the Dharmarja Yudhisth ira iseven
in disguise!the land is bound to be rich and bountiful and the cattle
thriving (4.27.1227; see also 29.910). Indeed, such is the prestige
of Matsya or Virtadea, the kingdom where Yudhisth ira lives while
in disguise, that it has been located by local legend at Dharapuram
in the Coimbatore district of Tamilnadu, at Sohagpur about a hun-
dred and forty miles east of Jabalpur in Madhya Pradesh, at Hangal in

Speiser, Descent of Ishtar, p. 55.
Among many authors who have assumed continuities between Indian Goddess
worship and Ancient Near Eastern traditions, the following have contributed important
insights: S. K. Dikshit, The Mother Goddess (A Study Regarding the Origin of Hinduism)
(Poona: International Book Service, n.d.); E. O. James, The Cult of the Mother Goddess:
An Archaeological and Documentary Study (London: Thames and Hudson, 1959);
Narendra Nath Bhatt acharyya, History of the kta Religion (Delhi: Munshiram Mano-
harlal, 1973); idem, The Indian Mother Goddess (2nd ed., Columbia, Mo.: South Asia
Books, 1977); F. J. Richards, Some Dravidian Affinities and their Sequel, Quarterly
Journal of the Mythical Society 7 (1917), 243284.
See Jan Gonda, Ancient Indian Kingship from the Religious Point of View (Leiden:
E. J. Brill, 1969), pp. 1120, 69.
46 chapter two

the Dharwar district of Karnataka, and at Wai below Mahabaleshwar

in Maharashtra,36 even though the Matsyadea referred to in the epic
was probably in the Punjab! Another look is thus in order at the idea,
found in the kernel interpolation of the Northern recension, that
Draupads garments are provided by Dharma.
Usually where the Mahbhrata speaks of garments, it speaks of
their color. The reconstituted disrobing passage makes no explicit ref-
erence to the color of Draupads garments. It even says that the gar-
ments which replaced the original had the originals form (tadrpam;
2.61.4 1), a term which could also imply identical color. The two-verse
kernel interpolation makes up for this uncertainty by saying in the
first verse that the sarees were various (vividha; 544*, 1.4), and then
in the second that Dharma protected her with garments of many
colors and whites (nnrgavirgn i vasanni; 553*, apud 2.61.41).37
Dharma thus clothes Draupad in many colors.
Such a thematic convergence of garments, colors, dharma, and the Earth
can best be explored in relation to other passages in the Mahbhrata.
During their period of disguise in Virtas kingdom, Arjuna is called upon
to defend Virtas cattle against a raid by Duryodhana and his allies. As
Arjuna sets out, Virtas daughter Uttar and her virgin (kany) friends
make a request:
O Brhannad [Arjuna], bring us bright garments for our dolls, fine,
colorful, and various (brhannade nayeth vsm si rucirn i nah /
pclikrtham sksmn i citrn i vividhni ca; 4.35.22cd23ab).
Arjuna then replies that he will bring back heavenly bright garments
(vsm si divyni rucirn i; 25, with variants). He soon overwhelms each
of his opponents, then stupifies them momentarily with the sound
of his conch, and sends Uttars brother to fetch their robes. When
Arjuna returns to Virtas court, he hands these garments to Uttar:
the very white (suukla) robe of Krpa, the resplendent yellow (ptam

On Dharapuram, see P. K. Nambiar and K. C. Narayana Karup, (eds.), Census
of India, 1961, Vol. IX, Madras, Part XI-D, Temples of Madras State, iii, Coimbatore
and Salem (1968), pp. 108109; on Sohagpur, see V. R. Ragam, Pilgrims Travel Guide,
Part II, North India with Himalayan Regions (Guntur: Sri Sita Rama Nama Sankirtana
Sangam, 1963), p. 262; information about Hangal and Wai obtained orally. Another
Karnataka site for Matsyadea is at Madugula: see Sir Walter Elliot, Local History, Vol. II
(manuscript: India Office Library, London), pp. 7274.
Following Nlakant ha, who takes virga, in the sense of without color, as pure
white (kevalaveta); Kinjawadekar (ed.), Shrman Mahbhrata (see n. 11), Vol. 1,
Sabhparvan, p. 111.
draupads garments 47

ruciram ) garment of Karna, and the blue (nla) robes of Avatthman

and Duryodhana (4.61.13).
As Heino Gehrts has seen, there is a continuity between the garments
removed from Draupad and the garments given to Uttar. Gehrts
interprets the garments in both scenes as symbolic of the womb and of
embryonic rebirth, corresponding to the use of garments to represent the
amnion and chorion in the Rjasya sacrifice.38 Draupads removed
garments are thus for Gehrts the precondition for the Pndavas
rebirth. The bestowal of the garments on Uttar represents that their
rebirth will be through her, rather than through Draupad. And indeed
it will, for Uttar becomes the wife of Arjunas son Abhimanyu, and
thereby mother of Pariksit, the sole survivor of the Pndava line.39
Actually, there may be further evidence of a connection between the
two episodes. Not only are the garments which the girls obtain vari-
ous, colorful, and even heavenly like those of Draupad. The girls
plea to Arjuna seems to contain an unmistakable allusion to Draupad.
The word they use for dolls is pclik. The literal meaning of the
term is of or relating to the people of Pacla, and is equivalent to
the meaning of Pcl as a common name for Draupad. Presumably
the people of Pacla were as well known for their dolls or puppets
as they were for their epic heroine. Actually, the word pcl itself
can also mean doll or puppet.40 The pointedness of the use of such a
term in the girls request becomes clear in relation to another passage
that connects Draupad with puppets. When she complains that suf-
fering and pleasure result from all creatures being manipulated by the
Placer (Dhtr) like puppets (3.31.2129), she uses an entirely different
termyoswith an identical meaning. It is thus possible to read the
girls plea as follows: O Brhannad, bring us bright garments for the
sake of what comes from Pacla (pclikrtham), fine, colorful, and
various. The term pclikrtham is open to several possible meanings,
but the basic ambiguity of reference to both dolls and Draupad
seems undeniable. For the sake of what comes from Pacla has
obvious reference to Draupads garments. The fact that the maiden

See Heino Gehrts, Mahbhrata: Das Geschehen und seine Bedeutung (Bonn:
Bouvier Verlag Herbert Grundmann, 1975), pp. 206207, 224225; J. C. Heesterman,
The Ancient Indian Royal Consecration (s-Gravenhage: Mouton & Co., 1957), pp. 92,
9798; I have discussed Gehrtss theories more fully in a review, in Erasmus 29, No. 34
(1977), columns 8691.
See Hiltebeitel [1976] 1990, 33653.
Vaman Shivaram Apte, The Practical Sanskrit-English Dictionary (Poona: Prasad
Prakashan, 1958), s.v. Pcl.
48 chapter two

Uttar asks for these garments of regeneration in the disguised

Draupads hearing is all the more poignant. Draupads disguise as
chambermaid (sairandhr) involves her wearing a single large black
am sumalinam mahat; 4.8.2),
very dirty garment (vsas . . . ekam krsn
no doubt an image of impurity and death. And at the end of the war,
Avatthmans curse will prevent her from having more children after
her own have been slain. But Uttar, who receives these variously
colored heavenly garments and then immediately thereafter marries
Arjunas son Abhimanyu, will, after the war and her husbands death,
give birth to Pariksit and thus to the regeneration of the Pndava line.
For her, the association of the garments with dolls no doubt antici-
pates the transformation from girl to woman and the forthcoming
processes of giving birth and becoming a mother.
The symbolism of regeneration, fertility, and auspiciousness associ-
ated with the garments at the miracle of the sarees is thus, as Gehrts
has seen, given a specific gynecological meaning in the gift to Uttar.
But Arjunas victory over the Kurus in the latter episode also involves
a symbolic foreshadowing of related themes in the war. One of the
most frequent similes used to describe the battlefield at Kuruksetra
is the comparison of the Earth to a woman. Among such passages, a
good number refer to the Earths covering of bodies, limbs, garments,
ornaments, and battle gear as being like a womans adornments. To
the vanquished the Earth appears stark. At the end of the fighting,
when Duryodhana makes his final retreat from the field, he sees the
Earth empty (drst v nym ca medinm; 9.28.16). Thus he says to
Yudhisth ira before his final duel with Bhma:
The Earth shorn of jewels with her warrior bulls slain, I do not wish to enjoy,
like a widowed woman (ksnaratnm ca prthivm hataksatriyapum gavm /
nbhyutsahmy aham bhoktum vidhavm iva yositam; 9.28.16).
As Draupad was shorn and symbolically widowed to the Pndavas, so
the Earth is shorn of jewels and a widow to the defeated Duryodhana.
But in the midst of battle the Earth has a different look:
The Earth shone forth like a wanton young woman (pramad) adorned
with diverse kinds of ornaments (vibabhau mah / nnrpair alam kraih
pramadevbhyalam krt; 6.92.65).
Wet with red blood sprung from the bodies of men, horses, and elephants,
the Earth was like an all-accessible resplendent girl attired in burnished gold,
garlands, and red garments (. . . bhmih / raktmbarasraktapanyayogn /
nr prak iva sarvagamy; 8.68.34).
draupads garments 49

and finally, most interestingly:

Or as a dark young girl would wear garments in white dyed with red saf-
flower, likewise was the Earth (yath v vsas ukle mahrajanarajite /
bbhryd yuvatih ym tadvad sd vasum dhar; 8.36.9).
The last passage is certainly the most suggestive. The Earth, like a dark
young girl clothing herself in white garments dyed red, reminds one of
Draupad in the sabh: the dark (krsn ) woman who clothes herself
with sarees of apparently various colors. As Nlakant ha indicates, the
tradition is that the colors were not only various, but included pure
white (kevalaveta).41 These three colorsblack, red, and white
have a long history in India (and elsewhere) as the three basic colors
of nature or prakti.42 The garments which Arjuna brings back for
Uttar are also white, yellow, and blue, the two latter being frequent
variations of red and black which seem to accentuate positive, aus-
picious, and even divine associations.43 It would thus seem that the
garments also represent the fundamental qualities (gun as) of nature.
Like Draupads garments, the guas are inexhaustible. As the three
constitutent strands of prakti, they form the basic fabric of mate-
rial continuity from creation to dissolution, from one universe to the
next.44 Draupads identification with the Earth is thus only one facet
of her fuller identification with prakti, nature. As the epics primary
embodiment of the Hindu Goddess, she represents all of nature, not
just the Earth. The Goddess represents prakti as composed of five ele-
ments: ether, air, fire, water, and earth. And these, in turn, are reducible
to the three guas.45 Insofar as Draupad is identified with the Earth,

See above, n.37.
On their use in color symbolisms, see Hiltebeitel [1976] 1990, 6076; Victor Turner,
The Forest of Symbols (Ithaca: Cornell University Press, 1967), pp. 5991; Brenda E. F.
Beck, Color and Heat in South Indian Ritual, Man 4 (1969), pp. 553572.
The major Hindu deities concerned with moksa are commonly shown with blue
skin: iva, Visn u and his avatras, and the Goddess.
According to the vetvatara Upanisad, prakrti with its three gun as is one of the
three unborns, along with the Lord (Ivara) and the soul. According to the Sm khya,
it is the red gun a, rajas, which stimulates the re-creation at each kalpa.
In the preceding interpretation I have attempted to delimit the discussion of gar-
ment symbolism to themes which either bear directly on the epic, or provide striking
parallels to pertinent aspects of the epic story. But this discussion of prakrti and the
gun as may remind one of Hellenistic notions that the soul is clothed in garments
of different colors as it descends to earth through the planetary spheres: see Franz
Cumont, Astrology and Religion among the Greeks and Romans (New York: Dover,
1960), pp. 108109. The closest Indian counterpart to this idea is the Jaina concept
of the colors of matter infiltrating the soul, a doctrine once attributed to Bhsma in
50 chapter two

however, one must appreciate the Mahbhratas attentiveness to

the goddess Earths mythology. The Earth is a more dynamic figure
in the Mahbhrata than the myth of her plea to the gods would by
itself reveal. In the passages which speak of her wearing as garments
the blood, limbs, and accoutrements of the warriors, there is an evi-
dent complexity to her character. In one aspect she is a promiscu-
ous girl, alluring, wanton or intoxicated (pramad), accessible to all
(sarvagamy). But in another aspect she is a goddess of destruction,
taking the warriors to her only in their death, and repeatedly glutted
in scenes of carnage with a proverbial river of blood that bears the
dead to the realm of Yama.46 In these images, one finds a fulfillment
of Arjunas prophesy that the earth will drink the blood (bhmih
psyati on itam; 2.68.31) of those who taunted Draupad in the sabh.
All this has much to do with Draupad, who leaves the sabh for exile
in a garment wet and smeared with blood (onitktrdravasan; 2.7
1.18) vowing to see the blood of her tormentors staining the garments
of their wives (1920).
Through all this emphasis on blood and gore, however, one should
not lose sight of the fact that the Mahbhrata war is a war fought
for the establishment of the dharma (dharmasam sthpana; Bhagavad
Gt 4.8), and that Kuruksetra is a dharmaksetra, a field of dharma
(ibid., 1.1). Twice during the war (three times in the Northern recen-
sion), Krsn a shows the battlefieldor more literally battle-earth
(ran abhmi, yuddhabhmi; 7.123,41; 8.41.59)to Arjuna, and tells
him at great length that the Earth which looks so ghastly is mysteriously
beautiful and resplendent (7.123.3041; 8,14.2659; 8, Appendix I,
No. 16). Krsn as revelations to the epics greatest warrior concern the

the Mahbhrata: see V. M. Bedekar, The Doctrine of the Colours of Souls in the
Mahbhrata: its Characteristics and Implications, Annals of the Bhandarkar Oriental
Research Institute 4849 (196869), pp. 329338. The earlier comparison of Draupads
disrobing with the descents of Inanna and Ishtar may also remind one of Salomes
dance of the seven veils. These two comparative suggestions were offered by W. Ran-
dolph Kloetzli. The ritual symbolism of garments and nakedness is also a comparative
theme that has been barely covered (see n. 3).
See 6.55.1112 and 121125, 85.25, 99.3338; 7.13.818, 20.3137, 68.4748;
8.33.8166, 36.2932, 55.3842, 57.2, 58.78; 9.8.2933; the river is compared to the
underworld Vaitarn (8.55.42, 58.7; 6.55.1 25, 90.38), but, insofar as it leads the dead
to Yamas realm or toward the realm of the Pitrs (8.36.31; 6.13.17; 9.8.33), it would
appear to be an earthly extension of the Vaitarn. The river flows toward the ocean
of the other world (paralokrn avamukh; 6.55.12). On the Vaitarn , see E. Washburn
Hopkins, Epic Mythology (1915; repr. New York: Biblio and Tannen, 1969), pp. 5,
draupads garments 51

divine nature of this battle-earth. Kuruksetra is a scene of a great

sacrifice of battle in which all the divine and demonic forces are
present to reestablish the sinking earth and to restore the dharma.
Concerning the dramatic interplay between dharma and the Earth, the
battle develops according to a sequence which I analyzed elsewhere.47
Yudhisth ira, the incarnation of Dharma, wins the war and reclaims
the Earth through a series of flawed triumphs over the marshals
(senpatis) of the Kaurava army: Bhsma, Drona, Karna, and alya.
These four are associated with a tetrad of colors: Bhsma with white,
Drona with red, Karn a with yellow, and alya with black. In Bhsma
and Karn as cases, this association is made specifically in terms of their
garments.48 The goddess Earth also figures prominently in the scenes
of the four marshals deaths, and actually becomes active in the final
two instances. Thus she swallows Karn as chariot wheel (8.66.59), and,
at alyas death, it was as if that bull among men was risen up to by
the Earth, out of love (9.16.53).49
This sequence would seem to mark the culmination of the epics
use of garments as symbols of regeneration and rebirth. The deaths
of the four marshals involve a behind-the-scenes intrigue between the
Earth, Dharma (in the person of Yudhisth ira), and Visnu (in the per-
son of Krsn a). It may also be noted that in the two deaths where the
Earth is not active, it is Draupads brothers who become the active
instruments whereby the marshals are slain: Bhsma by ikhandin,
Drona by Dhrst a dyumna.50 This configuration is not unlike that in the
sabh. As Draupad was reclothed by Dharma, and, as the tradition
developed, by Krsn a, so at Kuruksetra the Earth reclothes herself with
the garments of the slain, and, in particular, through the agency of
Yudhisth ira and Krsn a, takes back to herself the heroes who embody
the colors of continuity51 and regeneration. The miraculous sarees that
reclothe Draupad in the sabh and the dolls of Uttar are thus also
the garments of the Earth.

Hiltebeitel [1976] 1990, 28386.
On Bhsma, see 5.179.1014; 6.20.9; on Karna, see his yellow garments cited above.
See Hiltebeitel [1976] 1990, 270; a regrettable misprint there has . . .up to be the
earth; on the role of the Earth in other episodes and in the sequence as a whole, see ibid.,
pp. 251, 262, 279, and 284.
On ikhandins own connection with the Goddess, see Biardeau, Bhakti et
avatra, p. 210 n. 1 and 220221; Jacques Scheuer, iva dans le Mahbhrata: lhistoire
dAmb / ikhandin, Purusrtha, Vol. I, Part 2 (1975), pp. 6786.
The alteration of colors also pertains to the changes in the yugas and to the appear-
ances of Krsn a; see Hiltebeitel [1976] 1990, 6263, 283.



After over forty years of increasingly productive exploration of the

symbolism of the Mahbhrata,1 it is becoming clear that the Great
Epic possesses a remarkable coherence. This coherence results from
its bards and poets weaving a fabric of largely consistent symbols
and themes to convey their fundamental concerns through a sus-
tained narrative medium. Yet such a claim requires continuing sub-
stantiation, as the tides of epic research are not without their shifts.
This study will thus seek to lend further support to the argument
for coherence and will require repeating and rephrasing some of the
assumptions that have sustained such an approach methodologi-
cally, particularly as they have been presented in a series of essays of
which this one forms part of a sequence.2 But one premise should
be declared at the outset. It is here contended that the poets go to
the deepest level of their play with symbols in the epics fourth
book, the Virta parvan, which describes the period that the Pn davas
and Draupad spend unrecognized in Upaplavya, the city of King
Virta. The disguises which they adopt show the epic poets as true
symbol-masters, concealing and revealing the deepest identities
of their heroes and much of the purposeprimarily theological
of the roles they play in the epic narrative as a whole.
A brief review of the scholarship on the Virtaparvan will show that
the disguises have already been recognized as presenting a rich con-
centration of symbolic themes. Stig Wikander and Georges Dumzil
opened the interpretative discussion by observing that the disguises
of the five brothers conserve Indo-European tri-functional and
Indian caste-related traits.3 J.A.B. van Buitenen, moving beyond the
identification of isomorphic structures (relating to trifunctionalism

See Hiltebeitel 1979a, 83107, on this focus of interpretation since the 1930s.
See Hiltebeitel 1980a, 1981, 1982.
Stig Wikander, La Lgende des Pnda va et la substructure mythique du Mahbhrata,
trans. Georges Dumzil, in Jupiter Mars Quirinus (Paris: Presses Universitaires de France,
1948), 4:4849; Georges Dumzil, Mythe et pope: Lidologie des trois fonctions dans
54 chapter three

and caste), opened fresh insights by stressing narrative continuity

and context. He demonstrates the inadequacy of arguments that the
Virt aparvan is late or interpolated, and while not rejecting that
possibility for himself (as one might wish),4 he views the disguises as
burlesque inversions of the heroes roles and status in the epic as a
whole, a sort of bards version of Hol.5
Each of these approaches has provided suggestive insights and
grounds for extension.6 But each affords only an incomplete tableau.
Dumzil and Wikander stress connections only with the upper three
varn as. As we shall see, this misses much of what is essential. And
the Indo-European mythological prototypes include no model for the
disguise of Draupad, a lack which, in my opinion, is not corrected by
Daniel Dubuisson.7 And as van Buitenen himself recognizes, there are
no burlesque inversions in the pedestrian disguises of the twins.8
By way of working toward a comprehensive interpretation of the
disguises as a group, and of understanding the indissoluble connec-
tions between the Virtaparvan and the larger epic, the work of Mad-
eleine Biardeau proves indispensible. Moreover, it is but a short step

les popes des peuples indo-europens, 2d ed. (Paris: Gallimard, 1974), pp. 6263,
7173, 79.
See J. A. B. van Buitenen, trans. and ed., The Mahbhrata, Book 4, The Book of
Virta, Book 5, The Book of the Effort (Chicago: University of Chicago Press, 1978),
3:1821, discounting the arguments of E. W. Hopkins, The Great Epic of India (Calcutta:
Punthi Pusak, 1969), pp. 38283. Dumzil also lends weight to the interpolation theory,
but not convincingly; see Mythe et pope, pp. 89, n. 2, and 93. See also Dieter Schlingloff,
The Oldest Extant Parvan List of the Mahbhrata, Journal of the American Oriental
Society 89 (1969): 33438. This Kun a-period parvan list is too defective to yield solid
conclusions; also, if a parvan beginning with a or occurs where one would expect
Virt a, this could be ajta, incognito. In any case, the evidence of continuities
goes far beyond that cited by van Buitenen and makes the interpolation arguments
Van Buitenen 1978, 310, 2021. The amusing tone of the Virtaparvan is certain,
despite J. W. de Jongs review of van Buitenens Mahbhrata, vol. 3, in Indo-Iranian Jour-
nal 22 (1980): 5862. But a connection specifically with Hol, or with Phlguna month
(FebruaryMarch), breaks down. Arjunas disclosure at the years end occurs during
grsma, the hot season (MayJuly) (Mbh. 4.42.22; henceforth Mahbhrata citations,
all from the Poona Critical Edition, will not mention the text).
Thus the Indo-European perspective opens the possibility of comparing distinctly
epic themes: for example, one suspects a connection between the Virt aparvan account
and the disguises of Prince Goshtasp in Rum recounted in Iranian epic traditions. Van
Buitenens notion of inversions has also been extended; see Hiltebeitel 1981.
Daniel Dubuisson, La Desse chevelue et la reine coiffeuse: Recherches sur un thme
pique de lInde ancienne, Journal asiatique 166 (1978): 2913 10; for discussion, see
Hiltebeitel 1981, n. 5.
van Buitenen 1978, 7. See further now Hiltebeitel 2001a, 21640.
iva, the goddess 55

from her multifaceted analysis to the solutions proposed here. Since I

will make frequent reference to her studies, it will suffice for now to
indicate only the main threads.
First, the thirteenth year, which the Pn davas spend like creatures
dwelling in the womb [garbhavsa iva prajh] (4.66.10; cf. 246* apud
4.12.11),9 has the character of a dks, the consecration through
which one is reborn in the womb of the dks hut to the status of
sacrificer (yajamna).10 She further regards the dks theme as dou-
bled in the person of Arjuna, whose wrestling match with iva, when
the latter is disguised as a hunter (kirta), involves a transformation of
Arjunas offered body into a divine body, permitting him to ascend
to heaven and attain divine weapons to be used in the sacrifice of
battle.11 This dks-like latency during the period of disguise is further
reinforced by the names associated with their hiding place. Upaplavya,
the City to Be Overflowed, evokes the deluge.12 The name Virta calls
to mind the Vedic quasi-feminine cosmogonic principle virj.13 And
the name Matsya, fish, for the kingdom has both cosmogonic over-
tones (e.g., the Matsyvatra) and hints of chaos (the matsyanyya:
the big fish eat the little fish), and also associations with a recurrent
fish imagery that runs through the epic.14

On Mahbhrata references, see n. 5 above.
See Madeleine Biardeau, tudes de mythologie hindoue [henceforth referred to as
EMH] (IV), Part II. Bhakti et avatra, Bulletin de lcole franaise dExtrme Orient 63
(1976): 2078, EMH (V), Part II. Bhakti et avatra, Bulletin de lcole franaise dExtrme
Orient 65 (1978): 14957, 18788 (n. 3), Compte-rendu, Annuaire. cole pratique des
hautes tudes [henceforth referred to as EPHE], 5th sec., Sciences religieuses 82 (1973
74): 94; Madeleine Biardeau and Charles Malamoud, Le sacrifice dans lInde ancienne
(Paris: Presses Universitaires de France, 1976), p. 133.
Biardeau, EMH (V), pp. 14959, cf. EMH (IV), pp. 227, 241, 245.
Oral communication from Biardeau; see also Jacques Scheuer, iva dans le
Mahbhrata (doctoral diss., Universit de Paris, 3d cycle, 1975), p. 333. I received this
excellent workfor which grateful thanks to the authoronly after completing two
drafts of this article. I therefore cite it almost exclusively in notes. Many similar points
were reached independently, and our arguments often reinforce each other. However,
Scheuer largely ignores the Virta parvan; the main thrust of this article thus must
stand on its own.
Biardeau, EMH (IV), p. 208, n. 1, EPHE 82 (197374): 90.
See Biardeau, EMH (IV), Pp. 166, EPHE 85 (197677): 165, EMH (V), PP. 9299
(Yudhisth ira as Kanka, Heron, eater of fish), EMH (IV), pp. 21819, n. 1, EPHE 79
(197071), P. 143 (Satyavat, Uttar). Vernacular traditions also have Arjuna win
Draupad by shooting a fish.
56 chapter three

As regards the disguises themselves, Biardeau notes the element of

play (krda, ll)15 in the sports of the three eldest brothers, though
like van Buitenenshe sees nothing surprising in the disguises of the
twins.16 She has also caught sight of socio-ritual overtones of impurity
and untouchability in the disguises, particularly those of Draupad
and Yudhist hira,17 thus hinting at associations extending beyond the
upper three varnas. But her first concern is with Arjuna. In Arjunas
disguise as a eunuch, she sees the theme of sexual abstinence, brahma-
carya, in a form of renunciation suitable to the king and essential to
the exercise of royalty, though it eventually involves the renunciation of
royalty itself when Arjuna refuses to marry his pupil, Virtas daughter
Uttar.18 Insofar as Biardeau regards Arjuna as the ideal king, in
close rapport with the avatra, Krsn a, she also stresses the manner in
which Arjuna impersonates aspects of the avatra ideal.19 This ideal she
presents most succinctly as follows: the avatra displays the interplay
and integration of three paired formulations: destruction-cration,
rudraque-visn uique, ksatriya-brhmane.20 The point of departure for
this essay is the suspicion that in over-stressing Arjunas rapports with
kingship,21 Biardeau has so far underestimated his rapport with Rudra-
iva component of her most important avatra formulation.

Biardeau, EPHE 82 (197374): 91. See also her important discussion of Krsn as
cowherd disguise in EMH (V), pp. 20437, which is summarized with minor extensions
in Hiltebeitel 1989d.
Biardeau, EPHE 82 (197374): 92.
Biardeau, EMH (IV), p. 207, n. 1, EMH (V), p. 187, n. 2 (Draupad): on
Yudhist hira, see above, n. 14. See also her discussions of the corpse-bearing am
tree where the Pn davas conceal their weapons on entering Matsyadea: EPHE 82
(197374): 93, Mythe pique et hindouisme daujourdhui, Indologica Taurinensia 5
(1977): 4353; she also notes the possibility that Arjunas contact with iva the Hunter
involves impurity: EMH (V), p. 151.
Biardeau, EMH (V), pp. 18992; she seems to imply that Arjuna again doubles
for the Pn davas as a group, whose name she derives from pan da/pan dra, eunuch:
EMH (IV), p. 262.
Biardeau, EMH (V), p. 177 (and passim): lpope est la geste dArjuna et non celle
de Krsn a.
Biardeau, EMH (IV), pp. 18284.
Not to say he is not an ideal king, but Biardeaus discussion often narrows the
royal role of Yudhisth ira; see Hiltebeitel [1976] 1990, 192296.
iva, the goddess 57

Draupad and Arjuna

Certain methodological assumptions that will guide the remainder of

this essay are best indicated at this point. First, I regard the Mahbhrata
as a text which attempts a great synthesis of Indian civilization in the
name of Hinduism. By synthesis, however, something different is
meant from the confluence of Epos and Rechtsbuch stressed in
the last century by Joseph Dahlmann. The recent work leading to an
understanding of this synthesis has been carried out by scholars who
have stressed the transpositions or connections worked out by the
epic poets in relating the story to para-Vedic (in some cases Indo-
European), Vedic, Brhmanical, and Upanisadic symbols, myths, and
rituals, and also to the mythic material fully developed for the first
time in the background myths told in the course of the narrative
itself. But it becomes increasingly clear that a full understanding of
this synthesisand thus of the place of the Mahbhrata in the his-
tory of Hinduismrequires a recognition that the epic also evokes,
through its symbolism, certain cultural themes, myths, ritual practices,
and social norms that are not fully attested historically until post-epic
times, sometimes in later texts, sometimes even in contemporary folk
cults and practices. For example, there can be little doubt that the epic
poets know of a disheveled goddess of destruction akin to and prob-
ably identical with Kl. We know this, however, not because earlier
texts or the epic itself tell us myths about Kl, or even give us direct
allusions to her, but because the epic alludes to such themes through its
depiction of Draupad.22 Similarly, the Rmyana incorporates into its
main narrative the scenario of a buffalo sacrifice, alluding to complex
ritual details that are intelligible, not through myths about the goddess
Durgs conquest of the buffalo demon Mahisa (myths probably slightly
later textually than the Rmyana), but through contemporary
accounts of village buffalo sacrifices.23 This essay will observe another
instance of an epic theme that can only be understood from such
later sources. Nor are these isolated cases in either epic. One must
thus interpret the Mahbhrata not only retrospectively but, in a sense,
prospectively. Possibly the epic simply anticipates later forms. More
likely, however, it evokes forms which we know of only (or largely)

This argument is developed in Hiltebeitel 2001.
See Hiltebeitel 1980c, 187211.
58 chapter three

from later sources, but which are earlier than is commonly thought.
One would thus need to recognize the pliancy and selectivity of an
oral tradition in its symbolic articulation of some of the fundamental
continuities of Hindu culture, for which the epic is not only the first
great effort at synthesis but a means to transmit this synthesis through
the centuries, in India and abroad.
It is thus impossible to study the epic as a story frozen in its San-
skrit textual forms. For one thing, there are good grounds to suspect
that certain features of the story descend from an Indo-Iranian and
Indo-European past. But more than this, one must assume that the
epic poets made selective use of oral traditions and popular cultural
themes. Preposterous as it sounds, considering the immensity of the
text, one can pretty safely assume that the bards knew more about the
main story, both in terms of variants and underlying symbolism, than
they told. It is thus worth investigating whether what they left untold
but implicit, or what they alluded to through symbols, is not still
echoed in the vast oral and vernacular epic and epic-related traditions
that perpetuate the story to Indian culture to this day. I have come to
suspect that living traditions of and about the Mahbhrata are often
in close touch with traditional epic meanings that have escaped the
classically based literary scholars.
Nowhere is this point more vital than in the matter at hand: the
Pndavas and Draupads disguises. For one thing, the period in
disguise is immensely popular throughout India, as is evidenced by
the fact that Virta nagarthe city of their concealmentis locally
identified in numerous and far-flung places.24 More than this, as will
be seen, the theme of past disguise allows for a kind of continued
symbolic presence. Specifically, in the south Indian fire-walking cult
of Draupad (Tiraupatiyamman), which uses Tamil versions of the
epicprincipally Villiputtr l vrs Villi Pratam (ca. A.D. 1400)as
a cult myth and enacts the epic both as ritual and as night-long street
drama (terukkttu) with professional itinerant actors, the recital and
terukkttu enactment of the period in disguise are popular and imagi-
natively carried out, and mark the important transition to themes of
war and revenge for the festival as a whole. One point must now be
made about the Draupad cult, which I regard as presenting a perceptive

See Hiltebeitel 1980a, 105 and n. 36 (see n. 2 above). The original Virt anagar
is thought to have been at Raih, forty miles west of Jaipur.
iva, the goddess 59

and coherent interpretation of the epic. It is of course a kta cult, with

Draupad as a form of the Goddess. It was soon evident during my
study of the cult25 that much of its symbolism holds reference to akti
and iva. But what suddenly struck me one day, after weeks of noting
various pairings of icons in different ritual contexts, is that Draupad is
usually paired with one of her three eldest husbands and that, whereas
she is the Goddess, Arjuna is inescapably the foremost representative
of iva. This affinity of Arjuna and Draupad with iva and the God-
dess is an indispensible key to understanding the Draupad cult. Let it
also invite us to a fresh look at their portrayal in the Sanskrit epic.
Our first turn is, of course, to the disguises. As suggested earlier,
it is in their disguises that the Pn davas and Draupad reveal their
deepest symbolism. As regards Draupad, I have argued elsewhere
that her Sairandhr (chambermaid/hairdresser) disguise is treated
similarly both in her cult and in the epic. It involves references to
dra and outcaste roles, associations with extreme impurity, and evo-
cations of the Goddess in her destructive forms: Mrtyu, Kl, Klartri,
Durg.26 This last point is most instructive, because the epic identifies
Draupad as the incarnation, not of one of these destructive forms, but
of the auspicious goddess r-Laksm.27 It may thus be urged that her
disguise reveals her to be an embodiment, not only of r-Laksm, but
of the Hindu Goddess in her totality.28
If such is true of Draupad, one must look more closely at Arjuna.
In the epic, although r is eternally the wife of Visn u, she is peren-
nially the wife of victorious kings, foremost of whom is Indra.29 Does
Arjuna, Indras son, Draupads self-chosen (1.179) and favorite
(17.2.6) husband, reveal in his disguise deeper dimensions as a repre-
sentative of iva, just as Draupad does of the Goddess? The terukkttu
drama leaves no doubt. There the disguised Arjuna is an androgyne
and a clear evocation of iva as Ardhanrvara, the Lord who is half
woman. The actors left side has a breast, light rose-colored facial col-
oring, long hair, and anklets; his right side shows winglike epaulettes,

Fieldwork was carried out in 1975 under an American Institute of Indian Studies
grant and in 1977 on a National Endowment for the Humanities summer stipend.
See Hiltebeitel 1981.
See Hiltebeitel [1976] 1990, 6268, 8999, 14491, 20324.
See Scheuer 1975, 75.
See Hiltebeitel Hiltebeitel [1976] 1990, 9698, 15690.
60 chapter three

blue facial coloring, and a peacock crown.30 As already indicated, this

is but one of many associations between Arjuna and iva in this cult.
Are there similar correlations in the epic?
Three features of Arjunas disguise are most prominent: his name
Brhannad/Brhannal, his occupation as dancing and music instruc-
tor, and his appearance as a eunuch. With regard to the latter, the
epic descriptions leave it amusingly imprecise and ambiguous whether
Arjuna is physiologically a eunuch, a hermaphrodite, or simply a trans-
vestite. As we shall see, in effect he is described as all three. In any case,
it is not hard to show that nearly everything about his disguise holds
hints of identification with iva. An argument might be raised against
this that the iva thus evoked is post-epic. But this can be countered
on several grounds. First, as already proposed in connection with
Draupad and the Goddess, the poets evoke themes that they may never
describe. But that argument is not as essential here, since the poets are
less reticent on iva than the Goddess, and in most cases it is easy to
demonstrate that a rather full image of the classical iva is detectable in
the epic: first through the narrative itself, and secondarily in the late
epic ivasahasranmastotras, Eulogies of ivas Thousand Names.
Three such stotras are found in the nti and Anusana Parvans.31
Though they cannot, at least in their present form, be accorded the
antiquity of the material in or adjacent to the main narrative, they
provide us with information about iva that has value. As a style of
literature, such devotional lists of names have their Vedic precedent in
the atarudriya, a recitation of 100 (actually more) names of iva. The
late epic Sahasranmastotras may thus be regarded as a resurfacing
of a genre which probably continued to undergo development con-
temporaneously with the main body of the Mahbhrata itself. Thus,
even if they are late epic, they record images of iva that were no
doubt long in developing and provide complementary material to the

Wendy Doniger OFlaherty, discussing Indian Androgynes: Symbols of Inte-
gration and Disintegration at the Ninth Annual Workshop of the Conference on
Religion in South India, showed two paintings in which the male sides skin was blue
and the females either pink or gold. Actually, Arjuna here also evokes Krsn a with his
peacock crown: thus not only iva and Prvat but Visn u-Krsn a.
They occur in the critical edition at 12, app. 1, no. 28; 13, app. 1, nos. 46; and
13.17.30 ff. See Eugen Rose, Die ivasahasranmastotras in der epischen und purn ischen
Literatur: Eine religionsgeschichtltiche und kulturgeschichtliche Untersuchung (Bonn,
iva, the goddess 61

narrative. There is also an occasional affinity in language between nar-

rative descriptions of Arjuna and stotra descriptions of iva.
Let us then take up the three main features of Arjunas disguise. First,
Biardeau has perceived that Arjunas play as musician and dancer
echoes iva Natarja, though she also mentions Krsn a Kliyadamana.32
I would suggest that the identification with the dancing iva is the
primary one. The Mahbhrata provides at least one reference to
this aspect of iva in its main narrative, when it describes Bhma
as dancing like am kara (6.58.56; compare 8.21.15). And in the
Sahasranmastotras such references are numerous. Thus, You are
fond of dancing, always dancing, the dancing master, delighting in the
universe [nrtyapriyo nityanarto nartakah sarvallasah ] (13.17.50); or
again, Homage to the one disposed toward dancing, producing musi-
cal sounds with the instrument of the mouth, . . .conversant with song
and musical instruments [namo nartanalya mukhavditravdine . . .
gitvditraline] (12, app. 1, no. 28, lines 198199).33 Clearly the epic
poets must know iva as master of music and dance. Now compare
how Arjuna describes himself to Virta: I sing, I dance, I also play
musical instruments. I am good at dance and skilled at song [gymi
nrtymyatha vdaymi/ bhadro smi nrtte kualo smi gte] (4.10.8;
similarly, see 4.2.24).
Second, in Arjunas appearance as a eunuch or hermaphrodite (usu-
ally klba; also san dhaka [4.2.21]; trtym prakrtim, third sex [4.59*,
Northern variant]) dressed as a woman, it is hard to imagine that
the epic is not reminding its audience of ivas ambiguous or dual
sexuality. In one of its core narrative portions, it has iva oversee the
birth and career of the woman-man (strpumn [5.189.51) ikhandin-
ikhan din.34 And iva seems to be evoked in the cryptic language
that describes Arjunas condition. When Virta has Arjuna checked
to see if what the latter says about being a eunuch is true, he learns
that Arjunas non-masculinity was firm [apum stvam . . . sthiram]
(4.10.11). In modifying non-masculinity one cannot help but think
that sthirafirm, hard, solid, fixed, calmis an amusing reference
to the lingam and a reminder of iva, for whom Sthira is an occasional

Biardeau, EPHE 82 (197374): 91.
See also 12, app. 1, no. 28, line 299; 13, app. 1. no. 4, lines 5657, 74; app. 5, line 52.
See Biardeau, EMH (IV), pp. 22022; EPHE 84 (197576): 18384, 85, (197677):
13640; Scheuer 1975, 11837.
62 chapter three

epic name.35 Indeed, iva is the eunuch of the firm phallus. Says one
of the Sahasranmastotras: The one whose lingam is ever firm [sthira]
is therefore known as Sthnu [Pillar, Sacrificial Stake] (13.146.10).36
Further similarities in the ambiguity of Arjuna and ivas sexuality are
easily demonstrable. Arjuna appears as a great man [brhatpumn]
wearing the adornments of a woman (4.10.1). And when he appears
in his disguise to help prevent the Kauravas from raiding Virtas
cattle, the Kauravas remark that he has something of a man, some-
thing of a woman (4.36.30). Compare the following homages to iva:
To the one half smeared with sandal, half decked with unguents and
garlands (13, app. 1, no. 6, line 10); Who else has half his body occu-
pied by his wife; by whom is the bodiless one [Kma] conquered?
(13, app. 1, no. 5, line 50); Homage to the one who is female and male,
to the eunuch [strpum sya napum sya namah ] (13, app. 1, no. 6,
line 35); and most interesting: You are the Purusa whose seed is
gold, you are woman, you are man, and you are eunuch [hiran yareth
purusastvam eva tvam str pumm stvam ca napum sakam ca] (12, app. 1,
no. 28, lines 33940).37 When Virta puzzles over Arjunas disguise,
he similarly refers to him as a purusa in womans garb, adorned with
conches, braid, and earrings (4.10.5).
These last two citations, which might remind one that already in the
Svetvatara Upanisad iva is identified as the purusa (3.8.20) who is
male and female (tvam str tvam pumn asi [4.3]), serves to bring us
finally to the third main feature of Arjunas disguise, his name. Just as
iva is the male-female purusa (male) who is a napum saka (not-
male, eunuch), the eunuch Arjuna, as already cited, appears in
Virtas court as a great man [brhatpumn] wearing the adornments
of a woman. As Biardeau has seen, brhatpumn substantiates the
etymology which the commentator Nlakant ha perceived for Arjunas
cryptic name: Brhannal/Brhannad derives from Brhad-nara, great

See Soren Srensen, An Index to the Names in the Mahbhrata (Delhi: Motilal
Banarsidass, 1963), s.v. Sthira. San dha, Eunuch/Hermaphrodite, is also a name
for iva according to Lexicons cited by Monier Monier-Williams, Sanskrit English
Dictionary (Oxford: Clarendon Press, 1964).
The Mahbhrata knows of iva detaching his lingam in rage at Brahms infe-
rior creation (10.17.21). The Pine Forest castration may be alluded to in one of the
ivasahasranmastotras (13, app. 1, no.4, line 64): he sports with the daughters and wives
of the rsi s. Cf. also vetvatara Upanisad 6.9: iva is alinga, usually translated without
See also 13, app. 1, no. 5, lines 6970; 13,113* (after 13.14.102); 13, app. 1, no. 6,
line 16.
iva, the goddess 63

man, and is further equivalent to Mahpurusa. As already seen, pum s

and purusa are used interchangeably here to define the male side of
Arjunas female-bedecked appearance (4.10.1, 10.5, both cited above).
To be sure, we have here evocations of Arjuna as Nara in the latters
connection with Visnu-Nryan a.38 But there is probably also a reso-
nance of iva. Brhannad, a name in the feminine gender meaning the
great man, implies Arjunas completeness and echoes the homage to
iva as the Purusa who is male, female, and eunuch.
The scene of Brhannads appearance in Virtas court thus amounts
to a theophany. But it hardly stands without preparation or conse-
quence. Associations of Arjuna with iva follow a constant thread
throughout the Mahbhrata. Let us look first at another portion of
the Virta parvan, and then to the epic as a whole.
Still dressed as Brhannad, Arjuna reveals his identity to the Matsya
prince Uttara-Bhmim jaya when the two are left alone to defend
Virta s cattle from the Kaurava raiding party. With Uttara at the
reins, Arjuna defeats the Kauravas first singly and then collectively in
a battle that prefigures his triumph at Kuruksetra. It is here that the
symbolism of the dance first moves from the seraglio to the battlefield:
says the poet, it was as if Arjuna was dancing in battle [pranrtyadiva
sam grme] (4.57.9).39 This would seem a reminder of the destructive
dance of Rudra-iva, a suggestion reinforced by more direct allusions:
Thus having caused [the Kauravas] to see his raudra self, he of the
might of Rudra, held in check for thirteen years, Prtha, the son of
Pndu, roamed about releasing the terrible fire of his wrath on the
sons of Dhrtarstr a.40 Arjunas raudra (Rudra-like) fire, held in
check for the microcycle of twelve plus one years,41 can only evoke
Klgnirudra, Rudra-iva as the Fire of Time, lord of the pralaya or

Biardeau, EPHE 82 (197374): 9192, EPHE 84 (197576): 176: EMH (V), p. 189.
See Ramachandrashastri Kinjawadekar, ed., Shriman Mahbhratam with Bhrata Bha-
wadeepa by Nlakan th a (Pune: Chitrashala Press, 192936), 3:4 (apud 4.2.27 = critical
ed. 4.2.22). Nlakant ha says, from the non-difference of the consonants ra and la, and
da and la, one gets Nara: and he goes on to connect the name with Nara, Nryana,
and their hermitage at Badrinath. It is dubious that the feminine form holds dis-
tinctive reminders of such forms of Visnu as Mohin or Yoganidr, a suggestion of
Biardeau, EPHE 82 (197374): 91.
See Biardeau, EMH (V), pp. 18586, 197, 200.
darayitv tathtmnam raudram rudraparkramah
avaruddacaranprtho daa varsn i trn i ca
krodhgnimutsrjadghoram dhrtarstr esu pn davah (4.57.14; compare 8.32.14)
See van Buitenen, Virta, p. 4 (see n. 4 above).
64 chapter three

dissolution of the universe. One must also wonder at the following:

While Prtha was releasing his arrows, shooting with the right and
left hands, [his bow] Gn dva became, O king, like a whirling wheel
of fire [agnicakramiva] (4.59.12). How ancient are the sources for the
imagery of the medieval south Indian bronzes of iva Natarja danc-
ing in a circle of flames? Whatever one makes of that, there is one
unequivocal passage connecting Arjunas disguise and self-disclosure
with iva. Astonished but convinced that his effeminate companion is
none other than Arjuna-Pn dava, Prince Uttara overcomes his paraly-
sis at the prospect of battle with the Kauravas by exclaiming: I regard
you, in the disguise of a eunuch, as the roaming Holder of the Trident
[iva], equal to the king of the Gandharvas [Citrasena] or the God of
a Hundred Sacrifices [Indra] (4.40.11).42
Uttara thus sees in Arjunas disguise something of the identity of
each of these three figures.43 This triple connection is a precise one,
directing us to the links between Arjunas disguise and the rest of the
epic. For it is precisely these three who initiate Arjuna in three suc-
cessive episodes in the ran yakaparvan: iva the kirta (hunter)
teaching Arjuna the Pupata weapon and making it possible by his
touch for Arjuna to enter Indras heaven; Indra bestowing further
weapons and yielding to Arjuna his own throne; and Citrasena teach-
ing Arjuna music, dance, and the Gandharva Veda (3.81.14) which
involves the use of my.44 These three initiations are further rein-
forced by two successive confirmatory battles against asuras. First, to
satisfy Indras request for a daksin (gurus fee), Arjuna must fight
the Nivtakavacas. Setting out with Indras charioteer Mtali, the gods,
says Arjuna, thought I was Indra (3.165.16). He finally destroys these
foes with Indras vajra (169.1219). Then, on return to Indras abode,
Arjuna and Mtali see another Asura city called Hiranyapura, identi-
cal in name with one of the three cities which iva destroys in his
conquest of Tripura. Endangered by this enemy, Arjuna bows to the
god of gods Rudra (170.38) and fixes the Raudra weapon on his bow.

manye tvm klbavesen a carantam lapn inam
gandharvarjapratimam devam cpi atakratum (4.40.11)
Indra also has feminine forms, once cursed to have a thousand vaginas (see Wendy
Doniger OFlaherty, Hindu Myths [Baltimore: Penguin Books, 1975], pp. 9496)
and once born as Men, daughter of a king (see Rg Veda 1.51.13); but it would clearly
be difficult to connect these forms with his son Arjunas disguise.
Biardeau presents an illuminating analysis of this sequence; see EMH (V), pp. 14961
(with iva), 17583 (with Indra and Citrasena).
iva, the goddess 65

He perceives the weapon as a three-headed, nine-eyed man [purusa]

with three mouths and six arms, hair alight with blazing flames, his
head surrounded by serpents darting their tongues (170.39). Before
releasing it, he bows again to the three-eyed arva (170.41)a name
evoking Rudra-iva as hunter (from aru, arrow), and finally bows a
third time to the god who destroyed Tripura (170.50).
There can be no doubt that this pair of battles reinforces the dual
identification of Arjuna with Indra and iva.45 Arjuna completes his
novitiate with them by achieving a pair of conquests that allow him to
show his mastery of the two gods typical modes of triumph. It is not
merely that he uses their weapons, the vajra and the raudra (the latter
probably identical with the pupata46). In the first instance the gods
cannot distinguish him from Indra. And in the second, he bows to
iva as a hunterrecalling his recent intimate wrestling with iva the
Kirtaand as the destroyer of Tripura, the Three Cities, a duplicate
of one of which Arjuna has just effaced.
Concerning these sequences, the wrestling match with iva is the
most crucial for our purposes. With the foregoing analysis in mind, let
us reflect further on the interpretation of Biardeau. If, as she has con-
vincingly argued, the wrestling match with iva constitutes Arjunas
personal dks through which he is reborn directly into Indras
heaven, it is but a short step to recognizing that it is precisely through
this dks that Arjuna attains the identity with iva that is revealed and
concealed in the Virtaparvan. Biardeau has proposed that this dks
is a form of tm-yaja, the offering of oneself as oblation.47 But in
offering himself, Arjuna also becomes one with iva. This is implicit
in the sacrificial identity of offerer, victim, and recipient. Moreover,
iva shows Arjuna that as sacrificer and agent of destruction, the two
of them are one: their arrows strike the boar Mka at the same time.
As Biardeau has perceived,48 killing Mka simultaneously prefigures
the scene where Arjuna asks about a lance-bearing figure whom he
has seen before him in battle, the actual slayer of his foes, and Vysa

See Scheuer 1975, 220: Dans la lutte contre les Nivtakavaca, Arjuna etait
constamment identif Indra. Il est ici, dans un certain mesure, assimil a iva. Cf. also
Thomas C. Parkhill, The Forest Threshold: Princes, Sages, and Demons in the Hindu
Epics (doctoral diss., McMaster University, 1980), pp. 13637.
See Scheuer 1975, 21920, who notes that it is so identified by an interpolation
See Biardeau, EMH (V), pp. 15657.
Ibid., p. 158.
66 chapter three

tells him it is iva (7.173). There again Arjuna is one with iva in his
destructive role. It will not do to regard such episodes as isolated or
dismiss them as interpolations. Arjuna can defeat Bhsma only because
he is preceded by ikhandin, a personification of the Goddess but also
connected with iva. When Arjuna has only until the next sunset to
fulfill his vow to kill Jayadratha, Krsn a appears to him while Arjuna is
meditating in a dream and leads him to iva to guarantee that Arjuna
will recollect the pupata in this time of need (7.52.2 ff.). And finally
there is the contest of the doomsday weapons between Arjuna and
ivas other main protg Avatthman, which concludes the hostili-
ties at Kuruksetra,49 again showing a parity between Arjuna and iva.
Once this pervasive identification of Arjuna with iva is appreciated,
it sheds light on a number of other matters. Arjunas preparations for
his destructive capacity include another scene. In the only major epic
narrative concerning his identity as Nara, companion of Nryana
at their Badrinath hermitage, the two ascetics are challenged by the
battle-crazed tyrant Dambhodbhava. While Nryana remains inac-
tive, Nara humbles the king with his terrible reed weapon [ghoram-
aiskam] (5.94.28). The scene is in a sense prophetic: Nryan a-Krsn a
will be a noncombatant at Kuruksetra, while Nara-Arjuna will employ
the weapons of destruction. But Naras terrible reed weapon may be
related to Arjunas disguise. A straightforward etymology of the name
Brhannad is large reed or having a large reed (nada, nala: reed).50
Moreover, Krsn as noncombatancy, countered by ivas preceding
of Arjuna in battle as the actual slayer of his foes, calls to mind the
culminating passage of the Bhagavad Gt. Having just told Arjuna
that among Pn davas I am Dhanam jaya [Arjuna] (Gt 10.37) and
among Rudras I am am kara [iva] (10.23), having made Arjuna
witness to his pralaya-like theophany,51 Krsn a says: I am Time, cause
of the destruction of the worlds; . . . Even without you, they will all
cease; . . . Be the mere instrument [of destruction] [nimittamtram]
(11.3233). From the overriding Visnuite theological standpoint of
the epic, Arjuna is but the nimitta, the occasion or instrument of

See Hiltebeitel [1976] 1990, 31235. Cf. Scheuer 1975, 33543, a worthwhile cri-
tique, his most instructive point being that the battle cannot be said to occur till this
finale under the sign of Krsn a, exclusive of that of iva.
See van Buitenen 1978, 9; Biardeau, EPHE 84 (197576): 176.
See Biardeau, EMH (III), Bulletin de lcole franaise dExtrme Orient 58 (1971):
1937; Hiltebeitel [1976] 1990, 11440.
iva, the goddess 67

destruction, just as iva is in overseeing the destruction of the occa-

sional (naimittika) pralaya.52
It is also evident that the dual identities of Arjuna with Indra and
iva, and Draupad with r and Mrtyu, etc., have their widening impli-
cations. Insofar as royal sovereignty has temporarily been usurped by
Duryodhana, the crowned Arjuna (kirtin) and Draupad, Royal
Prosperity (rjar) incarnate, cannot appear together. In the kingdom
of Matsya, each goes their separate way. Speculatively, I have else-
where suggested that the disheveled Draupad-Sairandhr evokes the
unraveled prakrti of the dissolution (pralaya). In that connection,
it would be fitting that Arjuna should represent the asexual purusa in
isolation from her.53 In any case, during their period in disguise they
have only one exchange, accentuating their alienation. Draupad casti-
gates Arjuna for not being attentive to her suffering, to which Arjuna
replies: Brhannad, O blessed one, also obtains suffering unsurpassed.
She has gone to an animal womb [tiryagyonigat], O girl, but you
do not understand this (4.23.24). These harsh words not only stress
their dissociation form each other. Tiryagyonigat, gone to an animal
womb, is but another reference to hisand all the Pndavaspainful
rebirth in the womb of Matsya, the Fish.

The Eunuch in Media Res

It is now apparent that there are numerous links between Arjuna and
iva, and that it is through Arjunas disguise that they are brought
most sharply into focus. But one aspect of the disguise requires more
thorough discussion. From the moment at the dice match when
Duhsana taunts the defeated Pn davas by calling them barren
(or eunuch) sesame seeds [san dhatilh ] (2.68.8 and 1314, also 10:
klbh , eunuchs), the theme of eunuch-hood resonates through the
Mahbhrata. Thus, for example, Yudhisth ira, anticipating Karnas
death, will say: Those who were barren sesame seeds [san dhatilh ]

Cf. Scheuer 1975, 27880, n. 11, on similar uses of nimitta to those in Gt 11.33,
especially 3.199.3, used with reference to the dharmavydha, just slayer (concerning
a hunter); notes Scheuer: on noubliera pas que Rudra-iva est chasseur et invoqu
sous le vocable do vydha (
See Hiltebeitel 1981 on the dishevelment theme; and vetvatara Upanisad 5.10
on the soul: it is not female, not male, nor yet emasculate [naiva str na pumn esa na
caivyayam napum saka].
68 chapter three

there will now be sesame seeds [tilh ]! (8.52.16). But it is Arjuna as

eunuch-transvestite who brings this theme to central stage.
The subject of Indian eunuchs and transvestites is shrouded in con-
siderable obscurity. There has been little research, though what I have
found is of good quality.54 The scholarship covers data from essentially
three provinces: portrayals of eunuchs and transvestites in Sanskrit
drama and other classical courtly literature; classical legal and rit-
ual prescriptions; and folk practices. These three types of material are
arguably reconcilable. From studying them synoptically, a composite
picture of the eunuch-transvestite emerges that sheds a most surpris-
ing light on the significance of Arjunas epic disguise.
Some features of this composite, and its relation to Arjuna, are
relatively obvious. Eunuchs serve as trusted members of the seraglio,
the antah puraa role which Virta bestows upon Arjuna-Brhannad.
They have the reputation of being homosexuals, with a penchant for
oral sex,55and are looked upon as the very dregs of society.56 One
would stretch the epic evidence to insist that Arjunas disguise carries
such specific sexual connotations, but the social opprobrium is evi-
dent in Draupads disgust at seeing him in a disguise despised by the
world [lokaparibhtena vesena] (4.18.11). The most important details
in the composite, however, come primarily from the modern folk
practices. Though the evidence is scattered, there seems to be through-
out India a well-known class of eunuchs, appearing under different
names, but with certain common and persistent roles. These include,

Louis H. Gray, Eunuch, in Encyclopaedia of Religion and Ethics, ed. James
Hastings (New York: Charles Scribners Sons, 1912), 5:57985; N. M. Penzer, Indian
Eunuchs, in The Ocean of Story, Being O. H. Tawneys Translation of Somadevas
Kath Sarit Sagara, ed. N. M. Penzer, trans. C. H. Tawney (Delhi: Motilal Banarsidass,
1968), 3:31929; George T. Artola, The Transvestite in Sanskrit Story and Drama,
Annals of Oriental Research (Madras) (1975), pp. 5768; Wendy Doniger OFlaherty,
Women, Androgynes, and Other Mythical Beasts (Chicago: University of Chicago Press,
1980), pp. 283342. Surprisingly, only OFlaherty mentions Arjuna. See also Mircea
Eliade, Mephistopheles and the Androgyne: Studies in Religious Myth and Symbol,
trans. J. M. Cohen (New York: Sheed & Ward, 1965).
See Artola, Transvestite, pp. 6668, citing Vararuci, Ubhaybhisrik, ed.
A. K. Warder, trans. T. Venkatacharya (Madras: Sambamurthy, 1967), pp. 7071;
and Kmastra 2.9 (on oral sex).
Thus Mbh. 9.30.7071: Mlecchas [barbarians] are the dirt of mankind; rogues
[maust ikh] are the dirt of Mlecchas: eunuchs [an dhh ] are the dirt of rogues; those
whose sacrificial priests are warriors are the dirt of eunuchs. See Hiltebeitel [1976]
1990, 277. See also Morris Carstairs, The Twice-Born: A Study of a Community of
High-Caste Hindus (Bloomington: Indiana University Press, 1961), pp. 5961, 307,
323; eunuchs are also frequently regarded as degraded in the law books.
iva, the goddess 69

in addition to the opprobria already mentioned and, until recently,

associations with the harem (both Hindu and Muslim),57 three key fea-
tures: appearance at (1) festivals and temples, especially of the goddess;
(2) births; and (3) weddings.
Concerning the first, enough has been written about eunuch priest-
hoods to the goddess and the possible connections between such
orders in Asia Minor as the Galli, eunuch-transvestite servitors of
Cybele,58 and various Indian counterparts.59 The similarities are suf-
ficient to suggest, at least minimally, that both the Indian and near
Eastern cults involve extreme interpretations of a more widely held
practice of ritual transvestitism in service to the goddess, something
which at least in India is far more common and general60 than actual
eunuch priesthoods. Our concern with eunuch priests must be more
restricted, as it is unlikely that the epic poets have such a specific order
in mind in their depiction of Arjuna. But we are still faced with the
festival atmosphere of the Virtaparvan, first noticed by van Buit-
enen, but involving more complex ritualindeed sacrificialpatterns
than he saw. Not only does Arjuna play the eunuch-transvestite, but
Bhma pretends that he is Draupad (in vernacular traditions, dress-

One sometimes hears Hindus claim that the practice was introduced by Muslims,
who recruited by abducting Hindu boys, but clearly eunuchs served in the seraglio
before the coming of Islam. There are, however, reports of terrifyingly nefarious
recruitments; see Penzer, pp. 32225.
In addition to the articles of Gray and Penzer (see above, n. 54), see Sir James
George Frazer, Adonis Attis Osiris, 2 vols. (= The Golden Bough, vol. 5), 3d ed. (New York:
Macmillan Co., 1935), pp. 26376, especially p. 270, n. 1; Peter Tompkins, The Eunuch
and the Virgin: A Study in Curious Customs (New York: Clarkson N. Potter, Inc.,
Such as the eunuch priests of the goddess Huligamma in Bellary and Dharwar dis-
tricts of Karnataka, reported by Fred W. Fawcett, On Basivis: Women Who, through
Dedication to a Diety, Assume Masculine Privileges, Journal of the Anthropological
Society of Bombay 2 (1891): 34344; the eunuchs who still appear at the temple of the
goddess Catuhrng outside Pune, particularly at Dussera: see J. H. Hutton, Caste in
India: Its Nature, Functions, and Origins (London: Oxford University Press, 1961),
p. 165; the Alis (eunuchs) who come to festivals (especially Pongal) in Tamilnadu
(oral communication from C. T. Rajan); the Hijras (hijr) of much of northern and
western India, many of whom worship the Goddess (see Penzer, pp. 32125, with
citations from the various compendia on castes and tribes; Carstairs, pp. 5961); and
others of various names.
Henry Whitehead, The Village Gods of South India (Calcutta: Association Press,
1921), cites numerous examples; see also discussion in Hiltebeitel 1978, 791; and Marie-
Louise Reiniche, Les Dieux et les hommes: tude des cultes dun village du Tirunelveli
Inde du Sud (Paris: Mouton, 1979), p. 245.
70 chapter three

ing in her clothes) to slay the latters tormentor.61 It is thus not at all
inappropriate to suggest that Arjunas disguise is part of a wider set of
themes evocative of festival and sacrificial activities. But these must
be left somewhat imprecise. The epic itself indicates that the slaying
of Kcaka follows from a train of incidents that occur on a holiday
(parvin ; 4.14.5). One can only guess, what with the transformations of
Bhma and Arjuna (the latter stands aside until the end of this episode,
but his eunuchhood is a negative factor throughout), whether this
holiday was implicitly one that would highlight the goddess. If so,
it would provide one more link between Arjuna and iva, for it is, of
course, finally ivaor iva in union with the Goddesswhom males
impersonate as transvestite or eunuch priests.
The second and third appearances of eunuchs in the Indian context,
at birth and marriage ceremonies, can be discussed together. These
practices are widespread.62 There is room to suspect that their appear-
ances include an auspicious function, though in a highly ambiguous
setting. At marriage rites among the Bharvds in Gujarat, for instance,
a eunuch flings balls of wheat flour toward the four quarters to ward
off evil spirits.63 And when they appear at births and marriages, they
hold the power to bless as well as curse.64 But for the most part their
presence is regarded as a nuisance. They dance and sing (often abu-
sively) until they are paid to leave, and are usually not seen differ-
ently from beggars. Auspicious functions, if they exist, are not readily
acknowledged, and an inauspicious potential is also recognized, as
in the complaint of the cowherd Gang, in the Punjabi epic Pran
Bhagat: When I was in my mothers womb eunuchs danced at the
door. And so I am born lame, and have no hair on my head.65
It would be admittedly arbitrary to connect these practices with the
Mahbhrata were they reported only for recent centuries. But there
is clear evidence that they are much older than even the Mahbhrata,

On this scene and its sacrificial connotations, see Biardeau, EMH (V), p. 173,
EPHE 84 (197576): 176.
See Carstairs, pp. 60, 307, 323 (at both birth and marriage, in Rajasthan); Penzer,
pp. 32425 (Gujarat: at births of sons, particularly of formerly barren or sonless
women); among Alis in Tamilnadu, Hijras in Pune and Delhi, at both births and
marriages (various personal communications).
William Crooke, Indian Charms and Amulets, in Hastings Encyclopaedia of
Religion and Ethics, 3:44647.
Personal communication from Balaji Gopal.
R. C. Temple, Legends of the Punjab, 2 vols. (Bombay: Education Societys Press
Byculla, 18841900), 2:396.
iva, the goddess 71

at least as regards birth ceremonies. Atharvaveda 8.6 is a hymn sung to

guard pregnant women against demons. It would presumably have been
well known in later times, such as those which cover the period of the
epics composition, because it was employed in the smanta rite in the
eighth month of a womans pregnancy with binding on an amulet.66
The hymn mentions a herbal charm, baj, and, apparently different,
a yellowish talisman, ping,67 which are invoked to ward off a vast
array of birth-threatening demons, many of whose names, as Whitney
remarks, are in good part unknown elsewhere and untranslatable.68
What is evident and consistent, as if it were a leitmotif throughout
the hymn, is the connection of such demons with eunuchs. Verses 10
and 11, read together, show that the host of demons is said to dance
around the dwellings in the evening, making donkey-noises, dancing
like impotent men [or eunuchs; klvai va]. Other verses add slurs and
curses of the despised eunuchs and their companions: thrusters forth
of womens hips (8.13); pot-testacled, ays [impotent?] (8.15);
Womenless be the eunuchs (8.16); Gandharvas, women-seekers
(8.19); the hairy ones [that] devour embryos (8.23). The charms are
directed against two main ends. First, several verses seek to counteract
stillbirth: Whoever makes this woman one having a dead child, or
a miscarriage, him, o herb, do thou make disappear (8.9; similarly,
8.1820, 26). And the next to last verse, as a sort of finale, calls for
the male child to remain male: Ping, defend thou [the child] in the
process of birth; let them not make the male female (8.25). It would
seem that the eunuchs are perceived as a threat to the sexual identity
of the male embryo.
We have now neared the point to consider the relation between
the epics depiction of Arjuna and these century-spanning associations
of the eunuch. But first, one last verse from this Atharvaveda hymn
may impose the issue upon us, and in most intriguing terms: He who
lies with thee [the pregnant woman] in sleep, having become like a
brother and like a fatherthem, eunuch-formed, tiara-decked, let the
baj force from here (8.7). The phrase eunuch-formed, tiara-decked

William Dwight Whitney, trans., Atharva-Veda Sam hita, 2 vols. (Delhi: Motilal
Banarsidass, 1962), 2:493.
Turmeric, according to Lexicons mentioned by Monier-Williams, Dictionary.
Whitney mentions that the Ath. Paddh. [Atharvaveda Paddhati?] seems to prescribe a
talisman in the form of a doll made of red and yellow mustard plants (Atharva-Veda,
p. 494).
Whitney, ibid.
72 chapter three

is klbarpam tirtin. Whitney notes that the Kashmirian Paippalda

Recension reads instead klbarpam kirtinam. This accords with the
well-accepted view that the hapax tirtin in this passage is doubtless
identical with the later kirtinam, and again refers to some feminine
mode of dressing the head.69 Now of course Arjuna is precisely
klbarpam kirtin, the diademed one in the form of a eunuch, in
the Virta parvan.70 Moreover, to the Princess Uttar he becomes, in
the Atharvavedas words, like a brother and a father, or, in his own
words to Virta, when offered Uttar in marriage, I dwelt in the sera-
glio always seeing your daughter, secretly and in the open, and she
trusted me like a father [vivast pitrvnmayi] (4.67.2). It is evident,
minimally, that Atharvaveda 8.6.7 and the Virtaparvan rely on simi-
lar images of the eunuch, and not impossible that the epic poets depict
Arjuna with this paradoxical Atharvanic verse in mind.
In any case, Arjuna-Brhannad, in the Virtaparvan, is a eunuch
who presides over both a marriage and a birth. As just indicated, it
is he who arranges the marriage between his son Abhimanyu and the
Matsya Princess Uttar. Indeed, more than this, it is through his dis-
guise that he is able to prepare the bride, through a year of asexual
intimacy and instruction in song, dance, and music, for her marriage
into the Pn dava-Kaurava line. And it is in the context of preparing
Uttar for marriage that Arjuna also oversees, as a eunuch, the precon-
ditions for the birth of his grandson Pariksit, who will, after the battle
of Kuruksetra, be the sole male heir of the line and thus the last hope
of its renaissance. One might object that Krsn a oversees the actual
re-birth of Pariksit, reviving him after he is stillborn (a reminder of
the verses in Atharvaveda 8.6 that concern preventing eunuchs and
other demons from causing stillbirth); thus Krsn a actually effects this
renaissance.71 But it is Arjuna-Brhannad who prepares his future
daughter-in-law for her role as mother-to-be. When Brhannad and

Maurice Bloomfield, trans., Hymns from the Atharva-Veda, Sacred Books of
the East, vol. 42 (Delhi: Motilal Banarsidass, 1967), p. 538, n. 2; Manfred Mayrhofer,
Kurzgefasstes etymologisches Wrterbuch des Altindischen (Heidelburg: Carl Winter
Universitatsverlag, 1956), s.v. tirtah : he relates both kirtah and tirtah to Ardhamgadh
tirda, a coronet with three crests.
Klbarpa (4.10.7; 36.30). The kirta, apparently a diademed hair setting, is of
course absent from his disguise, but Draupad deplores the fact that a ven (braid) has
replaced his kirta (18.13); and when he reveals his identity to prince Uttara, Arjuna
explains Kirtin as one of his ten names (39.15).
See Hiltebeitel [1976] 1990, 34951.
iva, the goddess 73

Prince Uttara (Uttars brother) are about to set off to defend Virtas
cattle against the Kaurava raiding party, Uttar requests Brhannad
to bring back the Kauravas garments for her to dress her dolls
(4.35.23). Arjuna achieves this end, enabling her thus to dress her
dolls with the heavenly resplendent [divyni rucirn i] (35.25) white
robes of Dron a and Krpa, the resplendent yellow robes of Karna, and
the blue robes of Avatthman and Duryodhana (61.13). Clearly, in
dressing her dolls with these very garments, Uttar prefigures her
role as mother of a Kuru king.72 And, as pointed out elsewhere,73 her
use of the term pclik for dolls suggests that she will in effect
replace Draupad Pcl as the woman through whom the continu-
ity of the line will be assured. She thus asks for garments not only for
her dolls, but, more literally, for the sake of she who comes from
Pacla [pclikrtham] (4.35.2223).
Arjuna the eunuch thus prepares Uttar for both marriage and
childbirth. But in doing so, as Biardeau has seen in connection with the
marriage, where he refuses to wed Uttar himself, Arjuna, the ideal
king, symbolizes thereby his personal renunciation of sovereignty (see
above, n. 18). Such renunciation also applies to the issue of an heir. To
be sure, Arjunas impotence is temporary but it is consonant with
other mattersparticularly Avatthmans curse of the wombs of the
Pndava womenthat prevent the Pn davas and Draupad (and other
Pndava wives) from having further offspring. In this respect, there
is a fitting regulation found in the lawbooks. For example, accord-
ing to Yjavalkya Sam hit 2.14344; A eunuch [klba], an outcaste
and his son, one who is lame, a madman, an idiot, one born blind, a
person afflicted with an incurable disease and such others, must be
maintained without any allotment of shares [niram akah ]. But sons
of such persons, whether born of their own loins or on the soil [the
wife], being free from similar defects, shall obtain their fathers shares
of the inheritance [bhgahrin ah ].74 As a eunuch, then, Arjuna relin-
quishes his inheritance of sovereignty (among other things, as son of

One wonders whether to connect these dolls with the talisman, ping, men-
tioned in Atharvaveda 8.6, especially as the talisman may have taken doll form (see
above, n. 67).
See Hiltebeitel 1980a, 1981; Biardeau, EMH (V), pp. 19799.
See also Laws of Manu 9.2013, and Bhlers notes in Georg Bhler, trans., The Laws
of Manu, Sacred Books of the East, vol. 25 (New York: Dover Publications, 1969), pp. 372
73. Cf. Carstairs, pp. 5961: the hijras are but a half community, there being twelve
and a half in all (twelve being a traditional number for those obtaining shares in a
74 chapter three

Indra) for himself, but the inheritance passes through his slain son
Abhimanyu to his grandson Pariksit.75 Furthermore, in the Vjapeya
sacrifice, a eunuch (a long-haired purusa who is neither man [pum s]
nor woman [str]) is required during the Soma purchase to sell an
intoxicating beverage called the parisrut that is neither Soma nor sur
[liquor] (atapatha Brahmana These two latter drinks are
associated in the text with light and the gods (Soma) and darkness
and asuras (sur). The eunuchs ambiguous role in this soma purchase,
which in the Vjapeya is oriented toward attaining sovereignty, may
remind us that Arjuna acts as a eunuch specifically for the perpetua-
tion of the Soma (Lunar) Dynasty, whose essence is embodied in the
persons of Abhimanyu (the incarnation of Somas splendor [varcas])
and Pariksit.76
Given such a variety of converging themes, it will be well to sum-
marize their lessons for interpreting Arjunas disguise. As a dancer-
musician and eunuch-transvestite he evokes iva. In the former case
his ostensibly auspicious role carries with it the destructive overtones
of the dance and music of battle and the cosmic dissolution. In the
latter case, his ostensibly inauspicious role carries with it the promise
of the rebirth of the Kuru/Soma dynasty. Perhaps we may speculate,
by way of conclusion, that when eunuchs dance and sing at births
and weddings, they mark by their presence the ambiguity of those
moments where the nondifferentiation of the male and female is most
filled with promise and uncertainty: in the mystery that surrounds the
sexual identity of the still unborn child (Let them not make the male
female [Atharvaveda 8.6.25]) and in that which anticipates the re-
union of the male and female in marital sex.

sacrifice: e.g., the twelve ratnins in the Rjasya, and twelve communities in the south
Indian buffalo sacrifice, on which see Hiltebeitel 1980c, 196.
On the surprising matter of eunuchs siring offspring, and on different types of
eunuchs, some with such capacity, see Nrada Smrti 12.8018, discussing fourteen types
of eunuchs: Julius Jolly, trans., The Minor Law-Books, pt. 1, Nrada, Brhaspati, Sacred
Books of the East, vol. 33 (Delhi: Motilal Banarsidass, 1969), pp. 16669. Reference
thanks to Richard Lariviere.
See Hiltebeitel [1976] 1990, 33654.
iva, the goddess 75

The Pn davas Altogether

As we saw earlier, the various scholarly interpretations of the Pn davas

and Draupads disguises have all been partial. Common denominators
for the six disguises have not emerged. Our examination of Draupad
and Arjuna, however, has laid bare two underlying strands: associa-
tions with impurity, and representation of iva and the Goddess. Let us
take up these two topics separately, in each case testing their applica-
bility to the disguises separately and as a group.
As regards impurity, there is the notion, known at least to Nlakan th a,
that putting on disguises is itself impure. In a line that occurs twice
in the epic, parvakras are included along with soothsayers, friend-
harmers, and men involved with others wives (parvakra ca ci ca
mitradhrukpradrikah [5.35.39cd, 13.90.9ab]) as being in one instance
equal to Brahmin killers (5.35.42) and in the other inadmissable to
society, ejected from caste (apnkteya [13.90.5]). Curiously, Nlakan th a
glosses the first appearance of parvakra by arakrt, arrow-makers,
and the second by vesntaradhr, those who wear anothers dress/
disguise.77 But we have other indications of impurity that bear more
directly on the Pndavas disguises as a group. In Bhatt a Nryan as
Ven sam hra, Bhma says that the Pndavas and Draupad were con-
cealed by means of improper occupations [anucitrambhanibhrtam]
(1.11). No doubt this has most immediate reference to tasks inappro-
priate to their Ksatriya caste. But this itself involves taking on defile-
ment. Similarly, the theme of the brothers and Draupad taking on
improper tasks is unmistakably popular among participants and
audiences in the terukkttu dramas of the south Indian Draupad cult.
In fact, the Vanniyars, who constitute the most prominent caste in the
cult and who provide most of the actors for the dramas, are especially
attentive to such a theme. They are dras who view themselves as
Ksatriyas, as it were in a disguise forced upon them by history.78
It has already been noted that Draupads disguise and actions as a
Sairandhr (chambermaid/hairdresser) hold strong associations with
defilement and low caste (see above at nn. 17 and 26). Since these are
treated extensively elsewhere,79 it will be fitting to regard them from

Kinjawadekar, ed. (see n. 38 above), 6: 199 (same citation as for critical ed.).
See Hiltebeitel 1982.
See Hiltebeitel 1981.
76 chapter three

a new angle. In intensifying her defiling contacts during this last year
of exile from the throne, Draupad is cast in a manner that draws
from themes which today are associated with the village goddess. Note
the parallels with this myth about Selliyamman (from southern Chin-
gleput District, Tamilnadu). Forced to grant a demon various boons
and her own weapons, selliyamman
was now without power, without her shakti. The demon took over Selli-
yammans kingdom, and out of vengeance for the loss of his wife, he
made Selliyamman his slave [adimai], forcing her to do low domes-
tic tasks and field labor. Selliyamman called on her son Virabatra, her
brother Vishnu, and her husband Siva to defeat the demon, but none
of them could . . . So Vishnu took the form of Lakshmi, and the demon
was attracted to her. Lakshmi was then able to kill the demon, and to
return Selliyamman to her kingdom. During the time of her slavery to
the demon, Selliyamman had done all the unpleasant tasks in the village,
so the villagers were grateful to her and worshipped her thereafter.80
Like Draupad, estranged wife of at least one impersonator of iva,
during the reign of the demonic Duryodhana, the village goddess Selli-
yamman, wife of iva, must take on the lowest and most defiling tasks
during her period of slavery to a demon king.
Yudhisth iras associations with impurity are not immediately evi-
dent. His disguise as a Brahmin may be improper (though common in
the epic) for a Ksatriya, but it may not necessarily be thereby impure.
It is Biardeau who has found this indispensable piece of the puzzle;
Yudhist hiras name Kanka, Heron, connects with Dharmas disguise
as the crane (baka) who temporarily kills each of Yudhist hiras broth-
ers (3.296). Biardeau is surely right that when the epic refers to Dharma
as Yudhisth iras father, it is evoking Yama, although one cannot
say that the epic fully identifies Yama and Dharma.81 Yudhist hiras
name Heron thus carries connotations of Yamas own associa-
tions with impurity and death, and it can be no accident that success
during the period of concealment is promised to Yudhist hira by his
father Dharma-Yama, after the latter has dropped his crane disguise
(3.298.1619). The period of rebirth in the womb of Matsya thus
takes place under the auspices of death. Yudhisth ira enters Matsya,

Michael Moffatt, An Untouchable Community in South India: Structure and
Consensus (Princeton, N.J.: Princeton University Press, 1979), p. 273.
See Biardeau, EMH (V), pp. 99101; see Hiltebeitel 1980a, n. 17, discussing alternate
views on Dharmas identity.
iva, the goddess 77

Fish, as a Heron, an eater of fish, thus prefiguring his destructive

role at Kuruksetra, the battlefield where the crane and the heron will be
frequently mentioned, despite their aquatic habitat, among the carrion-
eating predators who will gather to consume the corpses of the fallen
With Bhma, the matter is less covert. In preparing food for Virta,
he handles its impurity for one of only apparent higher rank. But
Bhma is not merely washing vegetables and preparing curries. He
claims the role of govikartr, cow-slaughterer, the priest whose name
suggests the role of dividing up the sacrificial animal by cutting it
asunder (4.2.7).83 There can be no serious question that the epic poets
regard such a task as tinged with impurity. The Southern Recension
also clarifies matters by having Bhma identify himself as a dra
(4,191*) performing low karma (ncakarma [4.192*, line 1]).
Skipping Arjuna for the moment, the case of the twins is close to
Bhmas. On the surface, their assumption of Vaiya tasks, as super-
visors of horses and cattle, involves a lowering of caste rank and
thus some attendant impurity. Curiously, pauplas, those who tend
animals, are included in the same list that mentions parvakras as
inadmissable to society (13.90.6). But hidden behind these appar-
ently benign roles is again an identity as sacrificers. In a passage that
foretells the sacrificial roles which the major warriors will play at the
sacrifice of battle, Karn a says that Nakula and Sahadeva will act as
the amitr priest [amitram karisvatah ] (5.139.36); that is, as slay-
ers of sacrificial animals, holders of the priestly office that involves
putting the victim to death by suffocation or strangulation.84
For these four brothers, then, the disguises are tinged with impurity,
in connection with both their caste identities and their roles as sacrifi-
cers. Concerning caste, one must note thatcontrary to Dumzilthe
disguises embrace more than just the upper three varnas. With Arjuna
and Draupad included, the caste associations are with the total social
order, from Brahman to outcaste, the order which will be regener-
ated with their victory at Kuruksetra. As regards sacrifice, the primary

Biardeau, EMH (V), p. 104, p. 99, n. 1, and above, n. 14. See also Laws of Manu,
Compare EPHE 82 (197374): 9091, 95: Bhmas cooking utensils as sacrificial
See Citrabhanu Sen, A Dictionary of Vedic Rituals (Delhi: Concept Publishing,
1978), p. 110.
78 chapter three

concern with Draupad and all the brothers but Arjuna is with the
impurity of death. For Arjuna, however, the matter is different. He
whom Krsn a will teach to kill without sin, whom iva will precede in
battle, is concernedin his disguisenot with the impurity of death
but that of birth. It is Arjuna-Brhannad who hands on the garments
of rebirth-representative of the amnion and chorion85to Uttar for
her dolls.
As to our second strand, the associations of all the Pndavas with
iva, and of Draupad with the Goddess, it intertwines with the first.
Let us recall how, according to the epic, they all came to be born on
earth.86 In the overanxious maiden story, iva promises Draupad
in her previous life that, as she had prayed five times for a husband,
she would have five husbands. And in the Story of the Five Former
Indras, it is iva and Prvat playing dice who have consigned the five
Indras to a cave and forced the goddess r to weep and appear as one
of ill fortune. In both cases we see that it is iva who oversees the
rebirths of the Pn davas and Draupad. But it is the second myth
that is most astonishing in our present context. The gestation of the
five Indras in the cave and the appearance of r as ill-fortuned are
but foreshadowings of the Pndavas and Draupads concealment (and
in Draupads ease, her transformation from an image of prosperity to
one of ill fortune evocative of Mrtyu or Kl) in the womb of Matsya.
In each case we have the image of a dks in which the gods or heroes
offer themselves as victims to take on destructive sacrificial roles. And
the self-offerings in the Virtaparvan involve hints at association with
iva that go beyond the case of Arjuna alone.
The key here is a detail from the Rjasya sacrifice, a rite whose
structure and complexities have shed considerable light on many fea-
tures of the Mahbhrata, and in particular on the dice match.87 One
of the rites which are performed in the course of this sacrifice is the
offering to the eleven or twelve ratnins, the jewel-possessing digni-
taries of the realm. Each of the ratnins is connected with a deity, and
a number of them are also differentiated among the four castes. Now,

See Heino Gehrts, Mahbhrata. Das Geschehen und Seine Bedeutung (Bonn:
Bouvier Verlag Herbert Grundmann, 1975), pp. 2067, 22425; Hiltebeitel 1980a.
See Hiltebeitel [1976] 1990, 86101, 16991; Scheuer 1975, 94107.
See J. A. B. van Buitenen, On the Structure of the Sabhparvan of the Mahbhrata,
India Major (Leiden: E. J. Brill, 1972), pp. 78-83; Gehrts, passim (he rather stretches the
application; see Hiltebeitel 1977b for a review of Gehrtss book.
iva, the goddess 79

two of the ratnins are usually mentioned together: the aksvpa, or

surveyor of the dicing hall, and the govikartr, the cow-slaughterer.
Their caste is probably dra, and their deity is Rudra.88 The connec-
tion with Rudra is not difficult to perceive. iva is persistently fond of
dicing in Indian tradition.89 And it is he who takes on the impurity of
death in the classical Brhmanical animal sacrifice,90 here in connec-
tion with the office of sacrificial butcher.
Quite clearly, the two oldest Pndavas disguises are bared by these
convergences. Yudhist hira-Knka becomes the surveyor of Virtas dicing
hall; and Bhma, as we have seen, becomes literally his govikartr. Let us
now look more closely at the relations of each of these to Rudra-iva.
Biardeau, attempting to extend the link she sees between Yama-
Dharma and Yudhisth ira, has suggested that the latters newfound
mastery of dice suits a representative of Yama, given the link between
the daiva [fate] and the dice game. During the thirteenth year,
Virtas court will continue to play dice, and Yudhisth ira-Kanka will
not cease to win, announcing his forthcoming victory in the war.91
But Yudhist hira-aksvpa should not be confined to only one divine
dimension any more than Arjuna. Yudhisth ira shows a rapport with
both Yama and iva, as is evident from his dream in which he sees
Rudra, in a destructive and inauspicious form, facing toward the south,
the region of Yama (2, app. 1, no. 30). Though this passage is found
only in the Northern Recension, it signals essential themes. Occurring
before the fateful dice match in the Sabhparvan, it portends that the
dice match will not only be under the sign of iva, but that it will be
oriented toward destruction and death, that is, toward the south. It
is in this context that Yudhisth iras associations with Yama will, as
Biardeau has perceived, begin to be played out.92 Yudhist hiras mastery
of dice in the Virtaparvan, however, is not to be identified with Yama.
Here, as usual, dicing points to iva, but this time it is oriented toward
destruction and victory, and the mastery of fate and time which the four
throws of the dice represent in their correlation with four yugas.93

J. C. Heestermann, The Ancient Royal Consecration (s-Gravenhage: Mouton, 1957),
pp. 49 (with facing chart)-57; p. 55, n. 34 mentions the possibilty of dra rank.
Hiltebeitel [1976] 1990, 94101.
Biardeau, EMH (III), p. 80; Hiltebeitel 1978, 770.
Biardeau, EMH (V), p. 187.
Ibid., p. 105.
See Hiltebeitel [1976] 1980, 9495.
80 chapter three

As regards Bhma, a number of links with Rudra-iva amplify his

role as govikartr. As Biardeau has perceived, Bhmas aptitude for kill-
ing, in particular killing enemies with whom it is not always possible to
respect the rules of the perfect ksatriya, corresponds, in other terms, to
the rudraic aspect of the avatra.94 Thus he slays Draupads tormentor
Kcaka by compressing his limbs into a ball of flesh (mm sapin da) in
the manner of iva killing a sacrificial animal [paoriva pinkadhrk]
(4.21.5960).95 Bhma-govikartr is thus an image of iva-Paupati. As
cattle-slaughterer, he is prepared during his period of disguise to
undertake his own part in the slaughter of victimsultimately pau,
cattleat Kuruksetra.96 Here his disguise would seem to correlate
with that of the twins, who, as we have seen, not only tend horses
and cattle, but act as amitrs. This latter priestly office, which involves
administering the actual killing, is certainly tinged with impurity. But
more than this, as Scheuer remarks, noting a number of associations
between iva and the dangerous and impure aspects of the sacrifice:
One would be equally tempted to think that there was a particular
rapport between the samitr and Paupati; the ritual texts tell us noth-
ing, but doesnt one see Paupati, in the [Brhmana] myth, pursue and
wound Prajpati [the archetypal victim]?97 The twins sacrificial iden-
tities, which their disguises as animal tenders would seem to evoke,
would thus also be reminiscent of Paupati. And of course Arjuna is
the recipient of the Pupata, ivas doomsday weapon, named after
the same identity of the god.
The Pndavas and Draupads disguises thus show that the poets
conceive these heroes and their wife to have more than univocal mythic
associations. The Pn davas links with their trifunctionally arrayed
fathers are but one facet of the whole. They are also five Indras, a fact
of special importance with regard to Yudhisth ira.98 And we now see

Biardeau, EMH (IV), p. 233, cf. EPHE 82 (197374): 9596: he is to Arjuna what
iva is to Visn u.
See Biardeau, EPHE 84 (197576); 176; Scheuer 1975, 22328. In pretending that
he is Draupad, Bhma also acts in a manner that points toward the apparently post-epic
myth of iva in the form of a seductress with toothed vagina killing the demon di;
see OFlaherty, Hindu Myths, pp. 25161 (n. 43 above).
On the term pu used for victims in the Mahbhrata, see Hiltebeitel [1976]
1990, 32024.
Scheuer 1975, 283 (brackets mine); see Sen, p. 110 (n. 84 above): the amitr
also cuts [the victims] limbs with a sharp knife, citing pastamba rautastra 7.
14. 14.
See Hiltebeitel [1976] 1990, 8199, 169200.
iva, the goddess 81

that along with Draupad, they also evoke iva and the Goddess in her
destructive aspect. The full import of these textured heroic associations
with the gods requires fuller examination, particularly as it will affect
our understanding of Krsn a in the epic. For it is clear that the roles
of the epic heroes and heroines reflect a deep-leveled concern for
articulating complementary yet differentiated roles for figures closely
linked with Visnu, iva, and the Goddess.
Finally, it is noteworthy that the association of the Pndavas with
iva, which this study has emphasized, are also registered in folk tra-
ditions concerning the Mahbhrata. Throughout India one can find
caves, hills, and temples sacred to iva which are regarded as places
where the Pndavas spent their years of exile.99 It is thus implied that
during this period the Pn davas and Draupad worshipped iva, dupli-
cating as a group the special rapport with iva achieved by Arjunas
tapas and further anticipating the fuller rapport that they would have
with him during their thirteenth year disguised in the kingdom of
Virta. Their forest tapas thus prepares them for their dks-like ges-
tation in the womb of Matsya. And through the latter, they are in turn
reborn to perform the sacrifice of battle as sacrificers identified with
iva, the lord of destruction, yet also the god who, as Paupati, neu-
tralizes the impurity of sacrificial death for the benefit of the worlds
of gods and men.

The caves, Pn dulens, are often thought to have originally been Jain or Buddhist
ascetic retreats. I have visited one behind Ferguson Hill, Pune: attended by outcaste
priests, its major icon is a trident. Aivarmalais, Hills of the Five, are found throughout
Tamilnadu: one near Palani has caves with both Jain and aiva-kta Pndava and
Draupad appurtenances. I have visited Pndavevara temples at Hastinapura (behind
the old mounds) in Uttar Pradesh, and at Talegaon and Pandeshwar (near Jawalarjun,
a name meaning Arjuna is nearby, and Jejuri) in Maharashtra. Many other iva
temples without such names hold legends that the Pn davas and Draupad spent time
there during their forest wanderings. Several of our themes converge at the Chandan
Festival at Puri, which precedes the Jaganntha Ratha (Car) festival: images of Five
ivas from various iva temples around Puri are known as the Pacupandabas
(Five Pndavas). They are drawn about in a boat with an image of Balabhadra, all of
them together representing asceticism (in the epic Balarma is a forest wanderer and
trthaytrin), and attended on the boat by boys who dance as women! See Frdrique
Apfel Marglin, Wives of the God-King: Rituals of Hindu Temple Courtesans (doctoral
diss., Harvard University, 1980), pp. 26775.


The two Indian epics, the Mahbhrata and Rmyan a, are thoroughly
familiar with the two different domains of classification: purity-impurity
and auspiciousness-inauspiciousness.1 On occasion, either can be
found as the focus of a rather extended independent treatment, as for
instance in Karnas lengthy tirade against the impure (aauca) practices
of the Madras, the subjects of king alya (Mbh. 8.27.6673 and 8.30),2
or typical lists of mangalas, auspicious things, such as the twelve which
Yudhisth ira touches when he wakes up on the battlefield, including gar-
lands, well-adorned auspicious maidens (svalam krth ubhh kany),
and auspicious birds (mangalyn paksin ah ; Mbh. 7.58.19-21). One also
finds in both epics, but especially in the Rmyan a, frequent references
to auspicious and inauspicious omens (akunas) that preview a warriors
performance in battle, reunion with a loved one, or some other eventu-
ality. There is some considerable documentation and discussion of such
lists of the pure-impure and auspicious-inauspicious in P. V. Kanes
History of Dharmastra, with useful references to epic and other clas-
sical sources, so I refrain from discussing them further other than to
note that Varhamihira in his Yogaytra mentions dishevelled hair as
inauspicious, a point I will return to shortly.3
The two epics do not, however, always treat these two domains of
classification differentially. Many episodes highlight both, and bring
their relation to each other into strong relief. In doing so, the epics
provide important evidence for their study, for they portray their
interplay in a fluid narrative form. It will thus be less useful to extract
passages, like the one just cited, dealing with either matter alone, than

I would like to thank Frdrique A. Marglin for her work in clarifying these two
domains, and stimulating my research upon their relation to each other in the epics.
All Mahbhrata citations are of the Poona Critical Edition. All Rmyan a citations
are of the Baroda Critical Edition.
See Pandurang Vaman Kane, History of Dharmastra (Ancient and Mediaeval
Religious and Civil Law in India), 5 vols. (Poona: Bhandarkar Oriental Research
Institute, 19621975), Vol. 4, pp. 267333; Vol. 5, pp. 366367, 534540, 621622,
719814 and passim. The reference to Varhamihira, Yogaytra 3.14, on dishevelled
hair occurs at Vol. 5, p. 622.
84 chapter four

to observe the ways the epics treat purity and auspiciousness in the
round. This will involve keeping two things in mind. First, although
we may take the terms uddha and ubha as corresponding to our
basic pair purity and auspiciousness respectively, we must be attentive
to larger term-clusters and conceptual frameworks for each of the two
domains.4 Secondly, these two domains cannot be studied satisfactorily
in isolation from other domains. It will be useful to pose the question
in terms similar to those of R.S. Khare, who insists that an adequate
discussion of Indian food preparation must involve recognition of an
interplay between four axes: not only the pure and impure and the
auspicious and inauspicious, but the high and the low and festivity and
mourning.5 I will continue to follow the terminology of Dan Sperber
and refer to such axes as domains6: thus the domains of purity and
auspiciousness. But the epics are like food production in that their use
of the symbolism of these two domains involves a convergence with
the vertical and ceremonial domains which Khare cites, as also
with the domain of dharma. Altogether, these various domains also
evoke, through their convergence, a theological dimension to the epics
which I have begun to discuss in other papers.7
In focusing primarily on the domains of purity and auspiciousness,
this chapter will draw in part from the results of these other studies. But
it will also afford a chance to view these results against the background
of a larger issue that is one of my long-range concerns: the relation
between the two epics. The examination of the domains of purity and
auspiciousness in the two epics allows us to discuss important areas
in which the two narratives show significant thematic rapports. To
limit discussion, I will treat only two such areas. They involve matters
in which the two epics show important parallels and oppositions: the
violations of the heroines Draupad and St, and the preparations for
war. The logic of these choices is evident, for in dealing with women,

As Veena Das has pointed out to me (oral communication), uddha refers more
to purity of things and auca to purification processes, such as concern death and
menstruation. This essay was not written with such a distinction clearly in mind. It
would seem that it is auca more than uddha notions of purity that apply to Draupad,
but uddha notions that apply more to St. More work is thus needed on this point,
and I thank Veena Das for making this clear.
R. S. Khare, Culture and Reality. Essay on the Hindu System of Managing Foods
(Simla: Indian Institute of Advanced Study, 1976), p. 71 and passim.
See Dan Sperber, Rethinking Symbolism, trans. by Alice Morton (Cambridge:
Cambridge University Press, 1974), pp. 5254 and passim.
See Hiltebeitel 1980a, 1980b, 1980c, 198081, 1981.
purity and auspiciousness in the sanskrit epics 85

and with killing and death, we are dealing with two subjects that domi-
nate, perhaps more than any others, the Indian traditions ruminations
on the pure, the impure, the auspicious, and the inauspicious.
First let us deal with the women, Draupad and St. The discussion
of these two heroines representation of themes connected with purity
and auspiciousness can be designed around their portrayal through
three symbols: their garments, their hair, and in Sts case her jewels.
As I have dealt with these symbols elsewhere,8 a summary must suffice
here. In Draupads case, the pivotal scene is, of course, the dice match,
and the primary symbols are first the hair, and secondly the garments.
After Yudhist hira has reluctantly wagered her as his last stake in the
gambling match, she is dragged by her dishevelled hair into the Kaurava
mens hall (sabh). Her hair is dishevelled because she is menstruating,
and her impurity is accentuated by the revelation that she is wear-
ing but one garment, and that bloodstained (Mbh. 2.60.2532; 70.9).9
Moreover, she protests that she should not be brought before the men
in the sabh during this period of defilement. In this connection she
is not only impure but inauspicious.10
It is the accentuation on Draupads impurity, however, that charges
this scene with its great power and impact. But is she really impure?
Or, to put it differently, what does this impurity mean? First of all, it
is an oversimplification to say that menstrual blood is impure. For the
woman, at least, it is purifying and cleansing.11 Thus the outrage com-
mitted upon Draupad involves an interruption of her purification. For
a woman to resume her wifely duties and prerogatives, it is, moreover,
necessary for her to wash her hair after menstruation and bind it into
a braid. It seems clear from the Mahbhrata that Draupad avoids
doing this for the thirteen years that begin with this violation at the

Hiltebeitel 1980a, 198081, 1981.
One finds both classical as well as diffused ethnological prohibitions on hair-
braiding during menstruation; see Taittirya Sam hit 2.5.1, and P. Hershman, Hair,
Sex and Dirt, Man 9 (1974), 282283; cf. Gabriella Eichinger Ferro-Luzzi, Womens
Pollution Periods in Tamilnad (India), Anthropos 69 (1974), p. 128; Deborah Winslow,
Rituals of First Menstruation in Sri Lank, Man 15 (1980), 607608; see also idem,
608609 on changes of clothes after menstruation.
A womans dishevelled hair is a recognized symbol of inauspiciousness; see above,
n. 3 and the Tamil epic Cilappatikram 20, ven p 2: when the king of Maturai sees
the dishevelled heroine he dies.
See Winslow, Rituals of First Menstruation, p. 609: menstruation is sudu
veneva, becoming white or clean in colloquial Sinhalese; see also Hershman, Hair, Sex
and Dirt, p. 286; Hiltebeitel 1981, 203.
86 chapter four

dice match, and culminateafter her exile with her husbandsin the
battle of Kuruksetra. But she does not, in the epic, ever declare her
intentions to keep her hair thus dishevelled. And there is no mention
of her finally binding it up. There is thus neither an explicit vow of
dishevelment such as can be traced back to Bhravis seventh century
Krtrjunya (3.47);12 nor is there the further vow, traceable to Bhatt a
Nryanas eighth century Ven sam hra and now of pan-Indian popu-
larity, that she would keep her hair unkempt until she could braid it
with the blood of her tormentor Duryodhana (or of Duhsana, or
in some versions of both). Obviously these variant strands of tradi-
tion involve different shadings of the themes of purity and impurity,
and one must not be over-eager to arrange them into a chronology
or geography, as the Sanskrit epic itself seems to evoke, echo, or per-
haps anticipate post-epic and vernacular traditions).13 But it is clear
enough that the epic presents Draupad as wilfully undertaking a state
of extended symbolic menstrual defilement by continuing to wear her
hair loose during her twelve years of exile and her thirteenth year in
the ironic disguise of a hairdresser.
We have thus noted one way in which the epic story unfolds from
the issue of Draupads purity: it is not so much that she is impure, but
that her purification is interrupted in a way that requires the destruc-
tion of her violators. But there is a direct sequel to the hair-pulling that
presents a further development of the purity-impurity theme. This is
the famous disrobing scene, in which Duhsana attempts to remove
Draupads garment only to find it miraculously replaced by another,
and that in turn by others until he must give up in chagrin. The hair-
pulling and the saree-pulling are intimately related symbols, as Bhatta
Nryan a shows by repeatedly referring to them in a compound: the

Thus Bhravi has Draupad castigate her husband Arjuna as follows: My locks
hang loose, strewn with the dust of Duhsanas insult; having no one to care for them,
and left to the mercy of fortune, they reproach your valour. Ah, I hope you are still the
same Dhanajaya (winner of wealth) (duh sanmarsarajovikirnair / ebhir vinthair
iva bhgyanthaih / keaih kadarthkrtavryasrah / kaccit sa evsi dhanajayas
tvam)as translated and cited in Indira Viswanathan Shetterly (a.k.a. Petersen), Recur-
rence and Structure in Sanskrit Literary Epic: A Study of Bhravis Kirtrjunya,
Doctoral Diss., Harvard University, 1976, p. 187.
There are certain reasons to suspect that such themes find their way into the
Mahbhrata from southern or Dravidian influences. But the evidence is as yet
sketchy. See, however, Hiltebeitel 1981, 17981. Most scholars also regard Bhravi as
coming from the south.
purity and auspiciousness in the sanskrit epics 87

kembarkarsan a, the pulling of the hair and the garments.14 But

for our purposes, the most significant thing is that Duhsana begins
by pulling a garment that, as noted earlier, is bloodstained, and ends
up looking at a pile of sarees of many colors and white (2.544*).15
The Mahbhrata precedes this miracle with a debate over whether
Draupad is blameless, faultless (anindit) or, due to her poly-
andry, a whore (bandhak; 2.61.2336), and whether the Kauravas
have acted in accord with dharma by dragging her into the hall. The
miracle of the sarees is clearly an affirmation that she is blameless,
and that her polyandric dharma is beyond reproach. These are clearly
indications of her essential purity, something which henceforth the
epic audience understands even while hearing that she will wear her
hair dishevelled and don an impure guise during her period in dis-
guise. For in this period she begins by wearing a large black very dirty
garment (vsas krsn am sumalinam mahat; 4.8.2), and she assumes
a task that involves handling other peoples impurity as a hairdresser.16
The sequence of garments thus bespeaks the symbolisms of purity
and auspiciousness: the one blood-stained garment is replaced by gar-
ments that are white, an indication of her purity,17 and multicolored,
an indication also of her auspiciousness; and her dirty black disguise
is a means of concealing these essentially inviolate qualities.
If we turn now to Vlmkis portrayal of St, we may note at first
that her career parallels Draupads in many ways. Each heroine is
banished to the forest with her husband(s); each is violated in some
fashion by a male; in each case the violation spurs her husband(s) to
acts of revenge; and in each case the heroine is restored to queenly
status once that revenge is accomplished. Moreover, neither heroine,
once restored, retains her queenly status in its full glory. Draupad is
barren due to Avatthmans curse; and St suffers the reproaches of
the populace that leadat least in Vlmkito her second banish-
ment, this one by Rma. It is as if in each case the heroine has been
flawed by her earthly encounters. This is part of a larger pattern, found

See Hiltebeitel 1981, 18283, nn. 10 and 12.
See Hiltebeitel 1980a, 98101: I argue that this verse is part of the earliest level of
interpolation, earlier than the verses which say that Krsn a supplied the inexhaustible
On womens hairdressers roles after menstruation, see Winslow, Rituals of First
Menstruation, pp. 609610.
See idem, pp. 608609 on womens white garments and ritual purity.
88 chapter four

frequently in myths about the goddess, in which the human forms

which the goddess assumes, and the sufferings she endures, are taken on
as penance for some former fault, or in fulfillment of some former vow
of revenge. It might thus be formulated that the sufferings of heroines
in this world involve a necessary taking on of impurity, conditioned
by misfortunes of their previous lives as goddesses or saintly heroines.18
Despite such parallel episodes, however, St and Draupad are in
many ways each others opposites. And it is the symbols of hair, gar-
ments, and (with St) jewels that reveal this most clearly.19 Moreover,
the accent in Sts portrayal is not so much on the domain of purity
as on that of auspiciousness. These contrasts and their symbolic accen-
tuations emerge clearly from the parallel episodes just mentioned in
the two heroines careers.
First of all, although the scenes are similar, the first two occur in a
different order. For Draupad the violation precedes the exile; for St
the exile precedes the violation. This is a most significant difference,
for whereas Draupad enters the forest in a state of extended sym-
bolic impurity, St enters the forest with no question about her purity.
The main question that concerns Vlmki is how St can remain with
Rma, despite the reduced circumstances that the forest life will entail,
as a sumangal, an auspicious woman. His answer, in a scene that
has comic possibilities, is to first have Daaratha supply St with
enough garments and jewels to last her through her fourteen years of
banishment (Rm. 2.34.15); and then, toward the beginning of their
forest wanderings, to have the saintly Anasy supply her with addi-
tional accoutrements (2.110.1720; 111.1113). No one worries about
how this wardrobe is to be carried about. Perhaps it is but another
charge of the fortunate and ever uncomplaining Laksmana. In any
case, we are confronted with a heroine who enters the forest with her
purity assumed but her auspiciousness doubly guaranteed. She is not
disrobed, like Draupad, but robed to excess.
As to the violations, whereas in Draupads case the outrage is
focused on her purity, with St the central issue still remains her

For a suggestive initial discussion of this resonant theme, see Branda E. F. Beck,
The Goddess and the Demon: A Local South India Festival and Its Wider Context,
draft of paper to appear in Purusrtha 5 (1981), in press. By contrast, one will note the
rarity of suffering among heroes who represent Visn u and, to a lesser extent, iva.
This statement applies only to the Vlmki Rmyan a. The St of the Mahbhratas
Rmopkhyna is more like Draupad in these characteristics than she is like Vlmikis
Sta fact which I hope to take up in another article.
purity and auspiciousness in the sanskrit epics 89

auspiciousness. While Rvana bears her away to Lank, she drops

her auspicious (ubha) jewels and her auspicious (ubha) yellow
upper garment to mark the trail of her capture (4.6.9; 5.13.43; 52.2;
etc.) so that Rma will be able to find her. Thus whereas the violation
of Draupad involves a disrobing while she is impure, Sits abduction
occurs while she is fully auspicious. And while Draupads response
involves a wilful adoption of the symbolism of impurity to challenge
her husbands to revenge, St wilfully uses the symbols of auspicious-
ness to stimulate Rma. Most significantly, St retains a vestige of her
auspicious finery even during her captivity, and her status during this
time is that of a virahin woman separated from her husbandas
indicated by the wearing of her hair in a single braid (ekaven ).20
Here again St contrasts with the dishevelled Draupad. Moreover, it
is with her last jewel that she lets Rma know of her whereabouts by
sending it back with Hanumn for Rma to see it. It is no accident that
this jewel is her cdman i, the crest jewel with which she fastens her
hair and thus indicates what remains of her still auspicious virahin
status (Ram. 5.26.17; 36.52; 63.31; 65.30. etc.).
Further, as to the manner in which the two heroines challenge their
husbands, it is significant that Draupad does so by her constant pres-
ence among them. Indeed, she shares her experiences of impurity and
defilement with them, as we shall see. St, however, challenges Rma
by her absence, from afar, by speeches conveyed through Hanumn. It
is her separation and distance, not her companionship and presence,
that is the reproach to Rma. And Rma will in no way have to share
the impurity which her captivity inevitably entails.
As to their restoration, the Mahbhrata is largely silent about
Draupads reaccession as queen; but it is clear that, even though she
is barren, she shares the kingship with her husbands. The impurity she
has assumed and the vows (Mbh. 5.137.18) and tapas (9.4.18; 58.10)
she has undertaken are fulfilled even if they are not made explicit,
and are relegated largely to the past. And from here on she lives out
the rest of her life with them. On the contrary, when St is restored
to Rma, and hence eventually to the throne, it is only after she
has reluctantly fulfilled Rmas demand that she appear before him,
redeemed from her captivity, only with freshly washed and scented

A style of hair not really a braid (ven ): the hair is clasped singly in the back;
see Hiltebeitel 1981, 18485; 198081, 19899.
90 chapter four

hair and dressed up again in her auspicious finery (6.102.713). Here

the implicit question of her purity is finally addressed by the test
which Rma requires: her entry into the fire. Yet despite her emerging
from the fire unscathed, the question of her purity is only partially
resolved, relegated not to the past but reserved for the future as a sub-
ordinate (and seemingly ambiguous) issue among the deliberations
that lead to her second and final banishment. On that occasion, Rma
is convinced that St is sinless (appam; Rm. 7.44.6), of pure con-
duct (uddhasmcar; 8), and pure (stm uddhm: 9). But the
populace doubts this. What is crucial and predominant, however, is
Rmas concern that St be auspicious. It is this concern, and not one
for her purity, that leads Rma to decide on her banishment. Thus he
asks his ministers: What words, auspicious and inauspicious, do the
city folk say? Having heard, I will now do what is auspicious and not
what is inauspicious (ubhubhni vakyni ynyhuh puravsinah /
rutvednm ubham kurym na kurym aubhni ca; 7.42.10). Having
heard of the peoples rumors, the auspicious thing he does is to ban-
ish his wife, whom he knows to be pure. Moreover, the text shows
its concern for the auspiciousness of the scene by making it clear that
Rma has St escorted into her banishment by Laksmana of auspi-
cious marks (laksman a ubhalaksan a; 43.2)a frequent epithet for
Rmas brother (e.g. 6.79.1) which also reminds us that Rma is joined
in the forest not by just one representative of auspiciousness (St) but
two. Finally, St is replaceable in her role as representative of auspi-
ciousness in Ramas life. Once she is living in her banishment from her
husband, Rma replaces her in rituals by a golden statue (7.82.19).
Our discussion of the heroines thus reveals two different emphases
within the two epics. Whereas in the Mahbhrata the heroine and the
heroes follow a career where their purity and impurity are a central
issue, but one which involves them together in an integrated way and
includes themes of auspiciousness and inauspiciousness as well, in the
Rmyan a the heroines auspiciousness is revealed as indispensible to
the hero. But where the issue of purity is raised, it makes her inauspicious
even though she is pure. And the hero must thus repudiate her, his con-
cern for her auspiciousness being more fundamental than his concern
for her purity: and less for his wifes purity than his own. Whereas the
Mahbhrata heroes undergo impurity and face inauspiciousness like,
and together with, their wife, the hero of the Rmyan a must be kept
pure, and surrounded by auspiciousness, at all costs. This contrasting
purity and auspiciousness in the sanskrit epics 91

pattern can now be further examined with reference to the foremost

heroes of the two epics, Rma and Arjuna.21
Concerning the heroes, one finds further thematizations of the pure,
the auspicious and their opposites through symbols of hair and gar-
ments. But it is in connection with the wars in each epic that the issues
are most intensified. Each epic regards its battle as a sacrifice, and
the battles have many sacrificial overtones, including implicationsin
the Mahbhrata at leastthat those who are involved in killing take
on the impurity of bloodshed like the slaughterer and distributor of
portions of the animal in the Vedic sacrifice.22 Yet the battle, like the
sacrifice, also generates purity,23 andas Krsn a teaches Arjunakilling
can be aklista karman, unstained action, if done without the desire
for fruits. In fact, the term aklista karman is used frequently in both
epics, and primarily to characterize Rma and Arjuna.24 The battles
are also full of auspicious and inauspicious augurs, with the battlefield
itself suddenly and frequently shifting its aspect, in similes, from a
gore-begrimed river of blood to a host of stars on a beautiful moonlit
night.25 Here, however, we will focus not on the wars themselves, but
on the preparations for them. In particular, it is prior to the inevitabil-
ity of death and bloodshed in war that Arjuna and Rma each undergo
a sequence of transformations and interactions that involve them most
profoundly with the domains of purity and auspiciousness. We thus
turn to Arjunas disguise, donned during his period of concealment,
and Ramas passage through the hermitage of Matanga to his alliance
with the vultures, monkeys, and bears.26 Let us anticipate, at the out-
set, a sort of compensatory pattern. Whereas Draupad is concerned
primarily with purity and secondarily with auspiciousness, Arjuna is

I limit myself here for the most part to what concerns Arjuna. As far as themes
of impurity go, most of what concerns him also concerns his brothers; see Hiltebeitel
1980b, 16874.
See idem, p. 170.
This is indicated nicely by Indira Peterson, Recurrence and Structure, pp. 318319
(see n. 12 above) citing Mahbhrata references to war as sacrifice and the idea that
the warrior is purified by his weapons (14.60.23; 15.44.9).
On Arjuna, see Hiltebeitel [1976] 1990, 237, citing e.g. 3.39.1. For Rma see e.g.
Rm. 2.21.11.
See Hiltebeitel 1980a, 108.
I share the suspicion of J. A. B. van Buitenen, trans. and ed., The Mahbhrata,
Books 2 and 3 (Chicago: University of Chicago Press, 1975), Vol. 2, p. 835: while men-
tion is made on occassion of bears (rksa) the classification of monkeys is so predomi-
nant that one might wonder whether these bears are not really a kind of monkey.
92 chapter four

concerned primarily with auspiciousness and secondarily with purity.

And whereas St is concerned primarily with auspiciousness and sec-
ondarily with purity, Rma is concerned primarily with purity and
secondarily with auspiciousness. One can thus chart the accentuation
of our two domains in the two epics as follows:

Mahbhrata Rmyan a
primary secondary primary secondary
Draupad purity auspic. St auspic. purity
Arjuna auspic. purity Rma purity auspic.

As Madeleine Biardeau first saw,27 the year which the Pndavas spend
in disguise is presented in terms of the symbolism of the dks, the
consecration preparatory to a sacrifice: in this case, the sacrifice
of battle. There are several allusions to the Pndavas having become
like embryos in the womb during this period.28 The womb-like con-
dition of the dks is thick with associations with danger, death, and
impurity,29 and these affect the Pndavas and Draupad alike and in
general. Arjuna, in fact, seems most sensitive to such themes when he
tells Draupad of the unsurpassed suffering he has experienced while
adopting his disguise in the animal womb of Matsya, the kingdom
whose name means Fish.30
But whereas all the Pndavas encounter and take on impurity dur-
ing this period in disguise, Arjunas disguiseand the activities which
it involves him inare uniquely focused on themes of auspiciousness.
Arjuna dons the disguise of a transvestite/ eunuch/ bisexual named

See Madeleine Biardeau, tudes de mythologie hindoue (IV), Part II. Bhakti et
avatra, Bulletin de Lcole Franaise dExtrme Orient 63 (1976), pp. 207208; idem,
tudes [as above] (V), Part II. Bhakti et avatra, Bulletin de lcole Franaise dExtrme
Orient 65 (1978), pp. 149157, 187188 (n. 3). See also Hiltebeitel, Disguises, p. 149
and passim.
See not only the Virtaparvan passages mentioned in Hiltebeitel, Disguises, p. 149,
but also 9.55.30. The latter alyaparvan passage adds discouragement to the view that
the Virtaparvan is an interpolation; see Hiltebeitel, Disguises, p. 148, n. 4.
See Hiltebeitel, Draupads Hair, p. 194, n. 47.
See Hiltebeitel, Disguises, pp. 149 and especially 161, citing 4.23.23 where Arjuna
tells Draupad that she (Arjuna the eunuch/ transvestite) has gone to an animal
womb (tiryagyonigat) during his/ her painful period in disguise.
purity and auspiciousness in the sanskrit epics 93

Brhannad.31 As pointed out elsewhere, eunuchs have very ancient

functions in India.32 One of these, appearances at births, can be traced
to Atharvaveda 8.6. Another function, which might seem more recent
until one looks at the evidence for it in the Mahbhrata, is to appear
at weddings. In both of these circumstances, as well as more generally,
there seems little doubt that eunuchs are regarded as impure.33 But
what is interesting is that despite their impurity, their appearances at
births and weddings can be either auspicious or inauspicious. They
hold the power to bless or curse, and, in todays India, they may bless
with song and dance if appropriately paid, or may shower abuse and
gesture with obscenities until paid to leave. But even paying for their
removal is a way of assuring an auspicious result.
In Arjunas case, his eunuch role is clearly focused on auspicious-
ness. This is true even though, as Draupad tells him, his disguise is
one that is despised by the world (4.8.11). More particularly, it is as
a eunuch that he oversees the auspicious outcome of the wedding and
childbirth that result in the sole continuation of the Kaurava-Pndava
line: the marriage of his son Abhimanyu to the Matsya princess Uttar,
and the birth of their son Pariksit. First, as regards the marriage,
it is through his eunuch disguise that Arjuna has access to the royal
womens quarters, and particularly to Uttar. This access enables him
to be her instructor in the clearly auspicious arts of song, dance, and
music, arts which make her an ideal auspicious bride. And when
finally his identity is disclosed and he is offered Uttars hand in mar-
riage for himself, it is his eunuch disguise that gives Arjuna the pretext
for insisting that Uttar should marry not himself but his son: their
year spent together in asexual intimacy has made her like a daughter
to him; moreover, his adopted asexuality is a guarantee of her chastity,
and thus of her suitability as a bride for his son. Secondly, Arjuna the
eunuch also prepares Uttar for her role as a future mother. When
he sets off to defend the Kingdom of Matsya against a Kaurava cattle
raid, Uttar asks Arjuna to bring her the Kauravas garments for her
to dress her dolls (4.35.23). The garments Arjuna brings back for

See Hiltebeitel, Disguises, pp. 154156 on the ambiguities of Arjunas sexual
identity during his period in disguise.
See idem, pp. 161168.
On the proverbial impurity of eunuchs, see Mbh. 9.30.7071, and Morris Carstairs,
The Twice-Born: A Study of a Community of High Caste Hindus (Bloomington:
Indiana University Press, 1961), pp. 5961, 307, 323.
94 chapter four

her are the white robes of Drona and Krpa, the yellow robes of Karn a,
and the blue ones of Avatthman and Duryodhana (61.13). Once
again, as at Draupads disrobing, we have multicolored garments that
evoke purity (white) and auspiciousness (yellow, blue). It is clear that
in dressing her dolls with such garments, Uttar prefigures her role
as a mother-to-be. And the link with Draupad is intensified by the
ambiguity of the term used for dolls. Uttar asks for the garments
pclikrtham. This can be rendered for the sake of (my) dolls, or
for the sake of Draupad [she who comes from Pacla]. The auspi-
cious connotations of the garments which Arjuna the eunuch secures
for Uttar are thus a prefiguration of her becoming the mother of the
son who will continue the lineage, a privilege which Draupad, sonless
and barren after the war, must eventually relinquish. Thus the multi-
colored garments pass, in effect, from Draupad to Uttar. And while
Uttar becomes the focus of such auspicious prospects, the garments
which Draupad wears are the impure dirty black ones of her hair-
dresser disguise.
It thus appears that despite the impurity of his eunuch disguise,
Arjunas functions in that role are persistently auspicious: music, song,
dance, marriage, and birth.
In the Rmyan a, the preparations for battle which follow upon
the abduction of St are filled with equally bizarre encounters, and
equally significant evocations of the domains of purity and auspicious-
ness. It may also be that these encounters can be illuminated by a
similar correlation with the theme of the dks. But the Rmyan a has
its special twists.
Most of St, Rma, and Laksman as period of banishment in the
forest is a relative idyll. For the better part of their forest tenure, they
follow a southward path from one rsi s hermitage to another until
they reach that of the last of the great rsi s, Agastya, who is associated
with the southern boundaries of Aryandom, beyond which Rma is to
meet with sudden disasters. Not surprisingly, their path throughout
is described as auspicious (ubhe vartman i; 6.70.14). But the disas-
ters Rma encounters once he leaves Agastyas hermitage are indi-
cations that in Rmas career auspicious things sometimes take on
strange forms. Not only will St be abducted (something over which
the gods and rsi s rejoice); Rma will be guided and accompanied by
a bizarre concatenation of pure, impure, auspicious, and inauspicious
purity and auspiciousness in the sanskrit epics 95

To begin with, once he leaves Agastyas hermitage, he is guided to

Pacavati, where St will be abducted, by the vulture Jatyus. Not
surprisingly, vultures, no matter how friendly, are listed among inaus-
picious creatures.34 Then, once St is abducted, Rma must enter
Matangas Wood, where the rsi Matanga formerly had his hermitage,
in order to meet his monkey ally Sugrva and his monkey companion
Hanumn. First, quite surprisingly, given their popularity and promi-
nence in the epic, the Rmyan a seems quite definite in classifying
monkeys as inauspicious. When Rma is about to leave for the forest,
his mother Kausalys tearful parting words include this wish: May
there be no monkeys (plavag), scorpions, gnats, mosquitoes, reptiles,
or insects in the deep forest for you (Rm. 2.22.6). It is clear that
the concern is that Rmas path may be auspicious: Kausaly contin-
ues to list other animals which she hopes will not harm Rma, and
concludes may your arrivals be propitious/auspicious (gamste
ivh santu; 22.9). Further, when St, held captive, first sees Rmas
monkey messenger Hanumn, she thinks she must be dreaming and
regards the vision of a monkey as an ill omen: This I am surely
dreaming now, a disfigured (vikrtah ) monkey, a vision forbidden by
the stras (5.30.4ab). Hari Prasad Shastri translates vikrtah here
not as disfigured but inauspicious,35 which is clearly implied by
Sits next words: May all be well (svastyastu) for Rma, Laksman a,
and her father Janaka (5.30.4cd). St also soon adds, contemplating
the surprise she feels in finding Hanumns company pleasant rather
than disagreeable: Surely having seen a monkey in a dream is not
conducive to increase, but my increase is obtained (svapne drst v
hi vnaram na akyo bhyudayau prptum praptacabhyudayo mama;
5.32.21). The term here twice translated as increase (adhyudaya) is
certainly within the auspiciousness domain, providing thus further
evidence for the consistent inauspiciousness of the appearance of
monkeys. It is thus within this context that we should interpret
Hanumns first appearance before Rma in the disguise of a mendi-
cant (bhiksurpa; 4.3.3). It would be inauspicious for him to appear
as a monkey. Not only that. Hanumn soon turns out to speak perfect
(one might say pure) Sanskrit.

See e.g. Rm. 6.40.31.
Hari Prasad Shastri, trans., The Rmyan a of Vlmki, 3 vols. (London: Shanti
Sadan, 19621970), vol. 2, p. 408.
96 chapter four

As regards the rsi Matanga, though the Rmyan a does not say this
outright, he is well known from other sources as a Candla, an out-
caste. He has a place in the southern heavenly mansion (sabh) of
Yama beside Agastya, Time, and Death.36 Matangas associations with
impurity are thus multiple. Yet it is in the hermitage of such a rsi , and
in the extended surroundings of the hermitage on mount Rsy amka,
that Rma confronts the domains of purity and auspiciousness in a
fashion that reorients him toward his divine mission: the destruc-
tion of Rvana, now required by the latters abduction of St. This
reorientation, a sort of rebirth for Rma, thus holds some parallels
to the Pndavas sojourn in the womb of Matsya. As can be seen
already, Matangas hermitage is associated with death. In both epics,
one also finds the symbolic encounters with animals: the Pndavas in
the womb of the fish, Rma guided by his associations with vultures,
monkeys, and the bones of a buffalo demon named Dundubhi. It is
here that Rma makes his pact with Sugrva to kill Vlin in order to
obtain Sugrvas aid in recovering St. Let us thus look more closely
at the sequence of events in Matangas Wood and hermitage.
When Rma and Laksman a arrive at Matangas hermitage, they learn
that all the rsi s there (including Matanga) have passed away. But a men-
dicant woman named abar remains behind at Matangas instructions
to tell Rma briefly about the hermitage and its rsi s and then, benefit-
ting from Rmas sight, to ascend to heaven (3.70.1927). Before he
leaves this hermitage to seek out Sugrva on mount Rsy amka, Rma
performs ablutions there which he describes as destroying inauspi-
ciousness (aubha) and initiating good fortune (kalyn am; 3.71.45).
His visit to this hermitage is thus connected with the turn from inaus-
piciousness to auspiciousness, but also with an act of purification. All
this purity and auspiciousness are most paradoxical, because the place
itself is one defiled by a buffalo sacrifice, a specifically non-Vedic and
impure rite, that is linked with the story of Matangas departure from
his hermitage and Sugrvas refuge in its forests.37
The good or auspicious turn of fortune which Rma senses is,
of course, his forthcoming meeting with Sugrva and Hanumn on
mount Rsy amka. There the monkey exiles are protected from Vlin,

See Hiltebeitel, Rma and Gilgamesh, p. 204, citing Mbh. 2.8.26.
For a fuller discussion of what follows, see Hiltebeitel, Rma and Gilgamesh,
pp. 200211.
purity and auspiciousness in the sanskrit epics 97

Sugrvas brother, by Matangas curse. Here Vlin had formerly dis-

posed of the impure remains of the buffalo demon Dundubhi. This
buffalo asura had challenged Kiskindh, the monkey capital, provok-
ing Valins defense. Their fight lasted until Vlin grabbed Dundubhi
by the horns and crushed (nispista ) him until the blood flowed from
his ears. Vlin then hurled off the body: When it was thus impetu-
ously propelled, drops of blood from the wounds fell out from the
buffalos mouth and were lifted up by the wind toward Matangas
hermitage. Having seen them fall there, the sage, covered with drops
of blood, uttered a great curse against the hurler Vlin: This [hermit-
age/wood] here may not be entered by the thrower [of this carcass],
for upon entering it he will die (4.11.4042). As I have argued else-
where, the details of this buffalo slaying, including especially the role
of the outcaste rsi Matanga in handling the defiling blood of the
buffalos remains, are evocative of ritual details in the sacrificial cult
of the water buffalo.38
Vlin can thus not harm Sugrva while the latter remains in Matangas
Wood, on mount Rsy amka. Therefore, despite the places associa-
tions with impurity, it is auspicious for Rma as he has been told he
will need Sugrvas help in finding St. Mount Rsy amka is also where
St, seeing the monkeys below, had dropped her auspicious jewels and
garment to mark the trail of her abduction by Rvana. But Sugrvas
aid does not come to Rma without further indications from Vlmki
as to how these encounters with the ambiguities of purity and auspi-
ciousness affect the hero. It is certainly significant that Rma should
form his auspicious pact with Sugrva in such an impure spot. It is
a pact involving death, and not only that of the buffalo Dundubhi, but
the deaths of Vlin and Rvana. First, Sugrva points to Dundubhis
remains, a great pile of bones that shines like the peak of a mountain,
which Rma disdainfully kicks off to a distance of ten yojans with his
big toe. Rma does this ostensibly to impress Sugrva that he is Vlins
superior, but Sugrva naturally points out that when Vlin killed Dun-
dubhi the buffalo had flesh, making Vlins feat greater than Rmas
(4.11.4652). Yet Rmas dispensing of the bones is also a complemen-
tary act to Vlins hurling of the carcass, and thus one of a number of
indications that Rma and Vlins careers are to be viewed in parallel.

See idem, pp. 188211, especially pp. 204206, 210211.
98 chapter four

Indeed, Rma is in many ways Valins human counterpart:39 senior

brothers whose throne a junior brother has usurped and/or contested,
rightful rulers, descendants of the Sun, conquerors of demons while
in exile from their thrones, husbands whose wives are of questionable
(at least to some) fidelity. It would thus appear that in killing Vlin,
Rma is killing an animal substitute for himself. If so, this can be related
to the situation in the dks where the sacrificer undergoes symbolic
death and self-offering preparatory to performing as a sacrificer.40
What is most striking in these episodes in the hermitage and wood of
Matanga, however, is that though Rma clearly enters a symbolic realm
of death that is thick with themes of impurity, he is never recognizably
sullied by it as the Pndavas and Draupad are during their year in
disguise. The Vlmki Rmyan a takes considerable pains to maintain
Rmas purity. He never has to meet the Candla rsi Matanga, who
is indispensible to the buffalo-killing scenario.41 He touches only the
bones of this animal, and that only with his big toe! He has numerous
dharmic rationalizations for his killing of Vlin. His encounters with
such inauspicious animals as vultures and monkeys turn out to be
auspicious. But Rma is clearly unaffected by their potential impurity.
A vulture helps him from afar. He does not, like Yudhist hira, incorpo-
rate the name Heron (Kanka) into a disguised identity,42 or like all
the Pn davas suffer in an animal womb. Indeed, Rma never has to
enter Kiskindh, the monkeys cave city, whereas the Pndavas must
enter the city of king Virta to live like infants in the womb taking
on impurities as a condition of their rebirth. Yet Rma does spend the
rainy season outside of Kiskindh in a cave atop mount Praravana,
awaiting the cessation of the rains and the time suitable for war.
In Rmas case, then, the accent is clearly on his essential and
inviolable purity. He takes on no impurity, like Arjuna. Where it is

Of course, so is Sugrva. The two monkeys are animal doubles not only of con-
flicting aspects of Rma, but of Rmas relations with his own brothers. For this same
point viewed from a different angle, see J. Moussaieff Masson, Fratricide and the
Monkeys: Psychoanalytic Observations on an Episode in the Vlmkirmyan am,
Journal of the American Oriental Society 95 (1975), pp. 672678.
See Madeleine Biardeau, as cited in n. 26 above, and idem and Charles Malam-
oud, Le sacrifice dans lInde ancienne (Paris: Presses Universitaires de France, 1976),
p. 36.
This non-meeting with Matanga stands in marked contrast with the important
meetings with other rsi s earlier in Rmas forest wanderings.
On Yudhist hira and Kanka, see Biardeau, tudes (V), pp. 99101 and 104;
Hiltebeitel, Disguises, pp. 169170.
purity and auspiciousness in the sanskrit epics 99

encountered it is removed with the least possible contact. Yet there is

considerable paradoxicality in Ramas encounters with themes of aus-
piciousness. There seems to be an assumption that everything which
surrounds Rma must be auspicious, beginning with his wife and his
brother Laksman a, and even including such proverbially inauspicious
creatures as vultures and monkeys. The repetitive lists of auspicious
omens that greet Rma in the forest and in battle are but another
example of this intention. In some cases, such as in his forest encoun-
ters, it would seem that it is Rmas purity that renders things around
him auspicious; in others, such as with St, it would seem to be his
standards of purity that require her to be auspicious. In either case, the
Rmyan a seems to treat Rmas auspicious surroundings as a support
system for his purity.
To summarize, one has no trouble identifying instances in the two
epics where the domains of purity and auspiciousness have conver-
gent focalizations, but different accentuations. What is pure can be
inauspicious, as with Draupad. What is impure can be auspicious,
as with Arjuna. Rmas purity and Sits auspiciousness are sustained
throughout the Rmyan a, with only the slightest hints that Rma
might come into contact with the impure, and only the post-war, and
possibly late,43 developments concerning Sits inauspiciousness.
Where the negatives enter the picture in the Rmyan a, however, it is
in terms of Rma remaining pure but aligning himself with inauspi-
cious forces, and St remaining auspicious but being suspected of
impurity. One sees how differently the two epics treat these matters.
The Rmyan a idealizes its figures and presents the negatives only
through hints and innuendoes, and in instances which are largely
beyond the hero and heroines control. The Mahbhrata confronts
the ambiguities directly, and presents a hero and heroine who assume
their impure and inauspicious aspects by choice. Thus there can never
be any real question of Draupad or Arjunas purity or auspiciousness,
despite the extremes to which they go in donning opposite garbs.

It is frequently argued that the seventh book of the Vlmki Rmyan a is a late


In the immediate sequel to the famous dice match episode of the

Mahbhrata in which the heroine Draupad is the last stake wagered
and lost by her Pn dava husbands, her Kaurava captors first drag her
by the hair and then order her disrobing. This double theme involving
Draupads hair and sarees has preoccupied me for some time. My ini-
tial concern was to explore its rich cosmological and theological impli-
cations in classical sources.1 In this chapter I would like to address
some folkloric material bearing on the same twin subjects, and dis-
cuss it toward some additional ends, taking up the wider issue of pan-
Indian Mahbhrata folklores, and raising the question of the relation
between the distinctly Tamil folklore about Draupad that is found
in her cult and wider pan-Indian themes. Are Tamil and other south
Indian Mahbhrata folklores (some of which are almost certainly
older than the Draupad cult) a source of diffusion for similar themes
found elsewhere in India? Or does the classical epic just suggest com-
mon folk responses? Is there a sort of underground Mahbhrata,
one that is perhaps even reflected in the Sanskrit epic itself but also

In 1979 I wrote an article (Hiltebeitel 1979b in which I tried to bring into one
focus a discussion of Draupads sars and hair, themes that were treated separately
in Hiltebeitel 1980a and 1981. Some of the cosmological and theological formulations
worked out in those papers remain significant for the present discussion. I argued that
the epic Draupad is already an image of the goddess in her totality: not only as r-
Laksm, whom she explicitly incarnates, but as Bhdev (the goddess Earth), Klartri
(the Night of Time), Mla-Prakrti (primal matter), and with intimations of Durg
and Kl; in her relations to Visn u-Krsn a and to figures linked with iva; in her role
with respect to the turn of the yugas and the relieving of the Earths burden; in rela-
tion to the Earths potential dessication (it is the solar Karna who orders Draupad
stripped) and resubmergence at the occasional pralaya, the pralaya at the end of a
kalpa that burns and dissolves the Earth as the nucleus of the triple world through
the agencies of fire, wind, and water; and in connection with symbols that portray the
further potential for the unmixing and unleashing of all the elementsearth, water,
fire, air, and etherthat is prelude to the final dissolution or prkrta pralaya with its
unbraiding of the strands or gun as of matter (see especially 1981: 21011, which is
found now in Part C of chapter 1 above). I will argue here that Draupads sarees and
hair retain such theological and cosmological resonances in Draupads cult folklore,
especially in connection with the elements earth, fire, and fluids (a triad of elements
that also finds vivid staging in the Peter Brook Mahbhrata).
102 chapter five

different from it in certain basic accentuations concerning the god-

dess? What are some of the features that distinguish Mahbhrata
folklores from other Indian folklores? And how are folkloric themes
concerning the Mahbhrata related to distinctive modes of transmis-
sion and performance? How and why is the Mahbhrata linked with
certain regional folk epic traditions and not with others? And in such
regional folk epic traditions where there is a connection, how do we
understand their portrayals of virgin heroines at the center of conflicts
over land? Draupads sarees and hair provide a fitting entre into the
problematics of such questions.
That the Sanskrit Mahbhrata is itself a repository of folklore has
not, of course, gone unnoticed. But little consistency has emerged
from the appreciation of this fact. Walter Ruben (1944), for instance,
used a comparative method to try to explain away all the episodes
involving Krsn a as folkloric intrusions. N. B. Patil (1983) claimed the
status of folk motif for virtually any narrative element in the epic that
caught his eye. And Georges Dumzil tried to distinguish mythic, epic,
and what can be called folkloric levels in certain episodes: for instance,
at Draupads svayam vara, the myths that explain her polyandry by
the intercessions of iva in her former lives; the main epic-heroic
narrative itself; and the romanesque fatality of Kunts mistaking
Draupad for alms, and telling the Pndavas, her sons, to share it all
equally (1974: 10910). Though I once looked at these matters much
as Dumzil did (Hiltebeitel [1976] 1990: 2733) and would not dispute
that there may be some value in retaining such levels as heuristic
devices, I am now sensitive to their being rather arbitrary in general.
Moreover, in the specific case of the Mahbhrata, they become means
to deceive ourselves into missing the works integrity.2
In this vein, much has been gained recently by A.K. Ramanujans
placement of epic at the puram end of the folklore spectrum that runs
from akam, the interior of heart and home, to the puram exterior
of the public arena. As Ramanujan shows, at this puram pole, epic
has affinities with drama, and introduces tragic modes that are not

Dumzil treats these three levels hierarchically and, in effect, chronologically, with
myth the most privileged and formative, epic its transposition, and the romanesque
as afterthought, introduced to tie the other two together. There is no reason to accept
either the chronological or the hierarchical implications of such formulations in the
supposed formation of the text. As concerns Mahbhrata folklore, the categories
themselves would seem to collapse, or at least be continually permeable.
the folklore of draupad: srs and hair 103

found in akam genres (1986: 4151, 7172). All of this is very sug-
gestive for interpreting Mahbhrata folklores and their modes of
transmission and performance. Even the Sanskrit Mahbhrata itself
is proverbially not to be read in the home, as it arouses family con-
flict; whereas the Rmyan a, by way of contrast, should be read in
the home because it portrays ideal family conditions.3 And beyond
the Sanskrit Mahbhrata, one finds further suggestive ways in which
Mahbhrata folklore has distinctive features in relation to drama, and
to other mythic and epic folklores. According to Gustav Oppert
(1893: 9798), whereas the Rmyan a is favored above all by Brah-
mans, the Mahbhrata is adopted by dras.
The ways in which different folk dramas are patronized in India
suggests such a pattern. Best known are the Rs Lls of north India,
which are about Krsn as youth among the cowherds (Hein 1972;
Hawley 1981), and the Rm Lls, also from the north, performed in
connection with the fall festival of Dasar (Schechner and Hess 1997;
Schechner 1983: 238293). Mahbhrata drama cycles are less com-
mon and less well known. The Rs and Rm Lls seem to be largely
the expressions of high caste traditions of Brahmans and kings and of
pan-regional values.4 Mahbhrata dramas are sometimes sponsored
at these levels, as for instance with the kathakali and ktiytta m (Zarilli
1984). But they are sometimes also found with deeper roots in more
regionally intensive, cult-specific forms, tied in with the values of lower
castes, and in particular the dominant landed castes that sponsor most
of the festivals in which the plays are performed. Most similar to the
mix of drama, recitation, and ritual found in the Draupad cult are the
so-called Pn dav Lls celebrated in the mountains of sub-Himalayan
Garhwal. From the recent fieldwork of William S. Sax, we know that
the Pn dav Lls, or Pndava Plays, involve ritual dramas sponsored
by dominant caste Rjpts or Ksatriyas (warriors), who claim descent
from the Pndavas and regard their region as one in which certain epic
eventsthe Pndavas births, their Himalayan ascent to heaventook
place. The Pndavas are regarded as personal deities (ista devats).
The Lls are ways of worshipping Kl, for Draupad is regarded as

This is a well-known maxim, but I can cite it only from oral traditions such as the
mention of it by Robert P. Goldman and A.K. Ramanujan at different conferences.
The chief boy actors in the Rm Ll at Ramnagar (Varanasi) must be brahmans,
well-behaved, with good looks (Schechner 1983: 265); indeed, the Maharaja himself
is a Brahman (266).
104 chapter five

Kls avatra. They enact epic scenes, include a mix of bardic recita-
tion and dance-drama performance, induce possession by both actors
and audience, and evoke animal sacrifices through portrayals of death
scenes (Sax 1986; 1987).
Such features are all paralleled in the south Indian Draupad cult.
Draupad festivals are sponsored mainly by Vanniyars, a regionally
dominant landed caste in the South Arcot District of Tamilnadu where
the cult is centered. Vanniyars who are connected with the Draupad
cult claim to have originally been Ksatriyas and, like the Pn davas, to
have Draupad as their kuladevat or family deity. Draupad cult
martial traditions are in fact linked up with the royal fort and regional
mythology of Gingee, the medieval Nayak kingdom or viceroyalty
under the Vijayanagar empire. For instance, Gingee is often regarded
as the site where Draupad took a second birth to help one of Gingees
ancient kings, a close descendant of the Pn davas, to overcome the
demon of the Gingee Forest. Though Draupad is an incarnation of
r, the goddess of prosperity, as she is in the classical epic, she takes
on the form of Kl (Klrp) as well as traits of Durg in gaining
revenge for her violation by the Kauravas. Her festivals combine local
ritual, recitation of the Mahbhrata in Tamil, and nightlong plays. At
the largest festivals, there may be over sixty days of off and on ritual,
fifty afternoons of recitation, and eighteen nights of dramas. Shorter
festivals combine these elements in smaller but similar proportions,
also frequently emphasizing the number eighteen in some fashion (for
instance, eighteen days of recitation, or ten days recitation plus eight of
dramas), since the festival commemorates the epics eighteen-day war.
The ways that the Mahbhrata roots itself at these social and regional
levels need to be considered in connection with the way certain folk
epics in other parts of India link themselves with the Mahbhrata,
and involve similar mixes of drama, bardic recitation, and hero cult
ritual. I will return to this problem, or set of problems, in closing.

Draupads Hair and Sarees in Her Tamil Cult

In the Draupad cult, images of Draupads hair and sarees cannot be

dissociated from the primary natural elements of earth, water, and
fire.5 In what follows, one must keep in mind the cults ritual cycle,

Let me insist again on the continuity between classical and folk perceptions of the
Mahbhrata; cf. note 1.
the folklore of draupad: srs and hair 105

with its sequence of ceremonies that involve earth trampling, fire

trampling (tmiti), the crossing of a ditch of water called the milk
pit (pl kuli), and in which women who cross the coals often wear
brilliant yellow (macal) sarees and loosened hair. The mythology of
the Draupad cult fire-walk bears out the relevance of these images,
for among the accounts of how Draupad protects her devotees as they
cross the coals, one hears sometimes that she does so by throwing the
end of her sar over them, and sometimes that she does the same with
her unseen flowing hair (Hiltebeitel 1988a: 437).
The twin themes of hair and sars had already coalesced into a reso-
nant symbolic unit in Bhatta Nryanas Sanskrit drama Ven sam hra,
The Binding-Up of the Braid, probably from the early eighth cen-
tury. For with Bhatta , two things happen that recur in the Draupad
cult but are not found in the Sanskrit epic. First, Duryodhana replaces
Karna in ordering the disrobing, so that he becomes responsible for
commanding the violations of both the sars and the hair. And second,
Bhatt a initiates the device of alluding to the two violations together,
recalling them jointly and repeatedly, at least once in every act, as the
kembarkarsan a, the pulling of the hair and the garments (Hilte-
beitel 1981: 183 and n. 12).
This double accentuation recurs in the Tamil Mahbhrata of
Villiputtr l vr, a text composed around 1400 within the area of the
Draupad cult heartland, at the very point where Draupad makes her
celebrated vow:
He who brought me without fearing into the court of kings, touching
my saree, touching my hair [tukil tn ti alakam tn ti ] which was made
fragrant, its garland surrounded by swarms of bees, kingsnot until
the victory drums roll on the battlefield, having cut off [their] crowned
heads, smelling of raw meat, the hot blood falling, will I take up and bind
my dishevelled hair. (Villipratam 2.2.255)
As Richard Frasca (1984: 145), David Shulman (1985: 14) and I (1988:
198) have all sensed, the Villipratam is quite possibly responsive to
early developments of the Draupad cult itself. It is in any case most
definitely responsivelike the largely lost ninth-century Makpratam
of Peruntvanrto prior south Indian Mahbhrata folklores (Hilte-
beitel 1988: 15 and passim; Venkatesa Acharya 1981: 30731). More
to the present point, it is the text that is usually used for recitation
and exposition of the epic by the pratiyrs or prata piracankis, the
professional itinerant bards who sing from and expound upon this
Mahbhrata at Draupad festivals. And it is further the version of the
Mahbhrata that stands most directly behind the terukkttu dramas
106 chapter five

that are performed at Draupad festivals. I will not attempt here to

illustrate the many and often contrasting ways in which both the
pratiyrs and the dramatists introduce folk themes that go beyond
Villis classical text (itself enriched by folk traditions, as just noted).
Suffice it to say that just as each pratiyr makes his own folk-classical
weave (with variations sometimes stimulated by the particular require-
ments of local festivals), so does each drama vary in the same way,
with some far closer to Villi than others, but here again differing from
troupe to troupe, performance to performance, and also from printed
to handwritten to oral-performed variants. The drama that deals most
centrally with Draupads sarees and hair, Dice Match and Disrob-
ing, has a printed chapbook version by a poet named Irmaccantira
Kaviryar (1977), who lived sometime in the early nineteenth century
and seems to have played a major role in launching the transforma-
tion of this folk drama tradition into a chapbook folk literature (Hilte-
beitel 1988a: 15767). I cannot go here into the variance from such
chapbook texts that is found in performance; although it is consid-
erable, performed versions do seem to retain basic elements of the
structure of the chapbook versions as well as many songs and prover-
bial images.6 Irmaccantira Kaviryar stays closer to the Villipratam
in Dice Match and Disrobing than he does in certain other dramas
performed at Draupad festivals that he also authored. And in general,
he stays closer to Villi than other authors within the same genre. With
these points in mind, let us observe some of the ways that he situates
this classical theme in his folk theatre composition. Taking things in
order, we will look first at the hair pulling, then at the saree pulling,
and finally at the vow. Throughout this discussion, one must keep in
mind not only Irmaccantira Kaviryars text, but details and variations
in performance. The quite skeletal text tells us virtually nothing about
staging and props, matters that are clearly decisive for our subjects.
According to Irmaccantira Kavirayrs text,7 Duhsana seizes
Draupad by the hair. Actually, in performance he swings her by a
stick or a length of wound cloth that each holds at opposite ends,
Draupad holding the end of her hair to her end of the stick or cloth

Frasca 1984: 1400, translates a performed version of the first part of this play. I
discuss the relation between this performed version and the corresponding portion of
Irmaccantira Kaviryarss text in Hiltebeitel 1988a: 228, n. 7; 230231.
I thank Pon Kothandaraman of the Department of Tamil Literature, University
of Madras, for his help with my reading of these passages. Responsibility for all errors
is, of course, my own.
the folklore of draupad: srs and hair 107

to indicate that it is her hairs extension. The Duhsana actor will

often pause at this point to make a gesture of worship (prayer, devo-
tional salute, sometimes even lampwaving or dprdhan) toward
Draupads image (brought on a tr or chariot to watch the play) to
beg the goddesss forgiveness for this violation that he performs only
out of professional duty (Hiltebeitel 1988a: 234). Once Duhsana has
thrown her around the stage and to the ground, Draupad sings a deep
and defiant lament (pulampal), one that Irmaccantira Kaviryar indi-
cates is precisely the lament of Draupad when Duhsana touches
Draupads hair (turctanan turpataiyai mayirai totta ptu turpatai
pulampal; 1977: 67; henceforth referred to as CTU for the plays Tamil
title, Ctu-Tukilurital, Dice Match and Disrobing).
Alas, [I am] a sinner, O god
Can this injustice be, O god?
My kntal [hair knot] that has been decorated and stroked [or loosened;
kti] by my lords [my husbands/kings; mannavar], can this fool touch
it, O god?
My kntal that is praised and touched by my war heroes, can a low
person [pulaiyan] touch it, O god?
My kntal that is touched and unfolded by my husbands [loved ones;
kan n lar], can a base person [katalyan] touch it, O god?
That kntal that is arranged and stroked [kti] by my battle heroes, can
a fool touch it, O god?
My kntal that good pure heroes adorn with flowers, can an evil person
touch it, O god?
The kntal that smells sweetly from being adorned with fresh jasmine,
alas, it rolls in the dirt [man ], O god.
The kntal to which I applied campanki oil, alas, it rolls on the earth
[tarai], O god.
My kntal that shines after running through it with a comb, alas, it rolls
on the street [teru], O god. (CTU 67)
The kntal (colloquial kon ta i) is a powerful symbol and is a resonant
term, as we shall see, throughout Draupad cult folklore. The open-
ing four lines attribute the touching, arranging, loosening, adornment,
and praising of Draupads kntal alternately to the Pn davas and to
battle heroes. And the fifth and sixth lines bring in flowers, indicating
that it is specifically pure heroes who adorn it with jasmine. Obviously,
the Pn davas are war heroes and pure, but the alternation suggests a
ritual nuance as well. The touching, arranging, loosening, adornment,
and praising of Draupads kntal are also done by Draupads wor-
shippers: by the actors who act out her dramas, and by the pcris and
other temple officiants who handle and adorn her icons. Indeed, we
108 chapter five

shall soon discuss a myth that is precisely about a pcris handling

of Draupads hair. There is further an erotic-yet-defiling ambience to
the symbol. Draupad describes the loosening of her kntal in terms of
a love and affection between her and her husbands that is violated by
Duhsanas defilingindeed, outcaste-liketouch. But the defilement
works both ways, as the stick or cloth would seem to symbolize, for
Draupad is menstruating. Moreover, as Duhsana is quick to point
out in reply, the erotic connotation of Draupads marital statusher
polyandryinvites its own defiling interpretation. Says Duhsana:
As soon as five mens hands fall on you, you wicked cheat, does your
body emit fragrance? Will it grow [valarntupm]? As soon as the
hands of those five who keep you fall on you, having become proud,
if you experience my hand, will all the shining diminish? With five
[already], let it become six. What is that! Get up and come! (CTU
68). Also interesting in the lament is its conclusion: first the transition
from the fragrance of jasmines to the rolling of Draupads hair in the
dirt, on the earth, and finally in the street. The connection between
fragrance and the earth already typifies Draupad in the Sanskrit epic
from the moment of her birth (Mahbhrata 1: 155.43), and recalls
the association between smell and the earth that is found in classical
Indian philosophical systems (Biardeau 197172b: 41). The culminat-
ing line brings all this down literally to the street (or teru) leveland
smellsof the terukkttu.
As to the disrobing scene, we must again look beyond Irmaccantira
Kaviryars text to what is done in performance. In all cases, perfor-
mance of the disrobing scene is regarded as highly inauspicious, and
is sometimes omitted from a festival on this account. Draupad stands
on the musicians bench behind the stage screen (held by two mem-
bers of the troupe), and behind Draupad is Krsn a, either in back of
her on the platform itself, or above (sometimes flute in hand) on the
greenroom roof. He holds the end of a yellow or even gold saree,
macal in Tamil, that passes from his hands over Draupads shoulder.
Simultaneously, another macal clothsymbolically identical with the
saree that passes from Krsn a to Draupadmay be placed in front of
the processional icon of Draupad that would normally watch the
drama from her chariot at the back of the audience. The Draupad
actor folds her hands in the devotional salute and repeatedly cries
out Kvint! Kvinta! (the Tamil vocative for Govinda, a name of
Krsn a). The macal saree is said to represent her chastity (karpu) as
well as the source of all the other sarees that proceed from Krsn as
the folklore of draupad: srs and hair 109

graceful response to her prayers for his protection. The actual disrob-
ing is done in either of two ways. Usually, Duhsana reaches under the
screen and grabs hold of what he takes to be the bottom of her saree.
But he is foiled by what turns out to be a long chain of sarees tied end
to end, and eventually linked to form a revolving circle.8 The more he
pulls, the more the circle goes round and round, resulting in his baffle-
ment and exhaustion. Alternately, in a scene repeated several times,
one or more additional sarees may be placed loosely around Draupad,
and while Duhsana pulls violently as she prays, the Draupad actor
twirls as the sarees come offalways leaving, of course, the one saree
underneath to confront Duhsana with the impossiblity of this task.9 I
will return to the circle of sarees tied end to end, for it is a fitting meta-
phor for much that goes on in the terukkttu handling of Draupad
cult folklore. In either case, however, it is of interest that Duhsana
charges repeatedly that Draupad frustrates him by some magic craft
or use of mantras (CTU 8284; Hiltebeitel 1988a: 271272). The scene
has both theological subtlety and high dramatic intensity, often involv-
ing possession. But the audience also registers that either device is an
obvious sleight of hand, one which Duhsana is being portrayed as too
arrogant, violent, and ignorant to see through. The contrast between
what is obvious and what is subtle, or intangible, is fittingly portrayed
by this perhaps universal symbol of veiling, illusion, and fabrication
(Eliade 1963: 18082; Neumann 1963: 226234, 250, 284).
The disrobing scene also provides Irmaccantira Kaviryar an oppor-
tunity to rework into local idioms some of the cosmological themes con-
cerning the pralaya (dissolution of the universe) that figure in classical
versions (see note 1). Once again, the disrobing portends the disruption
of the ordered relation between the elements. When the miracle of the
disrobing ends and Draupad stands vindicated, the Kuru elders rise and
speak these softened words:

Richard Frasca indicates (personal communication) that in performances of this
play by troupes in North Arcot the chain of sarees was not linked up to form a circle.
As I have seen it performed at Tindivanam (1981) and Mlaccri (1986), the first
sars are pulled out straight, linked end to end, without forming a circle. Then, when
Duhsana tries again, the circle is formed.
Both versions can be seen in the longer version of my video film, Lady of Gingee:
South Indian Draupad Festivals, Part I (available through University of Wisconsin,
South Asia Films). See also Hiltebeitel 1988a: 236 and n. 22.
110 chapter five

To the five she is the goddess [or wife; tvi], to others, she is the
She is the goddess Earth [pmitvi], she is Fires self-manifestation
She is the goddess of this lineage [kulateyvam]. Is there any other like
(CTU 86; Hiltebeitel 1988a: 264)
In contrast, after the hair-pulling and just before the disrobing begins,
Draupad asks Dhrtarst ra:
O king of the Kuru dynasty who sees the greatness of the kula, Do you
allow this cruelty in your kula, O father-in-law?
If you permit these faults in the presence of the elders, my father-in-law,
Will the rain fall? Will the world survive [malai peyumm vaiyakam
(CTU 73; Hiltebeitel 1988a: 265)
There are repeated references to the fires that will erupt in the form of
Draupads vengeful anger (CTU 68, 85; Hiltebeitel 1988a: 271, 274).
Duhsana compares the impossibility of disrobing her to the use of a
well-sweep (erram) to drain the seven seas (CTU 86). But most inter-
esting, once again, are two contrasting evocations of Draupads rela-
tion to the earth. First Duhsana tells her that by stripping her he will
reduce her to an open ground, a desert (vetta veli; CTU 83; Hilte-
beitel 1988a: 272). But when he finds he cannot do this, he complains:
Millions and billions, beautiful and variegated, [crores of sun-bright
sarees of many kinds, emerging variously (?)], as the winged white ants
[cal] come out of an anthill [purru] in a great exodus [perum pkku],
how many crores of sarees must I remove? (CTU 85).10 Not to be
reduced to a barren ground or desert, Draupad is the earth in its most
resourceful and inexhaustible aspect, and in a form, moreover, that
connects her with other rural and village goddesses whose cult centers
on the plenitudinous mysteries of the anthill, or, more correctly, the
termite mound (Irwin 1982; Meyer 1986: 2930, 5859, 6, 121122,
and s.v. termite hill in index).
Finally, let us see how Irmaccantira Kaviryar handles the vow
Draupad makes to bring about her rebraiding. Having seen Duryo-

The bracketed addition is found in a different edition of Irmaccantira Kaviryars
play (1968: 80). The Tamil of the addition reads: koti criyap pirakcamn a claikal,
vitavita vcamakcaramy. The question mark raises my uncertainty about the clos-
ing part.
the folklore of draupad: srs and hair 111

dhana expose his thigh to her and having responded in dialogue to

his various taunts, telling him, among other things, that vultures will
peck at your thigh so that it will become a filthy place producing wrig-
gling worms, Draupads closing statement is as follows (first in meter
and closing in prose):
Having been made to stand in the royal presence, having touched my
kntal and saree [kntaluntukiluntn ti ],
the one who did this, growing weak in the future together with his kin
[vr: that is, roots] on the battlefield,
[until] your head [is] severed without beating the unequaled war drum
of victory,
I will not, taking it, braid my shiningly spread out kntal.
Listen, you kings in the sabh! If in the future, on the battlefield, I do
not tie up my hair [mayirai] standing on the chest of the man who
has done this disgrace of touching my hair and sars [kntalaiyum
tukilun tn ti ], having made me stand in the royal sabh, then I am not
Draupad. (CTU 88; Hiltebeitel 1988a: 30, 23637)
Irmaccantira Kaviryar in effect doubles the compound allusion to the
hair and sars that is found in, and no doubt drawn from, Villi. Be it
noted, however, that neither Villi nor Irmaccantira Kaviryar gives us
a fully operative festival version of Draupads vow. In fact, the passage
leaves it rather vague whether she is referring to Duhsana or Duryod-
hana. For a fuller and clearer expression of her own vow, we may turn to
another chapbook drama called Turpatai Kuravaci, one that is far less
classically based and therefore probably freer to bring out the cult-related
themes. Disguised as a gypsy fortune teller, Draupad tells Duryodhanas
wife and mother of the fate that awaits him and him alone.
She [Draupad] went to dwell in the forest, saying, On the battlefield
[patukalam] I will put up my kntal, which was spread out in the
great sabh.
She went to dwell in the forest, saying, In the war I will put up my
kntal, which was spread in the great sabh, O mother.
She went, saying, I, one woman, spread my hair and stood; as a result
everyone in the country had to keep their hair spread.
When the battle takes away the army on the eighteenth day, standing on
the king of the earth, the Amman will tie up her kntal.
Seizing the lord, aiy, the king who rules the country, standing on your
husband, the Amman will tie up her kntal.
Taking up a handful of blood, combing her hair, separating his ribs for a
comb, the chaste woman [pattiniyl], taking up the fourteen intestines
on both sides, gathering up her hair on the battlefield [patukalam], the
Amman will tie up her kntal.
112 chapter five

After the differences have vanished, Pcl will come, she said, after
crushing the decorated crowns of the hundred and one [the Kauravas
plus Jayadratha, their brother-in-law].
Having seized and cut off all the heads of her relatives [panklis], the
chaste woman, having put up her hair, will come to rule the earth
[pattiniyl talaimutittu pr la varuvl].
(Tanikcala Mutaliyr 1979: 44; Hiltebeitel 1988a: 3067)
It is in the final scene of the drama Patinett m Pr, or Eighteenth Day
War, the play that closes the cycle of festival dramas, that the hair-
tying, kntal mutittal, at last comes to pass in the fashion thus antici-
pated. I note only that in my experience it is invariably Krsn a who ties
the red or orange flowers into Draupads hair. The scene is then repli-
cated later in the same morning with the anointing of Draupads pro-
cessional icon atop Duryodhanas effigy on the ritual battlefield called
the patukalam, in this case with her hair tied by the hands of local vil-
lagers, or by the pratiyr. And often the rebraiding of the icon on the
effigy is complemented and doubled by a reappearance of members
of the troupe, who this time not only repeat Draupads rebraiding
but show the dishevelment of the now widowed Peruntiruval, wife of
Duryodhana, lamenting and carrying a winnow.
In fact, one could trace the double theme of hair and sars through
the drama cycle, and indeed the cult, as a whole. The drama cycles
often begin with the play Kan n an Jalakkirtai, Krsn as Water Sports,
which culminates in Krsn as stealing of the sarees of the Gops (who,
before bathing, loosen their kntals; Hiltebeitel 1988a: 18690).
The Krsn a who steals the sarees of the Gops replaces the sarees of
Draupad. Or alternately, a festival drama cycle may begin with the
play Turn ccri Ykaclai, Dronas Sacrificial Hall, which tells
about the births of Avatthman and Draupad, and introduces the
folk theme that Avatthman is born from a horse that is fed with a
plate of rice that iva has ejaculated into, after seeing Dronas wife
Krp appear before him naked and with her hair down as well (idem.,
190195). The disguise themes are enriched, with Bhma and Arjuna
in transvestite roles, wearing sarees (idem., 296298, 336343).
Indeed, again in terms of the theme of illusion and fabrication, it is not
insignificant that all parts are played by male actors. At some level it
must register on viewers that if Draupad (or the Gops, or Krp) were
actually stripped on stage it would be quite out of character. In both
the dramas and the Villipratam there is the prophecy by Sahadeva
that the only way to avoid war would be to bind Krsn a and shave
the folklore of draupad: srs and hair 113

off Draupads hair (idem., 313). Villi also introduces the notion that
Karn a has an instantly combustible saree that he uses to test whether
women who claim to be his mother are telling the truth: only Kunt
can wear it unharmed (idem., 31416). Throughout the battle scenes,
rolled-up cloths, and sometimes sarees, are used for heads. And when
Duryodhana is disemboweled, his guts can be represented by another
macal saree (idem., 421). Then there are some scenes involving the
end-pieces of sarees which I will turn to shortly.
Beyond the Villipratam and the dramas, one finds these themes
further articulated in popular cult myths. One sometimes hears that
from the beginning of Draupads forest exile to the end of the war,
the period that she wears her hair disheveled, she takes on the form
of Kl (klrp), her vivarp or Universal Form. While her hus-
bands sleep, she roams about the forests from midnight to 3:00 a.m.,
devouring whatever comes her way (wild animals, domesticated ani-
mals, even humans). And during the war she goes out at the same time
to consume the bodies of the dead. There are several stories that spin
off these themes that I will not go into here (idem., 289295). Suffice
it to say that they are linked with the disheveled Kl through their
evocations of the forest, the battlefield, and the crematorium.
As we begin to address the issue of Mahbhrata folklores outside
the Draupad cult, one more myth is instructive. It is a story known
in the immediate Gingee area, and apparently rarely if at all known
outside it. It supplies a sort of oral sthala purn a of Draupads origi-
nal temple (ti kvil, ti ptam) just north of the Gingee Fort in the
village of Mlaccri. This village, also known as palaiya ceci, or Old
Gingee, marks the site where Draupad is said to have taken a second
birth to come to the aid of a certain king Cuntan, a descendant of the
Pn davas, to rescue him and his Gingee kingdom from the Gingee
or Mlaccri Forest Demon (the forest is just west of the Mlaccri
Draupad temple). Different versions of this story give the king differ-
ent names, but I cite here the most complete and informed version I
found, that of the Mlaccri temples pcri, in which the king is the
same Cuntan, presumably later in life.
When Cuntan was ruling Old Gingee, there were ten acres of land near
the temple for a flower garden to supply the pcri with flowers for the
temple pjs. It was the custom that each day, after the flowers were
placed on the Amman, that the pcri would bring them before the king,
as prasda. One day, without the pcri knowing it, his concubine, see-
ing the beautiful flowers on the icon, took them off and placed them
114 chapter five

in her kntal, just to enjoy them. After some time, she removed them
and put them back on the Amman. Unaware what she had done, the
pcri then came and removed the flowers from the Amman and sent
them with various other pj articles to the kings palace.
When they reached the palace, there was one hair lying in the flowers.
The queen noticed it and asked the king: How can there be a hair in
the flowers? It is a stone icon [kal-cilai]. Does a cmi [a goddess] have a
kntal? The king immediately called back the pcri and demanded an
answer. The pcri had never seen the hair, and knew nothing of how
it got there. He was blinking, and could not answer. In rising anger, the
king said: I will give you a week. If you can answer my question in one
week, well and good. Otherwise you will be punished with death.
When the pcri returned to the temple, he unburdened his troubles
to the Amman. He wept, he prayed, he did daily and nightly observances
at the temple, never returning home, always imploring her: Somehow
you must show your hair to the king. At last the Amman appeared in
his dream, and said: On Friday let the king and queen come before my
image. If you do the regular apickas and pjs on that day, and place a
white screen [ven tirai] before me, then I will show my kntal.
The next Friday the pcri did a pj in the presence of the king and
queen. Only he could see the icon, as the screen was in front of it. Again
the king asked: Can a stone icon have a kntal? As soon as the rja
stepped forward to peer over the screen, the Amman flashed the dishev-
elled hair of her kntal out over the screen and into his eye. Yet at this
point only the pcri could see the Ammans real form. The king still
had a doubt: the pcri could have pulled a trick. It may be hair, but is
it the Ammans? He is going to lose his life, so he has resorted to some
trick. Thinking that it was artificial, the king touched the Ammans hair
and pulled one piece of it. When he did this, he heard the sound made
when a hair is pulled. The root of the hair came over the screen and
into his hand, and he saw that it was tipped with blood. Immediately
he became blind.
At once the king fell at the pcris feet, asking his forgiveness for
causing him so many hardships and doubting him, praising his devo-
tion as firmer than his own, and imploring him to Pray on my behalf
to Amman to restore my eyesight. The pcri prayed, and at last one
night the Amman appeared in his dream and said the king would get
back only three-fourths of his vision, not all of it, as he had tested her.
The next day, after the pj, the three-fourths of his vision returned. In
gratitude, the king then ordered the construction of various additions
to the temple.
The blind king is a symbol that has in fact one of its most powerful
expressions in the Mahbhrata itself, in the person of Dhrtarst ra.
Indeed, from a certain angle, it is Dhrtarst ras blindness that allows
the dice match, and the hair-pulling and disrobing that result from it.
The pulling of just one hair is sufficient here to evoke the dire conse-
the folklore of draupad: srs and hair 115

quences of the epic precedent. The hair tipped with blood recalls the
connection between Draupads menstruation and the hair-pulling in
the sabh. The blinded Cuntan is in this regard no less a descendant
of the Kauravas than the Pndavas. Just as it is the vow Draupad
makes concerning her hair that leads to the punishment of the kings
who violate her in the Mahbhrata, so it is again with her hair that
she punishes the king of Gingee for doubting her miraculous power
and violating the sanctity of her icon.
Aside from its evocations of the Mahbhrata, this myth is of inter-
est for its variations on certain south Indian mythic themes, and more
specifically as a variant of two myths that tell essentially the same story:
one from Puri in Orissa that has been set forth by Frdrique Marglin
(1985: 9293), and one from rkkulam in Krsn a District, Andhra,
that was kindly summarized for me by David Shulman.11 What is most
striking is that the Draupad cult myth is the only one where the icon
in question is that of a goddess. In the two others, the priests concu-
bine takes the flowers from an icon of Visnu, which (or who) must
eventually grow hair in a corresponding fashion to exonerate the priest
from the charges of his royal patrons. It would seem that in this case
we have an essentially royal myth that probably circulated during the
Nyak period outside the Draupad cult, and was adapted to the latter
only in a very local tradition. Yet how much more powerful an image
it is in the goddesss case!

Loose Ends and Closing Circles

In the rest of this essay, let us begin to explore some of the ways that
the Draupad cult is not on the receiving end, but part ofand in some
cases perhaps even the source ofwider networks of folk Mahbhrata
transmission. I return to the circle of sarees linked up end to end. We
have seen how this image of grace and illusion works in the text and
performance of Dice Match and Disrobing. One must appreciate,
however, that the end piece of the saree (in Tamil, muntnai, col-
loquial muntn i) is itself a powerfully operative symbol of Draupad
cult dramas and folklore. We have seen that Draupad drapes the end
piece over the coals to protect her fire-walking worshippers. In the

For summaries of the three versions, further discussion, and different consider-
ations, see Hiltebeitel 1988a: 9397.
116 chapter five

performance of Dice Match and Disrobing, Duhsana begins his

attempted disrobing by grabbing hold of what he thinks is the end
of Draupads saree (and in cases where the Draupad actor twirls,
Duhsana actually does hold her muntnai, though not her real
one). And at the other end of this endless circle, it is Krsn a, holding
one end of the special macal saree, who provides Draupad with its
other end as the piece to place over her shoulder, again covering the
Draupad actors real saree, but as if it were truly her own. Margaret
Trawick Egnor enriches our appreciation of this moment, comment-
ing on the use of the term a young girls garment for the shoul-
der-piece in the crying songs of the Paraiyar women: a young girlss
garment [citttai] is the top piece draped over the shoulder, worn by
girls come of age but not married. The same garment is tied on the
statues of female deities, because a deity is always young and always
a virgin (1986: 307). Draupad, whose virginity in her cult will be
discussed further, is thus in an iconic pose, with the loose piece of
macal saree over her shoulder being the only one untied in the whole
dramatic scene, and the one that accounts symbolically for the linking
of all the others. As with the hair, here too we have a symbol of bind-
ing and loosing, knotting and unraveling. Indeed, one senses that the
hair and the sarees are again complementary, for it is while the hair
is unraveled that the sarees are linked. Each binding and unraveling
further points to the collaboration of Krsn a and Draupad, Visnu and
the goddess: for just as Krsn a provides the macal saree at the disrob-
ing, so he resets the blood-symbolizing flowers into Draupads hair
when he reties her kntal.
I know of three instances in Draupad cult folklore where the
muntnai is used as a symbol, and in each case, I would argue, the sin-
gle unbound end pieces in these episodes evoke the linked muntnais
of Dice Match and Disrobing. In one instance, Arjuna disguises
himself as a woman (really a multiform of Durg, called Vijaympl,
Mother Victory) to seduce Prmannan (Draupads temple guardian
and servant-to-be) into giving up his pj implements so that they can
serve Draupad as the weapons by which she will be able to win the
eighteen-day war and retie her hair. In one variant of this story that I
have not found in any of its dramatic renditions, Prmannan accom-
panies Arjuna-Vijaympl who says she has to urinate. Prmannan
insists on holding the end of her saree, so as to be sure she does
not disappear. Arjuna then disrobes behind a bush, ties the other
the folklore of draupad: srs and hair 117

end to the bush, and makes off with the pj items.12 In this comic
inversion, Arjuna, not only a multiform of Durg but a stand-in for
Draupad, disrobes himself and leaves Prmannan, standing there
with the end of a sar, as demonic and dumbfounded as Duhsana
(they are made up identically in the dramas and often played by the
same actors). A second use of the muntnai is made in the dramas.
At the end of the Pn davas year spent incognito in Virtas kingdom,
Virta interrupts his dicing with the disguised Yudhist hira to protest
the latters seeming belittlement of Virtas son. He throws the gam-
bling pieces at Yudhisth ira and wounds his forehead. Draupad then
steps forth and stanches the bleeding with the end of her saree. As it
stands, the use of the saree here can be traced to the Villipratam, for
in the Sanskrit epic Draupad uses a golden bowl. The important thing
in all cases is that she keeps the blood from touching the ground, for
as Yudhist hira soon explains, had it done so, it would have destroyed
Virt as kingdom. This alone is a sufficient evocation of the war to
come (Hiltebeitel 1988a: 300301). With the muntnai replacing the
golden bowl, however, we have also an evocation of the link between
sarees and blood at the disrobing. And as our third piece of muntnai
folklore will make clear, it also ties in with the miracle of the sarees
as well.
The play that precedes Dice Match and Disrobing in Draupads
festival cycles is the play Irjacya Ykam, Rjasya Sacrifice. In fidel-
ity to the structure of the Mahbhrata, it portrays the great sacrifice
of royal consecration that entities Yudhisth ira to paramountcy over
the other kings, and spurs the embittered Duryodhana to initiate the
dice match. At the Rjasya, a pivotal episode involves the vilification
of Krsn a by iupla that results in Krsn as beheading of this incarnate
demon with his discus in the very midst of the sabh. Draupad cult
folklore has made this scene the prelude to Krsn as rescue of Draupad
at the disrobing. In the drama, when Krsn a throws his cakra he cuts
his finger. Out comes Draupad. Holding the muntnai of her saree,
she offers it to Krsn a to stanch his blood. Delighted with her show of
compassion, he promises that should she ever need help, he will find a
way to return her favor in kind. The drama cycle once again reminds
us that what will link the ends of the sarees together at Dice Match

My thanks again to Madeleine Biardeau for supplying me with this amusing vari-
ant; see Hiltebeitel 1988a: 344 and n. 15.
118 chapter five

and Disrobing will be the ominous theme of blood. Indeed, let us

recall King Cuntan and the blood-tipped piece of Draupads hair.
There is also a popular and less ominous variant of this story that is
known outside the drama cycle.
Once while the Pndavas and Krsn a were swimming in a tank, wear-
ing only their loincloths, Krsn a lost his in the water. When everyone
finished bathing, he remained there, hesitating and feeling shame. No
one understood why he wouldnt come out. Meanwhile at a neighboring
tank, Draupad and her maids were also bathing. She observed the scene,
understood, tore off a piece of her muntnai, and threw it to Krsn a, from
one pond to the other. Krsn a caught it and used it as his loincloth. Then
he came out of the water and told Draupad: In my life, I wont forget
this kind of help. Out of gratitude, I will repay it in kind when the occa-
sion comes. (Hiltebeitel 1988a: 227)
Here it is not the somber theme of end pieces tipped in blood that
links Krsn a and Draupad but the more amusing byplay on sexual
modesty (vetkam). Krsn a will make his loincloth from the portion of
the saree that covers Draupads breasts. For the tiny piece of cloth that
Draupad throws to him, she will get in return the inexhaustible flood
of sarees. Indeed, one is reminded further of Krsn as theft of the sarees
of the bathing Gops, from the play that often starts the drama cycles.
How appropriate it is that Draupad should be promised unending
sars for rescuing Krsn a from the same impasse in which he had left
the Gops (Hiltebeitel 1988a: 226227, little modified).
I have gone to such length in trying to show the interconnected-
ness of this folklore within the Draupad cult not only because that is
the best way to appreciate it, but because it provides us with a slender
entre into the problematics of transmission, by which from here on
I mean diffusion.13 The variant of Krsn as promise that is found in the

On diffusion, see chapter 6. With regard to the classical discussion of folklore
diffusion by Kaarle Krohn, the following reservations may be noted as bearing on dis-
tinctive features of Indian Mahbhrata-related regional folklores: (i) The statement
that the most nationalistic creation of a people, its heroic poetry, seldom spreads
across linguistic boundaries (1971: 146), would be inadequate to describe the Indian
situation, and may likewise be inadequate to describe the Indo-European situation
(see Hiltebeitel [1976] 1990: 5759). (2) The use of terms like hallucinatory forms,
fantasy, and superstition: (Krohn 1971: 59, 138, 16871) are inadequate to account
for the fluidities within what Krohn calls the basic form of this folklore; the distinc-
tive features of this basic form are the presupposition of a classical Mahbhrata
and the armatures of ritual. (3) The emphasis on mechanical laws of thought and
imagination in the rich variation of oral tradition over and against creative pro-
the folklore of draupad: srs and hair 119

heart of the drama cycle is known in a north Indian variant. Susan

Wadley has found the northern variant told in connection with the
ceremony in which sisters tie a protective string called rkh around
their brothers wrists: One time Bhagavan Krsn as hand was cut and
bleeding. When Dropadi saw this she immediately tied a piece of cloth
from her dhoti [sic] on her brothers hand. Because of this tying, Shri
Krsn a saved Dropadis honor at the time of Dusharsans taking her
sari (Wadley 1976: 158). The folk epic setting is not stipulated, but
one sign suggests that this northern variant has a southern source.
It draws upon the apparently southern theme, richly underscored in
the Draupad cult, of Draupad being Krsn as tankai, his younger sis-
ter (Hiltebeitel 1988a: 226227). Since this concatenation of themes
is so richly interwoven in the Draupad cult, it seems that the best
argument one could make here is that this fragment of north Indian
Mahbhrata folklore has traveled not only up from the south, but
from the Draupad cult itself.
There are no doubt other cases where Mahbhrata folklores from
diverse parts of India could be traced with similar probability to
Draupad cult sources.
At present, however, I know of no other instances where this is likely
outside of the south (that is, in fact, outside Tamilnadu and Andhra
Pradesh). Karthigesu Sivathamby argues that the various south Indian
and Tamil Sri Lankan drama forms that draw on the Mahbhrata
(he includes terukkttu) indicate a common prototype that itself
results from a fusion of epic material with local legends (1981: 360).
I doubt, however, that there is a coherent folk epic prototype outside
the Sanskrit epics themselves. Even the theme that recurs most promi-
nently in the south Indian vernacular retellings and folk dramatiza-
tions of the epicDraupads vow to rebraid her hair with bloodis
told very differently from region to region: sometimes it is Duryod-
hanas blood, sometimes Duhsanas and sometimes the mingling of
both (Hiltebeitel 1981: 180, n. 3; 1988a: 409, n. 19). The prominent
differences in the ways that the Mahbhrata is treated in vernacular
retellings and regional drama forms suggest that it is not so much
a common folk prototype that shapes them as the different regional
mythologies (Hiltebeitel 1988a: 148). Yet south Indian folk variants

duction (9798) seems to push what are otherwise good insights to an unnecessary
superorganic extreme; cf. Dundes 1965: 219220.
120 chapter five

of the Mahbhrata would seem to have a relatively strong intercon-

nectedness that lessens as one moves to the north. The themes I have
highlighted in this chapter do not seem to have the same prominence
in the Pn dav Lls of Garhwal. Similarly, the tantric accentuations of
Draupads connections with hair and blood in Nepalese Mahbhrata
folklore are certainly different from what one finds in the south: in
a Nepalese account, when Duhsana is decapitated two spouts of
blood shoot from the wound, one for Bhma to fulfill his vow to drink
Duhsanas blood, and the other for Draupad to fulfill her vow to
wash her hair in it (Anderson 1971: 235; Hiltebeitel 1988a: 409). There
is, however, need for further research and information to clarify these
relative consistencies.14
In one respect, though, we are now, I think, at a point where we
can begin to say something more significant on the relationships
between regional folk Mahbhrata traditions throughout much of
India, including north and south. Here, however, it is not a matter of
vernacular and regional reworkings of the epic itself such as one finds
in the Tamil and Garhwali Mahbhrata cults, but of the ways in
which the Mahbhrata impacts upon distinctive regional folk epics.
As numerous scholars have noted, a number of regional Indian folk
epics are linked with the Mahbhrata. Such linkages would seem to
be more widespread than with the Rmyan a, although the latter are
not unknown, as John D. Smith has shown in his discussions of the
Pbj epic from Rajasthan (1986, 1989, 1991).15 Where it is a question
of regional folk epics linked with the Mahbhratasuch as the Tamil
Elder Brothers Story, the Telugu Epic of Palndu, and the lh of the
Hindi-speaking heartland of north Indiathe epics seem to be con-
nected with regionally dominant landed caste traditions, although this
is far clearer in the two southern examples, where regionally domi-
nant castes are prominent in sponsoring the hero cult festivals that
celebrate these folk epics, than it is for the lh, which is without such
festivals and without any single caste or caste cluster that predomi-
nates among its reciters or listeners (Schomer 1984). In any case, the
lh evokes a past and lost age of regional Ksatriya dominance. Let us

More work needs to be done to see whether a similar interconnectedness can
be found in the north relative to the south. So far, from what is known of folk
Mahbhrata traditions in Garhwal, the Kuruksetra area, Nepal, and through the lh
(see below), there is nothing to suggest a comparable contiguity.
I thank John D. Smith for letting me see this latter work in advance of publication.
the folklore of draupad: srs and hair 121

note that what I am suggesting here about the social and landed back-
ground of regional folk epics linked with the Mahbhrata is what
we find also in such Mahbhrata cults as the Draupad cult and the
Pn dav Lls. Folk epics linked primarily with pastoralist castes, such
as the Ktamarju Kath discussed by Velcheru Narayana Rao and the
above-mentioned Pbj epic, by contrast, do not seem to be connected
with the Mahbhrata.16 Upon reflection, however, if further research
bears out these patterns, it should not be surprising. The Mahbhrata
is precisely the classical epic that concerns the issue of Ksatriyas and
the land. It is thus quite logical that regional folklores about dominant
landed castes, those castes that translate Ksatriyas values into local
and regional terms, should link themselves with it.
So far, however, though linkages between regional folk epics and
the Mahbhrata have been noticed, they have not been adequately
interpreted. All too often one hears it implied that they are meaning-
less afterthoughts, tagged on to what is allegedly genuine about the
regional folk epics, whether it be their true historical cores or their
regional counterculture ideology (Roghair 1982: 9293, 136137; cf.
Hiltebeitel 1984b). Or the connections made by the folk epics them-
selves are seen as a kind of superficial register of deeper connections:
motifs that recur in the Indo-European heroic life cycle, and are vir-
tually universal (Beck 1982: 122123, 12628, 132). To be sure, the
connections which the folk epics make with the Mahbhrata may
look superficial and rather arbitrary. There is no obvious consis-
tency from one regional folk epic to another in the ways they link
up with the Mahbhrata. The Epic of Palndu, for example, occurs
at the juncture between the dvpara and kali yugas, thus replacing
the Mahbhrata war in that interval and relegating it back to a time
within the dvpara yuga (Roghair 1982: 97, 108, i4., 165, 320). The
great battle of the lh, on the other hand, comes to an end when the
kali yuga is one-fourth over (Waterfield and Grierson 1923: 273). Cer-
tain Mahbhrata themes and episodesdice matches, disguises, vari-
ous types of births, marriages, and deathsrecur in the regional folk
epics, but with no repeated pattern. And certain heroes and heroines
of these regional folk epics are said to be incarnations of Mahbhrata

The Ktamaraju Kath is associated with Golla herders (Narayana Rao 1986:
141, 144), the Pbj story is patronized primarily by the Rebr caste of shepherds
and camel herders (Smith 1986: 53).
122 chapter five

heroes and heroines, but each folk epic seems to make its choices of
who gets incarnated from the Mahbhratas vast cast of characters
quite differently.
Yet scholars cannot afford to treat such correlations as superficial
or random, any more than the catalogues of divine incarnations in
the Mahbhrata. The correlations are consistently different, but they
are consistently there, and we must find the right questions to ask of
them. To my mind, it is the failure to formulate one pair of questions
that is at the center of this current malaise. Those questions are as fol-
lows: What is the role of the goddess in the Mahbhrata in relation
to Draupad? And what is the role of the goddess in the regional folk
epics in relation to the heroines who are linked with Draupad? Oth-
ers have appreciated the centrality of the goddess in relation to the
heroines of the regional folk epics (Beck 1982: 4952, 122, 128136;
Roghair 1982: 9192, 121123; Smith 1989), but without appreciating
that this centrality involves a reinterpretation of what are essentially
bhakti images of the heroine-as-goddess in the classical epics. The
oversight is most acute in John D. Smiths case, when he observes that
goddesses are relatively unimportant in the Mahbhrata and the
Rmyan a, but they play a major role in many vernacular epics, often
bringing about a war of destruction that will annihilate the heroes
(Smith 1989, 182). What about Draupad? Or even St? In suggesting
that these are the right questions, I am, of course, aware that I am gen-
erating them in part from what I have found in the Tamil Draupad
cult, which in effect stands midway between the classical Mahbhrata
and such regional folk epic traditions. I am further suggesting that
amid the various and seemingly arbitrary clusters of incarnations in
the different folk epic traditions, there is one central consistency. That
is a heroine who represents the goddess through an imagery that clari-
fies itself, at least in some small but significant part, in relation to the
imagery of Draupadnot the Draupad of the classical epic, however,
so much as the Draupad of the Draupad cult.
In the folk Mahbhrata of the Draupad cult, Draupad is a virgin,
a kanni: she has no sexual relations with her five husbands, who are
only nominally the fathers of her five children (Hiltebeitel 1988a: 293
95). In fact, Tanaka (1987:409, n. 2) reports from a Draupad festival
in Tamil Sri Lanka that Draupad is said to wear a yellow saree because
it is the colour of a virgin. She also has what I call low-status ritual
service companions or guardiansone of them the above-mentioned
Prmannan and the other a Muslim who ostensibly help her Ksatriya
the folklore of draupad: srs and hair 123

husbands win the Mahbhrata war, but also serve the purpose, as
guardians of the virgin goddess, of handling the impurities of blood-
shed required by the Mahbhratas sacrifice of battle. These minimal
features recur in the three regional folk epics just mentioned that link
up with the Mahbhrata. There are the low-status ritual service com-
panions (Cmpka in the Elder Brothers Story; Kannama and others
in the Epic of Palndu; the four Banphars and the one Muslim, Mr
Tlhan, the reincarnation of Bhma, in the lh) who serve not only
the dominant caste heroes who represent the Ksatriya ideal, but the
heroine from that real or purported Ksatriya caste who represents the
virgin goddess. In each case this virgin heroine is the pivotal female
figure in bringing about the climactic episodes of the respective great
battles. And in each case, it is the nonconsummation of her marriage
that stands as one of the primary fatalities that shapes the deaths of
the major heroes: the real or purported Ksatriyas no less than the
low-status service companions. Finally, each of these Ksatriya virgin
heroines is linked directly with Draupad. Tankl in the Elder Broth-
ers Story and Bl in the lh are reincarnations of Draupad.17 And
Mncla in the Epic of Palndu has a pet parrot named Draupad,
the parrot itself being one of Draupads iconographic emblems and
a metonym for her forest wildness in the Tamil Draupad cult (Hilte-
beitel 1988a: 26364).
The consistencies in the ways these three regional folk epics tie such
Mahbhrata-related themes together are, of course, enlightening for
their contrasts with the Mahbhrata. In the regional folk epics, the
links between the virgin goddess and the land require the regional
heroes to authenticate the sacrifice of battle in different, but still vic-
torious, terms: ultimately, that is, through their own pure self-offerings
to the territorial goddess.18 It thus makes a large difference whether

Beck 1982: 182, indicates that Tankl is considered an incarnation of Draupad
only in written versions of the Elder Brothers Story. The fact that this relation is not
found in the bardic version she recorded does not, however, guarantee that it is not
found in oral variants at all.
I would like to thank Arjun Appadurai for his editorial reflections on the anthro-
pological implications of the argument here, suggesting that the underlying struc-
ture which accounts for the hook-up of Draupad as a goddess to the problem of
landed dominance, has to do with the link of purity, power, and status, and more
particularly that the purity of women is crucial to land-holding groups as an index
of such groups control of, and rivalries over, the power and fertility of the soil (per-
sonal communication). For background, see Yalman 1963; Tambiah 1973: 92110;
Hart 1973, Rajam 1986. See also Narayana Rao 1986: 143 and, for some supportive
124 chapter five

Draupad is a sexually active wife who incarnates the goddess primar-

ily in the form of the fickle and favor-bestowing r, or a virgin who
incarnates the goddess primarily in the destructive capacities of Durg
(with shades as well of Kl). In the first case, she is an image of the
prosperity of the sacrifice that will always accrue ultimately to the
rightful kings, who happen to be her husbands. In the latter, though,
she is an image of the goddess of victory to whom warrior heroes will-
ingly and chastely dedicatethat is, sacrificetheir lives in battle.19
In short, and in closing, there may well be an underground folk
Mahbhrata. But it cannot be monolithic. It has no prototype outside
the Sanskrit text (which can never be assumed to have fallen out of
the folk epic frame of reference). If such a folk Mahbhrata exists,
however, it would seem to be centered on images of the goddess and
the control of the land. Its lines of transmission and adaptation are
too vast to ever trace fully. But those lines that do emerge suggest the
crossing of many geographical and linguistic boundaries, and symbols
and motifs that recur in a wide spectrum of reflexive and interpen-
etrating genres: from Mahbhrata vernaculars to folk dramas, from
folk dramas to ritual idioms, from ritual idioms to temple tales, from
temple tales to sisters tales, from sisters tales to regional folk epics,
from regional folk epics to Mahbhrata vernacularizations.

data and thoughts of my own, Hiltebeitel 1988a: 89, 3239, 7475, 1012, 22223,
29394, 39798. The folk epics of Telugu trader castes also accentuate the role of
virgin heroines, but in significant contrast with the emphases in the Telugu martial
epics of landed castes; see Narayana Rao 1986: 13449, most notably the central
theme contrast between protection of caste integrity in the trader caste epics with
the control of territory in the martial epics (140) and the corresponding heighten-
ing of control of women by men in the martial epics (146). Cf. the virginization of
Kannaki in folk variants of the non-martial Cilappatikram (Beck 1972).
At this point, further discussion in the original of this chapter relating these three
virginal heroines to themes of sarees and hair is omitted, since it is more developed
in Hiltebeitel 1999a, and since the lh heroine Bl is discussed at length in chapters
16 and 17 below.


Fieldwork on the Tamil cults of Draupad and Kttn t avardeities

who are also heroic figures in the Mahbhrata, Indias classical San-
skrit epic2suggests that if one is to reflect on how they might open
onto a wider exploration of Indian folk religion, no topic is more
basic and in need of thinking (there is not enough to call it rethink-
ing) than that of diffusion.3 Looking at the index I prepared for my

This essay is a revision of the keynote address to the Regional Workshop on Folk-
loristics sponsored by the Ford Foundation for the Department of Folklore, St. Xaviers
College, Palayamkottai, Tamilnadu. I delivered it on 27 June, 1994 to launch a series
of twenty-one lectures on the workshop theme of folk religion. I thank S. D. Lourdu,
A. Chellaperumal, A. Dhananjayan, N. Ramachandran, A. Ramanathan, A. Sivasubra-
manian, and Bhaktavathsala Bharathy among the faculty at the workshop, and many
of the students, for their stimulating comments on this paper and the ones that fol-
lowed it. I thank Linda Lam Easton for prompting an earlier draft of this paper for a
1988 American Academy of Religion panel, and Jonathan Z. Smith for his thoughtful
comments as respondent.
My main studies of these cults pertinent to this article are Hiltebeitel 1988a, 1991a,
1995b, 1998, 1999a, and 1999b. On the Kttnt avar cult, see now chapters 1114
See Bell 1989, analyzing various scholarly strategies for treating folk or popular
religion in China. Marking an increasing preference for the latter term, Bell leans
toward the most recent of three approaches in which interpreters avoid two-tier bifur-
cations and do not isolate religious institutionsor religion per seas the data of
analysis; rather they focus on symbols and rituals in which they see the dynamics of
culture played out. Thus this approach implies a distinct theory of religion as a fully
embodied cultural system, while culture itself is presumed to involve the internal
generation of both distinctions and unities, and its holism is described as a function
of either underlying structures of some sort or the imposed limits of geography as they
moderate the degrees of similarity and difference (43). I think, however, there is at
least for the study of Indian folk religion a little too much stress here on the holism
of culture, not to mention its explanatory power for interpreting religion, for this
approach to be easily adopted in its entirety. I also think there are good reasons not
to abandon the term folk religion. Even though it was handed to me as a theme for
the workshop, it is for India a more focused term than popular religion, which as
Fuller has shown, can and must mean just about anything (1992: 5; his topic is defined
as the practical religion of ordinary Hindus taken as 83 percent of Indias popu-
lation). For this article, I try to formulate wherever possible a discussion of Indian
folk religion without prejudging it to be folk Hinduisma point urged on me by
S. Ravindran which I think well worth exploring further; cf. e.g. Bayly 1989: 173 on
commonalities in south Indian folk religion that cut across its Hindu, Christian, and
126 chapter six

first volume on the cult of Draupad (it is primarily this cult that I will
be discussing here), there are two rather ample references under the
heading Diffusion of Draupad cult that reach a total of twenty-two
pages (Hiltebeitel 1988a: 2131 and 3951). This usage in the book had
a critical edge to it. I intended it as one of those red flags you dont
really expect anyone to notice. Among historians of religions the term
diffusion has been in some disfavor since recognition of the excesses
of what Robert Lowie calls the extreme diffusionists (1937: 158;
cf. Harris 1968: 37982)the pan-Babylonianists, pan-Egyptianists,
and Kulturkreislehre theorists of the Vienna schoolto which can be
added the pan-Indianism of Max Mllers hypothesis of the Vedic
origins of mythology and Theodore Benfeys theories of the diffusion
of the folktale (see Claus and Korom 1991: 5760). At the time that
Marvin Harris could seek to expose the fundamental sterility of the
attempt to explain cultural differences and similarities by appealing
to the non-principle of diffusion in anthropology (1968: 377), I, as a
graduate student in history of religions at the University of Chicago,
was continually urged not to underestimate independent invention
or polygenesis as an alternative to the competing comparative strat-
egies of diffusion, universal archetypes, synchronicity, and something
like pure random coincidence. This was not only the emphasis of my
professor Mircea Eliade, as I read it, but of Stig Wikander and Georges
Dumzil, who, despite their obvious reliance on a model of diffusion
for studies in Indo-European society, thought, myth, and ritual, devel-
oped their arguments as a check on the supposed excesses of diffusion
theory, and always stressed the creativity of each Indo-European cul-
ture in reworking the diffused Indo-European material.
For post-60s (since my time) reflection at Chicago on the vague
language of diffusion, and, relatedly, that of syncretism, we have Jon-
athan Z. Smith to thank.4 Smith seems to call for greater clarity or pre-
cision on these terms, or perhaps improved alternatives: among other
things, because they have been misused in connection with we-they
theories, and because their explanatory value has been underestimated
in treating areas of high culture contact with primitive religions,

Muslim forms (Bayly puts only Hindu in quotes in this context, which strikes me
as problematic).
Smith 1978: 232; cf. 243, and, relatedly, 67 on syncretism; 1982: 2728.
orders of diffusion in indian folk religion 127

most notably in connection with certain erstwhile high gods (1982:

6689; 1987a: 115).
In Indology, with the notable exception of two monumental works
Gananath Obeyesekeres The Cult of the Goddess Pattini (1984: 26, 33,
474, 53038) and Madeleine Biardeaus Histoire de poteaux: Variations
vediques autour de la Deesse hindoue (1989: 111), both of which I will
return tothe term diffusion has usually gone unused if not avoided
in the literature, including that on regional folk cults comparable to
Draupads.5 And in some of the better recent Indology it seems to be
a term of opprobriun (Dumont and Pocock 1959: 4054, as cited in
Tanaka 1987: 17, 123): a point I will get back to shortly. In any case, as
my studies have developed, it has seemed important not to avoid the
term, though I have used it so far only in a very limited sense: that of
the diffusion, or alternately dispersal (similarly, Obeyesekere 1984:
33), or diaspora, of the Draupad cult itself. In this paper, after an
examination of some of the reasons why the vocabulary of diffusion
is particularly problematic within Indian studies, I will discuss what I
have called orders of diffusion in the Draupad and Kttn t avar cults
(although Im afraid that many other prepositions would also apply
here: for example, into, through, around, from, and beyond these
cults). My aims will be to seek some clarification of such usages with
regard to Hinduism generally and folk religious cults like Draupads
and Kttnt avars in particular, and to think through some of the the-
oretical issues raised by the study of folklore in folk religion.

Diffusion Beneath the Confusion

Many of the standard constructs of Indian religious studies can be

looked at as strategies for addressing issues of diffusion, presumed or
real, though most often without actually using the term. Most obvi-
ous are constructs positing Aryan or Dravidian influences: these have
been vastly over-used, and have unfortunately been polarized into a
mutual we-they scholarly mythology that has been as seductive to
non-Indians as Indians. It has taken two main forms: one focused on

See Meyer 1986; I have not reread the whole book to confirm that the term is not
used, but it is unused at points where it might be (e.g., 6264, Moving of the Goddess
from one place to another (i.e., transplantation of her cult); 15758, Geographical
Considerations; cf. Werbner 1977: ix, xi, iii, 197200.
128 chapter six

origins via archaeology and prehistory (for example, the largely aban-
doned Aryan invasion theory versus the so-called Dravidian sub-
stratum theory); the other positing models of influence within history
(for example, Aryan colonization versus The Occurrence of Tamil
Themes in Indo-Aryan6). Also, certain studies of the goddess have
been prone to posit some kind of neolithic and early urban diffusion
relating the supposedly Dravidian Indus Valley culture to the Medi-
terranean and the Ancient Near East, and to contrast the prominence
of goddesses in the Indus Valley with the relative lack of goddesses
in Aryan (including Vedic) traditions. But these constructs are frag-
ile. We seem to have moved farther and farther away from judicious
and meaningful application of these terms, at least in religious studies,
and closer and closer to a recognition that the only areas of accept-
able usage are in language and kinship.7 The Dravidian substratum
theory (see e.g. Zimmer 1951: 5960, 18486, 2172 1, 24142, 25
153, 31415) is one of a numberothers are the aboriginal substra-
tum theory (see e.g. Eliade 1958: 105, 29358), and the argument for
the Hinduization of tribal traditions that often disguises a scholarly
tribalization of Hinduism8that combine geographical diffusion
with a diffusion upward, in one form or another, from the bottom
or fringes of Hindu society to its presumed top or core. In reverse, one
finds the downward diffusion models that, as Dumont, Pocock, and
Tanaka have recognized, lie at the heart of M. N. Srinivass influential
but much-criticized concepts of Brahmanization, Sanskritization, and
their many sequels such as Kshatriyazation, Puranicization, and, yes,
Hinduization. I condense Tanakas useful summary:
According to Srinivas, Hinduism among the Coorgs can be classi-
fied into four levels: All-India Hinduism, Peninsular Hinduism,
Regional Hinduism and Local Hinduism . . . Spread is Srinivass
key concept for understanding the above four levels. There are two
types of spread. One is horizontal, that is the geographical spread
of a particular form of Hinduism. The other is vertical, that is the
penetration of a specific complex of beliefs and rites among differ-

The latter heading is from Hart 1975: 28591.
See Trautman 1981: passim. One can perhaps add to this the use of Dravidian as
an architectural term.
As in Eschmann 1978 and Eschmann, Kulke, and Tripathi 1978, presenting the
arguments of the joint German-Indian Orissa Research Project. See Hiltebeitel 1991a:
9799, 12528, 15559.
orders of diffusion in indian folk religion 129

ent castes within a particular language zone. Srinivas considers that

Sanskritic and non-Sanskritic beliefs and rituals are disproportionally
distributed among various castes. As we move down the caste hier-
archy from the Brahman to the Untouchable, the Sanskritic elements
gradually diminish (Tanaka 1987: 1516).
As Tanaka recognizes, the very concept of spread reminds us of
diffusionism, and poses an analytical inconsistency between a func-
tionalist holistic approach to the village level (that he traces in Srini-
vas to Radcliffe-Brown), and an explicitly diffusionist approach to the
levels beyond it (1987: 17). Moreover, as R. P. Werbner observes,
level-specific approaches (like Srinivass) show a bias toward nest-
ing relations rather than cultural variability, and would thus skew
any discussion of diffusion or spread in that direction (1977: ix, xi).
Meanwhile, in reverse of functionalist holism, structuralist holistic
approaches to Hinduism, which begin, at least implicitly, at the All-
India level, are uneasy with the language of diffusion as a means of
moving from that level to others, and stress instead the recurrence of
structures and replication of ideology without emphasizing or even
positing diffusion at all.9 Thus Madeleine Biardeau suggests that the
emotive aspects of bhakti come not from the Dravidian South or from
popular or tribal origins (standard conflicting views), but from a
sort of trickle-down effect of values from the top of the social scale to
the bottom, whence they percolate back up, suffused with new emo-
tion, to permeate the whole.10 Or along similar lines, in discussing a
certain type of post found widely in south Indian folk religion, she
argues that its disjecta membra across the terrain suggest not only a
dispersal through diffusion compounded by a loss of meaning, but
cumulative evidence for a continuing link with the primary source
of inspiration, with the culture savante des brahmanesa virtually
omnipresent vertical relation that revitalizes the symbols forms,
functions, and utilizations (1989: 11112). I will mention this post and
Biardeaus reflections on its diffusion again shortly.
Two more devices that scholars have drawn on to characterize
what they take to be peculiarities of Hinduism have made any seri-
ous discussion of diffusion within Hinduism, and within Indian folk

See Biardeau 1989a [1982], 1989b, and, applying similar principles at the regional
and local levels, Reiniche 1979.
Biardeau 1981: 176; see discussion in Hiltebeitel 1983: 207.
130 chapter six

religion, virtually impossible. One is the much overrated insistence

on Hinduisms absorptiveness and tolerance, which becomes an
excuse to say virtually anything, with Greek and Iranian influences
being especially invoked in the allegedly syncretistic, catch-all epic
period from the Mauryas to the Guptas. And the other is the latest
fad of arguing that there really never was such a thing as Hinduism
until the Muslims, or the British, named and invented it, which
becomes an excuse to say nothingother than that culture contacts
have had their impact, obviously much over-rated here, and that such
impact continues as the naming process of scholarship continues.11
Needless to say, the situation is a mess. Clearly one must avoid letting
intangible and poorly researched megaconcepts (Obeyesekere 1990:
25768) like great tradition, Sanskritization, Aryan, Dravidian, toler-
ance, absorption, and invention dictate the terms by which one seeks
to trace (or defuse) diffusion (cf. Claus and Korom 1991: 68), and
rather find terms that have tangible and practical value in relation to
what is being traced on the ground. But it must be said that most of
these theories, approaches, and evasions, no matter how shop-worn
they have become, have raised issues that must be addressed if we are
to seek greater clarity in talking about Indian folk religion, since nearly
all these theories are posed with folk religion as one of the problem-
atic areas to be explained. While I will not attempt to review them
again, they are all pertinent to what I will now have to say about the
Draupad and Kttn t avar cults.

Orders of Diffusion in two Tamil Mahbhrata Folk Cults

First, to resume briefly the aforementioned pages indexed under dif-

fusion of Draupad cult in Draupad I, the argument is that the cult
consolidates between the late thirteenth century and the end of the
fourteenth century (probably in the early fourteenth century) in the
region of Gingee (South Arcot District, Tamilnadu), that the Gingee
region is thus the core area of the cult, and that the dispersal of the
cult has gone through three main historical phases:
1. A diffusion throughout the Gingee core area itself, coincident
with the emergence and self-definition of Gingee as a Hindu kingdom

For a telling critique of this implicitly shallow standpoint, see Pollocks discus-
sion of deep orientalism (1994: 9697).
orders of diffusion in indian folk religion 131

(or, more precisely, a Vijayanagar Nayakate), with a high concentra-

tion of this dispersal involving the carrying of the cult into rural vil-
lages, with the founding therein of new temples. Let me add that it is
not possible to establish conclusively that the Draupad cult actually
begins in the core area. It can be shown, however, that it is the area
of its most intense regional diffusion: both horizontally across the
terrain, and vertically through the caste system.
2. The carrying of the cult beyond the core area. This no doubt had
already begun during the period of Nayak rule in Gingee, but it inten-
sified after that rule was brought to an end, in about 1648, with the
takeover of Gingee by a succession of increasingly distant outside rul-
ers: Bijapuri Muslim rule from 164877, Maratha rule from 167797,
and Mughal Rule, increasingly nominal, from 16971801. From 1718
on, the Gingee Fort was abandoned as a regional political center, with
Mughal overlordship centered in nearby Arcot while the real centers
of regional power continued to shift to French Pondicherry and Brit-
ish Madras. In these conditions, which were extremely harsh for the
Gingee area and resulted in depopulation within it and migrations
away from it, Draupad cult diffusion was directed primarily to the
larger towns and cities in surrounding kingdoms and districts within
present-day Tamilnadu and neighbouring states (areas in which the
cult has had minimal rural village spread, as distinct from its diffusion
through the core area).
3. Under colonial conditions, the cult extended its diaspora to areas
beyond south India, such as Fiji, Singapore, Malaysia, Reunion Island,
South Africa, and Sri Lanka. With the likely exception of Sri Lanka,12
none of these migrations seems to have occurred before the nineteenth
These three phases allow us to identify certain effects of diffusion
within, and through, the cult. Though it is not possible to be sure
what elements of the Draupad cult go back to its constitutive period
in the Gingee core area, certain elements today are either not found in
the Draupad cult beyond the core area, or are either more highly con-
centrated in the core area Draupad cult than they are in the Draupad
cult beyond the core area, or have singular features within the core

Tanaka 1987: 337 says that the Draupad temple at Udappu (Puttalam Dis-
trict, North Western Province, Sri Lanka) is traced by contemporary informants to
seventeenth-century migrations from south India.
132 chapter six

area. Most of these core area elements pose the possibility of discrete
studies of diffusion that range beyond the Draupad cult, since only
one of them is unique to it. I can only indicate the outlines of what
such a study might entail, and have chosen three examples. In select-
ing, I have chosen not to discuss the Mahbhrata itself, though it
does pose a number of interesting issues: as a medium for the diffu-
sion of Indo-European themes into Hinduism (as argued by Wikander
1960, Dumzil 1974, and Hiltebeitel [1976] 1990); as a means of the
diffusion of Hinduism into Southeast Asia and Indonesia; in the ways
and possible reasons that the epic story itself registers symbolically the
presence of outsiders (Yavanas or Greeks, Iranian akas, even tributes
from Rome and Antioch and China) within its narrative purview; and
in vernacular localizations, retellings, and oral folk epic reenplotments
of the classical epic within India from the medieval period on.13
In choosing these three examples, I wish to proceed from here on
by stressing two interrelated points. My criticism of, and skepticism
about, many of the implicit and explicit theoretical models for dealing
with diffusion in Hinduism that have now been discussed are based on
my conviction that much of the aforementioned mess can be at least
constructively re-sorted, if not cleared up, by a better understanding
of two things: what Hinduism draws from the Vedic revelation, and
how it extends that revelation through bhakti: the Mahbhrata being
one of the main sources and mediums of this extension, another being
the cult of the goddess. Indian folk religion has usually been studied
as if it allowed one to posit a preserve beyond these constructs, which
are assumed to be Sanskritizing influences imposed upon it. But
the notion of a hypothetical unSanskritized preserve is, I would say, a
dubious one, and fails as a working premise to account integrally for
Vedic elements and bhakti expressions that are found in Indian folk
religionunlike megaconceptson the ground. In this paper, I pur-
sue these matters by way of select Draupad cult examples that would
seem to draw from these inspirations.14
1. The unique feature just mentioned is a myth, known in the core
area but unknown beyond it, that describes how Draupad, heroine of
the Mahbhrata, was born a second time to come to the aid of an

This latter topic will be a subject in Hiltebeitel 1999a. Cf. Ramanujan 1991 on
retellings of the Rmyan a.
For further discussion, see Hiltebeitel 1983, 1988a, 1991a, and in 1999b for pre-
liminary discussion of the role of the horse at certain Kttnta var festivals.
orders of diffusion in indian folk religion 133

ancient king of Gingee to help him defeat a demon of the Gingee Fort
or Forest (Hiltebeitel 1988a: 7688; 1991: 48389). The myth identi-
fies her second advent with the stabilizing of the Gingee kingdom in
a remote past, well before what can be identified as Gingees actual
history, and explains the founding of her original temple in a vil-
lage just outside the capital. I have stressed the way this myth reflects
a Hindu little kingdom ideal that finds similar or at least compa-
rable expressions in the regional cults of other Indian folk deities
of royal and dominant caste clans (that is, kuladevats, clan dei-
ties), whose cults, like Draupads, seem to have been consolidated in
the twelfth to fifteenth centuries as part of a response to the incursions
of Islam (Hiltebeitel 1988a: 21, 1002). One might further see these
developments as locative traditions in the sense proposed by Jona-
than Z. Smith, involving archaicization of native traditions, revival
of myth, emphasis on the saving power of kingship over and against
monstrous demons (a hundred- or thousand-headed one in this case),
and, moreover, within the core area, the ritual and mythic localization
and poetic vernacularization (by Villiputtr l vr, ca. 1400) of the
Mahbhrata itself (cf. Smith 1978: xiii). It would not only be pos-
sible but fruitful to investigate these developments as part of a folk
religious bhakti continuum, with such locative expressions in ten-
sion with more universalizing utopian sectarian, pilgrimage, and cult
traditions, whichespecially in north Indiabecome increasingly dis-
sociated from the possibilities and values of royal implementation.15
Moreover, over time and space, it can be shown that the Draupad
cult loses this locative dimension. In the second phase of diffusion
outside the core area, not only is the Gingee setting and mythology
forgotten, but the myths and rituals are shaped to conform with more
Brahmanical and pan-Hindu models (indeed, such terms as Sanskriti-
zation, Brahmanization, and Puranicization are all serviceable at this
level). And in the third phase of diffusion outside of India, one finds,
at least in Fiji, the cult transformed into an expression of what could

In Maharashtra, an example of such a tension is found in the locative and
utopian cults of Khandob and Vithob respectively; see Stanley 1987: 5167 on
extended contrasts between the two traditions. Similarly, see the contrasting royal
mythologies of the Draupad (locative) and Kttn ta var (more utopian) cults in the
Gingee core area (Hiltebeitel 1999b). On uprooted utopian forms of bhakti under
Muslim sway in the north, see Hein 1986: 296318; Hiltebeitel 1999a (on the lh and
Pbj oral epics); 1995a (on Rs-, Pndav, and Rm-lls).
134 chapter six

certainly be called a utopian vision of pan-Hindu, and even pan-

Indian, unity (Brown 1984).
2. More frequently found at Draupad temples in the core area
than beyond it are representations of Draupads two guardians, Pttu
Rja and Muttl Rvuttan. In the fashion just indicated, the farther
one gets from the core area, the more Pttu Rja is fused with or
replaced by the pan-Hindu guardian deities Bhairava or Vrabhadra,
both violent manifestations of iva.16 Pttu Rja figures in the myth
of Draupads second advent. Draupad cannot sever the demons
last head unless someone, i.e., Pttu Rja, keeps it from touching the
ground: a mytheme that Pttu Rja almost certainly inherits from the
head-holding Bhairava (Hiltebeitel 1988a: 37282; 1989a: 36467). In
terms of myth and iconography, what is most interesting about Pttu
Rja is that he combines traits of both Bhairava and the buffalo demon
Mahissura, slain by the goddess Durg. Pttu Rja/Pota Rju (Tamil
and Telugu spellings) means the buffalo king, and seems, although
there is no myth to actually tell the story, to be the form taken by the
buffalo demon once he has been (sacrificially) slain and transformed
by his death at the goddesss hands (a bhakti theme) into her guardian
or demon devotee. Pota Rju is a popular guardian god for village
goddesses in coastal Andhra Pradesh. Indeed, it is the disjecta mem-
bra of his form as a post in front of village goddess temples in Andhra
that Biardeau has traced throughout so much of the Deccan and south
India, and in so many mythic, ritual, and iconographic multiforms
(1989; cf. 1984). His role in the Draupad cult can be shown to have
diffused southward from that center.
Muttl Rvuttan, on the other hand, is a Muslim, a Muslim turned
convert-devotee of Draupad the goddess. In Muttl Rvuttan we find
the incongruity of a Muslim devotee of a goddess whose Mahbhrata
mythology is (usually thought to be) pre-Muslim. Since he figures
most prominently in the core area, and is incorporated into some
Draupad cult Mahbhrata folklore, I take it as a sign that Islam was
a catalyst in the Draupad cults core area consolidation. He too is a
guardian of other gods (e.g. Aiyanr, Kutti Anta var) and goddesses
(e.g. Mriyamman) beside Draupad, but has so far been found as
such only within the southeastern parts of the Draupad cult core area

On both figures, see Hiltebeitel 1988a: 37273; on Vrabhadra (Vrapattiran) in
the Draupad cult south of the core area, see Hiltebeitel 1991a: 11216, 40629.
orders of diffusion in indian folk religion 135

and in the Tanjore District and beyond to the south.17 His incorpora-
tion into a regional Mahbhrata and goddess cult seems to represent
not so much a borrowing into Hinduisn from Islam as an adjust-
ment within a system of Hindu representations whereby he stands,
as a Muslim, in place of other traditional mythic and ritual embodi-
ments of outsiderhood and impurity.18 As suggested earlier, the classi-
cal Mahbhrata has similar mechanisms. But the importance of both
these figures for our present discussion lies not on the level of myth,
but that of ritual.
In coastal Andhra, Ptu Rja is represented by a type of post that
in Orissa, just to the north, is called a ypa. More specifically, what in
Oriya, an Indo-Aryan language, is called a ypa is in Telugu, a Dra-
vidian language, called a pota rju. Ypa is the term for the sacri-
ficial stake in the Vedic animal sacrifice. The Orissan ypa and the
Andhra pota rju are both used in animal sacrifices. These are not
Vedic animal sacrifices, however, which admitted only pastoral graz-
ing animals (horse, bull, ram, he-goat), but buffalo sacrifices. In Orissa,
these are connected with the worship of Durg, and the post is called
Khambhevar, Lady of the Post, a name for this aspect of the god-
dess, while in Andhra, as already noted, it is called buffalo king. In
other words, in Andhra the post is identified with the sacrificial victim,
or, more specifically, with a form that the victim takes after his death
to become an instrument of death for other sacrificial victims. This
transformation has its likely analogy in myth in the transformation
through death of the Buffalo Demon into the same Buffalo King. But

See Oppert 1893: 482 (connected with Mriyamman); Hiltebeitel 1988a: 103;
1989b{ADD b}: 368 n. 4 (with Kutti Anta var; A. Dhananjayan [personal communi-
cation, July 1994] has also found Muttl Rvuttan connected with Kutti An ta var in
southeastern South Arcot). He is found along with Pavatairyan and Maturai Vran
as guardians facing Aiyanr at the Aiyanr temple on the road between Nellikuppam
and Cuddalore in the same general southeastern South Arcot area.
Nonetheless, one does find Draupad temples where Muttl Rvuttans shrine is
attended by Muslims: e.g., at Palayamkottai (Palayamkottai Taluk, Tirunelveli Dis-
trict), where, as Muttalip Appa(rah), his green-cloth-draped tomb (dargah) is attended
by a Muslim family that makes offerings there of roti, plantain, flowers, perfumes,
and sweets (note the absence of the more toxic things he is normally offered else-
where (Hiltebeitel 1988a: 12526; 1989b: 36365). He is supposed to have worshipped
Draupad by circumambulating the temple after going to war, and then died at the
spot, and the cooperation of Hindus and Muslims at the temple has been recently
cited by a Tamil newspaper article as an instance of Hindu-Muslim unity. I thank
Mr. Vincent S. Paul for this information (personal communication, July 1994) and for
his assistance, along with Mr. S. Ravindran, at the temple.
136 chapter six

the sacrificial logic also has its Vedic precedent. The Vedic ypa, when
cut, becomes the first victim of the animal sacrifice, and then becomes
an instrument of death for other victims. The ritual placement of the
pota rju post at the outer eastern boundary of the goddesss temple
(like the Lady of the Post of Orissa) also has its precedent in the
position of the Vedic ypa on the eastern boundary of the great altar
(mahvedi) used in animal sacrifices. And, most crucially, the shape of
the pota rju post is itselffollowing Biardeaus argumenta stylized
modification of the Vedic ypa. Moreover, through procedures that
one might well identify by Claude Lvi-Strausss terms of parcelling
out and repetition (1981: 67274), both the Vedic ypa and the pota
rju stake are bundled with other sacrificial implements that extend
and replicate its functions in mobile-portable forms. Thus the first
chip from the cutting of the Vedic ypa tree is tied to the ypa, then
handled in the ritual as a sort of mobile ypa, at one point jointly with
the sacrificial knife; and the piercing function of the knife is then par-
celled out to the spits or prongsone a double spit (dvila) and the
other a single spit (ekala)that pierce the animals omentum, the
ultimate offering-part of the animal, while it is held over the sacrificial
fire: all of which finds its counterpart in ways that Pota Rju is ritually,
mythically, and iconographically linked with ritual weaponsmost
notably, the sacrificial knife, trident (trila, which I take as a variant
of the three prongs of the two Vedic spits that pierce the omentum),
and whip (a variant of the cord that binds the victim to the sacrificial
stake)that are used in Andhra in the buffalo sacrifice, and in the
Draupad cult (Draupad is said to obtain these from Pttu Rja) to
ritually reenact the Mahbhratas sacrifice of battle as a victory of
the goddess Draupad.19
If we can trust these constructs, we can identify two orders of diffu-
sion which, as we have seen, Biardeau formulates in a kind of tension.
One is a diffusion of Vedic elements into folk Hinduism. In this vein,
it must be observed that while the pota rju of Andhra retains (and
probably begins with) the character of a post connected with animal
sacrifices, Draupads Tamil Pttu Rja is usually represented anthro-
pomorphically, and thus rarely as a post,20 and is also only rarely con-
nected with animal sacrifices (whose symbolism is transposed to the

See Biardeau 1984; 1989: 1112; Hiltebeitel 1988a: 33393; 1991a: 11765.
See Hiltebeitel 1991a: 10116 for a discussion of all cases known to me.
orders of diffusion in indian folk religion 137

epic sacrifice of battle). But as if by compensation, the Draupad cult

picks up on another Vedic paradigm that is unrepresented in rela-
tion to the Andhra pota rjus. It places its icons of Muttl Rvuttan,
the outsider who handles impurities, to the northeast of Draupad
temples: a spot that corresponds in the Vedic sacrifice to the place
where a low status officiant called the amitar does the actual kill-
ing and butchering of the sacrificial animal outside the mahvedi (see
Hiltebeitel 1988a: 12527; 1989b: 36466; 1991a: 14247). This dif-
fusion can only be ill-described by such constructs as Aryanization,
Brahmanization, Sanskritization, or little and great tradition. Nor can
one resort easily to the reverse construct of the so-called Dravidian buf-
falo sacrifice promoted by late nineteenth-and early twentieth-century
researchers, who included Pota Rju in their purview.21 There does not
appear to be anything solidly Dravidian about this Pota Rju complex.
To rephrase Biardeaus argument, it is not simply a question of trac-
ing spread from one place to another, but of understanding how
the Brahmanical source of inspiration in the Vedic revelation diffuses
itself. This has been accomplished in many ways, but one that has gone
largely unattended is diffusion from region to region and cult to cult,
through ritualand specifically sacrificialidioms, forms, and struc-
tures of folk Hinduism. Brahmans and Sanskrit are ingredients in this
process, but not such as to supply its name, or justify a name that is
intended to exclude them.22
For the second type of diffusion, for which Biardeau uses the image
of disjecta membra, one might turn Lvi-Strausss terminology to
advantage. This type of diffusion is in large part the extension through
space of the parcelling out or fragmentation (Lvi-Strauss 1981:
673) of the ritual implements connected with the pota rju stake and
the am tree, from whose wood Andhra pota rjus are made. This
wood is also used for one of the Vedic fire sticks, and Biardeau has
brilliantly traced such connections and transformations through many
folk religious contexts that draw on associations between Pota Rju,
the am, and fire (1984; 1989: 5062, 28186, 299317). Some cults,

Most notably Oppert 1893: 461, 476; Elmore 1915: 1831, 8588, place facing
p. 92, 99, 129; Whitehead 1921: 3940.
See further the comments of Zvelebil 1994 on my surprising thesis that the
singular and durable features of what is ostensibly the ritual of a folk or popular cult
are precisely the ones that take us beyond folk traditions, as such, to the classical and
even Vedic ritual systems of Hinduism (quoting Hiltebeitel 1991a: 3).
138 chapter six

including the Draupad cult, even have myths about the migration
(diffusion?) of such elements.23 It is, however, important that we be
careful about notions of loss of meaning on the one side, and Vedic
knowledge on the other. This is no doubt a documentable process (see
e.g. Biardeau 1989: 11922). But there are too many types of knowl-
edge, ignorance, and forgetting to let such a formulation stand alone.
To begin with, one may ponder that at least two of these cults have
folklores that parody Pota Rju-Pttu Rja as the gluttonous, stupid,
and ignorant Brahman.24
3. In the core area, battlefield rituals that reenct the Mahbhrata
war take place on a terrain called the patukalam, the place of dying
or of lying down. It is important to stress that this terrain replicates
much that can be found in the temple, including multiforms of what
has been discussed above. This term patukalam occurs in other god-
dess cults in Tamilnadu, and is a variety of what can be called kalam
rituals that also extend into Kerala, Andhra, and apparently Tulunad.
Kalam itself denotes both battlefield and threshing ground, pointing to
a connection between war and rice (and grain) cultivation. Core area
Draupad cult patukalam rituals differ markedly from Draupad cult
rituals that reenact the Mahbhrata war outside the core area, but
are similar to patukalam rituals for other goddesses, also outside the
Draupad cults core area (but possibly associated with what might be
called core areas of the cults of these other goddesses).25

See Roghair 1982: 16071 on the bundling of weapons and ritual implements
with Pota Rjus am tree multiform on the people of Palndus migration to the
south. Cf. Hiltebeitel 1988: 33667, 38293.
See Roghair 1982: 20414; cf. 332 on the ignorance of destiny of the Brahman
Anapotu Rju, Older-Brother-Potu Rju; Hiltebeitel 1988a: 35461. On Pota Rjus
gluttony as a likely Brahman trait, see Roghair 1982: 215 n. 12; Hiltebeitel 1991a: 125,
where I did not note that Roghair had preceded me in making this suggestion. It may
be further proposed that Pota Rju parodies not only the Brahmans demand for food,
but the sacrifices demand for bloodshed.
See Hiltebeitel 1991a: 16682; 283438. According to Peter Claus (1992), the
garodi, or gymnasium, is probably cognate with the Malayalam kalari where (the
Kerala martial art of) kalaripayattu is practiceda connection with Malayalam and
Tamil usages of both kalari and kalam that extends from gymnasium to ritual ground
and battlefield (e.g. patukalam, alukalam). See the review by A.V. Navada of Bannanje
Babu Amin and Mohan Kotian, Tulunaada Garodigala Saamskritika Ashyayana: A
Cultural Study of the Garodis of Tulunad Karnataka Folklore Newsletter 2,4 (rest of
citations incomplete): The authors argue that the guru stambha, seen even today in
the garodis, is actually a wrestling pillar (mallakambha). Originally centres for training
in battle techniques, the garodis became memorials to the heroes, Koti and Chenayya,
after their tragic death on the battlefield. Gradually, they became places of worship.
orders of diffusion in indian folk religion 139

Now one of the most important rites included within the patukalam
complex in Draupad cult core area rituals (but not in Draupad fes-
tivals outside that area), and also included in the cults of other god-
desses outside the Draupad cult core area, is an enactment of the
death and revival of one or more of the goddesss male relatives. In
Draupads case, the theme of revival is unevenly found at her core
area festivals, but where it is found, the revival is of her five sons (who
in the classical Mahbhrata die, but are not revived, thus underscor-
ing the significance of this ritually inspired transformation of the
story) (Hiltebeitel 1991a: 33954). In other cults, where, with no older
story to alter, the revival theme is more consistently pronounced, it
is the goddesss brother(s) or husband who dies and is brought back
to life (ibid., 35566). Here we have one of the classic problems for a
theory of diffusion, since it would be a mistake not to recognize that
the revived heroes, in their cults, are dying and rising gods.
Obeyesekere has taken one route in addressing these issues in con-
nection with the cult of Pattini. The oldest version of her story, the
Tamil Cilappatikaram, probably dates from around 400600 A.D., and
Obeyesekere regards it as our oldest source on the dying and rising
pattern in South Asia. The heroine Kan n akis (Pattinis) lamentation
momentarily revives her unjustly slain husband Kvalan so that she
can speak to him before he goes to heaven, where she will soon join
him after carrying out the vengeful rectifying acts that establish her as
a goddess. Obeyesekere opts for an argument of diffusion from West
Asia, not ruling out ancient contacts via the Indus Valley Civilization
(1984: 530), but choosing rather to limit the argument to continuous
and intensive contact between West Asia and Kerala, so as to render
plausible the hypothesis of the West Asian origin of the Pattini cult
(ibid., 53031). The argument has more recently been taken up by
Fynes (1993), who has interesting things to say about Isis and about
Red Sea-to-India trade in general, but no new research on Hinduism
to back his comparisons.26
Obeyesekere argues that Pattini is originally a Jaina-Buddhist deity,
fitting his hypothesis that the diffusion from West Asia was carried
out through coastal trade routes that penetrated into the popular

The garodi type of hero shrine is distinctive of the Billavas. My thanks to Purushotam
Bilimale for this information (oral communication, July 1994).
The next three paragraphs repeat and partially revise and update material from
Hiltebeitel 1991a: 36668.
140 chapter six

[Obeyesekeres italics] religion of the new local entrepreneurs to be

grafted onto the worship of the goddess that was, according to him,
a preexisting feature of Dravidian folk religion (ibid., 53135; quote
from 533). The Jaina-Buddhist argument is buttressed by the com-
ment that the drama of death and resurrection is alien to Hindu-
ism, a claim supported by the admission that, As far as I know the
dramatic enactment of the death and resurrection theme is not found
anywhere in Hindu ritual, though the theme of the wife resurrecting
her husband is found in myth (ibid., 530). The point is reiterated by
Fynes (1993: 385): The cult of a mother goddess who resurrects a
dead god is foreign to Indian religion. He has clearly gone no further
than Obeyesekeres discussion for information on India, and uncriti-
cally accepts it to argue for what he aggressively calls the non-Indian
origins of the original elements of the Pattini cult (38586).
This is all very shaky. I leave aside the issue of whether the cat-
egory of dying and rising gods is a misnomer for the ancient Near
Eastern and Mediterranean pattern argued for by Sir James George
Frazer.27 Although Obeyesekere calls upon Frazer for background sup-
port, his argument allows for an interpretatio Christiana of the dying
and rising theme by the time it would have allegedly influenced the
Pattini cult (1984: 532). I am still not clear what the evidence is for a
pre-existing Dravidian folk mother goddess cult, not to mention the
evidence that it did not have dying and rising gods. As far as I know,
all evidence on popular Dravidian goddess cults comes from no earlier
than the medieval period. Nor am I convinced that Pattini, even in the
Cilappatikaram, can be claimed as originally Jain-Buddhist but not
Hindu.28 Indeed, the Cilappatikaram itself is also about the Pandyan
king of Madurai and especially the Cera king of Vaci (Zvelebil 1994:
120) who seem to be described in ways that are more Hindu than Jain
or Buddhist: the first, who proclaims Pattini as a goddess, and the
second who establishes her cult. But more to the point, the drama
of death and resurrection is hardly alien to Hinduism, or undocu-
mentable in Hindu ritual. It is found in south Indian kalam rituals,
including the Draupad cults patukalam ceremonies. More classically,
one finds it in the iconography of the revival of iva by the dancing

For an argument that the category dying and rising gods is a misnomer prior
to the interpretatio Christiana, see Smith 1987: 52127; 1978: 72.
Cf. Zvelebil 1994: 120 in agreement, especially on the Cera King Cenkuttuvan,
and Fynes 1993: 386, relying on Obeyesekere.
orders of diffusion in indian folk religion 141

goddess in Tantric iconography. And it probably has deeper roots,

and, again, possibly Vedic precedents, in the capacity of the sacrificers
wife to revive the sacrificial victim, who symbolically represents her
sacrificer-husband (see Hiltebeitel 1991a: 36680).
Ethnographic evidence for kalam and patukalam rituals is, of course,
relatively recent, and none of the cults in question can be traced back
prior to the twelfth-to-fifteenth centuries. There is, moreover, no guar-
antee that they included these dying and rising rituals from the begin-
ning. One could thus argue that the Pattini cult has diffused this theme
into medieval and later folk religious traditions: something that is,
indeed, possible, at least as a contributing factor.29 There is, however,
an important difference between Pattini traditions as we find them in
the Cilappatikaram, and kalam and patukalam traditions. Whereas the
Cilappatikaram centers its story on a hero and heroine from merchant
families, and is written by a Jain (all of which feeds into Obeyesekeres
Jain-Buddhist hypothesis), kalam and patukalam ritual traditions, and
the epic and folk epic traditions connected with them, are centered
on, and can indeed be said to be the property of, landed dominant
castes. Moreover, I suspect that the same could be said for the ritual
reenactments of the Pattini myth in Sri Lanka, which are village ritu-
als (gammaduva) with a royal prototype (Obeyesekere 1984: 3435,
119) that contain an ideology like that in south India where dominant
landed caste rituals on the village and regional level stress the rap-
port between kingship and the prosperity of the land. In other words,
patukalam rituals of death and revival, like the Sri Lankan gammaduva
for Pattini, are expressive of a rapport that is difficult to explain by or
trace back to the interests of merchants. Here we can simply turn the
tables and point out that the Cilappatikaram is not evidence, or at least
direct evidence, of a ritual. It is a myth: one behind which it would be
just as easy, and no less hypothetical, to read a folk goddess cult that
already had dying and rising rituals with a royal-landed ideology as it
would be to read a merchant ideology that had savored and borrowed
a dying and rising drama from West Asia.
In any case, we can pose a diffusionist explanation of why Draupad
cult patukalam ceremonies include such dying and rising rituals in the

I would, however, continue to stress that arguments for one-to-one influences,
including the replacement of the Pattini cult by such other cults as Draupads, must
be viewed suspiciously; see Hiltebeitel 1988a: 17 n. 11, 149, 201 n. 25. The latest to sign
on to Draupad as a replacement for Pattini is Fynes (1993: 382).
142 chapter six

core area, but not outside it. It is because those who transplanted the
Draupad cult away from its heartland could not retain their status in
these new areas as dominant landed castes. Away from the core area,
such rituals lose their meaning, and are replaced by others (Hiltebeitel
1991a: 3034, 32438, 399429). One must thus propose that whereas
patukalam rituals of dying and rising gods have diffused from cult to
cult, they do not diffuse within such cults (at least in the case of the
Draupad cult) beyond their core area.
Yet we should not be too quick to dismiss the larger model of diffu-
sion from the Near East out of hand. One can compound the problem
by one more set of considerations. Core area Draupad festival rituals
involve a number of reciprocities between what goes on in the temple
and what goes on at the patukalam. One of these is a correspondence
between the dying and rising rituals just described and the ritual per-
formed in and around the temple of raising sprouts of nine grains,
usually for ten days, before submitting the premature seedlings either
to desiccation beneath the hot sun or submersion in water. Though one
will hear from some informants that thats it, they just die, I have also
seen a ritual where their continued growth by the side of the tank is
said to be a miracle that results from calling upon Krsn a. There is also
a Draupad cult myth which tells that the original seeds for the nine
grains were blackened by roasting to prevent their seeding, and then
miraculously revived by Krsn a (Hiltebeitel 1988a: 3019; 1991: 5378).
Here we have the Indian counterpart to the Gardens of Adonis that
Frazer drew into his discussion of dying and rising gods (1907: 182
216); and at least in the Indian case it would be a mistake to say the
rituals contain no hint of rebirth (Smith 1987b: 522) or that they lack
symbolic connections with agriculture. In this case, missing links of a
much older diffusion between the Ancient Near East and India have
been proposed from the Indus Valley Civilization and Indian tribal
sources (Haekel 1972). However, the tribal rituals in question have a
structure that can clearly be traced to the Hindu rituals of Navartra
and Dasar: that is, the fall royal festival to the goddess whose village
festival counterpart often connects Pota Rju with village goddesses,
nine grains ceremonies, etc. Again, it is also possible to identify proto-
types for the Indian forms of these rituals in the Vedas: in the case of
the sprouting rituals, in the rite of devas offerings that involve quickly
sprouting plants that form part of the preliminaries to the royal con-
secration within the Vedic Rjasya sacrifice (Hiltebeitel 1991a: 57,
77). The great Vedic royal sacrifices have clearly left their traces in
orders of diffusion in indian folk religion 143

the rituals of folk Hinduism via their transformations in bhakti fes-

tival practices centered on the goddess. But in this case, I would not
go so far as the anti-comparativism of Marcel Detienne (1977) or
M. Zanen (1978)both of whom have contributed important studies
of Mediterranean Gardens of Adonisin ruling out diffusion on the
grounds that India and the Ancient Near East are totally different
culture areas.30 Here we may someday have enough evidence to argue
for an ancient diffusion that includes the Vedic examples as traces of
something still older.

Folk Religion and the Diffusion of Folklore

Drawing on these three examples, it is possible to move to some wider

reflections on what folklore studies can learn from the study of folk
religion. The first and most obvious lesson is that, important as they
are, one must go beyond the ethnographic present and the perfor-
mance context. With regard to Indian folk religion, one must be able
to think in depth about Indian religions: Hinduism especially, but,
as we have seen, also Islam, and all the rest as well.31 And one does best
to expect that rather than solve problems, the conventional scholarly
dichotomiesgreat tradition/ little tradition, Aryan/ Dravidian, San-
skrit/ vernacular, Vedic/ classical/ folk, Brahmanical Hinduism/ folk
Hinduism, north/ south, high/ lowwill more often than not create
problems. This is not to say, however, that they are always false prob-
lems; just that they will always be complex, political (note the use of
capital letters), and, where at all interesting, generative of new ques-
tions. Folk religion is an area of permeable boundaries in which such
dichotomies are interesting because folk religion shows them at work
in conditions that are changing and alive.
A second lesson is that folk religion has specific kinds of folklore.
As Propp suggests, one wants to get to the idea of a work of folk-
lore, which, in the case of Russian bylina, will, he says, always be
connected with the struggle of the people at a definite time (1984:
158). But it is all too often that we find folklore, and especially folk

Zanen 1978: 1. Cf. Harris 1968: 37477, critiquing the concept of culture area
as a component of diffusionism.
For the possibility of a major Buddhist carry-over into Draupad cult folklore,
see Hiltebeitel 1991a: 42729.
144 chapter six

religion, interpretedby Propp includedin ideological fashion, and

indeed defined as ideology.32 In the Indian case, folk religion must,
I think at least usually, be understood as bhakti. And as such, it draws
on bhakti traditions of considerable variety and depth, including not
only medieval sectarian ones, but the Mahbhrata and the Rmyan a
in classical, vernacular, and folk forms that involve recurrent bhakti
themes and practices that include sacrifice, possession, and distinctive
understandings of land and the goddess. If bhakti is an ideology, that
is one thing. But it must be understood as a distinctive ideology. If
one invokes ideology as an explanation of a folk religious production
without some sense of the depth and specificity of its bhakti (which
can be quite varied), one runs the risk of serious distortion.
Third, in studying the distinctive folklore of folk religious cults,
one must think inter-cultically. As we have seen in discussing kalam
and patukalam rituals, different cults have variants of common ritu-
als and myths related to them. Again, it is a question of permeable
boundaries, or, as I have argued in one Kttnta var cult study (Hilte-
beitel 1999b), of frames and what is beyond the frame. Whereas
the Mahbhrata frames certain consistent but different themes for
both Draupad and Kttn ta var festivals, the two cults differ in what
is beyond the Mahbhrata frame: at Draupad festivals relatively lit-
tle (see Hiltebeitel 1988a: 139); at Kttn ta var festivals a great deal
and, as one should expect, particularly things that relate Kttnt avars
worship to the cults of goddesses: in South Arcot District, especially
Kl iyamman and Anklamman; in North Arcot District, especially
Kenkaiyamman (Gang); in Salem District, especially Mriyamman;
in Coimbatore District, Mriyamman and Pommiyammanand usu-
ally, in all these regions, also Draupad. Performers and participatants
move freely between folk cults, and their folklores reflect this.
Finally, one should also think interregionally. Pota Rju can be
traced from Telugu to Tamil, from Andhra Pradesh to Tamilnadu,
from the cult of the heroes of Palndu to the Draupad cult, and indeed
through much of the Deccan. This is not the only such connection,

The term is used variously by Narayana Rao 1986 (referring mainly to castes);
John D. Smith 1989 (exposing an allegedly hidden theodicy); Claus and Korom 1991:
67 (critiquing forms of urban and regional chauvinism that have become a substi-
tute for thoughtful analysis of the real ideology which lies behind folk traditions);
cf. Nirmala Devi 1987: xixiv, speaking of class/caste struggle without using the term
orders of diffusion in indian folk religion 145

and indeed, like any good connection, it opens the way to others. Such
a figure carries a folklore that links up with other folkloresincluding,
as the Draupad cult shows us, a folklore of the Mahbhrata. This is
a folklore that ties in with distinctive forms of folk religion not only
in south India but north India.33 One of the most successful outcomes
of those who have divided India between great and little traditions,
or pan-Indian (Sanskritic) and regional and local (vernacular) tradi-
tions, is the prevailing view that folk religion is never more than local
or regional (see Blackburn 1989). There are good reasons to challenge
this view.

See for now Hiltebeitel 1989b: 41321. I will take this issue up in greater depth
in Hiltebeitel 1999a.


This chapter will discuss some of the reciprocal issues raised by the
ll traditions of the Tamil Draupad cult for the study of ll tradi-
tions more widely: in particular, in the north Indian dramas of the
rm, rs, and pn dav lls. The accent will thus be on lls as dramas,
looked at primarily in their theological and ritual contexts. Draupad
cult dramas are normally designated by two terms evoking dance-
drama: one Sanskritic (ntakam, used in Tamil, but from the Sanskrit
word for drama), the other Tamil (kttu, with dictionary glosses of
dance and drama).1
There is, however, also one more term for the dramas that are per-
formed at Draupad festivals. The authors who composed chapbook
editions of some of the most important plays, from the early nine-
teenth to the early twentieth centuries, called them vilcam. The term
includes much the same semantic range as ll: dalliance of men and
women, a kind of dramatic rhyme (Fabricius 1972, s. v.); sport, play,
pastime, pleasure, diversion, dramatic composition (Tamil Lexicon,
1982 s. v.). One of the dramas to be called a vilcam is r Kan n an
Jalakkirtai yennnnum Kirusn a Vilcam (r Krsn as Water Sports,
called Krsn a Vilcam). This play is about Krsn as youthful sports: from
his birth to his water sports or water play (jalakkirtai)2 with the
Gops, referring to the episode when he steals their sars while they are
bathing. The cycle of Mahbhrata dramas thus begins with Krsn as
youthful vilcam-ll-krdahis play-play-playand extends to
other vilcams in which he is at play with the Pndavas. The impli-
cation is that his play extends from his youth to his epic adulthood,
in which, among other things, he restores the sars of Draupad just
as he stole those of the Gops (Hiltebeitel 1988a: 183 90).3 Finally,

There is also, however, a folk etymology bringing kttu close to the meanings
play, sport; see now chapter 13 below.
Sanskrit jalakrda: krda, Sanskrit play, sport.
Other than the drama about Krsn as youth, whose author is unknown, the
only other Draupad festival dramas I know of that are called vilcam are all by
Irmaccantira Kaviryar: Dice Match and Disrobing, Arjunas Tapas, Bending of
148 chapter seven

at Draupad temples across the Tamilnadu-Andhra border, in Chit-

toor District, Andhra Pradesh, the preliminary drama on the youth of
Krsn a is just called r Krsn a Ll (Chandra Shekar 1961:108).
It is my sense, however, that for all the ll-drama traditions we
are discussing, ll is more than drama. For rs lls it is a whole atmo-
sphere of imaginative evocation of religious moods and sentiments
(Hawley 1981: 16, 40, 239). For rm ll and pn dav ll it would seem
that ll means a whole festival, with mutually reinforcing recitative,
dramatic, and spectacular components, or as Richard Schechner has
put it, texts (Schechner 1983: 239, 278). Nor, of course, are these
categoriesatmosphere versus festivalmutually exclusive in the
instances cited. The threefold composition of the rm and pn dav lls
is basically congruent with the three components of a typical Draupad
festival, which involves (1) recitation of the Tamil Makpratam by
Villiputtr l vr (called prata piracankam), (2) terukkttu dramas,
and (3) ritual (Hiltebeitel 1988a: 139; 1988b, Part 1). The ritual high
points involve comparable spectacles to those at rm lls, as we
shall see.
A Draupad festival is not called a ll. The usual Tamil word for
festival is used: vil or tiruvil . But there are special reasons to sin-
gle out Draupad festivals as being more directly comparable to our
three examples of north Indian lls than most other south Indian,
or at least Tamil, festivals. First, as just mentioned, Draupad festival
drama cycles often begin with a play about Krsn as childhood, fusing
his youthful lls with his adulthood among the Pndavas. One might
thus see the Draupad festival dramas as a fusing of rs ll elements
(though minus that particular scene: Krsn as circle dance with the
Gops) and pn dav ll elements. I would not, however, suggest that
the similarities result from any direct influence either way. Second, as
I will discuss further, the Draupad festival, pn dav lls, and rm lls
share affinities through common connections with Dasar. Before tak-
ing up that issue, however, I would like to present an overview of some
of the more instructive points of connection, comparison, and contrast

the Bow for Draupads Marriage, and The Poison Pond (Naccuppoykai, in which
the four younger Pn davas drink water from a pond and fall dead, until Yudhisth ira
saves them by answering a Yaksas questions) (Hiltebeitel 1988a: 45657). It is
Irmaccantira Kaviryar who seems, in the early nineteenth century, to have played
the seminal role in reworking traditional oral and palm-leaf versions of the plays into
authored chapbook publications (15767). It may be that he called the plays vilcams
to link them with the Krsn a cycle and/or give them a Sanskritic tone.
draupad cult lls 149

between Draupad cult lls and these northern counterparts, and

then focus on the chief point in a Draupad festival where recitation,
drama, and spectacle converge in a ll-like atmosphere: the death
of Duryodhana.

Avtars and Icons

It can hardly be accidental that lls of the type we are discussing con-
cern themselves primarily with three textual themes: Mahbhrata,
Rmyan a, and the youth of Krsn a (textually, that is, first Harivam a,
then Bhgavata Purn a). The first thing this should tell us is rather
obvious: that ll traditions are rooted in bhakti, as are these texts: no
less in their Sanskrit versions4 than in their vernacularizations, which
our lls enrich and build upon most directly: the Rmcaritmnas of
Tulsidas and the great Hindi poets (Hein 1972: 15657) for the rm
and rs lls, the Tamil Makpratam of Villiputtr lvr for the
Draupad cult.
It may be that one can find partially or even wholly comparable
phenomena elsewhere, particularly in connection with folk epic-based
hero cults (Beck 1982; Roghair 1982) and goddess cults, as, for instance,
especially in the Pattini cult described by Obeyesekere (1984). But
whatever the affinities, these traditions do not rely ultimately on what
I have suggested elsewhere (1988a: 185) are the three great founding
mythological texts of the Sanskrit bhakti tradition. As far as I know,
there is no ll tradition based on any of the aiva Purn as, or on
the Dev-Mhtmyam, the founding Sanskrit bhakti text for the god-
dess (though Hein mentions something comparable, the nirtakali in
Bihar, where, as reported by Francis Buchanan, the impersonation of
iva and Prvat was done entirely by boys [Hein 1972: 266]and,
of course, Durgpj in Bengal and Orissa is in some ways a ll-like
Dasar variant of rm ll).5 The obvious point here, however, is that

It is gratifying to see Pollock come to the formulation that the divinity of Rma
as divine king has been integral to the Rmyan a since it took on its monumental
form under the hand of the poet or poets we call Vlmki. Pollock interprets this as a
current of political theology that runs through the text (1991: 1554, 6465, 68). It
remains to explore the relation between this political theology and what Madeleine
Biardeau calls the universe of bhakti, which she sees pervading both epics and the
Harivam a (Biardeau 1976, 1978).
Coburn (1991: 168) shows how a modern-day devotee of the Dev-Mhtmyam
can interpret the main action of the text as the Mothers cavorting in, and as, the
universe (ll). But this involves no dramas.
150 chapter seven

our main ll traditions have not only to do with bhakti but with
avataran a and, more specifically, the human descents of Rma and
Krsn a (young and middle-aged).6
This is nothing new; all I am attempting to do is alter the depth of
field. As Hawley has nicely said, It is consistent . . . that the theology
of the avatar is usually conceived in dramatic terms . . . . In Vaisn avite
thought the whole world process is conceived as a drama. All the world
is literally a stage and God (Visnu/Krsn a), in many forms, is the actor.
He creates, if it can be called that, in play only (the word is ll) and
the form of what emerges is dramatic. Then, in a special role, he enters
the drama himself: that is what happens when Krsn a is born (1981:
60). In descending to earth from time to time he plays various roles:
these are his avatars (16). But avataran a, as we know from Madeleine
Biardeau (1976, 1978), involves not just Visn u but iva, the goddess,
and indeed the whole Hindu cosmology, including its social order.
Thus when Krsn a is born, his indispensable dark foster sister is the
goddess Yoganidr.
Looking at ll dramatic traditions comparatively, rs and rm lls
are certainly the most prestigious and widely known. Mahbhrata
drama cycles are less common, or at least less publicized. Rs and
rm lls seem to be largely the expressions of high-caste traditions
of Brahmans and merchants, and of panregional values. The central
actors, the svarps, who impersonate the principal figures of devo-
tion (Rdh and Krsn a; Rma, his brothers, and St), are Brahman
boys: chosen in both ll traditions for their good looks, good fam-
ily, and so forth (Schechner 1983:265; Hein 1972:13536, 158, 227;
Swann 1990b:228, 234 [noting exceptions]): a selection process that
appears to go back to the very founding of these lls in the sixteenth
to seventeenth centuries (Hein 1972: 1089, 22427, 274). Rs lls at
Brindavan are patronized by the towns Brahman temples (159), while
seeming to attract the occasional royal patron such as Jai Singh of
Jaipur during his tenure as governor of the Moghul province of Agra
(17211728) (Hawley 1981: 277; cf. Swann 1990a: 210). The maharja
of Banaras, meanwhile, is a Brahman himself, like the svarps (Schech-

See now Couture 2001, showing that such formations from avatr in the
Mahbhrata and the Harivam a correspond to a technical usage in classical Indian
theatre designating the entrance or descent of a character on stage.
draupad cult lls 151

ner 1983: 266).7 Elsewhere, rm lls seem to be patronized above all

by merchant classes (Sax 1991b: Swann 1990b: 21819), also by Brah-
mans, and sometimes by municipal governments (Hein 1972: 9596).
No doubt numerous castes and even Muslims (9596) are among
those who patronize the rm ll in various cities and towns, but the
Ramnagar rm ll certainly realizes an ideal, if it does not necessarily
set a norm.
Mahbhrata dramas, at least in the two cases under discussion,
seem to be more regionally intensive and cult-specific, andthough a
full spectrum of castes, including Brahmans, may have festival roles
they are tied in primarily with the values of the dominant landed castes
who sponsor most of the festivals in which the plays are performed.
Rather than focusing on relatively pure male deities who should be
impersonated by Brahmans, they focus on female deities whose violent
and impure side is registered by their associations with Kl.
As Sax informs us, the pn dav lls involve ritual dramas sponsored
by dominant-caste Rjpts, who are regarded by themselves as well as
other castes as Ksatriyas. They claim descent from the Pndavas and
regard their region as one in which certain epic eventsthe Pn davas
births, their Himalayan ascent to heaventook place. The Pn davas
are regarded as personal deities (ista devats). The lls are ways of
worshiping Kl, and Draupad is regarded as Kls avatra (Sax 1986,
1987). In one case where a predominantly Brahman village sponsored
a pn dav ll, it was regarded by Rjpt informants as a matter of
false impersonation and was considered as something of a botched
job (Sax 1987: 15).
Most of these features are paralleled, with instructive south Indian
variations, in the Draupad cult. Dramas are sponsored mainly by
regionally dominant Vanniyars (in South Arcot) or Vllars (in North
Arcot), who claim either descent from Ksatriyas (Vanniyars) or cer-
tain Ksatriya prerogatives (Vll ars) (Hiltebeitel 1988a: 3239). Brah-
mans and other castes do not, however, regard either community as
having legitimate Ksatriya status, such as is attributed to Rjpts in
Garhwal. With the prevailing sentiment being that there are no legiti-
mate Ksatriyas or Vaiyas in Tamilnadu, the general perception is that
the chief sponsoring and participating castes at Draupad festivals are

Sax (1991b), however, indicates that the fact that Vibhuti Narayan Singh is a
Brahman is an anomaly (he was adoptedis he technically a Brahman anymore?).
152 chapter seven

dras. Accordingly, prominent registries of high and low run

through various aspects of the cult and its festivals, bearing not only
on participants but on the different forms taken by the deities.
Krsn a is repeatedly said to be higher than Draupad (Hiltebei-
tel 1988a: 27577). Draupad is worshiped as a manifestation of the
supreme goddess Parakti (6, 79, 81; 1991a: 46), as an incarnation of
r (Hiltebeitel 1988b, Part 2), and as a multiform of Durg (Hiltebeitel
1988a: 36882; 1991a: 227, 242, 288, 327, 360, 458). But her rituals and
myths center most often and most directly on her affinities with Kl,
whose form (klrp) she takes on in her most violent and impure
aspects (Hiltebeitel 1988a: 28995, 434; 1991a: 397416, 430, 437, 473).
Draupad is also said to have taken on a special regional aspect in her
Gingee avatra: that is, her second birth in Gingee, the Vijayanagar
Nayakate capital, to rescue an ancient Gingee king who descended from
the Pndavas (Hiltebeitel 1988a: 34, 89, 6592; 1991a: 48389).
Actors are mainly Vanniyars, or in some areas Tampirans (a subgroup
of Pant rams, non-Brahman temple priests) or Van nrs (washermen)
(Hiltebeitel 1988a: 138; Frasca 1990: 19098). I cannot imagine a Tamil
Brahman village wanting to perform a Draupad festival. But some
Draupad temples and festivals outside the Gingee-Tondaimandalam
core area have been variously Brahmanized or Sanskritized, as, for
instance, in Thanjavur District just to the south of the core area
(Hiltebeitel 1991a: 11216, 314, 388, 40829), and at Dindigul, far-
ther south, in Madurai District (Hiltebeitel 1988a: 4351). Within the
Gingee core area, Brahmans often play significant roles at Draupad
festivals opening ceremonies of flag hoisting, fire offering, and wristlet
tying (Hiltebeitel 1988b, Part 1; 1991a: 7986), and also, more rarely,
serve as Mahbhrata reciters called pratiyrs or prata-piracankis
(Hiltebeitel 1988a: 78 n. 11). Meanwhile, though most drama troupes
will have one or two youths, who often achieve great audience appeal,
there is no special group of youths like the svarps, and the young-
est actors rarely play the chief devotional roles of Draupad, Arjuna,
and Krsn a. These are usually reserved for seasoned and accomplished
actors who, in any case, are never Brahmans.
The roles of the svarps in the rs and rm lls also serve to point
up some other instructive comparisons. It is clear that in both of these
northern traditions the svarps are comparable to icons. In the rs
lls, people know perfectly well that these [the svarps] are normal
children . . . The mystery is quite the other way around. These are plays
in which children enact the naturalness of childhood in order to stir
draupad cult lls 153

the imagination of adults: the best Krsn a is the one who acts most like
himself, a child unbridled (Hawley 1981: 18). In well-known compa-
nies especially, the boys playing Krsn a and Rdh are apt to receive
special deference and affection (18). At Ramnagar, across the Ganges
from Varanasi, the five rm ll svarps are mrtis, carried to keep their
feet from touching the ground while they are in costume after a perfor-
mance. Like temple icons, they are put away for the night (Schechner
1983: 261, 263). In rm lls, other roles may be performed by mem-
bers of any of the four castes (Hein 1972: 72). Indeed, one would like
more information on whether there are differentiations among those
who play forest Rsi s (in the Rmyan a mostly, but significantly not
all, Brahmans), kings, queens, monkeys, bears, and Rksasas. But the
svarps must be Brahman boysindeed, local boys, carefully selected,
as we have seen, and surely pure as well as charming.
Certainly Hein is right that the use in these lls of prepubescent
boys (who may in the rm lls grow a little beyond this) is a trans-
formation of earlier folk drama traditions in which the actors were
adults (1972: 72, 23031). Hein probably goes too far in suggesting
that the kumr-pj (virgin girl worship) of ktism lies behind this
transformation (26571), but the two are no doubt parallel develop-
ments. The intention was surely from the beginning to create a devo-
tionally inspired theater in which actors, playwrights, and producers
would suffer none of the contempt that ancient India heaped upon
the stage profession (274). More suggestive is Heins remark that
the five svarps of the rm ll must be Brahman because when they
appear in costume and crown as the very embodiments (svarps) of
the divinities, even brahmans will bow down to them and worship
them (72). To explain this phenomenon historically, it would be bet-
ter to look at the regional sociology that fed into such transforma-
tions locallyagainst the background of Muslim rulethan to seek
an explanation from outside influences.
In the Draupad cult things are rather different. The drama troupes
are itinerant, and they contract for each festival at which they perform.
Their status is not high, being lower than that of the Mahbhrata
reciters. Pratapiracankis sometimes disdain the dramatists for vul-
garizing the Mahbhrata, and there is certainly no closely cued cor-
related alternation between epic textual recitation, which takes place in
the afternoon, and the dramas, performed at night, as there is between
the simultaneously performed recitation of the rmyan is who sing
from Tulsidas and the svarps who speak most of the dialogues
154 chapter seven

(sam vds) at Ramnagar (Schechner 1983: 239, 27273, 278; cf. Hein
1972: 79). Similarly, trustees often treat the terukkttu actors poorly in
forging contracts and in making local arrangements for their stay in
a particular village. Normally they have only the thatched green room
behind the stage to sleep in during the hot days after their nightlong
performances, and no one puts them to bed like icons.
Yet at one level that is what they still are, though in ways whose
differences require us to begin thinking about the nature of village
goddess cults and of the Mahbhrata. All of these ll traditions
give pause and movement to what can be called iconic bhakti tab-
leaus, moments of living darana, often with surprising ironies and
inversions (see, e.g., Schechner 1983: Rmas playful, even ironic
omnipotence is shown by the way he not only breaks ivas bow but
exposes it as a stage prop). They also actualize the mythological set-
tings (Vrindvana, Ayodhy, Lank, etc.); at Draupad cult dramas the
audience is continually addressed as members of the court (Sanskrit
sabh; Tamil capaiyrkal), which is especially powerful at the dis-
robing of Draupad. Then there are those beautiful liminal moments
of transition when dramas begin, and especially when they are over,
when the actors receive worship and the audience (in Hawleys closing
words) walks into the play (1981: 274; cf. 226).
Here, however, we can note some distinctions. At Brindavan the
svarps are offered rat (waving of light), and spectators come to
touch their feet (Hawley 1981: 226; cf. Hein 1972: 90, 138). At Ram-
nagar, after the last drama, even the maharajas family worships
them, washing their feet (Schechner 1983: 261). When Draupad cult
plays conclude at dawn, the status of the actors is, as we would now
expect, lower. They do not receive rat from the audience (though the
Draupad actor may have earlier received the tray onstage to offer rat
to the Krsn a actor), and no one touches or washes their feet. Rather,
upon the conclusion of the drama, the one or two actors who have
played the leading devotionally significant roles (Krsn a, Draupad,
Arjuna, iva, Prvat, etc.) circumambulate the processional icons that
have come on a chariot (tr, ratha) to watch the play after touring
the village in the late hours of the previous night. The actors perform
diprdhan (offering of light) to the processional icons and then dis-
pense prasda in the form of turmeric powder to those who have come
forth from the audience (Hiltebeitel 1988b, Part 2).
The actors, still in costume and makeup, are an iconic presence but
one that is decidedly low, and can actually be ranked in relation to
draupad cult lls 155

others. Highest are the icons in the temple sanctum, the mlavars.
Next are the processional icons (those of five metals [pacaloha] being
higher than those in wood, if a temple has both), which are them-
selves like avatras in that they represent the forms in which the dei-
ties descend from the temple onto the village stage and ritual grounds
where the Mahbhratas conflicts are reenacted (Hiltebeitel 1991a:
3952, 14265). Then one comes to the actors, or rather the roles they
embody. Before certain killing scenes, the slayer and victim will go
together to worship the mlavar in the sanctum (Hiltebeitel 1988a:
418). Or if the stage is not near the temple, the Duhsana actor, when
he is about to attempt the disrobing of Draupad, may beg forgiveness
of the portable icon of Draupad on the chariot that faces the stage
(facing east, like the temple icon), after which the processional icon
may be screened by a yellow cloth so that she (the higher wooden
or five-metal processional form of Draupad) does not have to see her
own (the lower Draupad actors) disrobing (Hiltebeitel 1988b, Part 1).
Lower still than the actors are those in the audience who become pos-
sessed during such scenes of sacrificial violence and sexual violation,
and who thus embody Draupad and perhaps other Draupad cult dei-
ties in this additional way (Hiltebeitel 1988a: 276).
One could also go on to various pots used in procession that embody
the goddess, though these do not fit into this chain of descent in any
readily recognizable way and are most important in connection with the
cults fire-walking rituals (Hiltebeitel 1991a: 45158). Clearly, though,
taken as a whole, descents into the mixed and multicaste world
of the Mahbhrata are rather different from those into the worlds
of the cowherds and Rma. If one wanted to represent and preserve
the ideals of pure Brahmanic Hinduism, there would be no point in
dressing up little boys as Draupad, Arjuna, and Krsn a-Prthasrathi
at Kuruksetra. But then, of course, the Draupad cult does not repre-
sent pure Brahmanic Hinduism but a popular Hinduism sponsored
primarily by communities that Tamil society at large regards generally
as drasthough for present purposes it must be stressed again that
the cult is not without its Brahmanic components.

The Death of Duryodhana

Against this background let us now take up the scene at Draupad

festivals that, as I described it earlier, has the most to tell us about
156 chapter seven

continuities between the Draupad cult and the rm and pn dav lls.
The scene is the death of Duryodhana, which I will introduce by way
of some comments about improvisation before turning to the Dasar-
rooted themes that relate the three traditions just mentioned.
One of the most attractive things about lls is their improvisational
character, which Hawley has caught most vividly in his description of
the rs ll, with its emphasis on imagination (1981: 1718). Let us
note in particular the spontaneity achieved through temporal anach-
ronisms: the irrelevance of the order of the rs lls as a cycle (104,
1067) and the supposedly later Mahbhrata events that are some-
times referred to as if they had already happened (21920). In the
rm ll, pn dav ll, and Draupad cult traditions, there is generally a
fidelity to the familiar epic temporal sequences8 but no lack of novel
episodes, whose generation is fascinating in all cases (see especially
Hein 1972: 7475, 156, 274; Hiltebeitel 1988a: 153448). According
to my two main terukkttu actor informants, their true guru is their
imagination (karpanai) (Hiltebeitel 1988b, Part 1; 1992). As to rm
lls, standardization has apparently resulted in improvisation being
more restricted at Ramnagar (Schechner 1983: 278) than elsewhere.
In any case, it was in making my film of a Draupad festival at
Mlaccri village, Gingee Taluk, South Arcot, in 1986 (Hiltebeitel
1988b) that I was most deeply struck, after writing most of a book
on the subject (Hiltebeitel 1988a), by the improvisational vitality of
Draupad cult dramas. I take as an example the play that closed the
Mlaccri festival cycle, with its enactment of events building up to
the death of Duryodhana and the rebraiding of Draupads hair. To set
the stage, one must know that on the previous afternoon the prata-
piracanki had brought his recitation of Villiputtr l vrs Tamil
classical version of the epic to its close, with a narrative version of
Duryodhanas death. All that night the drama called Eighteenth-Day
War carried the same basic narrative through the deaths of akuni
and alya to the same culminating point. At last, at dawn, the final
scenes, to be captured on video, were to be played out. I had seen four
versions, all different, of this finale, two by this same troupe. None,
however, included the two following novelties.9

Epic chronology is generally kept more faithfully in the drama cycles than in
the rituals. See Hiltebeitel (1988a: 16566; 1991: 22930, 28788, 32829, 404). For
discussion of an exception, see now chapter 9.
For a fuller discussion of these innovative episodes, see Hiltebeitel 1992a.
draupad cult lls 157

First, when Krsn a learns that Duryodhana has made Avatthman

his fifth marshal and commissioned him to kill the Pn davas in their
sleep, he concocts a new scheme to trick Avatthman into killing the
Pndavas children instead. Rather than having the Pn davas move
camp on some pretext, he calls forth one of the terukkttu chorus,
introduces him as the Sage Who Never Lies (poy-colta-munivar), and
tells him that when Avatthman comes looking for the Pn davas, the
truthful rsi should tell him they are under his asshole. This, Krsn a
explains, will not be a lie because he (Krsn a) will indeed dig a pit for
the Pn davas to hide in. Krsn a then leaves the rsi to squat over the
alleged pit until Avatthman arrives. When Avatthman hears the rsi
tell his outrageous and unbelievable truth, he of course goes on look-
ing elsewhere, and soon finds and kills the Pn davas children by mis-
take (Hiltebeitel 1988b, Part 2). On this bawdy and irreverent material,
quite typical of the terukkttu, I would just note that it is rather dif-
ferent from the tastes of the rm ll, in which, according to Hein,
nothing indecent is even hinted at (1972: 98).10 I know of nothing to
suggest that it is not similarly different from the rs, with its emphasis
on solemnity and piety (Swann 1990a: 199). Sax, however, indicates
that bawdy humor is also typical of the pn dav lls (1991b).
Now the five Pndava children whom Avatthman kills are, of
course, Draupads children. Tamil Mahbhrata traditions, which the
terukkttu follows, actually have Duryodhana lament their deaths, since
the five are his nephews, whose killing portends, as far as he knows,
the end of his family and its dynastic line (Hiltebeitel 1988a: 42324;
1988b, Part 2; Shulman 1985: 14445). He thus curses Avatthman
for his mistake. In some of the versions I have seen of the end of this
play, Draupad starts to put up her hair in anticipation of fulfilling her
vowuttered after her violation at the dice matchof oiling it with
Duryodhanas blood. But when she sees her slain children, she laments

Kumar (1988: 18097), however, describes an exception in the crude actions
and verbal obscenities (190) that formed part of one episode of the rm ll of the
Chaitganj section of Varanasi from about 1890 until reform forces achieved its sanita-
tion in the 1930s. The ll in question, the nkkatayy, or cutting off of the nose by
Laksmana of Rvan as sister rpanakh, results in the gathering together of Rvan as
demon army for an all-night procession of costumed grotesqueries involving mainly
lower castes. This can be instructively compared to some of the nighttime demon
army processions that form part of larger festivals, in Kerala (cf. Hiltebeitel 1991a:
17677). There may be other examples where both the rm ll and rs ll include
raw humor, but probably not, I suspect, with the same normalcy as the terukkttu or,
I suspect, the pn dav ll.
158 chapter seven

bitterly, again unfurls her hair, and makes a second vow that she will
now not put up her hair until she is brought Avatthmans head.
Eventually, thanks as usual to an intercession by Krsn a, who knows
Avatthmans connections with iva prevent any possibility of killing
him, Draupad finally accepts Avatthmans tiara (with which Duryo-
dhana had installed him as his last marshal) in lieu of his head. Only
then does she finally put up her hair (Hiltebeitel 1988a: 41344).
Now, however, something unexpected occurs. In the version cap-
tured on video, Draupad never laments her children or demands the
death of Avatthman. Rather, she and Krsn a engage in a repartee that
handles the themes of birth, death, the motherhood of the goddess,
and even the final message of the Mahbhrata, in other but no less
striking ways.
As Duryodhana lies dead on the musicians bench, Krsn a calls for
Draupadhis sister Ykatvi, the Goddess of the Sacrificeto
come forth to fulfill her vow. He leads her, circling the battlefield-
stage, pointing out the dead warriors, and finally shows her Duryo-
dhana and invites her to tie up her hair with his blood. At this point
Duryodhanas corpse makes a surprising move, raising its right arm
from the elbow with the hand closed in a fist. Draupad draws back,
Draupad: But I cannot do it. Why is his fist closed?
Krsn a: Yes. A man is born. He had his hand closed. When a person
dies, the fist is stretched open. How beautifully my younger sister asks
this question.
Draupad: But this man has one fist closed. There is a reason for it. What
is the reason?
Krsn a then tells how Duryodhana died counting five things on his fin-
gers, any one of which, had he availed himself of it properly, would (he
thought) have won him the war. We need only note here the dialogue
over the fourth finger:
Krsn a: Parimukanar cnpati, the horsefaced marshal. Pari means horse.
Avatthman was born out of a horses womb.
Draupad: True.
Krsn a: [Duryodhana thought] If only I had given the marshalship to
Avatthman earlier, he would have killed the Pn davas. I gave it to
him at dusk, the time the light is gone. Without realizing it, he killed
the young Pacapndavas.
Draupad: True. [In this version, this is Draupads only response to the
death of her sons, which she seems to know about already.]
draupad cult lls 159

Krsn a: If only I had thought about it, Avatthman would have killed
the Pn davas for sure. Like this, Duryodhana thought about it in his
mind. Thus the fourth finger.
Finally, having gone through all of Duryodhanas five closed fingers and
having reassured Draupad, with the inevitable smile, that he had five
counter-measures in store anyway, Krsn a convinces her that Duryod-
hana is dead, and that she can go stand on his chest unafraid and finish
her vow, which she does. What is striking here is that Draupad feels she
cannot fulfill her vowwhich involves not only dressing her hair with
Duryodhanas blood but using his ribs as a comb and his intestines as a
garlandbecause Duryodhana reminds her of a baby that has just been
born. Indeed, in this version Duryodhana, who appears like a baby, is
the only one to have lamented Draupads own slain children. As she
prepares to dress her hair with his blood, she is on the verge of appear-
ing like Kl as the mother who delights in the blood of her own chil-
dren. For not only have her five sons now died in the war their mother
required; as the supreme goddess appearing in Klrpa, the form
of Kl, Draupad is the mother of all, including Duryodhana. One is
reminded of well-known myths in which Kl ceases her wild dance
when she sees iva in the form of a baby (Hiltebeitel 1991a: 364). Here,
Krsn as reassurances have the similar purpose of dispelling this image of
Draupad as Kl, so that her klrpa can in effect be deactivated when
she ties up her hair (Hiltebeitel 1991a: 364, 392, 405, 475; 1992a).11
Once this play is over, the shift is completed from recitation to drama
to ritual spectacle. The latter takes place at a ritual battlefield site called
the patukalam, separate but often not far from the terukkttu stage. It
involves the construction of a huge prone clay effigy of Duryodhana
(facing up, from forty to over a hundred feet in length) surrounded
by pits lined with margosa leaves for the reenactment of the deaths of
Draupads children. The spectacle, however, is all-inclusive, involving
roles for the prata-piracanki, who recites at key points and may serve

The theme of the five closed fingers has a famous counterpart in the story of
the great r Vaisnava theologian Rmnujas arrival before the corpse of his precur-
sor, Ymuncrya. Seeing three fingers closed on Ymunas hand, Rmnuja asked
whether Ymuna had died with any unfulfilled wishes and was told by the latters
disciples that there had indeed been three: to write a Vaisn ava commentary on the
Brahmastra, to express his gratitude toward Vysa and Parara, and to show his
great affection for the Tamil saint Namml vr. When Rmnuja said he would try to
fulfill these wishes, the dead Ymunas three fingers straightened out, and Rmnuja
was acclaimed his successor (Carman 1974: 30).
160 chapter seven

as a sort of master of ceremonies, and actors from the drama troupe. A

vast crowd, equaled only by the one that will return for the afternoons
firewalk, gathers by the time the effigy is finished. For our purposes the
main order of events can be outlined as follows:

1. Two actors replay the final duel between Bhma and Duryodhana.
The Duryodhana actor finally lies on the thigh of the effigy at a
point where a hole has been filled with reddened water (= blood),
and the Bhma actor cuts the effigys thigh.
2. As Duryodhana (the effigy) lies dying, five men carry out traditional
roles, often passed down through their families, of representing the
five Pn dava children slain in their sleep.
3. After the five have risen up, with various connotations of their revival,
Draupads processional icon (pacaloha if possible, otherwise of
wood) is brought on her portable palanquin-chariot, and as it is car-
ried over Duryodhanas prostrate form, blood is taken up from the
hole in the effigys thigh and smoothed into the Draupad icons hair
while the hair itself is tied, with red or orange flowers, into a knot.
4. An actor impersonating Draupad now mounts the effigys chest
and replays the rebraiding scene. Sometimes the Draupad actor
gets possessed and laments her sons, who may be ritually revived
but are still narratively dead. In such cases there is a contrast: the
icon represents the goddess in her victorious aspect, atop her fallen
foe; the actor represents the suffering goddess, the heated goddess,
the goddess as victim who suffers for her children.
5. Finally, in a scene that has no precedent in the earlier prata-
piracankam or drama, Duryodhanas wife comes forth with her hair
disheveled, beating her breasts in a mock lament, and then holding
a winnow. Her entrance fulfills another dimension of Draupads
vow: that upon the rebraiding of her hair, the Kaurava women
would suffer this reversal of fortunes.

Clearly it is at this point that the three scripts of recitation, drama,

and ritual converge with the greatest intensity and symbolic multi-
valence. Indeed, they are rebraided like Draupads hair. At the fire-
walk that follows, the actors are normally absent, or at least have no
scripted role.12

Portions of the preceding summary are incorporated from Hiltebeitel (1991a:
draupad cult lls 161

All this activity around Duryodhanas effigy invites, and indeed

requires, as I see it, some discussion of parallels with rm ll rites
and the relation of both configurations to Dasar. Before we can turn
to such parallels, however, we must note a feature of Dasar that has
left traces in both rm lls and Draupad festivals that has so far gone
largely unnoticed.
The chief matter to concern us at Dasar is the am pj, the wor-
ship of the am tree.13 At royal celebrations of Dasar or Vijaydaam,
am pj is part of a complex with two other rites: smollanghana, or
crossing of the boundary of the capital, and aparjit pj, wor-
shiping the goddess under the name Aparjit, the Invincible. The
three rites form part of one procession to the am tree, which is itself
invoked as Aparajit. And the crossing of the boundary occurs either
before or after the am pj, depending on whether the am tree is
inside or outside the capital limit.
In addition, the yudha pj, or worship of the royal weapons,
which usually occurs on the previous ninth day before Dasars
tenth, may also be done conjointly with the am pj: the weap-
ons being taken to the am, worshiped with it, and then carried back
from the am to the palace. The worship of weapons and crossing of
the boundary under the protection of the invincible goddess opens
the season of military campaigns at the end of the rainy season, and
may further involve a rite calling for the king to shoot arrows in each
of the four directions, or at a symbolic enemy, or set off a fusillade
of guns (see, among others, Biardeau 1984: 813; 1989a: 299317;
Rmakrishna Rao 1921: 3058; Fuller and Logan 1985: 85, 89; Masila-
mani-Meyer N.d; Kane 1975, Vol. 5, Part 1: 17795; Tod 1972, vol. 1:
467; Kinsley 1986: 10611).
Now rm ll is a Dasar without a king. Or, more precisely, it is
a Dasar that is usually performed without a king or, more exactly,
with Rma as the substitute king. The main popularity of the rm ll
is in villages and towns, and within wards of cities, much as is the
case with Draupad festivals. Ramnagar is a rare exception, the festival
there apparently having been royally appropriated and reformed in the
first third of the nineteenth century (Schechner 1983: 256, 258; Hess
1988: 4).

Portions of the following summary are incorporated from Hiltebeitel (1991a:
142 54).
162 chapter seven

On Dasar day the maharja performs the yudha pj in the court-

yard of his fort, which is also the staged environment for Rmas fort-
capital of Ayodhy, and then in an extraordinary and magnificently
theatrical procession of elephants, he sets off on the more than five
kilometers to Lank, to the southeast, where Rvana is represented in
his fort by his huge standing effigy. Before he reaches Rvanas fort,
at what would appear to be a boundary point between the two forts,
the king comes to a am tree, which he circumambulates and wor-
ships.14 Now the origin of the am pj is often credited to Rma,
though it does not appear in the Rmyan a (Biardeau 1984: 6; 1989a:
3023), just as it is often given a source in popular versions of the
Mahbhrata (Biardeau 1984: 7; Hiltebeitel 1991a: 15256, 316). The
king then rides up, over, and past the battleground outside Rvan as
fort, whereupon, after a brief ten minutes, he turns around to make
his way back to his own fort.
Schechner senses the mysteryit is the only time in the Ramll
that the Maharaja literally invades the performing spaceand rightly
calls attention to the combination of yudha-pj and smollanghana,
though not mentioning the related am pj: the royal practice of
kings to march their armies to the borders of their domain, proclaim
the territory as theirs, confront their opposing number across the bor-
der, and go home (1983: 24950, 27071). Significantly, the maharja
also explains the brevity of his stay in Lank by his reluctance to see
the killing of a king (269; cf. Hess 1988: 40). The maharja claims to
differ in this aversion from his great-grandfather, who stayed to watch
the enactment of Rvanas death. Hein describes what is surely a more
typical enactment (one that is reminiscent of the well-choreographed
dance-duels between Bhma and Duryodhana around Duryodhanas
effigy): Two carriages bearing impersonators of Rma and Rvana
circled round and round in lively imitation of the tactical gyration of
the chariots of the two champions in combat. A great shout went up
as Rvana was struck down, after which Lank is stormed and the
effigies burned (Hein 1972: 7677). At Varanasi the Rma actor shoots
thirty-one arrows at the effigy: one to dry up Rvan as navel (his vital
point), the rest for his twenty arms and ten heads. The Rvan a actor

I thank Philip Lutgendorf and Linda Hess for their information on this point.
There is no mention of a am tree at Ramnagar, or other rm lls, in any of the
sources I have consulted.
draupad cult lls 163

then shows these effects by removing his twenty arms and ten heads
and bowing before Rma, symbolizing his union with the lord in
death. After nightfall the effigy is then burned to represent Rvan as
cremation (Hess 1988: 41).
The maharja is thus not the only king to enter Lank. The other is,
of course, Rma himself. As far as I can see (though the information is
sketchy), virtually all rm lls include a nine- or ten-day period that
culminates with the death of Rvana on Dasar or Vijaydaam (Hein
1972: 9697). But whether others include a am pj on Dasar day is
uncertain. Considering the textual footing of the rm ll, the absence
of information should not be surprising. Rma performs no am
pj either in the Vlmki Rmyan a or in Tulsdss Rmcaritmnas.
Tulsds, however, does include an episode unknown to Vlmkione
that seems to have pan-Indian recognition due to its association with
the famous pilgrimage site of Rmevara. It is an apparent equiva-
lent of, and possibly even a substitute for, the am pj: Rmas linga
pj at Rmevara. There are traditions that Rma offers a linga pj
both before he crosses the Palk Strait to Lank from Rmevara and
after he returns.15 But most interesting for our purposes is the for-
mer tradition, given pride of place by Tulsds (Hill 1952: 36768),
in which Rma initiates this ritual, while his monkey allies build their
bridge, before crossing the ultimate boundary of India to invade the
kingdom of Lank and confront his enemy Rvana. Leaving aside for
now the question of how a lingam could be a substitute for a am
tree, let it suffice to say that what is interesting about the Rmevara

The tradition at Rmevara is that Rma could not touch food without daran of
iva. Because at the seacoast such a daran was difficult, he had to make an earthen
image of iva for darsan before taking food. iva then blessed him with the promise
of his success. Then, after killing Rvana, upon return to Rmevara, St performed
her fire ordeal there (!) and Rma relieved his sin of brahmahaty (brahmanicide)
for killing Rvan a, which required a ivalinga. Hanumn set off to bring one from
Mount Kailsa, but when the auspicious time for the worship was in danger of pass-
ing, St made one out of sand that miraculously turned solid and immovable when
Hanumn, annoyed at returning too late and at the neglect of the lingam he had taken
such troubles over, tried to lift the sand lingam with his tail. Rma then ordered that
Hanumns lingam be set up to the left of the main one, to receive the first worship.
It is also at Rmevara, before building the bridge, that Rma challenges the ocean
king Sagara, threatening to kill him with his arrows if he does not dry up the ocean
(Das 1964: 7074): a reminder of the challenge to enemies, real or substitutional, that
is also part of Dasar. See Jagadisa Ayyar (1982: 492): there are actually eleven lingas
there, one established by Nala, the monkey architect of the setu (bridge); cf. also Shul-
man (1980: 5051).
164 chapter seven

lingam is its boundary position.16 In oleograph print depictions of this

scene, which represent the most popular aspects of this popular tradi-
tion, Rma and Laksmana perform the linga pj before a stone votive
lingam with yoni pedestal while the monkeys construct the setu toward
Lank in the background (see Vitsakis 1977: 31, plate 11).
In any case, when one looks to popular Rmyan a traditions
beyond the Rmcaritmnas and the rm ll, the am pj reap-
pears in structurally much the same position as Rmas ivalinga pj.
Indeed, the am pj is itself often explained with reference to the
Rmyan a. Rma does not, however, do his founding am pj in
sandy Rmevara but in Lank itself, and not before he has killed
Rvana but after. Thus in one such account, Rma, unable to get the
better of the ten heads of Rvan a which grow back as soon as he cuts
them off, addresses a prayer to Durg. Durg grants him the prayer
of killing Rvana precisely on the day of Durgsta mi, the day of the
buffalo sacrifice [of Navartra]. On the day of Dasar, Rma offers a
pj to the am and sets off back to Ayodhy on his celestial chariot
(Biardeau 1984: 6). The boundary of Rmas realm has, as it were,
been extended into Lank itself, especially now that Rvan a has been
defeated. There are also numerous Purnic and popular traditions link-
ing Navartra and Durg Pj more generally with founding prayers
and offerings by Rma to Durg: again, both just before and just after
the killing of Rvan a (Kinsley 1986: 1089). Here, however, one often
finds no reference to the am pj, which seems to be typical of Durg
Pj celebrations. Like rm ll, the Durg Pj seems to be another
substitute for Dasar without a king. Or, more exactly, Durg Pj is
a celebration in which Durg (rather than Rma) substitutes for, or at
least takes priority over, the king (see, e.g., str 1980: 3397).17
It seems, finally, to be a reasonable supposition that both rm ll
and Durg Puj are what can profitably be called popular substitutions
at the Brahmanical level for Dasar celebrations without a king, or at
leastas with Durg Pj, at which the role of regional kings is more

One should better say lingams in light of the complexities and variants of the
Rmevara myth, which differs at key points from the traditions found in Tulsds;
see note 15.
Coburn (1991: 153) refers to rm ll and Dasar as royal festivals with which
Durg Pj overlaps. I assume he is contrasting a nonroyal Durg Pj with the
royal character of the other two. Durg, however, is as royal as Rma. The real contrast
is between a festival requiring real human kings, and festivals for which a deity may
substitute for a human king.
draupad cult lls 165

in evidence (Eschmann 1978: 279 and passim; Biardeau 1989a: 6779,

270, 281, 300)with Hindu kings very much in eclipse. They are
found in the great central belt of north India: from Orissa and Bengal,
in the case of Durg Pj; and through Bihar, Madhya Pradesh, Uttar
Pradesh, and Panjab, in the case of rm ll (Hein 1972: 101). And I
think that what Hein (100), Schechner (1983: 25657), and Hess sug-
gest about rm ll would also apply to Durg Pj as well: that these
rites developed against the background of Muslim, and then British,
rule as symbolic expressions of the ideal of an alternate mythical and
ritual Hindu rj. Indeed, one could regard the recent Rmjanmabhmi
agitation as an attempt to sponsor a multimedia pan-Indian rm ll
against the background of a secularism that devolves precisely from
Islamic and colonialist antecedents. In all such rm lls, the energy
comes mainly from wards of cities and towns, where their popular
character becomes increasingly evident. But they retain their Brah-
manical structure.
As to Draupad cult lls and pn dav lls, they are popular,
as we have seen, at a largely non-Brahmanical level: that is, among
regional landed dominant castes. These castes, however, are the ones
that take on symbolic royal functions at the regional and village level.
And though they transform them, and often change their calendrical
timing (Draupad festivals seem to be variants of a spring Navartra;
pn dav lls take place in the winter), they seem to have at one of
their fonts the ritual inscription of Navartra-Dasar. In the Draupad
cult, when Draupads chariot is led over the effigy of Duryodhana,
the chariot itself may be led by the portable icon of Draupads guard-
ian Pttu Rja, or the Buffalo King, who (as has been shown else-
where) is not only mythically the reformed transfiguration of the
Buffalo Demon but ritually a multiform of the am, the lingam, and
the sacrificial stake (Biardeau 1989a; Hiltebeitel 1988a: 7688, 33393;
1989b; 1991a: 10356). The fact that in the pn dav lls centrality is
given to the ll in which the Pndavas hide their weapons in the am
tree (Sax 1986: 811) encourages us to look in the same direction: for
one thing, the amor more exactly its pine tree substituteis in
evidence at the epics postwar avamedha (see Hiltebeitel 1988a: 133).
Indeed, as Biardeau has shown (1984: 7), the story of the Pn davas and
the amnot the hiding of the weapons alone but the retrieval of the
weapons before the waris another piece of Dasar folklore, and, as
with the explanation through Rma, another popular explanation for
the origin of the am pj.
166 chapter seven

Finally, it is also clear that the rm ll and the Draupad cult

patukalam have similar elements beyond the am pj. The imper-
sonation by local men of the death and revival of the Pndavas five
young (and unmarried) children bears a ritual analogy with the func-
tion of the five rm ll svarps. The five men who impersonate the
dying and rising Young Pacapndavas at Draupad festivals are usu-
ally senior temple officials, typically Vanniyars, and in any case never
Brahman. But their impersonation of these sixteen-year-old children
requires preliminary purification rituals that include sexual abstinence
and vegetarian food, or just milk (Hiltebeitel 1991a: 34546). While
some temples perform this Young Pacapn dava ritual, others per-
form an alternate ceremony involving sword pressing (katti cvai)
by real youths (kumras) (351). Ceremonies similar to the Young
Pacapndava rite, but involving or recalling five young or immature
children, are also found in popular Dasar traditions in Maharashtra
and Karnataka and in a buffalo sacrifice in Andhra (36970).
Further parallels are evident in the construction and demolition
of effigies of the chief epic demons (Duryodhana and others in the
Draupad cult; Rvana and usually Kumbhakarn a and Indrajit in the
rm lls), and in the rituals and myths found in both traditions about
the confrontation between the kings of two forts, and the construc-
tion of a battlefield fort as the main ritual terrain (Hiltebeitel 1991a:
29496, 33238, 399432).
The founding of the rm and rs lls seems to go back to the six-
teenth to seventeenth centuries (Hein 1972: 1089; 22427, 274): a
time that could well also be that of the consolidation of Draupad cult
patukalam rites in the Gingee core area. Not only are both epics thus
tied into the popular mythology of Dasar; they have supplied the myths
for similar battlefield rituals that in the rm ll and the Draupad cult,
and apparently in the pn dav ll as well, involve popular transposi-
tions of Dasar rites into idioms of ritualized epic conflict.



Since it is high spectacle and usually the culmination of a south Indian

Draupad festival, it is not surprising that firewalking was the main,
and often sole, focus of the briefer notices of the Draupad cult that
appeared sporadically from the late eighteenth to early twentieth cen-
tury in travelers accounts, journals, government reports, district gazet-
teers and manuals, and scholarly notices. Until recently, this literature
was minimally descriptive, rather than interpretative, and, not until
F. J. Richards 1910 article on two Draupad festivals in Bangalore and
his brief 1918 account of a festival in Salem district, was serious notice
taken of the relation between the firewalk and the Draupad cults most
distinctive ceremonies, those linking it with the Mahbhrata.1
The earliest available western language, or at least western-spon-
sored or western-inspired, sources on the Draupad cult, are instruc-
tive for the issues they raise about changing colonialist constructions of
Hinduism. This paper will trace these sources through their first three
phases: the late eighteenth to early nineteenth century emergence of
orientalism; the early to mid-nineteenth century handling of religious
policies under the Company Raj; and the treatment of caste, especially
through the Census, in the mid to late nineteenth century under the
crown. My main focus will be on a document from the end of the sec-
ond, Company Raj phase, called Reports on the Swinging Festival,
and the Ceremony of Walking through Fire (Reports . . . 1854).
A fuller paper could go on to document the brief turn-of-the-cen-
tury emergence of anthropological description, for which Richards is
exemplary; and then the almost total lack of interestthe exception is
Sarat Chandra Mitras 1936 article comparing a Draupad cult firewalk
with one in Biharfrom the time of the last gazetteers (up to 1918)
and Bishop Henry Whiteheads Village Gods of South India (1922) to,
and through, the early years following Indian independence in 1947.

See Hiltebeitel 1991a on the relationship between the firewalk and more specifi-
cally Mahbhrata-related rituals in the Draupad cult (especially chapters 1, 815).
168 chapter eight

It should further be noted that since 1947, with the exception of brief
notices by Carl Gustav Diehl (1956), F. A. Moses (1961), M. Fazlul
Hasan (1970), and Franois Gros and R. Nagaswamy (1970, 11923),
the course of modern Draupad cult studiesto coin a phrasemust
be traced through Sri Lanka (Somander 1951; Raghavan 1961; Tanaka
1987), Reunion (Blaive, Penaud, and Nicoli 1974), Singapore (Babb
1974), and Fiji (Brown 1976, 1984), before it finds its way back to the
cults south Indian heartland in the mid-nineteen-seventies and eight-
ies (Biardeau 1976, 14344, 15051; 1989, 1322; Hanur 1983; Frasca
1984, 1990; Reddy 1985; Shideler 1987; Hiltebeitel 1982, 1988a, 1991a).
The long silence before this recent flowering is itself symptomatic of
trends that have governed scholarship on Tamil religion through the
twentieth century. The prestige of the Tamil classics and Brahman-in-
spired interest in the great tradition in its sectarian transformations
have left Tamilnadu one of the last regions of India, and the scholars
who study it one of the last academic interest groups, to catch up, or
catch on, in the study of popular folk religion.2 One encouraging sign,
however, is that this belated flowering has placed the study of Tamil,
and more generally south Indian, folk religion at the center of recent,
and, one must hope, forthcoming scholarly debates concerning the
relation between popular or folk Hinduism and Hinduism more
One instructive game that can be played with the early sources is to
imagine that they are all we could know about the Draupad cult: that
if we wanted to know anything about it, we would have to reconstruct
it from them. This would be to take seriously the prognosticative and
evaluative intentions of the 1854 report just mentioned. From the mid-
nineteenth century on, firewalking was expected to have only a short
future, and cults connected with it were expected to die out. This per-

Not surprisingly, it is the folk religion of Bengal, and then that of the Bombay
Presidency, that received the earliest attention, mainly in such serials as the Journal of
the Royal Asiatic Society of Bengal and the Journal of the Royal Asiatic Society of Bom-
bay. For early developments, see Kopf 1969, 94, 149. On the Dharma cult in Bengal,
which included (like the Draupad cult) both firewalking and martial rituals, see such
early and continuing notices as Sen 1833, Sarkar [1917] 1972, especially Chattopad-
hyay 1942, and further bibliography in str 1980.
Very selectively, see Biardeau in Biardeau and Malamoud 1976, 13853; Reiniche
1979; the essays in Hiltebeitel 1989 (especially the Introduction, pp. 118, attempting
to draw out the implications of debate between the various contributors); Biardeau
1984, 1988, 1989a, 1989b; Hiltebeitel 1988a, 1991a, 1991b; Fuller 1992. Cf. Heester-
mann 1985, 56 on the agonistic festival and its Vedic and epic prototypes.
colonialist lenses on the south indian draupad cult 169

spective allows us not only to identify the slants of the sources in terms
of their foci and stated interests, but, from hindsight, to observe their
selectivity in what they trivialized or left out. This is not, of course, to
imply that recent research has uncovered the cult just as it was when
it was first reported from 1782 to 1922. Changes are an important part
of the story. But the earliest sources assure us that most if not all of
the main features one finds today were part of the cult from the late
eighteenth century on.

A. Mapping, Binoculars and Snapshots

Behind our first phase then, is the emergence of early French and Brit-
ish colonialist orientalism. The first generation of Indologists, which
included the orientalist founders of the Asiatic Society of Bengal and
the College of Fort William in Calcutta, were reared on Enlightenment
ideals of natural religion, tolerance, and classicism, and known as ori-
entalists for their attempt to appreciate Indian culture, or at least what
they thought to be the best of it, as a spur to its own inner renewal.
According to C. A. Bayly, Eighteenth-century Europeans were still
impressed by the richness of the high traditions of Hindu and Muslim
civilizations in India and intrigued by their popular forms (1988a, 39).
Hindu literatures, both Sanskrit and vernacular, were sought out and
studied with an eye to determining their chronology and the history
that lay behind, and was embedded within, them. For south India and
the Draupad cult, one may look at two sources against this back-
ground: Pierre Sonnerats 1782 Voyage aux Indes orientales and the
Mackenzie Manuscripts, collected between 1795 and 1810, by Colin
Sonnerats description of a Draupad cult firewalk is accompa-
nied by a lithograph of four firewalkers, curiously led by a woman
(I know of no parallel for this). The men in file behind her carry
implements (a sword, a karakam pot) across the coals, in cue before
a palanquin bearing the icon of the goddess. Sonnerats treatment is
noteworthy for its placement of Draupad within a discussion of the
wider pantheon that includes not only pan-Hindu deities, but other
Tamil folk deities such as Mriyamman, Aiyanr, Mannrcuvmi,
and Kttavaryan (1782, 1: 207, 24447). He defines such deities as
Dieux subalternes, and observes that, with the exception of Aiyanr,
they are disavowed by Brahmans, who regard them as profane
170 chapter eight

(impies) (ibid., 243). A brief discussion of Draupad cult ritual is

focused entirely on firewalking, and set back-to-back with a discus-
sion, also with a lithograph, of hookswinging for Mriyamman. He
attributes the origin of the firewalk to a story that Draupad did it
originally to purify herself each year when leaving one of her five hus-
bands to pass into the arms of another (pour passer dans les bras
dun autre) (ibid., 248; cf. Hiltebeitel 1988a, 438 and n. 4). One may
note, however, that while Sonnerat emphasizes the subaltern char-
acter of these gods, and links their cults with the lower castes, he also
takes pains to show that there is an interpenetration between their
mythologies and those of the Brahmanical pantheon. Such a broad
view would not be attempted again until the work of Gustav Oppert
(1893), but by that time Oppert would organize his discussion around
the Aryan-Dravidian issue, trying to place Draupad and her cult in
the Dravidian camp (1893, 89101).
Mackenzie worked closely on his long tours mainly, but not solely,
with Brahman assistants, with whom he had a kindly rapport, and
is credited with passessing [sic] that conciliatory turn of mind that
soon reconciled all sects and all tribes to the course of inquiry fol-
lowed in these surveys (Mahalingam 1972, ix). Yet his collection
by no means anticipates the work of romantic-nationalist folkorists.
Mackenzie gives us a glimpse of how The eighteenth-century concern
with beliefs and systems of value [as with Sonnerat] gave way to the
empirical documentation of known facts, the creation in social stud-
ies of analyses and taxonomies which distantly reflected the norms of
Linnean botany (C. A. Bayly 1988, 89). His work on the collection
had to be frequently interrupted by his more professional duties as a
mapper-surveyor, military engineer, and occasional secret agent (in
Java). The collection no doubt had the same motivation as his map-
ping surveys of the Nizams dominions of Hyderabad and the Guntur
Sircar, which, as an 1810 citation put it, was to supply information
useful for military, financial and commercial purposes (Mahalingam
1972, xi). The collection itself also provided a mapping of south India,
but of its past: in particular, yielding information on the authentic-
ity or legitimacy (as the British would see it) of land grants, claims
to temple privileges, and the dynastic lineages of royal, poligar, and
regional landed dominant caste little kingdoms of south India (see
C. A. Bayly 1988, 23). As the British were wont to claim in the late
eighteenth century, most contemporary Indian rulers were tyrannical
usurpers of previous dynasties and rights (C. A. Bayly, 81). Whereas
colonialist lenses on the south indian draupad cult 171

[Warren] Hastings had often worked in the dark his successors after
1792 had at hand detailed family histories of most of the ruling fami-
lies of India, assessments of their military capabilities and notes on
their commercial resources (ibid., 87). Mackenzies collection sup-
plied rich material for such assessments, and it is no accident that it
was catalogued in 1828 by H. H. Wilson, the Sanskritist who mined
the Purnas for similar documentation of the pre-Islamic history of
dynastic India (Mahalingam 1972, Preface; Kopf 1969, 16777).
It is from two Tamil manuscripts in the Mackenzie collection that
we get important glimpses into the Draupad cult during this period.
One, the Karntaka Rjkkal Cavistra Carittiram, composed around
18023 (Dikshitar 1952), is virtually a commissioned history of the
Gingee area by a knowledgeable Gingee resident named Nryanan
Pill ai, himself a member of the Knr herder caste that traditionally
supplies the priest (pcri) of Draupads Gingee or Mlaccri temple.
In this text there is a precious allusion to a myth about the hair of
Draupads stone icon at this temple, and an assignment of the temples
building to the reign of the founder of the Gingee Nayak line, Tubakki
Krishnappa (150921). According to Nryanan Pill ai, it was to this
ruler that, revealing herself, Draupad-amman showed the hair in the
flower that had been fastened to her (Hiltebeitel 1988a, 89; Dikshitar
1952, 25). Fieldwork has shown that variants of this myth are still told
at this temple, most notably by its present-day Knr pcri (from
Nryanan Pilla is community), in which the king (whose name varies
from version to version) is blinded when he doubts the authenticity of
Draupads stone icons hair. To get his sight back, the rja undertakes
to build or improve her temple. Within what I have called the Gingee
core area of the Draupad cult, this temple is regarded as Draupads
original temple, the site of her post-Mahbhrata second advent
(Hiltebeitel 1988a, 88100).
The second Mackenzie manuscript (Mackenzie Collection n.d., no.
68) is an account, probably by two pcris of a Draupad temple in
Dindigul, far south of the Gingee core area, of an eighteen-day festi-
val structured around a sequence of episodes from the 14th-century
Tamil version of the Mahbhrata of Villiputtr l vr rendered in
Tamil songs and enacted in street processions. Recent fieldwork found
that this type of festival, which can probably be traced to the Gin-
gee core area, is no longer performed at the Dindigul temple, having
been replaced in the intervening years by a highly Brahmanized type
of festival highlighted by daily adornments (alankra) of the deity
172 chapter eight

borne on a succession of different mounts or vhanas (Hiltebeitel

1988a, 4351, 15153). The two manuscripts together allow us to trace
continuities in core area myths and rituals back to the beginning of
the nineteenth century. We should appreciate that though they were
collected by Mackenzie to meet his antiquarian and taxonomic inter-
ests, their historical and ritual information is native, having preexisted
the process of its collection, and that though the form it took was no
doubt influenced by Mackenzie and his assistants, the authors prob-
ably supplied it without misgivings about its potential for political,
economic, or military purposes. The two texts give a glimpse of what
could have been learned further about the Draupad cult had Macken-
zie, or someone else, studied the manuscripts closely and asked some
more interesting questions. But the Draupad cult was not, of course,
an issue that was highly on anyones mind.
Finally, one other source reflects this early period, but from a differ-
ent angle. The 1961 Census of India volume on temples in Chingleput
District and Madras City gives the following notice on the Dharmarja
[Draupad] temple on Mundakanniyamman Koyil Street in Mylapore,
Madras: The temple was originally located in Sullivan gardens. At the
time of the firewalking ceremony, an English-man Sullivan by name
was looking at the crowd through his Binocular when he lost his eye-
sight. He built the present temple in Mundakanniyamman Koil Street,
Mylapore some 200 years ago and got back his eyesight (Nambiar
et al. 1965, 203). The story is also handwritten in a notebook at the
Mundakanniyamman temple, whose Hindu Religious and Charitable
Endowments (HRCE) Board government-appointed officers now over-
see the declining Dharmarja temple, which hasnt held its festival in
at least fifty years. Nothing further was known about the story, even
after an HRCE officer made some phone calls.
The Sullivan of Sullivan Gardens is by all accounts Benjamin Suli-
van (sic), who stopped off at Fort St. George, Madras, in 1778 on his
way to Calcutta, where he intended to practice in the Calcutta Bar.
As the only regularly educated Barrister on the spot in Madras, he
was consulted on an important legal case, and then decided to stay on
there, first as a Standing Counsel, then, amid considerable and lasting
controversy over his salary, as Attorney General, from 1781 until his
appointment as a Judge of the Supreme Court in 1801 (Love [1913]
1968, 3:140, 164, 167, 301, 378, 47980). He is also said to have been
the Originator of the Madras Post Office (Muthiah 1987, 143). A
substantial house and ornamental grounds are shown in the survey of
colonialist lenses on the south indian draupad cult 173

1798 (Love [1913] 1968, 3:572; cf. 370 n. 2) in the area around what is
now Luz Corner in Mylapore. This appears to have been his residence:
the only considerable mansion in the locality from Teynampet and
Royapettah to the Adyar River, a considerable tract that, aside from
a few parcels of private property, was for the most part under wet
cultivation (ibid., 537). As we might expect, historical sources tell us
nothing of Sulivans blinding. But there is an intriguing detail that
may have fed into our story. In 1787 Government received a letter
from ten gentlemen resident in the Luz, of whom Benjamin Sulivan
was one, complaining of the nuisance caused by kilns in the locality.
We, the underwritten Inhabitants of the Luz, humbly beg leave to
represent to you that between the Houses of Major Alexander and
Mr. Sulivan there is a small Village of Potmakers, who are constantly
employed in burning Pots and Tiles to the common annoyance of all
the Neighborhood. We are informed that some years past there was a
representation of this Nuisance made by the late Dawson [ne] Drake,
Esqr., and that the kilns were in consequence destroyed by order of
the Honble Board, but that these people, understanding the Order was
not entered in the Minutes of Consultation, had the boldness, on the
first subsequent change of Government, to reerect them (ibid., 370).
From potmakers kilns to firewalking crowds is at least a plausible
transition for an Indian folklore about British responses to indigenous
In any case, the convertedand often blindedBritisher is a stock
motif. As we have just seen, such a story is told about blinded kings at
Draupads original temple at Gingee. One sees here how the popu-
lar culture of Madras was quick to insert the British into the tradi-
tional kingly role of temple protector and converted bhakta. Similarly,
at the festival for the goddesses of the Gingee Fort, there is now a
storyin the absence of the Britishthat revolves around the blind-
ings and recovery of the local police inspector (Hiltebeitel 1985, 184;
198, n. 46). Indeed, I have become the subject of such a story myself.
In 1982, while I was photographing the pulling of a large processional
chariot at a festival for the god Kttn t avar (the son of Arjuna and
his serpent wife Ulp), I was so engrossed that if someone had not
pushed me out of the way, I might very well have been hit by one of
the large chariot wheels. I was, of course, standing on the ground.
Upon my return to the same festival in 1990, local informants identi-
fied me to my two companions, my friend Lee Weissman and assistant
and friend J. Rajasekaran, as the person who, some years past, had
174 chapter eight

been taking pictures of the procession from high up in a tree. I had

fallen down, was blinded, and had my sight restored only after I gave
some money to the temple.
Sulivans blinding through his binoculars is, in any case, a vivid
image, suggesting that the goddesss power is channeled through them.
I always imagine an improvement to the story; that Sulivan was looking
not at the crowd when he was blinded, but at the firepit. Indeed, the
improvement is especially tempting, considering the following story,
from the same 1961 Census publication, about the nearby Kolamu-
zhi (on the site, this goddess is known as Klavili) Amman temple:
The deity is the village Goddess of Mylapore. There is a story that
when an English man took a photograph of the fire-walking festival
in this temple, he lost his eyesight. As advised by the temple trustee,
he made a fervent prayer to the Goddess and got back his vision.
To commemorate this event, the Englishman used to take the festi-
val deity of the Goddess to the Fort St. George and present her with
Koorai, a silk saree and Thali, a gold jewel worn by married women
(Nambiar et al. 1965, 201; thanks for the reference to Lee Weissman,
personal communication, May 1990). In fact, at the Draupad temple
in Singapore, there is a widely circulated story that brings all these
themes together: sometime early in this century one of the British
officials of Singapore expressed skepticism about timiti [the firewalk].
He was persuaded to come and watch, and when he looked at the pit
through a pair of binoculars he saw a beautiful woman draping the
end of her sari over the coals. He was then struck blind and did not
regain his sight until he had made a substantial financial contribu-
tion to the temple (Babb 1974, 4041). The skeptical British official is
blinded seeing the goddess protect her devotees with her saree, while
the converted British bhakta gives the goddess sarees. In every case,
binoculars and cameras are dangerous ways to see a Hindu deity in
any manifestation.
For the colonial period, such stories about British officials as con-
verted bhaktas have their practical counterpart in the increasingly
problematic assumption by the Britishbeginning in the 1820s and
continuing to evolve into the present-day HRCE Board of the secu-
lar Tamilnadu state governmentof the vacated royal protective
role in Hindu temple administration (see C. A. Bayly, 114; Appadurai
1981, 7071, 115, 137; Presler 1987, 1536 and passim). Moreover,
prior to the full licensing of Protestant missionaries in 1813 (Kopf
1969, 144), there is much to suggest that the British and French had
colonialist lenses on the south indian draupad cult 175

a more receptive attitude to Hindu power goddesses (to use Susan

Baylys term [1989, 2734]) than they did after that time. It is said that
the goddess Klikampl, whose temple is now in Muthialpet (part of
what used to be called Blacktown) was formerly the goddess of Fort
St. George until missionary pressures forced the British to relocate her
temple outside the fort (Lee Weissman, personal communication, May
1990). Indeed, the British missionaries came along with British women
for British men to marry, a combination that must surely, along with
the evangelical impulse, have helped to curtail any ongoing attentions
by British men to Hindu ammans.

B. Hooks and Fires

The surge of this evangelical tide sets the tone for our second period,
in which utilitarian and missionary reform impulses introduce great
cleavages into the orientalist construction of Hinduism, and into
Hinduism itself. From this second phase on, the divide-and-conquer
agenda was no less scholarly than political; there was a pre-colonial
Hinduism, which Orientalists of this second phase did as much to
deconstruct as to reify and invent. As C. A. Bayly has shown, though
the Brahmin interpretation of Hindu society was theoretical rather
than actual over much of India as late as 1750, only to be firmly
ensconced a century later, the colonialists did not create this inter-
pretation of India; rather they speeded up and transformed social and
ideological changes which were already in train. [T]he missionary
and utilitarian critique of Hinduism, which became more bitter after
1813, had the effect of concealing here also the vitality of developments
which arose from within the three main categories of Hindu religious
activity-ritual, devotion to god, and the recension of knowledgeand
which were often only lightly influenced by the West ([1988] 1990,
15859). In other words, Bayly argues that the traditional three paths
maintained a Hindu sub-surface continuity through this period (and
later ones). The Draupad cult provides a good bhakti and ritual exam-
ple: a cult whose own survival tactics can often be read between the
lines of the colonialist reportage that was so ready to dismiss it. This
brings us to a singular document from this period: the aforementioned
report, commissioned from Fort St. George itself, and mentioned ear-
lier, concerning hookswinging and firewalking practices throughout
the Madras Presidency.
176 chapter eight

In a query sent out on 12 September 1853 to the Chief Magistrates of

each district of the Madras Presidency (which covered areas of Tamil-
nadu, Kerala, Andhra Pradesh, and Karnataka), four questions were
asked about the practices of hookswinging and firewalking (extract
from the Minutes of Consultation, No. 861):
1. Are they more or less frequent now than formerly?
2. In what way are they connected with the religious faith of the
3. Are there accidents of frequent occurrence?
4. Has the Magistracy ever interfered with or discouraged the obser-
vance? (Reports . . . 1854, 910, 25).
Normally the Draupad cult would involve firewalking, and not hook-
swinging, so the documents treatment of the former will be our
main concern. Unless otherwise indicated, the ceremony referred to
will thus be that of firewalking. But for three reasons we must keep
both ceremonies in view. First, the two rites elicit comparative judg-
ments about them by the respondents. Second, they are not mutu-
ally exclusive, as is registered by E. F. Elliot, the Madras Magistrate,
who mentions hookswinging (Chedul Feast) at a Dhurmaraja Kovil
(ibid., 3). Similarly, W. Knox, the Acting Magistrate of Ganjam Dis-
trict, indicates that at one village named Chicacole, firewalking is done
immediately after the swinging festival, the same parties performing
in both (ibid., 21). Further, it is mentioned that in Nellore District
hookswinging is done in commemoration of suttees (ibid., 1314),
while in Rajahmundry District firewalking is accompanied in some
cases by self-inflicted stabbings and woundings (ibid., 29), and by
flesh-piercing in Masulipatam District (ibid., 34). The common prin-
ciple is in each case a combination of sacrificial posts and/or weapons
on the one hand, and sacrificial fires (a prominent Draupad cult myth
tells that Draupads firepit makes the Kaurava widows into suttees;
Hiltebeitel 1988a, 44042) on the other.
Let us take up the four questions in order.
1. Frequency. E. B. Thomas, Coimbatore, and F. B. Elton, Nellore, say
firewalking is frequent in their districts (Reports . . . 1854, 28, 31), but
most others agree that it is in a state of decline. However, H. Forbes,
Magistrate of Tanjore District, illustrates this alleged decline by say-
ing that though it is done to a lesser extent than formerly, it is fol-
lowed now in 415 places (ibid., 26). He does not indicate the deities
for whom it is done, but Draupad temples are very common in that
colonialist lenses on the south indian draupad cult 177

district and the firewalk remains there one of Draupads main ritu-
als. Current-day informants estimate that she has about a thousand
temples in Tanjore District. Similarly, E. Maltby specifies that there are
489 Dhurmah Rajah pagodas in the district of South Arcot, which
we now know to be the heartland of the Draupad cult core area. But
he expects the rite gradually to disappear as the wealthier and more
educated part of the community do not join it further than as occa-
sional spectators (ibid., 3233). Meanwhile, in Madras city, E. F. Elliot,
Chief Magistrate and Superintendent of Police of Madras, regards the
firewalk as decreasing in its frequency. The narooppoo teroonaul, or
Fire Feast, is still performed in sixteen Pagodas, within the jurisdiction
of the Supreme Court, generally in the month of Audee, or July. It is
less frequent now than formerly. It is observed, with one or two excep-
tions, at a particular class of Pagoda named after Dhurmarajah. Out
of these sixteen places, seven are in the villages north of Black Town,
five in the villages west of Black Town, and four in the Town itself
(ibid., 20). It would appear that some of these temples have indeed
discontinued the practice since then, and even in a few cases ceased
to exist. I know of only eleven Dharmarja or Draupad temples in
Madras today, eight of which are in the Black Town-Parrys area. At
least two of these continue the firewalk (the Dharmarja-Vinyakar
temple at Lingi Chetty Street, and the Draupad temple at Sowcarpet,
where I saw the ceremony for the first time in 1977).
2. Connection with religion.4 H. C. Montgomery, the Chief Secretary
in receipt of the various reports, in summarizing their contents indi-
cates that hookswinging does not seem to be in any way connected
with the religion of the observers, but to be performed in the fulfill-
ment of vows (ibid., 1). He seems to share this judgment with, or
base it on, the report of the Canara Magistrate, Chamier, who observes
that hookswinging forms no part of the religious worship of the Hin-
doos, and is merely the fulfillment of a vow made in time of extreme
sickness (ibid., 1718). The criterion here for religious authenticity is
whether the practice connects with worship. We may recall, however,

Dirks 1987, 32432 shows how the British nineteenth century land policy, in
distinguishing inherited private property from land tenures for charitable and reli-
gious purposes, allowed the British to define both categories while proclaiming non-
interference in the religious sphere so defined: . . . the government appropriated to
itself the right to determine whether religious objects were being met, and indeed to
define the domain of the religious itself (ibid., 332).
178 chapter eight

that the questionnaire puts the accent on religious faith. This is

picked up by C. J. Bird, Magistrate of Tinnevelly District: Though the
ceremony is in no way connected with the religious faith of the natives,
yet it is observed by certain persons in performance of their vows
(ibid., 27). Similarly, T. Conway, Kurnool Magistrate, says firewalking
is not connected with either the Muslim or Hindu faith (ibid., 37).
In any case, reflecting the Protestant Reformations attack on vows as
detrimental to true religion, the authorities exclude them from both cat-
egories: worship and faith. Or else they are treated ambiguously, as by
H. Forbes in Tanjore District: It is only so far a religious ceremony that
persons suffering from sickness either in themselves or their relatives,
make a vow to pass through fire if recovery should ensure (ibid., 27).
E. F. Elliot, Madras, says: Since the receipt of the letter from the gov-
ernment, I have had opportunity to converse with the assembled War-
dens [= tarumakartts or chief trustees; Appadurai 1981, 120] of the
said Pagodas, and they admit that the ceremony is in no way connected
with their religious faith, but they plead the usage of their ancestors
for the continuation of the performance of this silly, but certainly not
dangerous practice,for I am not aware that there is a single accident
arising from it on record (ibid., 20).
Similarly, Montgomery says, the ceremony of walking through fire
like the Swinging Festival is of only partial occurrence, and can scarcely
be called a religious observance, being for the most part in fulfillment
of vows voluntarily made. Moreover, he adds: The practice does not
appear to be acceptable to the higher classes, and, on the whole, it is
apparently falling into disuse, a result which must be accelerated by
the increasing intelligence of the people. The Governor in council, he
says, advises that discouragement in the manner suggested with the
Swinging Festival will suffice for its extinction (ibid., 19).
Some of the reports widen the criteria for suggesting that firewalk-
ing might be non-religious. In Rajahmundry, T. Prendergast mentions
that a class of Lingadharloo (Lingyats) regard fire treading as an
efficacious observance for recovering their sanctity if by any chance
they lose their lingam, but even amongst them it is not considered as
an essential ceremony (ibid., 29). In other words, it may be religious,
but it is not essential to the religion. H. V. Conolly, Malabar, says: The
best informed parties allow that the ceremony though of ancient date
and looked on as efficacious by many, is not one to which the Hindoos
are bound by the precepts of their religion (ibid., 28). More specifi-
cally, firewalking for village goddesses in Ganjam District, according to
colonialist lenses on the south indian draupad cult 179

Acting Magistrate W. Knox, is performed to avert some expected evil

from the village, or from the performer. It is reported by all to be much
on the decline, and is nowhere ordered by the Shastras (ibid., 21).
Similarly, in North Arcot, the Magistrate J. D. Bourdillon, writes: This
observance is not concerned with the Hindoo Religion by being pre-
scribed in any religious books (ibid., 23).
However, F. B. Elton, Nellore Magistrate, mentions that firewalkers
for Durmarauzooloo (Dharmarja) state that it is done according
to a work called Augamashaster (no doubt gama-stra) (ibid., 31).
Similarly, J. Bird, Magistrate of Trichinopoly District, indicates that
it is prescribed by the religious books of the lower orders (ibid.,
25). One may suspect that textual authority is often claimed on
the basis of hearsay, and probably sometimes invented. But having
voiced such a suspicion recently myself (Hiltebeitel 1988a, 78 n. 11)
about a rumored text called the Tirpatai Mnmiyam or Draupad
Mhtmyam, I stumbled across this very text at the Pondicherry
Draupad temple in April 1990. Bearing a 1901 title page, it contains
not only the story of Draupads Gingee advent, as I had been led to
expect, but her instructions on how to align the deities in her temple,
to form their icons, and to perform her festival (Ilatcumanappil llai
1902; cf. Hiltebeitel 1991a, Appendix).
Along with the criterion of textual authority, Bourdillon, the North
Arcot Magistrate, deepens the discourse on caste that has already sur-
faced in other evaluations. The only classes who take any part in these
proceedings are some of the Soodra castes, and for the most those
of the least consideration and of the least range of intelligence. The
Bramins [sic] have no concern with them; nor have many of the most
respectable Soodra castes (ibid., 23). H. A. Brett, Magistrate of Salem
District, also reports that the ceremony is practised by the lower castes
only and not by the Brahmins (ibid., 24). On the other hand, F. B.
Elton, Nellore, mentions that it is done, among others, by Brahmans
(ibid., 31). In any case, there is a recurrent implication that lower caste
religion is barely worth the name: says J. H. Cochrane, Chingleput Dis-
trict Magistrate: This practice being observed only by the lower orders
of the people (Hindoos) cannot strictly be considered as connected
with their religious faith (ibid., 37). The reports return repeatedly to
this theme, referring the practice to the lowest and most ignorant
classes, castes, orders, or grades of people (ibid., 20 [Madras],
2122 [Ganjam], 25 [Cuddapah], 27 [Tanjore], 28 [Coimbatore], 29
[Malabar], 32 [South Arcot]). According to the Madras Magistrate,
180 chapter eight

the same individuals exhibit annually like any other class of Jugglers
(ibid., 20). The range of terms for caste suggests a common adminis-
trative attitude behind a confusion of administrative terminology.
Bourdillon, however, does not see these criteria as sufficient to
assess the rite as non-religious: . . . it must be regarded as a religious
observance, inasmuch as it is always associated with the worship of
the deity styled Durmaroyswamy by the Tamil people and Veerbudra
Swamy by the Teloogoo. He goes on to describe ten- to eighteen-
day festivals, with dancing, singing, processions of the icons, as well
as firewalking through a shallow pit of half a foot deep and several
yards broad and long. He notes that the festivals usually occur at no
set times, but only occasionally, when the people of the village or
neighborhood subscribe to raise the necessary funds, being from 100
to 200 Rupees, or when an individual incurs the charge in fulfillment
of a vow (ibid., 2223). Clearly he is talking mainly about Draupad
festivals. Similarly, for South Arcot District, the Draupad cult heart-
land, E. Maltby says that Dhurmah Rajahs annual feast derives its
name from the [firewalking] ceremony which is popular among the
poorer orders who regard it as a religious duty which is efficacious in
averting evil and assuring good (ibid., 32). For Salem District, Brett
observes that the rite is in decline among Hindus, implying that it is
a feature of their religion, as it is also among Muslims in his district.
He is the only official in this document to register an awareness of the
Mahbhrata: The practice is traced to some legend connected with
Dropadee, the wife of the Pancha Pandavas, or five brothers, whose
story is related in the Mahabharat (ibid., 24). An obvious and unfor-
tunate indifference is registered as to the legend.
Just as hookswingingwhich one magistrate urges should be sup-
pressed on Sundays, or at least removed from public sight! (ibid., 4)
probably evokes tacit reminders of the cross, religious overtones of
firewalking are also registered by acknowledgment of affinities with
Islam. Thomas, the Coimbatore magistrate, after his remarks on
hookswinging and firewalking, refers additionally to a practice often
causing bloodshed, and always a risk of serious disturbance and colli-
sion. I allude to the Mahommedan and Hindoo festivals which occur,
often simultaneously, at this season of the year. The two processions
occasionally meet in the streets, and a public disturbance and often
bloodshed ensues. He mentions an affray having nearly taken place
lately at Ootacamund. It is not clear what Hindu and Muslim festivals
he is referring to (as concerns the season in question, his letter is
dated March 6). He urges that It would be very advisable and easy, to
colonialist lenses on the south indian draupad cult 181

restrict the two parties, (to any ceremonies or exhibitions they liked),
within their own limits and houses, but forbidden to parade on the
public thoroughfaresa single order of Government would stop it in
every district, but of course the order should be generalthese hurt-
ful and often dangerous exhibitions are chiefly got up by the lower
ordersthe more respectable and intelligent of each class, Mahom-
medan and Hindu [sic], would gladly see them stopped (ibid., 910).
The general issue of rights during processionsnot only of conflicting
religions, but of conflicting castesis frequently attested in nineteenth
century sources, with the British often prompting legislative wrangling
while shying away from direct arbitration in the manner of traditional
kings (Appadurai 1981, 162; Dirks 1987, 34445, 35862; C. A. Bayly
[1988] 1990, 114).
The Muslim-Hindu similarities are, however, more specific, for, as
a number of the magistrates mention, firewalking is done by south
Indian Muslims at Muharram. Thus F. B. Elton observes that it is done
during Muharram at 123 villages in Nellore District (Reports . . . 1854,
31). Others mention its occurrence in Masulipatam (ibid., 34), Ganjam
(ibid., 20), Canara (ibid., 36) and Kurnool districts (ibid., 37). Mean-
while, Muharram with firewalking was sponsored in Madras city, from
this period to the present day, by descendants of the Nawab of Arcot
(S. Bayly 1989, 174, 182, 229). Coincidentally, while Muslim elites,
especially Sunni, sought to reform the Shia festival of Muharram
(C. A. Bayly, 166), a reform Hindu movementthe Satya Narayani
movement founded by Swami Narayan (c. 17801830)had con-
demned animal sacrifice, feasting and fire-walking (idem, 161) as part
of a suppression, approved of by the British, of tribal and low-caste
forms of religion which persisted on the warrior and nomadic fringes
of the society of Gujarat (idem, 161).5 Affinities in south India between
ten-day Muharram festivals and Draupad festivals modeled on the
ten-day Navartri-Dasar deserve further study. Both festivals recall
epochal battles with heroic deaths through rituals involving displays

Raymond Williams work on this movement (1984) does not mention a firewalk-
ing prohibition. He indicates, however, that such a prohibition would not be surpris-
ing, given the movements general orientation on popular practices, including suttee
(personal communication, November, 1990). In a letter (3 December, 1990), however,
he mentions a reference in the Sikshapatri, one of their sacred texts that contains 212
rules, that may apply to firewalking: No one should mutilate any part of his body or
that of others with a weapon, as a punishment for performance of misdeeds either by
oneself or others, in a state of excitement, or as an atonement of misdeeds. (16). The
text is translated in Dave 1974.
182 chapter eight

of weapons, their use for self-inflicted forms of self-offering, lamenta-

tions, and firewalking. Indeed, two facts suggest that the Islamic and
Hindu traditions have influenced each other at this level. Draupad
has a Muslim guardian, Muttl Irvuttan, suggesting that Islam was a
catalyst in the formation of her cult (see Hiltebeitel 1988a, 1721, 42,
51, 10127). And firewalking, to the best of my knowledge, forms part
of Muharram only in India.6
3. Dangers. Most magistrates, including South Arcots (Reports . . .
1854, 32), record little in the way of harm caused by firewalking. But
Thomas, the Coimbatore magistrate, mentions several accidents,
nearly causing death, having happened (ibid., 28). Forbes, the Tan-
jore Magistrate, mentions one instance of a man having fallen in the
middle of the fire, and been very seriously injured (ibid., 27), while
deaths occurred at Muharram firewalks in Nellore and Tinnevelly Dis-
tricts (ibid., 31, 28). In Tinnevelly, It was reported that the accident
occurred from the individual being under the influence of liquor at
the time when he fell in the pit. The true cause of his falling . . . was
not ascertained. The practice was then discontinued in both villages
(ibid., 2728). On the other hand, six years after two boys died in a
firewalk for Dharmarasuzooloo (Dharmarja) in Nellore District, the
practice continued despite the order by the former Magistrate, named
Smith, that security be taken every year from the principal persons
interested, that they be responsible that no accident will occur (ibid.,
31). H. V. Conolly, the Malabar Magistrate, and Cochrane, the Chin-
gleput Magistrate, each voice the no doubt justifiable suspicion that
accidents do occur, but that they were not being informed about them
(ibid., 2829, 37). Suffice it to say that firewalking can be dangerous,
and a few injuries and one death have occurred in ceremonies I have
witnessed (see further Hiltebeitel 1991a, chap. 14 and n. 48).
4. Government interference. Hookswinging is generally regarded as
more condemnable than firewalking, and is thus more frequently a tar-
get for suppression (Reports . . . 1854, 7, 910, 18). Thus E. B. Thomas,

My colleague Seyyed Hossein Nasr (personal communication, January 1991) con-
firms what I have inferred on this point from studies of Muharram in West Asia. For
other Indian examples of Muharram firewalking, see Derrett 1979, 287; Jaffri 1979,
225; Hjortshoj 1987, 296; Nambiar, Karup et al. 1968, 58. 176364 marked the sup-
pression by the British of the campaigns of the rebel commandant Yusuf Khan,
whose rebellion was commemorated in a ballad that extols his martyrdom and dis-
memberment in terms that show the affinities between popular Islam and Hinduism
at the regional level (S. Bayly 1989, 194215).
colonialist lenses on the south indian draupad cult 183

the Coimbatore Magistrate, says that orders have been given for its
entire discontinuance, and thus prohibition will be strictly enforced. He
adds: I would beg leave to submit for the consideration of Government
whether a similar festival celebrated annually both by Hindoos as well
as Moossulmans, and more dangerous, though less disgusting than the
Swinging Feast, might not also be discontinuedthe walking and run-
ning through fires (authors italics), before temples, or in public places.
It will be found on enquiry in every district, and is frequently attended
with dangerous consequences, and even loss of life (ibid., 910).
A few voices spoke for a more permissive attitude toward firewalk-
ing. Thus Bourdillon, the North Arcot Magistrate who recognized the
practice as religious, encouraged a more condescending approach: I
cannot learn that Magistracy have ever interfered to check or discour-
age the practice of firewalking, nor do I think that such interference
is called for; certainly not in the way of prohibition. Though seeming
very dreadful at first sight, the thing loses this appearance on closer
examination. It appears to involve little or no personal pain, and no
danger of more serious injury; and it seems to me proper to be classed
less among brutal and inhuman practices than among idle and fool-
ish. It appears to be certainly on the decline, in this District at least,
and it can hardly fail to continue so as knowledge and enlightenment
advance; especially, eschewed and discouraged as it is already, by
the most intelligent and influential classes of the community (ibid.,
2324). Needless to say, many North Arcot Draupad festivals still
continue the practice vigorously.
Most magistrates, however, favored that firewalking be discouraged,
and at least contained if not prohibited. For Madras, E. F. Elliot states,
This ceremony has been discouraged by the Police, in as much that
the people are no longer allowed to perform it in streets or thorough-
fares, but only within the enclosure of Pagodas. From my last inter-
view with the Wardens [i.e., tarumakartts], I expect that several of
them will cease to celebrate the feast altogether, perhaps by substitut-
ing some other observance (ibid., 20). Thurston mentions a sugges-
tion by a Collector in Tanjore district that the coals be replaced by a
pit of flowers (1906, 484).7 In any case, in Madras today, firewalking is
still performed in the streets, with police permission.

Firewalkers often compare the coals to cool flowers; see Hiltebeitel 1991a, chapter
14 and n. 44.
184 chapter eight

T. Prendergast, Rajahmundry, recommends government prohibition

of firewalking rather than mere discouragement, supporting this by
describing instances where firewalking is combined with self-inflicted
stabbings and woundings and the invocation of the deity . . . when
in the impotence of uncontrollable passion . . . The mere fire tread-
ing is looked upon as the act of a desperado but when it is attended
by wounding and effusion of blood the exhibition is disgusting and
alarming (Reports . . . 1854, 29). A. Purvis, Guntur District Acting
Magistrate, finds it only in one village and recommends forbidding
any such proceedings in the future (ibid.).
In Tanjore the Magistrate H. Forbes introduces a voluntaristic
strain that runs through several accounts. He called the managers
[again, probably tarumakartts] of all the 78 pagodas where hook-
swinging was done, and received from them a voluntarily . . . written
engagement to discontinue it, so that this cruel and revolting practice
may, in this province, be considered as for ever at an end (ibid., 18).
A. Robertson, Magistrate in Vizagapatnam District, was assured on
enquiry that suppression of both practices would be agreeable to all
respectable classes of natives throughout the District, and that they
owed their observance only to vows uttered by the most ignorant or
the impositions of the designing (ibid., 26). In South Arcot, The
Magistracy has confined itself to preserving order at the festivals and
preventing any compulsion being used to induce individuals to per-
form the ceremony (ibid., 32). From Bellary District, C. Pellys report
similarly accentuates the volitional, though this time in the negative,
along with a contractual commitment to discontinue the practice:
There seems to be no wish on the part of the people to continue this
ceremony, and many of those who have engaged in it have appeared
before me and executed written engagements agreeing to discontinue
it altogether, and I expect those who have not yet come in, to do the
same, which caused me to defer my reply (ibid., 3536). Considering
the lingering pace, one suspects that Pellys expectations must have
had a long wait before it was met. He adds, however, that similar
engagements have been voluntarily given to discontinue the Swinging
Festival (ibid., 36).
This voluntaristic emphasis combines not only with the elicitation
of written contracts, but with legislative ingenuity on the part of sev-
eral other magistrates. Our 1854 document falls within a period from
184178 in which Arjun Appadurai marks an increasing drift toward
court litigation of temple matters (1981, 15764). W. Knox, Acting
colonialist lenses on the south indian draupad cult 185

Magistrate of Ganjam, feels pretty confident that if Government were

to put a stop to it [firewalking], and the disgusting practice of swing-
ing, the order would be approved by the great mass of native society,
who do not attend these festivals; but perhaps the better way would be
not to run the risk of exciting any ill-feelings, even among the lower
classes, and merely to issue an order forbidding its introduction into
any village, where it does not exist. As the best check under exist-
ing circumstances, I have issued orders that before either swinging,
or passing through the fire, the actors shall appear before the nearest
Police officer, and satisfy him, that they are voluntary agents in the
business, and further have warned those taking part in these festivals,
that in the event of death ensuing, they will be liable to punishment
for homicide (ibid., 2122). Similarly, the Canara Magistrate F. N.
Maltby advises no interference, but concludes: if any serious injury
were to result on any occasion the parties aiding and abetting would
under the present law be liable to punishment (ibid., 36). The same
legal tactic has recently become law in the aftermath of the Roop Kan-
war suttee case.
Indeed, Forbes, the Coimbatore Magistrate, compares firewalking
instructively with suttee: It is not really an object of religious faith,
but only practiced by the lower orders; and might, with the swing-
ing and other like unnatural, unreasonable, and dangerous practices,
be easily abolishedindeed after the abolition of Suttee, it is to be
regretted that these minor evils and abuses, have so far been allowed
to continue (ibid., 28).
We need not over-scrutinize the enlightened rhetoric of reform
that has such difficulty answering whether these rites are religious, and
whether they should thus be suppressed despite the Companys gen-
eral policy of non-interference and respect for religious differences. As
Susan Bayly has shown, it is part of an attack on folk religion not only
in Hinduism but in Islam and Christianity (1989, 193215, 29099
and passim). The Magistrates are caught in a bind on these issues.
What is most striking is that the bind is sustained by a kind of willed
blindness to certain realities of these cults and ceremonies. They are
ceremonies of the lower orders, but no recognition is given to which
of the lower orders it is who are specifically most involved.
Firewalking and hookswinging rites are patronized, generally, by
village headmen in Andhra; by landed dominant castes, often in village
headman roles, throughout much of Tamilnadu; and, in the case of
the Draupad cult, by local leaders of the Vanniyar caste, a traditional
186 chapter eight

landed warrior peasantry. Similarly, there is only the most minimal

acknowledgment that both Hindu and Muslim firewalking festivals
commemorate great tragic mythologies of battle, from Karbala to
Kuruksetra. It is precisely the martial ideology behind these rites that
is being subverted, ridiculed, and nullified in these magistrates reports
and recommendations.
We see here the tail end of a long process carried out by the Brit-
ish in south India, and, as we have seen, paralleled elsewhere in the
subcontinent. This was achieved by the neutralization, through reduc-
tion and appropriation, of the traditional Hindu and Muslim protec-
tive royal prerogatives: first, through the eighteenth century, of Hindu
and Muslim rulers themselves; then of regional little kings such as
the poligars at the end of that century; and finally, through the first
half of the nineteenth century, of the last vestige of the little king
ideal, that vested in the caste and village headmen who, among other
things, sponsored village and local dominant caste festivals (see C. A.
Bayly 23, 2728, 67, 107, 109, 14850, 174, 194, 205) for royal god-
desses such as Draupad and kingly gods such as Aiyanr (Dirks 1987,
298304, 37475). Indeed, as Appadurai has shown so nicely, the Brit-
ish could never attune themselves either intellectually, religiously, or
politically to the elegant and symbiotic division of sovereignty in
which, in traditional terms, The sovereign deity is the paradigm of
royal authority. By serving this deity, in the form of elaborate royal
gifts that generate special royal honors, and by protecting the redis-
tributive process of the temple, human kings share in this paradig-
matic royalty (1981, 51).
As C. A. Bayly has shown, after they had warred down intermedi-
ary chieftains and magnates in order to deal with the village elites,
and thus end the system of military land tenures which had pre-
vailed under the poligars, the British negotiated with village head-
men beginning in the 1820s, enhancing their status as an entree to
the economies of rural south Indian village culture ([1988] 1990, 107,
6768). But by mid-century, the village headmens traditional mirasi
rights and status (cf. Appadurai 1981, 141) had been monetized,
subdivided, and eroded through the increasing stress on cash crops,
division and selling off of family (and temple) lands to create small,
individual plots, and the decline in openings for military service. The
result was that farmer-warrior communities, not to mention so-called
criminal castes like the Maravars and Kall ars, were absorbed and
reduced into the wider peasant body (C. A. Bayly 14850) of what
colonialist lenses on the south indian draupad cult 187

our magistrates, writing in 185354, referred to, without differentia-

tion, as the most ignorant among the lower orders of dras, the
lower ranks and classes, and so on. Not coincidentally, 185354 marks
what Dirks has called the last gasp of the old regime in Pudukkottai,
the one remaining Poligar kingdom amid this sea of Madras Presi-
dency districts. There the British had to make a stand of force to sup-
press a rebellion led by a dismissed member of the palace guard
(riyakrar) that fed on resentment of rural taxes and discontent over
the profligacy and extravagance of the Tondaiman rja, himself thor-
oughly beholden by then to Brahman advisors and indebted to the
Company. As one would expect, the rebels made efforts to appropriate
symbols of the royal function (Dirks 1987, 30916, 321). Meanwhile,
for the north, just three years later, C. A. Bayly sees the loss of sta-
tus and political rights at the village level as one of the powerful
incentive[s] to revolt in 1857: that is, of course, in the so-called Sepoy
Mutiny [1988] 1990, 150).

C. Farmer-Warriors after a Century of Contempt

This brings us to the last phase of colonial Draupad cult documenta-

tion that we will attempt to cover: the late 19th century imperial con-
cern with the concept of caste, which, as J. C. Heesterman observes,
was hardened by administration, law courts, and especially the census
into a system of tight compartments, each stereotyped by its customs,
beliefs, attitudes, means of subsistence, governing institutions, and
other attributes. Like the village republic, each caste came to be seen
as a separate world complete in itself (1985, 182, cf. 180202; Appa-
durai 153 on the myth of the self-sufficient village). T. Aiyakannu
Nayakars Vannikula Vilakkam: A Treatise on the Vanniya Caste
(1891) addresses this situation.
The purpose of this well-researched treatise is to note that in the
former Census reports the Kshatriyas of Southern India were held in
contempt, and to urge that the present Census superintendant make
new inquiries and be more receptive to Vanniyar claims of Ksatriya
status (ibid., 6). Drawing equally on Tods Annals and Antiquities of
Rajasthan, which documents Rjput Agnikula Ksatriyas (ibid., 2428),
and Tamil sources, including Villiputtrs Makpratam (ibid., 1617,
37), which identify the Ceras as ancient South Indian Agnikula kings,
Nayakar is particularly keen on establishing the authenticity of a
188 chapter eight

Fire-born lineage of kings along with the more widely recognized

Solar and Lunar lines, or dynasties, since the Vanniyars claim to be
Ksatriyas of the Fire-lineage. Nayakar also draws on the opinions of
various Brahman pundits and Madras academics that his statements
are consistent with and do not seem in the least contrary to the
Shastras (ibid., 15). Here he is appealing to the same authorities as
the British magistrates 40 years earlier. The Ksatriya claim is further
buttressed by reference to Gustav Opperts view that the Vanniyars,
also called Pall is, descend from the Pallavas (Oppert 1893, 89101).
As I have mentioned, Oppert draws these dubious connections into
the Dravidian/Aryan divide. But Nayakar goes on, citing other Brit-
ish as well as Tamil authorities, to show that Vanniyars, who served
as local chieftains in Chingleput and Ramnad Districts, as princes in
Sri Lanka, and subordinate soldiers for the poligars, are now reduced
to the status of cultivators (Nayakar 1891, 2035; Appendix, p. 2).
Though it is not his main focus, he is also well aware that Draupad is
widely worshiped among Vanniyars, and that her own birth from fire,
along with that of her epic brother Dhrst adyumna, represents the pur-
est and holiest form of Ksatriya birth: one that occurs only to effect a
certain object (ibid., 810), as in the Mahbhrata, where the object
is the destruction of unruly Ksatriyas and the renovation of the earth.
Among special Vanniyar customs, he mentions that As a relic of
the origin of the Vannikula Kshatriyas from Fire, the Fire Pot
(neruppucatti or neruppukarakam), which comes in procession on a
fixed day during the occasion of the annual festivities of Draupadai
and other goddesses, is borne on the head of a Vanniya. Also in dra-
matic plays the king persona has always been taken part by a Kshatriya
who is generally a Vanniya (ibid., Appendix, 10). Here he refers to the
terukkttu dramas performed at Draupad festivals on Mahbhrata
themes drawn from Villiputtr.
It is easy to see that this document attempts to redress the leveling
to an undifferentiated peasant status of warrior-farmers, and of vil-
lage and caste headmen of landed dominant castes, that had occurred
earlier in the century. It is also ironic that current-day Vanniyars are
now agitatingin ways that much of the rest of Tamil society consid-
ers quite violentfor backward class status toward the end of gaining
government affirmative action concessions in jobs and education.
For our purposes, what is striking is that despite the predictions
of its decline that continue among some of Draupads devotees even
today, the Draupad cult has retained a vitality that such sources as
colonialist lenses on the south indian draupad cult 189

these, including Nayakars, with his reference to relic customs, con-

sistently fail, or refuse, to acknowledge. Perhaps we may view the
adaptability of such cults, in Ashis Nandys terms, as a triumph of the
dionysiac Kshatriyahood of the uncolonized mind (1983, chapter 2,
especially 7479). Indians crossed this cultural gap in ways the British
could only find confounding:
From the fire pit to the cricket field is a long jump, and Mr. Rajagopal
Moodelliar, who took part in last nights fire-walking, plays at todays
cricket match in Chepauk. His fire-walking performance is one instance
which goes to establish the truth of the observation often made, that the
average Hindu, while keeping abreast of modern times and benefiting
from the civilising agencies of the West, holds fast to his ancient super-
stitious practices. (Firewalking . . . 1900, 320, describing a firewalk for a
goddess called Thanthoni Amman near Madras.)
Among other factors, it was convenient and indeed purposeful for the
British to ignore that the Draupad cult, and the cults of other power
goddesses, represented access to a type of power that continued to
draw on persistent but also changing images not only of the relations
between the goddess and the little kingdom, but also her blessing of
the little king or queen in every firewalker.
Finally, one more note on what is left out of all this early docu-
mentation. It is not until W. Franciss 1906 South Arcot gazetteer and
F. J. Richards 1910 article on Draupad festivals in Bangalore that
we begin to hear of Draupads chief guardian, Pttu Rja. Francis
sets the tone: At South Arcot Draupad temples, he says, is often a
figure of Pothuraja, or the king of buffaloes, a person of ferocious
aspect. The stories accounting for his connection with Draupad are
conflicting and puerile, and need not be set out (1906, 99; Hiltebeitel
1988a, 76). Francis registers surprise that Vanniyar peasants know the
Mahbhrata uncommonly well. But he dismisses as puerile what
we may call the Draupad cults innovations in its folk Mahbhrata.
Today we know that this Buffalo King is Draupads link to Durg,
slayer of the Buffalo Demon (Hiltebeitel 1988a, 7688; 33393). And
we also know that Draupads other guardian, the Muslim Muttal
Rvuttan, whom none of our sources acknowledge until 1975 (Babb
1974, map 1, with no discussion), has much to tell us about how this
folk cult passed beneath the noses of so many early observers, or,
as our Hindu sources tell us, blinded those who looked too closely,
but only so that they might realize the power and grace of what they
had seen.


Landscapes of Urban Memory: The Sacred and the Civic in Indias High-
Tech City.
By Smrti Srinivas. Minneapolis: University of Minnesota Press, 2001.
376 pp. $57.95 (cloth); $22.95 (paper).
This important book signals the value of bringing theory in the human
sciences to bear on the study of Indian complexities, which is to say
that it is a work of disciplined interdisciplinary reflection on some
fascinatingly interrelated topics. It is about urban sociology and urban
history, focusing on Bangalore and comparing it with other cities, from
its staid and royal neighbor, Mysore, to its far-flung partners in the
global village. It explores different kinds of sacrality, from that of local
goddesses to the sacrality of urban sprawl. It studies a religious cult,
tracing the fortunes of Draupad worship by Bangalores Vahnikula
Ksatriyas, particularly at the Big Dharmarja Temple in Tigalarapet
Draupad is the heroine of the Mahbhrata; Dharmarja is the eldest
of her five Pndava husbands). It concentrates on this temples annual
festival, the Karaga jatre, named after the procession of a decorated
clay pot (karakam) through city routes. It is about ecology: how this
community, upon settling in Bangalore, obtained lands near water
tanks and and transformed its martial ethos into a profile of urban
gardeners selling vegetables, fruit, and flowers to the city. It traces the
effects of municipal and corporate plans that pushed back Bangalores
green space and underwrote the despoliation of most of the tanks and
wells that sustained these peoples gardens. It is an account of politi-
cal mobilization by the Tigala alliance that the Vahnikulas formed
with two affiliated fire-born castes in order to seek government res-
ervations and compensation for their lost garden lands. It is about
contested urban images: a ritual fort versus a sports stadium, the jatre
versus the National Games, Draupad versus Miss World (Bangalore
hosted this pageant in 1996), the informal sector and migrant and local
labor versus development plans and capital investment (p. 251). Tying
such things together, Srinivas seeks to understand the Karaga jatre
in its function as a mnemonic, its relationship to urban form and the
body and to constructions of the civic (p. 27).
192 chapter nine

Drawing on urban studies, religious studies, work on memory and

the body, classical Tamil landscape theory, globalization theory, and
performance theory, and ingeniously adapting Mikhail M. Bakhtins
literary concepts of heteroglossia and the chronotope to oral narrative
and processional ritual, Srinivas poses a question: Why has this jatre
grown, while the group that sponsors it has declined through dis-
placement, disenfranchisement, and marginalization (p. 246)? Her
answer seems to be that the jatres performance is ultimately politi-
cal (p. 246); it acts as a kind of newspaper, gathering up events of
the past year(s) and staging them (p. 34). Srinivas also suggests that
each years journal lies on a palimpsest of previous years, decades, and
centuries, and that it can be read as a depth charge into Bangalores
pasts. This argument is made on three main fronts: memory, bodies,
and oral epic.
Memory-wise, the jatre runs counter to a kind of civic amnesia
by drawing attention to groups who regard themselves as legitimate
subjects of urban history (p. 251). It succeeds in waking the city while
the new middle class sleeps in its highrises through a newer jatre in the
Doopanahalle suburb (p. 178). Here I have a small reservation. More
attention could be paid to forgetting, to the selectivity of memory, and
to the contrivance of false histories that Vahnikula advocates turn up
as recovered memories or invented traditions, some of which are
put in evidence but bypassed as a mythic universe (13237).
The book is especially persuasive in terms of somatic memory,
which Srinivas finds in cultural resources that are sedimented on
and stored in the bodies of citizens through particular gestures and
practices. She detects this historical deposit in movement, music,
rituals and certain styles of clothing (p. 30); in regimes of the body
concerned with the control of male and female bodies of players and
the flows between food, blood, and semen (p. 246); in the relation
between goddesses bodies and the boundaries of city and forest (p. 203);
in Draupads phases of incarnation, personification, and localiza-
tion (p. 246) and her embodiments in the arenas of the temple, fort,
and gymnasium, where her priest and other worshipers train among
wrestlers (p. 155); in the Karaga itself as Draupads body, and also
her priests body while he carries her. Srinivas coins the useful term
kinetograph here, suggesting a kind of dance notation, in order
to describe representation of place and time through movement
(pp. 13940, 203): alternation between sites such as the in-laws house
(temple) and the mothers house (fort); circumambulation of bathing
review of landscapes of urban memory 193

and bygone garden spots; procession to boundaries of the old fort-city;

and a suburban cycle connecting the city jatre with partial replicas in
the suburbs and surrounding towns, recalling the ring of forts with
tanks and gardens around Bangalore at its sixteenth-century founding
by Kempe Gowda (pp. 18299).
The oral epic, the Karaga Purana, consists of Tamil songs and
Telugu prose, with Kannada mixed in the latter (pp. 2078). It is
sung late in the tenth night after virtually all of the jatres kinetics
are over, and with practically no one besides ritual specialists attend-
ing (pp. 20910). Srinivas views this new story line as a recita-
tive topos . . . alongside the kinetographs (p. 203), and relates it to
Bangalores past. But, the oral memories argument is less convinc-
ing than the somatic argument. Srinivas mentions a family resem-
blance between the Karaga Puranas story about Potha Raja and his
fort-city of Shivananda Pattana and narratives from the Telugu Epic
of Palndu and the Draupad cult in Tamilnadu (p. 206). Bangalores
Vahnikula Ksatriyas are Tamil migrants to the city and are clearly the
main source for the Karaga Puranas Potha Raja story and another of
its stories about Draupad disguised as a gypsy. Srinivas argues that
the Karaga Purana could very well describe the Bangalore of Kempe
Gowdas time, but with two subtle absences: it does not mention
gardeners within the fort-settlement or tanks (pp. 235, 248). Trac-
ing the texts landscape to a period that extended from the ninth or
tenth century to the sixteenth, Srinivas argues that it provides another
model in addition to the ones that came into existence in Bangalore after
the sixteenth century, and as a reminder of an earlier history of urbanism
in south India (pp. 23435).
I believe there is something to this latter point, since the Potha Raja
story may recall the city of Kalyan in northern Karnataka that lies at
the twelfth-century origins of reformist Viraaivism. But this would
mean that the Karaga Purana is only tangentially connected with Ban-
galore, even if, as Srinivas presents it, its audience makes Potha Raja a
local Bangalore hero (pp. 155, 172, 17576, 229). Srinivas observes that
the Karaga Puranas topos includes forest, cave dwellers, forest sages,
occupations with wood, cowherds, nomads, huts, a walled fort with
a prison, and a palace. But, one attribution is dubious: [t]here are
also forest gardeners, it is clear, from the Vahnikula ksatriyas origin
(p. 234). Gardening comes up in only one Karaga Purana sequence
one that cannot be traced to Draupad worshipers outside Bangalore.
This story tells how Draupad devours the demon Timirasura, who
194 chapter nine

can regenerate himself from his own blood drops. Although other Ban-
galore sources connect this story with Draupads ascent to heaven at
the Mahbhratas end (pp. 15152), the Karaga Purana assigns it to
her forest exile. While still in her cosmic form after devouring Timira-
sura, Draupad gives birth to her chief Vahnikula followers (includ-
ing the Karaga Puranas singers) from her breasts. Then, as these new
sons sing her praises, she turns back into Draupad, gives ritual
images of herself and the Pndavas to these new children (presumably
for the Dharmarja temple), and urges them to become gardeners
(pp. 22627). This Bangalore story seems to mark the Karaga Purana
as an adaptation of an otherwise Tamil/Telugu cycle, gearing it the
Vahnikulas transformation into gardeners through the sanction of
the goddess they would have brought with them to the city. It is both
a myth of occupational origins and a kind of loose addendum to a
jatre that ties in much more extensively to memories of urban gardens.
There is nothing in the myth to indicate that gardening is the Vahni-
kulas original forest occupation.
As Srinivas notes, slippages of various kinds take place in the
Karaga Purana: it reverses the sequence of events of the classical
Mahbhrata and provides metonymic substitutions (pp. 3334).
Draupads life is not told consecutively in either the Karaga Purana
or the jatre (see pp. 20910), as it is at Kalasapalyam (p. 109), one of
Bangalores three other Draupad temples. Again in contrast to Kala-
sapalyam, this festivals Mahbhrata ends with Draupads marriage
to Arjuna and ignores the Mahbhrata war. Nevertheless, the Vah-
nikula Ksatriyas who gave shape to this would probably have known
a fuller Mahbhrata that was not the Sanskrit Mahbhrata. And if
that is so, what is interesting is that they would have elected to disag-
gregate it, not only separating ritual from myth but also emphasiz-
ing an out-of-sequence surplus (p. 206) like the Karaga Purana and
disjointed ritual fragments. In this regard, Srinivas no doubt points to
what was important for them: giving prominence to the Karaga as an
embodiment of urban power (akti) in order to implement strategies
and opportunities that were more pressing than the integrity of their
Mahbhrata. Over time, this has involved not only remembering but
forgetting, both supplemented by self-inventing mythic histories.


A surprising thing is observable in scholarly work on Indian classi-

cal and folk epics: their goddesses, and their heroines divinity, we
are told, are not part of the original stories. Goddesses are one thing,
and scholars who have favored the notion that epics begin as heroic
tales about real men have not treated them that much differently
from male deities: both have been viewed as afterthoughts in the epic
process, mythologizations of the historical kernel, latecomers to the
presumed epic amalgam. The only difference, and this has been said
of Indian religion more generally, is that goddesses are supposed to
come a little later into the epic mix than gods. They are presumed to
be the very last thing an epic would really or originally be about. But
what about heroines and their associations with divinity? If goddesses
do not belong to authentic epic, then a heroines links with divinity
need not be seriously considered at all.1 Yet Indias classical epics and
many of its folk epics have central places for heroines who are under-
stood to have some kind of divinity. Draupad, the heroine of the
Mahbhrata, for instance, is from birth dark like the goddess Earth
and an embodiment of r, the goddess of Prosperity. And indeed,
Draupad is worshiped as a form of the Great Goddess by certain com-
munities in Tamilnadu, south India (see Hiltebeitel 1988a, 1991a). In
studying her cult along with the classical epic, I have, over the years,
wanted to insist that the scholarly pattern of occluding the goddess,
which I have tried to show is textually untenable for the epic precisely
because of the way it portrays the heroine,2 has also taken the form of
another scholarly occlusion of the feminine, which one may take as
implying that of women.
Here, however, under the title Draupads Question, I would like to
take the matter of the heroine not in the direction of the goddess, but
in the direction of real flesh-and-blood historical women, both those

See my criticism (Hiltebeitel 1999a: 3743) of Blackburns views of epic develop-
ment, of which these comments are an extension.
On the textual issues, see Hiltebeitel (1980a; 1981; 2001, ch. 7).
196 chapter ten

for whom a heroinewhatever her divinitymight have spoken in

the time of the epics composition and as its text unfolds in history,3
and for whom she might still speak today. I would like to engage these
issues initially as they are at play in both the epic text itself and in
modern scholarship on the epic, and then look at what has become
of this question in an instance of contemporary fiction. Between god-
desses and flesh-and-blood historical women, what a heroine ques-
tions can be of interest in thinking about whether the goddess is a
The scene is the epic dice match between the rivaling Pn dava and
Kaurava cousins. Having bet all his wealth and brothers, Yudhist hira
is prodded into losing himself and then Draupad, his wife, who is also
the wife of his four brothers. The wager lost, Duryodhana orders that
Draupad be brought to the Kauravas as the slave of her new masters,
and Draupad asks her question to the messenger, refuses to come
with him, and challenges him to ask it in the gambling hall, which
he does. Says the messenger, As the owner of whom did you lose
us? So queries Draupad. Whom did you lose first, yourself or me?
Here, and for the rest of the scene, Yudhisth ira did not stir, as though
he had lost consciousness (Mbh 2.60.9).
It is the insolubility of the question, and the impasses it opens,
that provoke the two violent scenes of Draupads violation, her hair-
pulling and disrobing. The latter, the very last straw, is ordered by
the threefold response of the usually noble Kaurava ally Karna, who,
unbeknownst to himself and almost everyone at this point in the story,
is the Pndavas oldest brother. Karn a says: (1) Yudhist hira could bet
Draupad because she is included within his total property (61.31);
(2) he may have been prodded, but he bet her audibly and uncon-
testedly; and (3) One husband per wife is ordained by the gods . . . ;
she, who submits to many, is for certain a whore . . . strip her (61.38).
Textual contestation is especially rich here. How was she protected
by inexhaustible sarees? In reducing its reconstituted text, the Poona

For a comparable approach to questions raised by heroines in the putatively
male world of Western epics, see Suzuki (1989). I view the Mahbhrata as from
about 200 B.C.E.
Is the Goddess a Feminist? was the lead title of a 1994 American Academy of
Religion panel that I organized, from which a book followed with the same lead title
(Hiltebeitel and Erndl 2000). This essay (Hiltebeitel 2000) derives from my contribu-
tion to the original panel.
draupads question 197

Critical Edition makes it an unexplained wonder.5 In possibly the old-

est interpolation, they come from the concealed dharma. Cynthia
Leenerts, one of my recent graduate students who has written on the
Sanskrit epics and twentieth-century Indian fiction (1997), proposes
that Draupads own dharma as a devoted wife protects and vindicates
her. The Critical Edition editor Franklin Edgerton thinks cosmic jus-
tice is apparently implied (1944: xxix). Or was it the story everybody
knows, including Draupad and Krsna later in the Critical Edition
(5.58.21; 80.26), that Draupad prayed to Krsna?
We cannot follow the full debate over Draupad s question, the
insults, and the terrible vows of revenge.6 It is overtly a question of
dharma in the mens gambling hall or sabh as a courtroom. Dire
omens finally interrupt and prompt the Kauravas blind father to offer
Draupad boons that allow her to choose Yudhisth iras freedom, and
Karna recognizes that she has been the Pn davas salvation (nti) and
their boat to shore (64.13) by asking a question that now hangs unre-
solved over the entire epic.
How then does Draupads question relate to our books question?7
For the epic poets, Draupad is asking a womans question, if not the
womans question. Other heroines in the MahbhrataDamayant,
St, Ambalso raise it in varied forms: Is it really you who is doing
this to me? We see what subjugates Draupad in the mens hall in the
sarcastic words of Karna: There are three who own no property: a
slave, a student, and a woman are not independent (asvatantra). You
are the wife of a slave, his wealth, dear, without a lord, the wealth
of a slave, and a slave (yourself ). It is the dharma that subjugates
Draupad: these words echo a famous verse in the Laws of Manu that
describes women as never independent (asvatantra). Yet her appeal
to the dharma is also what saves the Pn davas and, in some sense,
herself. As we have seen, Draupad is also the incarnation of a goddess.
Is she a feminist? Granting definitional, historical, and intercultural
cautions, if we say yes, we have our best support from a speech she
begins as follows: These Kurus stand here in the hall, lords of their
daughters and daughters-in-law, all considering even my wordanswer

Thus Edgerton (1944: xxiv): No prayer by Draupad; no explanation of the mirac-
ulous replacement of one garment by another; no mention of Krsna or any superhu-
man agency.
See especially Mehendale (1986), and my analysis in Hiltebeitel 2001, ch. 7.
See n. 4 above.
198 chapter ten

this question of mine the proper way (61.45). She speaks about, and
perhaps for, women as a class, and challenges the men to consider a
question that questions their lordship over and ownership of women
in contexts of patriarchy. But we also have reasons to say no. Draupad
is a literary figure, not a woman. The Dead Indian Sanskritizing Males
(DISMs) who wrote the epic have made her a voice for what is at
best their sympathetic understanding of a womans question. And
what a strange image and voice! Let us look at the moment that she
is wagered and lost.
akuni, Yudhist hiras deceitful opponent at dicing, addresses Yudhist hira
with the probing verse: There is surely your dear lady (priy dev), one
throw unwon. Stake Dark (Krsn ) Pcl. Win yourself back by her.
The first meaning of dev is of course goddess, which would be an
overtranslation but hardly an overestimation of the lady. More than
this, dev is the goddess as she who plays (Biardeau 1985: 17), and
Pcl is a name for Draupad that is used with heightened frequency
in this scene (Biardeau 1985: 11, 1314), meaning the puppet. As
Yudhist hira sets to wager her, he speaks from what feels like a revery:
She is not too short or too tall, not too black or too redI play you
with her . . . Doll-like, iconic, she is bet and lost, and Yudhisth ira will
say nothing until her question wins him his freedom. Later, Draupad
will say we are all puppets, our strings pulled by the Creatorto which
Yudhisth ira will reply that she is eloquent with passion, but heretical.
As Pcl, Draupad is thus the doll or puppet who speaks, who even
recognizes herself as such; as dev, she is the lady who plays who is also
played. She raises the feminist questions of a Barbie dollwhich can,
of course, be interesting.
It is clear that in all this talk about betting oneself, Draupads ques-
tion is a philosophical one about the nature of Self. As a philosophical
question, it is compounded by legal issues of property, ownership, and
slavery in the hierarchical context of patriarchal marriage, and sym-
bolized around the figure of the ultimate lord, master and owner, the
king, in relation to the subjecthood and objecthood of the queen, his
wife. These themes are in the Sanskrit, as are those of possession of Self
vs. possession by the madness of dicing, and, I submit, the theme of
love and abandonment, of love tested to, and maybe past, the breaking
pointlove between six people in one and the same marriage! Listen
to what Yudhist hira says when he bets himself just before he wagers
Draupad: I am left, so beloved of all my brothers. Won, we shall do
draupads question 199

work for you when the self is itself a deluge (upaplave; 58.27). Slipping
into his loving revery when he bets the dear lady whom all five broth-
ers hold dear, Yudhisth ira descends into silence once he has lost her.
Here we see Draupads question from a new angle. Yudhist hiras loss
of self appearsit is only described so by othersto be a loss of con-
sciousness. Is it the higher Self that is ultimately at stake in Draupads
Here we seem to be in an agonistic multi-dialogical situation that
reverberates with Upanisadic scenes in which fathers and sons, gurus
and disciples, and, in a few memorable passages, especially men and
women churn the oppositional languages of rivalry and status to release
the saving knowledge of the Self that is one. Garg, whose questions to
Yjavalkya are like arrows; Yjavalkya to Maitrey, renouncing the
world and saying goodbye forever to his dearer-than-ever knowledge-
discoursing (brahmavadin) wife: Not for love of the husband is a hus-
band dear, but for love of the tman . . . ; not for love of the wife . . .
But in the Mahbhrata, the language of such questions and answers
has been compounded by the terminologies of Sm khya, Yoga, and
bhakti. Here Al Collins has been immensely helpful in revealing how
a sovereign self, male (purusa), replicates itself in other selves through
a scale of forms that presents a problem for male identity formation
(1994: 34; cf. 2000). Man (and, for that matter, woman), as mind-
ego-intellect, is feminine matter, prakrti, livingultimately uncon-
sciouslyfor the sake of purusa, of man as conscious self or soul. As
Collins says, The problem is not limited to kings who top this scale
of forms, but is universal: it is highly dangerous to claim to be a self
(1994: 4). How beautifully this describes Yudhist hiras predicament;
no wonder he is silent.
But what about Draupad? In a context where the feminine is
unconscious matter, can this be a feminist question and the questioner
be feminist? Madeleine Biardeau, also insistent that a Sm khya prob-
lematic underlies Mahbhrata portrayals of heroines, gives us our
best route to a negative answer. For her, the heroine-goddess repre-
sents prakrti as unconscious matter, blind ignorance given to obsti-
nancy; matter that unknowingly yet somehow inerrantly works on
behalf of purusa through blind initiatives; heroines whose ignorance
is unknowing in particular about dharma (1984: 263). There is more to
this than I can go into, but I think it is one-sided. Biardeau says noth-
ing about ignorance when it is a question of the heroines ingenuity
200 chapter ten

in posing questions that are riddles that prompt the heros anagnorisis
and draw him out of his concealment, riddles about the intimacies and
pains of love, riddles that are a lifeline allowing him and her to use
their wits and to play, or as David Shulman puts it, fight for time
(2001, 42). Draupad certainly knows enough about dharma to ques-
tion it. It is not just her obstinancy that makes her persist. As another
of my recent graduate students, Lena Taneja, says, Draupad never
seems to doubt for a moment that she is truly free. It is perhaps her
sense of this freedom that keeps her sticking to the question that will
also free her husbands.
I cannot trace the Mahbhratas tale of love (and other things)
between Draupad and Yudhisth ira any further here. Suffice it to say
that from the dice match on, there is no humanly happy ending. But
even on the level of Sm khya ultimates, there are strange goings-on for
epic love stories to activate. Here again Collins is eye-opening, showing
that there are ways in which prakrti as buddhi (the intellect or faculty
of awareness) becomes indistinguishable from purusa (1994: 2122).
It is worth asking whether Draupads sense of her freedom might be
a reflection of what Gerald Larson and Ram Shankar Bhattacharya call
the intuitive capacity of the buddhi for freedomthe capacity finally
of discriminating the presence of contentless consciousness (1987:
8283). Pure contentless consciousness as witnessing passive pres-
ence (1987: 7781) is, however, probably not that attractive a mate, so
it is not surprising that prakrti, as Collins says, makes a very ambiva-
lent feminist.
So Draupad too must be an ambivalent feminist. Perhaps the most
we can say is that her question is one that an interculturally sensi-
tive feminism might find interesting, and that it is not surprising that
Indian feminists sometimes find it worth writing and thinking about. I
refer especially to Mahasveta Devis Bengali short story Draupadi and
Gayatri Spivaks translation and discussions of it (1988).
Draupadi is Comrade Dopdi Mejhen, a Santal tribal woman who
has participated in Naxalite resistance to landlord oppression, survived
a police raid, escaped to the forest, and joined book-educated young
gentlemen revolutionaries, for whom she is a courier between forest
and village. She is being hunted down by army Intelligence and its half-
breed Santali informers. Mahasweta Devis story keeps us on course;
she writes: Government procedure being as incomprehensible as the
Male Principle in Samkhya philosophy, the hunt for Dopdi is first led
by a captain whose diabetes was inflamed when he did not find her
draupads question 201

and her husband in the body count after the raid. He soon turns the
case over to Senanayak, the elderly Bengali specialist in combat and
extreme-Left politics, who respects the opposition, hoping to write
eventually about his theory that In order to destroy the enemy, become
one (Spivak 1988: 199).8 He has much in common with the passive
witnessing incomprehensible male principle. Mahasweta Devi satirizes
this principle as analogous to Archimedes fulcrum point: one of the
tribal specialists runs in with a joy as naked and transparent as Archi-
medes and says, Get up, sir! I have discovered the meaning of that
hende rambra stuff. Its Mundari language(p. 190).
Dopdis husband Dulna is killed in ambush. She loved Dulna more
than her blood. No doubt it is she who is saving the fugitives now, says
a narrative reminder of Draupad. When Dopdi is captured, she is offi-
cially Sanskritized to Draupadi for Senanayak, who questions her for
about an hour, breaks for his dinner, and orders her gang rape: Make
her. Do the needful (1988: 195). In the morning, rather than wash
and wear the white cloth that has been kept aside for her to appear in
before Senanayak, Draupadi stands up. She pours the water down on
the ground. Tears her piece of cloth with her teeth. Seeing such strange
behavior, the guard says, Shes gone crazy, and runs for orders. Sena-
nayak comes and, at the end, is terrified: What is this? He is about to
bark. Draupadi comes closer. Stands with her hand on her hip, laughs
and says, The object of your search, Dopdi Mejhen. You asked them
to make me up, dont you want to see how they made me? He sees
her raped nakednessWhere are her clothes? Wont put them on, sir.
Tearing them. He receives her defiant questions, spoken as her lips
bleed. She shakes with an indomitable laughter that Senanayak simply
cannot understand . . . Whats the use of clothes? You can strip me, but
how can you clothe me again? Are you a man? She spits blood on his
shirt, and says, there isnt a man here that I should be ashamed. I will
not let you put my cloth on me. What more can you do? Come on.
Counter mecome on, counter me? Draupad pushes Senanayak with
her two mangled breasts, and for the first time Senanayak is afraid to
stand before an unarmed target, terribly afraid (1988: 196).
So Dopdi has three, four questions, but the last one is really left
hanging. What more he can do to her is an unpleasant matter to con-
template. This Senanayak can do a lot. There is little doubt that he

Spivak italicizes the Bengali usages of English words in her translation.
202 chapter ten

will kill her. His fear might not even prevent what Dopdis comrades
call her next stage: interrogation? torture? There are no suppressed
husbands and descending deities left. This is not that story. And this
story does not go on (and on . . .). We know that Senanayak cannot
clothe her either as a man or a god. He is neither. And we know that
Dopdi has no fear before him.
Spivak writes, It is when [Dopdi] crosses the sexual differential into
the field of what could only happen to a woman that she emerges as
the most powerful subject, who, still using the language of sexual
honor, can derisively call herself the object of your search (1988:
184). She is the only one who uses the word counter as in killed
by police in an encounter, the code description for death by police
torture. Dopdi does not understand English, but she understands this
formula . . . What is it to use a language correctly without know-
ing it? (186).
Let us remember Draupads class-action appeal for daughters and
daughters-in-law. It calls into question two kinds of male lordship:
that of kinship and family, and that of the dharmic politics of kingship
in the sabh or mens hall. Dopdi Mejhen is captured because she is
caught between two kinds of loyaltyone, the tribal community of
Santals; the other, the comradeship of the gentlemen revolutionaries.
Here Spivak can return us to the questions of territoriality, the occlu-
sion of women, and the scale of forms that allows patriarchal patterns
to replicate themselves from one political situation and scholarly situ-
ation to another:
The figure of the exchanged woman still produces the cohesive unity of a
clan, even as what emerges is a king . . . [Through all the heterogeneous
examples of territoriality and the communal mode of power, the figure
of the woman, moving from clan to clan, and family to family as daugh-
ter/ sister and wife/ mother, syntaxes patriarchal continuity even as she
is herself drained of proper identity (1988: 21920).
If Sheldon Pollock (1994) can question whether there is a deep Ori-
entalism in classical Sanskrit sources, perhaps Draupad can be a deep
feminist. If not, she has thought up a good question, one that seems
to have been good for feminists to think with.9 We are left with the
question: What, amid the philosophical ignorance or the linguistic

The next two sentences and beginning of the last one are newly introduced
draupads question 203

knowing without knowing ascribed to her, could Draupad-Dopdi

know? Could she have a feminist consciousness before and after all?
Since one can only go so far in questioning literary characters, perhaps
it is sufficient to revive Yudhisth iras silence, and thank Spivak for one
last question, and the answer she gives it: can men theorize feminism,
can whites theorize racism, can the bourgeois theorize revolution and
so on. It is when only the former theorize that the situation is politi-
cally intolerable (1988: 253).



Two Tamil Mahbhrata folk cults that I have studied recreate the
Indian epics sacrifice of battle in their festivals, and rethematize, in
the symbolic registers of south Indian village goddesses and regional
kingdoms, its meditations on death and regeneration (Hiltebeitel [1976]
1990: 245360; 1988a: 394435). One is the cult of Draupad (the epics
chief heroine); the other that of Kttn ta var, son of Arjuna (the epics
chief hero) and the serpent-woman Ulp. I will focus in this and in
the next three chapters primarily on the latter. Kttnt avar is a Tamil
deity whose myths and rituals orchestrate multiple, multiform deaths.
The enormity of death that is incorporated in either cults perfor-
mance of the Mahbhrata war is, in the Kttnt avar cult, condensed
into the reiterated deaths and revivals of this one hero, and expressed
metaphorically through mythic and ritual forms of body-building
involving icons, effigies, and personsthat construct, transform, dis-
mantle, and reconstruct bodies and heads. Although the metaphor
could also apply to Draupad cult rituals, it is especially appropriate
for Kttnt avar, who is built up like Kirilov in Dostoyevskys The Pos-
sessed. What is surprising in his cult, and of moment for this chapter,
is that this primarily martial metaphor has been appropriated by Alis,
Tamil transsexuals or eunuchs.
Kttn t avar is more widely known as Aravn, the name used in
both cults, and in Tamil versions of the Mahbhrata. In the Sanskrit
Mahbhrata he is called Irvat. The name Kttnt avar is used only in
his own cult. For the moment, we may say that Aravn and Kttn t avar
are two names for one deity, although I will concentrate on a South
Arcot tradition in which the two names refer to him in two successive
lives. The Kttn t avar cult has its greatest concentration in the South

This and the next two chapters have been slightly revised, mainly to avoid overlaps,
but also in view of rounding off my studies of Kttn ta var with a new chapter 14 for
this volume. The four chapters appear here in the order they were written.
208 chapter eleven

Arcot and North Arcot districts of Tamilnadu. This is also the case
with the Draupad cult, which, I have argued, probably developed in
a core area centered on the medieval kingdom of Gingee (Ceci),
which included these two districts and the adjacent Chingleput Dis-
trict to the northeast. Kttnta var temples, however, are much fewer
in number and more limited in their diffusion. Draupad temples are
found with heavy diffusion into Tanjavur District to the south of the
core area, and have extended diffusion throughout Tamilnadu up
into southern Andhra and nearby Karnataka. Kttnt avar temples
are found only in a belt from South Arcot to Coimbatore, and have
not extended into Tanjavur District (see map). The two cults provide
glimpses of contrasting ways and contexts in which the Mahbhrata
has been activated as a text for Hindu revivalism.2
1. Alnturai (Alandurai), South Coimbatore Tk., Coimbatore Dt.
2. Annr, North Coimbatore Tk., Coimbatore Dt.
3. Anumantrttam (Hanumanteertham), Uthangarai Tk., Dharmapuri
(now Krishnagiri) Dt.
4. Cmiypuram, Harur Tk., Dharmapuri Dt.
5. Crppapptt u, Chengam Tk., NA (now Tiruvannamalai) Dt.
6. Cinkanallr (Singanallur), South Coimbatore Tk., Coimbatore Dt.
7. Clavaram, Vellore Tk., NA (now Vellore) Dt.
8. Kalankiyam, Gopichettipalayam Tk., Coimbatore (now Erode) Dt.
9. Kacappalli, Avanashi Tk., Coimbatore (now Tiruppur) Dt.
10. Kilvanampati, Chengam Tk., NA (now Tiruvannamalai) Dt.
11. Konalr, Gingee Tk., SA (now Cuddalore) Dt.
12. Kottattai, Chidambaram Tk., SA (now Cuddalore) Dt.
13. Kvilr, Palacode Tk., Dharmapuri Dt.
14. Kumramankalam, Vaniyampati Tk., NA (now Vellore) Dt.
15. Kuricci, South Coimbatore Tk., Coimbatore Dt.
16. Kottampatti, Avanashi Tk., Coimbatore (now Tiruppur) Dt.
17. Kvkkam, Ulunderpet Tk., SA (now Villupuram) Dt.

Cf. Pollocks mythopolitical reading of the Rmyan a as a revivalistic imaginary
that served as the privileged instrument for encoding or interpreting the empirical
realities of the twelfth to fourteenth centuries (1993: 28081), and subsequent cen-
turies as well. Pollocks important insights need, however, to be further historically
and regionally contextualized in relation to analogous activations of the Mahbhrata
(such as the Draupad cult), and also carried back to a rereading, in this light, of the
Sanskrit epics themselves, which Pollock seems to take as originally depicting a once-
actual (rather than also imaginary) political model (1986: 924). I pursue these issues
elsewhere (Hiltebeitel 1999a).
dying before the mahbhrata war 209

18. Madhukarai, Pondicherry State.

19. Mmpatti, Harur Tk., Dharmapuri Dt.
20. Manntipatti, Uthangarai Tk., Dharmapuri (now Krishnagiri) Dt.
21. N. Tattakkal, Krishnagiri Tk., Dharmapuri (now Krishnigiri) Dt.
22. Otukattr, Vellore Tk., NA (now Vellore) Dt.
23. Palaiyacramankalam, Salem city and Taluk, Salem Dt.
24. Panaimatal, Attur Tk., Salem Dt.
25. Pannrippatti, Uthangarai Tk., Dharmapuri (now Krishnagiri) Dt.
26. Pppirettippatti, Pappireddipatti Tk., Dharmapuri Dt.
27. Paraiyappatti, Harur Tk., Dharmapuri Dt.
28. Pelr, Salem Tk., Salem Dt.
29. Pennaivalam, Tirukoyilur Tk., SA (now Villupuram) Dt.
30. Pillaiyarkuppam, Pondicherry State
31. Pulimtu, Vellore Tk., NA (now Vellore) Dt.
32. Putr, Vellore Tk., NA (now Vellore) Tk.
33. Pvanakiri, Chidambaram Tk., SA (now Cuddalore) Dt.
34. Tailpuram, Vanur Tk., SA (now Villupuram) Dt.
35. Tat trmatam, Tiruchendur Tk., Tuticorin (formerly Tirunelveli) Dt.
36. Ttavr, Attur Tk. Salem Dt.
37. Tvanampattinam, Cuddalore Tk., SA (now Cuddalore) Dt.
38. Tvanr, Tiruvannamalai Tk., NA (now Tiruvannamalai) Dt.
39. Tiruvetkalam, Chidambaram Tk., SA (now Cuddalore) Dt.
40. Tutiyalr (Tudiyalur), North Coimbatore Tk., Coimbatore Dt.
41. Vatavalli, North Coimbatore Tk., Coimbatore Dt.
42. Vellayampati, Vaniyampati Tk., NA (now Vellore) Dt.
43. Vellalr (Vellalur), South Coimbatore Tk., Coimbatore Dt.
44. Vtantavati, Tiruvannamalai Tk., NA (now Tiruvannamalai) Dt.
45. Vranantal, Chengam Tk., NA (now Tiruvannamalai) Dt.
Map of Tamilnadu showing Kttn ta var and Aravn Temples visited
in fieldwork, with alphabetical list of temples corresponding to num-
bers on map:3

This map is newly prepared for this volume to account for all Kttn ta var temples
discussed in this and the next three chapters. Cf. Hiltebeitel 1995b: 449 for the map
of 32 temples prepared for the initial version of this chapter (while under preparation
by the Journal of Asian Studies, the temple numbers were incorrectly alphabetized);
and 1999b, 277 for the map of Kttnta var temples mentioned in chapter 13. Tk. =
Taluk. Dt. = District. NA = North Arcot. SA = South Arcot. Older district names are
mentioned along with new ones since the older ones are most useful in mapping broad
cultural areas. For maps and discussion of the wider extension of the Draupad cult,
see Hiltebeitel 1988a: 2331; 1991a: 46, 1116.
210 chapter eleven

I have visited almost every one of the Kttn t avar temples mapped
here, and have found that their ritual and mythological differentiation
by region is profound. These regional variations supply pieces for the
puzzle we must try to assemble.

Aravn in Sanskrit and Tamil Sources

One learns of Arjunas brief affair with the Ng or serpent woman

Ulp in the first book of the Mahbhrata, the di Parvan (1.206),
but nothing is said there about the conception of their son, Irvat.
dying before the mahbhrata war 211

What we learn about Irvat comes from the scene of his death dur-
ing the eighth day of battle in the Bhsma Parvan. Here we find that
Ulp was a childless widow, her Nga husband having been slain by
Suparna, that is, the heavenly bird Garuda (6.86.7). Having become
cheerless, she was bestowed on Arjuna by her father Airavat, the ser-
pent king (but possibly elephant king, ngarja; 67: serpents and
elephants having an important interchangeability in Aravns subse-
quent mythology). Her son with Arjuna, Irvat, was thus begotten
upon the wife of another (paraksetre). This note of illegitimacy is
compounded by one of family discord among the Ngas, which parallels
212 chapter eleven

that between the Kauravas and the Pn davas: Irvat was abandoned by
his (unnamed) fathers brother (pitrvya), who hated Arjuna, and was
raised by his mother in Ngaloka (9). It can be no accident that his
father is called Kauravya. The discord in Irvats Nga family reflects
that between the Kauravas and the Pndavas.
Although there is no mention of it in the ran yaka Parvan account
of Arjunas sojourn in Indraloka (after his tapas to iva), the Bhsma
Parvan also tells us that Irvat went to visit Arjuna, his father, when
Arjuna was with Indra in Indraloka. His visit thus parallels Arjunas
ascent to Indraloka to meet Indra, his own father. By reminding
Arjuna of his affair with Ulp, Irvats mother, and filling Arjuna with
joy and paternal affection, Irvat lays claim to Arjunas paternity just
as Arjuna had done with Indra. In return, Arjuna calls upon Irvat to
assist the Pn davas in the coming battle. In the Sanskrit epic, this is
all there is to motivate Irvats siding with the Pndavas, and, likewise,
the only hint that he might have done otherwise.
So now, with this immediate background, Irvat comes to battle
on the eighth day, and engages in a fierce cavalry fight. When only a
remnant (esa; 22) of the horsemen is left, the six younger brothers
of akuni, called Saubalas, ride out, supported by akuni, leading the
Gndhra warriors in the charge against Irvat. The Saubalas surround
him on all sides (30). Pierced like an elephant (iva dvipah ; 32) with
lances and hooks, Irvat is bathed in blood, but his natural firmness
(dhairyam; 33) is still unchecked. He tears the lances off his body and
rushes at the Saubalas on foot, bearing a sword (khadga), while they
remain on their horses. They surround him again, and he kills them
by severing their arms and other limbs (1). Only one Saubala, named
Vrsa va, escapes, badly wounded.
Then Duryodhana sends out ryarngin (= Alambusa), a terrify-
ing Rksasa descended from the singular (one-horned) sage Ryarnga
(he is repeatedly called ryarngi), who bears a grudge against the
Pndavas for Bhmas slaying of Baksura, one of his Rksasa rela-
tives. Each using his my, the two rise into the sky, and Irvat severs
Alambusas limbs repeatedly, which only grow back with new youth
(yauvanam; 5859). Cut again and again like a tree (iva drumah ; 62),
Alambusa tries to seize Irvat, enraging him. Then an unnamed Nga
of Irvats mothers lineage (anvayo mtrkas tasya) arrives surrounded
by many other Ngas and assumes a mighty form like the serpent
Ananta (dadhra sumahad rpam ananta iva bhogavn; 6667). This
Ananta-like Nga momentarily covers (chad) the Rksasa with Ngas
dying before the mahbhrata war 213

until Alambusa reflects (dhytv), assumes the form of Sauparna-

Garuda, and devours all the snakes. When the Ananta-like kinsman
of his mothers line is devoured by Alambusas my, Irvat becomes
confounded (vimohitam), and Alambusa finally beheads him with
his sword, leaving his severed crowned and earringed (sakun dalam
samukutam) head to fall to the ground resplendent like a lotus or
the moon (padmendusadraprabham; 70). The warriors on both
sides continue to fight on with heightened intensity, as if possessed
(vist h ) by Rksasas and Bhtas (85).
One might trace forward a number of themes in this account into the
Tamil reworkings of the tale. The name Irvat, as Madeleine Biardeau
has suggested, may reflect the Vedic id as the principle of sacrificial
fecundity vied over by gods and demons, and now by the Kauravas
and Pndavas (Hiltebeitel 1988a: 320; Biardeau and Malamoud 1976:
14344; Biardeau and Pterfalvi 1985: 187; Heesterman 1993: 2013).
As possible prelude to Aravns kalappali (battlefield sacrifice), Irvat
is noted for his firmness while bathed in blood. He is compared to an
elephant (which in Tamil versions sometimes serves for the Kauravas
counter-kalappali: that is, after the Pndavas offer Aravn, the Kaura-
vas offer an elephant (Francis 1906: 375); or, in Villiputtr lvrs
Tamil Makpratam, an elephant is offered along with Aravn).4 He
is involved in dismembering scenes, although it is he who dismembers
rather than being dismembered himself. Similarly, the Ngas come to
his aid in the battle by covering one of the combatants by their my:
in the Sanskrit epic, they cover Alambusa; in popular Tamil traditions,
they encompass Aravn himself. In each case, the Rksasa dispels this
illusion by assuming the form of Garuda, the eagle mount of Visnu
and perennial terrifier of snakes. In the Sanskrit epic, Alambusa does
this upon his own reflection; in Tamil versions, he assumes the form
of Garuda at the devious prompting of Krsn a. The severing of Irvats
diademed head is followed by closing scenes of possession (vea)
that could prefigure possession scenes connected with Aravn in our
two cults. And there is the role of the Ananta-like snake who comes
momentarily to Irvats rescue, perhaps foreshadowing the appearance

Medieval Tamil Paran i poems speak of parts of elephants killed or captured in
battle being made into feares of Kl temples, and of iva offering Kl elephant skin and
intestines (Nagaswamy 1982: 2527). iva, of course, revives Gan ea with an elephant
head offered to appease Prvats outrage at the death of her son and transformation
into Klrp.
214 chapter eleven

of Ananta himself as Aravns grandfather in the latters South Arcot

folklore (see Hiltebeitel 1988a: 32930; 1991: 286, 30819).
Indeed, the Sanskrit account unfolds a triple bird-snake opposition.
First, Irvats father Kauravya is slain by Garuda. Second, as Biardeau
points out (198081: 23637), the six brothers of akuni, whose own
name is that of a vulture-like bird of ill omen, are repeatedly referred
to as the Saubalas. Subala, after whom this name derives, is one of the
six sons of Garuda. Irvat, half-Nga himself, thus gets to kill five of
the six Saubalas, who perhaps represent a negative side of the lineage
of Garuda. Third, Irvat is slain by Alambusa when the latter takes on
the form of Garuda himself. There is thus an alternation: first, bird
over snake, accounting for Irvats matrilocal upbringing, unusual for
a Ksatriya; then snake over bird, accounting for his fleeting victory in
battle; and finally bird over snake, accounting for his death.
But even adding all these things together, one is left with the ques-
tion why Irvat? How is it that the Tamil tradition has taken this
one very minor hero, this hero for a day, and given him a mythology
that far surpasses anything that the Sanskrit epic can rightly be said
to anticipate?

Tamil Variants of Aravns Mythology

It is probably not that difficult to explain Irvats transformation from

a minor Sanskrit figure to a major Tamil one. Early Tamil interest in
Irvat can be traced to the ninth century of the Pallava period, when we
find not only the composition of the oldest surviving Tamil rendition
of the epic by Peruntvanr, which supplies Aravns literary debut,
but the beginnings of temple recitation of the Mahbhrata, possibly to
instil a martial spirit (Hiltebeitel 1988a: 14); the development of temple
iconography of head offerings and other forms of dismemberment to
Durg (31820); and the emergence, clear already in Peruntvanr and
intensified in tenth- to twelfth-century Paran i poems, of Kl as Tamil
goddess of the battlefield (Nagaswamy 1982: 2228; Hiltebeitel 1991a:
31213). From a low status and outlying figure, he becomes in Tamil
traditions the focus of a new martial spirit linked with devotion to the
goddess and a dedication to royal superiors so noble and complete
that he is willing to sacrifice himself painfully to the goddess for the
victory of either sides king. Tamil traditions, and especially our two
cults, only unfold upon these primary themes, finding in them further
ramifications to play out in terms of regional, village, caste, familial,
dying before the mahbhrata war 215

and gender tensions, which, as they are expressed through forms of

popular bhakti or devotion, are the subjects of this chapter. We will
begin to trace these uniquely Tamil unfoldings in terms of Aravns
accumulation, from classical to popular sources, of three boons, the
last of which, from the folk traditions of our two cults, is his prewar
Peruntvanr tells us nothing of Aravns birth, presumably because
the early parts of his work are lost. Just before the battle, however,
Duryodhana asks Bhsma for the best astrologer to fix the beginning
of the war and to recommend the ideal victim for kalappali, battlefield
sacrifice. Bhsma recommends that Duryodhana consult Sahadeva.
Duryodhana goes to Sahadeva, learns the ideal hour and the ideal vic-
tim, Aravn, and then goes to Aravn to request that he offer up his
blood and flesh to Kl on the forthcoming amvcai, or new moon
night. Aravn happily promises his compliance, and both go home.
Krsn a, with his tiruvulla m (his holy mind), knows what is afoot. Pro-
voked, he tells Dharmarja (Yudhisth ira) that the Pn davas are foolishly
giving boons to the Kauravas, and that Sahadeva has fixed things for
Duryodhana exactly as he, Krsn a, had planned them for the Pndavas!
Having already planned that Aravn would offer himself as a battle-
field sacrifice for the Pndavas, now that his plans are thwarted, he is
ready to abandon them and go off in a huff to Dvrak. Dharma hum-
bly requests his pardon and help. Krsn a agrees, and sets out to convert
the fourteenth day (caturda) of the dark fortnight into an amvcai
so that Aravn can sacrifice himself a day early for the Pn davas. This
he achieves by getting Brahmans to do amvacai ceremonies on the
caturda. Technically, an amvcai occurs when the sun and moon
meet (by definition, it is their coming together, Sanskrit, sryendu
sam gamam (Venkatesa Acharya 1981: 141]). Candra and Srya are
surprised that the Brahmans are performing deceased ancestor cer-
emonies (tarpan am, etc.) a day early, but knowing that Brahmans are
infallible, they come to look, and thus the sun and moon meet, mak-
ing it an amvcai. Krsn a now orchestrates Aravns kalappali so that
it will occur on this preponed amvcai for the Pndavas instead of the
Kauravas. Yudhisth ira is stupefied at the thought of human sacrifice,
especially that of his brother Arjunas son, and in extended terms, his
own (cf. Shulman 1993, on the theme of filicide). But he finally agrees
when the resolute Aravn offers the compromise that he will make his
initial kalappali of flesh and blood on the caturda (fourteenth), and
then give what remains of it to Duryodhana the next day.
216 chapter eleven

Aravn now demands his own boon in return. He wants to die a

warriors death on the battlefield. Krsn a grants him life up to the eighth
day, even without his flesh and blood. The sacrifice then occurs with
flowers, incense, and waving of lamps; the warriors roar, and Krsn a
touches their weapons. On the wars eighth day, Aravn fights bravely.
Duryodhana asks the Rksasa Alampucan (Sanskrit Alambusa) to kill
him, but Aravn outmatches him. Seeing this, Krsn a goes to Alampucan
and tells him that he can kill Aravn in the form of a bird of prey, the
brahmany kite (kalulavuruvukkon tu ), the bird that Tamil folk tradition
identifies with Garuda. Doing so, Alampucan kills Aravn (Venkate-
sha Acharya 1981: 13744; Manivannan 1984: 1314). Thus Aravns
first boon, historically, is nothing more than a narrative requirement if
he is both to offer kalappali, his new charge in the Tamil epic, and to
retain his eighth-day death. Krsn a is now in charge both of Aravns
kalappali and of the reflections by Alampucan that enable him to
bring about Aravns final fall.
Further innovations in the Tamil tradition can then be traced around
the boons that Aravn obtains from Krsn a for his compact to perform
kalappali. In the Prata Ven p, he obtains only one boon by which, in
effect, he virtually dies twice: once by his kalappali before the war, and
again on the eighth day. By the time of Villiputtrs fourteenth-century
Makpratam, he asks a second boon of being able to watch the war
for a few days (cila nl), and then to die in battle after killing many
enemies (5.7.6). Aravn cuts off his limbs one by one according to the
(kta) Ymala rite. His head survives like a radiant lamp (malarnta
tpam ena; 8), and seems to be all that survives to fight and be slain by
Alampucan on the eighth day.5 This second boon of a few days of
war-vigil does not suggest, but also does not rule out, that the head will
survive the slaying by Alampucan. But that is what Krsn a promises in
the folk traditions of our two cults. Here, this second boon promises
Aravn that he will be able to see the whole war. This requires a third
death: that of his head (or at least eyes and ears), which are now graced
with life till the eighteen-day wars end. This war-vigil is regularly ritu-
alized in the Draupad cult, where Aravns effigy overlooks the ritual
battlefield (patukalam); but it can also be played out in various ways
at Kttnt avar festivals (see chapters 1214). Finally, Draupad and
Kttnta var cult folk traditions also add a third boon: Aravn does
not want to die a bachelor, since dying unmarried would prevent him

I thank David Shulman for clarifying this section of the Villipratam.
dying before the mahbhrata war 217

from receiving the ancestral rites of a Father or Pitr. Krsn a, asking,

Who would marry someone just about to die, solves this difficulty by
taking the form of Mohin, the Enchantress, and marrying Aravn in a
last-minute wedding. This is the version of Aravns marriage reported
in scholarly literature on south Indian popular Hinduism (Whitehead
1921: 27; Shulman 1980: 3067). But it turns out that it is known only
in South Arcot. Elsewhere, Krsn a solves this predicament by arrang-
ing entirely different prewar marriages for Aravn. This chapter, how-
ever, will be concerned only with the Mohin variant and its ritualized
importance in the Kttnta var cult of (eastern) South Arcot, where
his festivals are most spectacular, and where the myth seems to have a
special connection with the increasing involvement in the cult by Alis,
so-called eunuchs.
To prevent confusing the historical order of Aravns three boons as
his story develops with the narrative sequence of his three deaths in the
fully developed story, the following correlations can be kept in mind:

Historical Order of Boons Narrative Order of Deaths

1. battlefield sacrifice (day 1)

1. die a great hero ------------------- 2. slain by Alambusa (day 8)
2. see entire war ------------------- 3. of surviving head (day 18)
3. marriage to Mohin

While the first and second boons correlate with the second and third
deaths, the third boon relates to the first death.

Kttn ta var and the Alis

Alis participate in South Arcor Kttn ta var festivals at a number of

villages (Tvanampatta nam near Cuddalore, Pill aiyrkuppam near
Pondicherry, Tiruvetkalam and Kottat t ai near Chidambaram), but
the Kvkkam festival is their main convergence point. They do not,
at least as a group, participate in Kttn t avar festivals elsewhere in
Tamilnadu.6 At Kvkkam, they arrive in large numbersperhaps up

This statement was made on the basis of field inquiry at twenty-nine Kttnt avar
temples, at only one of which, Pulimtu in North Arcot, was there any hint at the
possibility of an exception. Informants at that village were uneasy over what appeared
218 chapter eleven

to a thousand, not only from throughout Tamilnadu but from Tamil

Ali communities in cities all over India, and also from Singapore.
They come, however, only for the climactic events of the eighteen-
day festival, from its fourteenth through sixteenth days. According to
informants at Kvkkam, Alis have participated in the festival there
for many years, but have come in large numbers only for the last
thirty or so, since an Ali named Tvi (Dev) became a great devotee of
Kttnta var there, and spread the word to other Alis (see also Nrull
1990: 41).
Writing in 1906, W. Francis, in his South Arcot Manual, describes
a Kvkkam festival, quite different from todays, as:
one of the most popular feasts with Sudras in the whole district. . . . [It]
occurs in May and for 18 days the Mahabharata is recited by a [Vanni yar],
large numbers of peopleespecially of that casteassembling to hear it
read. On the eighteenth night a wooden image of Kuttandar is taken to
a tope and seated there. This is the signal for the sacrifice of an enor-
mous number of fowls. Everyone who comes brings one or two, and the
number killed runs literally into thousands . . . While this is going on, all
the men who have taken vows to be married to the deity appear before
his image dressed like women, offer to the [Vanniyar] priest a few annas,
and give into his hands the talis [tli: marriage pendant] which they have
brought with them. These the priest, as representative of the god, ties
round their necks. The god is brought back to his shrine that night and
when in front of the building he is hidden by a cloth being held before
him. This symbolizes the sacrifice of Aravan and the men who have just
married to him set up loud lamentations for the death of their hus-
band (Francis 1906:376).
If Francis is describing anything accurately, whether from observation
or hearsay, the festival has changed dramatically. Above all, nothing
is said about Alis. The men who dress like women, wear the tli, and
lament the gods death are presumably from the Vanniyar and other
dra communities that Francis mentions. The climactic rituals occur
on the festivals eighteenth rather than, as now, its sixteenth day. And
there is nothing today that would explain how Kttnt avar could be
seated by a tope.

to be conflicts among the temples headmen and supplied contradictory information.

The group of informants that indicated a presence of Alis had heard of the Kvkkam
festival and gave the impression of thinking I would appreciate a positive answer.
Other informants denials were simple and straightforward. As we shall see, however,
in chapter 14, another exception emerged.
dying before the mahbhrata war 219

Francis also went further to mention that similar vows are taken
and ceremonies performed at the other South Arcot Kttnt avar fes-
tivals, and left it for Whitehead to visit them and retrieve our earliest
account of Aravn becoming Kttn ta var. At Tvanampatt anam, Cud-
dalore Taluk, the Kttnt avar temple pcri (priest), told Whitehead
a story I did not find in circulation there or elsewhere. Indra, having
committed Brahmanicide, became incarnated as Kttn t avar with a
curse upon him that his body should rot away, leaving only the head.
Anyone who married him in the morning would have to cut her tli
in the evening, so no one would do so. Krishna, however, took pity
on him, assumed the female form of Mohin, and consented to be
married to him in the morning, and then, as he vanished all but the
head, the tli was cut in the evening (1921: 27). Whitehead adds that
men dress as women to commemorate this marriage by the sacrifice
of fowls and goats to the god a little distance in front of the shrine,
where there is still today an offering stone on the edge of the beach.
This variant, which either de-Mahbhratizes Kttn t avar or repre-
sents a pre- or non-Mahbhrata-linked version of his cult, is curious
for its connection with Indra (and Brahmanicide) rather than with a
son of Arjuna, Indras son. Mohin resolves the predicament, but with-
out any hint that she is an Ali. Currently Alis are said to come to the
Tvanampatt anam festival, but only in small numbers. Whitehead also
mentions a tradition that Kttn ta var got his name from an Asura,
or Demon, whom he killed, but does not give the story or locate his
source for it. One of the objects of my fieldwork was to learn about
this demons story. I found only two informants who knew it, both
of whom circulated in eastern South Arcot. Ko. Elumalai (1989: 34)
gathered another version at the 1988 Pill aiyrkuppam Kttnta var fes-
tival in the same area. Otherwise, except for the pcri at the Tvanr
temple in western South Arcot, who had heard a version but could not
remember it, I found nothing but likely fragments. This discussion of
the Kvkkam festival will be framed with the two accounts.
I cite first the account by the Pratiyr (Mahbhrata-reciter)
Muttucmi Pilla i, who performs mainly at southern South Arcot
Draupad festivals, but also, as of 1990, had been the Pratiyr at the
Kvkkam Kttn ta var festival for nearly ten years. His version is
important for its linking of Kttn ta vars myth with the Kvkkam
temple and festival.
After the eighteen-day war, Arjuna, Bhma, and Draupad began boasting
about having won the victory. Krsn a listened and said, We cant come
220 chapter eleven

to such conclusions. Lets go to the battlefield to ask Aravn. When

Aravn saw all his fathers and uncles and everyone approaching him, he
was lying on the battlefield with no strength left. Krsn a asked him, Who
won the eighteen-day war? And who was the cause of the battle coming
to an end? Aravn replied: It is not because of my fathers; the victory
didnt come because of them. O lord (ayan), your conch shell held the
heads of those sens (soldiers, enemies, armies), and your cakra (discus)
cut their necks. This is what I saw with my eyes. Bhma got angry: How
could Aravn say this? Hes just saying it because of Krsn a. He raised
his mace. Fearing that Bhma would wound Aravn, Krsn a called Garuda
and ordered him to leave Aravns body in the Carapanka River.
When Aravn was left in this river, a city developed there, which became
very famous, called Centirakiri (Candragiri). Having taken the form of a
child, he started floating down the river. The king of Candragiri, in the
forest hunting, came down to the river to quench his thirst and saw the
floating child. The Candragiri king had never had a child, so he rushed
down to the river and held the child in his arms. The child cried Kuv
Kuv. That is why this (village) is called Kvkkam. The Candragiri Raja
decided to call the child Caraplan [Reed-child: caram, reed + plan,
child, possibly reminiscent of Murukans birth from the caravanam,
thicket of reeds]. Day by day, hour by hour, he grew into a man.
At that time Kttcran made a digvijaya (a conquest of the regions).
The king of Candragiri fought against him. Caraplan saw his father
being wounded by Kttcran, and rushed to the field. He fought so well,
everybody admired it; and he defeated Kttcran. After defeating him,
he came running to the feet of Parakti: You are the supreme mother
of the world, Kanni (maiden) of the world, a good woman (Uttami), Dev
Pcl, the undeflowered Pattini (aliyata pattini: Pattini is the Tamil
goddess of chastity), Goddess of the Five (Aivar tvi). Om akti, Icc
akti, Kriy akti, di akti, Parakti. Now I need a temple (layam) in
this world. I should be able to repose (amara vn tu m) in the temple. I
should be able to give my blessing (arul)7 to my devotees (atiyar). So give
me your blessings (arul potu), Amma. Parakti said, You have killed the
valiant Kttcran; so from now on you will be known as Kttn t avar.
Thus at Kvkkam they built this holy field (tiviya ksettiram) with a big
temple for him.
Clearly, Parakti, the goddess who answers Kttnt avars prayers to
establish his cult at Kvkkam village, is Draupad. The details of the
floating head and its babys cry, kuv, are among the fragments of
the Kttn ta var myth that I have found elsewhere (see chapter 13). But
only in South Arcot does one find them as parts of a myth linked with

On arul as presence, and more often translated as grace, see Shulman 2002
and chapter 14 below.
dying before the mahbhrata war 221

Candragiri. And only at Kvkkam itself did I find a version linking

the myth with Draupad and the origins of the Kvkkam temple.8
These two cults overlap in their South Arcot core areas, and several
villages there have both Draupad and Kttnta var temples with inter-
related festivals (E.g., at Tvanampatta n am, Tailpuram, Madhukarai,
and Pillaiyarkuppam. See also Elumalai 1989: 34, 3845). Kvkkam
has no nearby Draupad temple, and, apart from this myth, which is
recited by an itinerant Pratiyr and not a villager, it seems to have
developed its festival in ways that are less close to Draupad cult pat-
terns than those of other villages. One thus begins to detect significant
contrasts not only in the ways the two cults activate and interpret the
Mahbhrata, but in the ways that the Kvkkam festival is open to
further variations on the Kttnta var cult pattern.
The second account of Kttn ta vars rebirth from Aravns baby-
head will be richer on the matter of the head. But before we examine
it, we must look at the heads treatment in the Kvkkam festival.
It is here that we may look to answer the question of possible influ-
ences of the Alis in reshaping the festival since Franciss time. At the
very least, we can detect certain accommodations to their involvement,
and points within the complex of myth and ritual that would seem
to reflect that involvement intensely. These have to do with Aravns
third boon: his marriage to Mohin.
According to Ali informants interviewed at Kvkkam, Mohinin
this appearanceis Krsn as Ali avatra. (This is, of course, one of the
least widely known myths of Mohin. See Oppert 1893: 508; Dumont
1957: 4012; 1959: 8081; OFlaherty 1973: 2829, also 32, 5051, 228
29; 1975: 27480; Shulman 1980: 30714, for other Mohin myths).
El umalai reports that Alis speak of this avatar as one of mixed gender:
Bhagavan himself, having assumed this guise (vtittu: or costume) in
the situation of a male-female mixture (n pen kalappu), has made this
avatar (1989: 54). r. Nrull further contextualizes the identifica-
tion, quoting indirectly from Tvi, the Ali who inspired other Alis to

El umalais account from Pill aiyrkuppam has no reference to Kvkkam. It places
theological emphasis not on Draupad but iva, and maternal emphasis not on Draupad
but the Candragiri kings wife Kirupaci: When the demon Kttcran drove them
away from their kingdom, they did tapas to iva. When they were gathering water
from a river [unnamed], Aravn, who had died performing kala ppali, having taken
the form of a childs head (ciracu kulantai arukkon tu ), came floating down the river.
Kirupaci took this child (akkulantaiyai etuttu) and raised it. When it became a big
person (periyavan), it fought Kttcran and won (1989: 34).
222 chapter eleven

come to the Kvkkam festival: Krsn a came taking an Ali avatram

(tn ali avatram etuttu vantr); when he married Aravn, Krsn a
came in the form of an Ali (ali uruvil, used twice) (Nrull 1990: 43).
The account claims an almost certainly unfounded northern purn ic
source for the story and for the honor it allegedly bestows on north-
ern Alis, while at the same time lamenting the disrepute of Alis in
the south.9 It identifies the appropriate night for the Alis castration
operation as the full moon night of Cittirai (AprilMay), because the
Mahbhrata says that on that night alone, Krsn a took the Ali avatra
and married the hero (vran) Aravn (61). The term vran in Tamil
folk religion connotes not simply hero, but guardian deity (Hilte-
beitel 1991a: 48, 1078, 11213; Biardeau 1989a: 123, 13132).
Like the Alis, Krsn a has thus undergone a sex change, from male
to female. The Alis affinity with the self-mutilating Kttnt avar are
no doubt transparent and significant, though it must be insisted that
there is never the slightest hint that the penis is one of the thirty-
two bodily parts that he dismembers among his offerings to Kl (cf.
Hiltebeitel 1988a: 327; 1998). As the perfect male being whom the Alis,
impersonating Mohin, marry, Aravn evidently retains his vran mas-
culinity. But the Alis certainly focus on the third boon, which links
the deitys self-mutilation to Kl with his marriage to Mohin. It is,
of course, problematic to ascribe the genesis of particular symbols in
the cult to the Alis. But it is striking that so many of the cults major
symbols, which can be found at festivals where they do not participate
and which probably anteceded their increased participation even at
Kvkkam, have deepened resonances for them. Since I will discuss
this matter, and its psychoanalytic implications, more extensively else-
where (Hiltebeitel 1998; see now chapter 12), a few words must suffice
on the Alis castration operation. According to accounts of Alis inter-
viewed at Kvkkam, when the penis is cut off, the cut piece (tun tu )
and blood are placed in a bucket of water or a new pot. No blood
should touch the ground. The tun tu swells up and bobs around or
quivers (tuti) with a half-life (pti uyir). Later it is taken and bur-

The purn ic story (purn ak katai) about Krsn a taking an Ali avatra is more
popular in the north than in the south. In the south, people consider Alis to have a
vulgar (koccai) birth and to be worthy of ridicule. But in the north, people are said
to worship them as a divinity if they see Alis. This honor (perumai) comes to them
because of Krsn as avatra as an Ali (Nrull 1990: 44; cf. El umalai 1989: 59, with a
similar account of the higher status of north Indian Alis).
dying before the mahbhrata war 223

ied in a small pit. Alternately, again reflecting information from the

Kvkkam festival, Nrull reports: The Alis hold the new pot which
contains the blood with his member, place it on their heads and go
dancing and singing continuously in a procession. Then they go and
release it in the sea (katal) (1990: 63). However momentarily, like
Kttntavars head, the severed penis takes on a new life of its own,
and can be danced in procession.

Kttn ta vars Processions

To attend the Kttnt avar festival today (which I did in both 1982 and
1990) leaves little doubt that the Alis have become part of the cults body
language about Kttnt avars various deaths and births. But perhaps
only a limited part. Among the welter of rites in the festivals eighteen
days, I restrict discussion to those that concern his gradual embodiment
and disembodiment through processions. At Kvkkam, he is taken on
procession in three ways: first, danced as a head; second, his head car-
ried on a platform (ktyam); third, embodied on his full chariot (tr).
The first occurs after the nightfall of the first day, and is repeated
through the first six days. To music and firecrackers, the head is brought
outside the temple and danced through the villages inner rectangle of
streets. Every house offers the touring cmi (deity) a pj of lamp-
waving, coconuts, etc. Goats and fowl were offered up until 1987, but
this was discontinued due to pressures against animal sacrifice from
the Tamilnadu State Government Hindu Religious and Charitable
Endowments (HRCE) Board, whose officers have been administering
the temple since about 1980.
Before turning from the northward street to head east, a passage-
way allows one of the dancers to dance briefly on a lane that extends
further north. This points toward the weeping ground or alukalam,
the site of the rituals that will take place on the sixteenth day, com-
memorating Kttnta vars death, or (more impressionistically) deaths.
After being danced through Kvkkams four main streets, the head
is returned late each night to the temple. Only Vanniyars dance, nine
of them in relay, and only when the Harijan or Paraiyar drum band,
the rja mlam (kings band), plays. The dancers claim they feel no
vcam (possession). But they are in a conventional cmiyti (pos-
sessed god-dancer) role, since they claim that Kttnta var, a great
dancer, makes them dance spectacularly. Their bold strides are said to
recall the style of medieval Vanniyar warriors.
224 chapter eleven

From the thirteenth day, preparations begin for the gods fully
embodied procession. These involve multiple forms of ritualized death
and revival. The chief thirteenth-day rite allows Kttnt avars life to be
transferred from his head into a pot (kalacam) for the period that the
head is repainted in a shed behind the temple. The fourteenth day cen-
ters on the evening rite of kampam niruttal, the standing of the post,
in which the Kvkkam village headman (Ntt nmai) sets the post in
the tr with the HRCE-appointed Tarumakartt (chief temple trustee)
in attendance. Once this occurs, men who vow to marry Aravn may
begin to purchase turmeric-dyed tli strings at two rupees apiece from
HRCE officials under the pantal (temporary roof) extending from the
front of the temple. A piece of turmeric root tied into the tli forms a
pendant that hangs from the neck in place of the usual gold-piece. The
pcri (representing Aravn) then ties the tlis around their necks in
the sanctum. Harijans also sell tlis outside the pantal for fifty paise,
mostly to other Harijans. The stripping and repainting of Aravns
head and the carpentry work on the tr are undertaken by the artisan
caste of cris, who often serve as priests in low caste village rituals.
On the fifteenth day, a festive atmosphere builds up from evening
on, centering on the activities of the Alis, who arrive in increasing
numbers from the fourteenth through sixteenth days. They have year-
by-year arrangements to rent the houses of local villagers on the south-
ern street of Kvkkam for the nights of their stay. Others construct
temporary shelters in nearby fields, or fill the hotels in neighboring
towns, especially Villupuram.
By nighttime, it is a carnival. In 1990 (unlike 1982), a temple ele-
phant was brought in, and a sideshow tent housed what a loudspeaker
repeatedly billed as good clean family entertainment: comedy shows,
disco dancing, and slides of MGRs death. Until well past midnight,
the Alis dominate the scene. They dance and sing, playing kummi, a
stick-dance usually performed by women, and karakam-dancing with a
flower-karakam (punkarakam) said to have Kttnt avars akti. Some
also offer themselves as prostitutes in nearby fields. A rate quoted for
oral sex was ten or twelve rupees, fifty for Americans. Some admit
to finding men attractive (Nrull 1990: 25). Unlike Indian women,
flirtatious Alis are unrestrained in making eye-contact with men in the
crowd and are equally free with more physical overtures.
After midnight, while many revelers sleep on the ground outside
the temple, in a ceremony I was not permitted to attend, the final
touches are applied to the painting of the deitys head. By now it is the
dying before the mahbhrata war 225

sixteenth-day morning. According to an cri assisting in the rite, at

about 3:20 a.m., a Brahman, along with about nine elders, goes to the
temple sanctum for the kalacam that has held the gods uyir (life,
soul) since the thirteenth day. Ten minutes later the vital vessel is
brought into the shed close to the Cmis head. The cri and the
Brahman perform somewhat parallel rites that in effect double the
efficacy of the revitalization of the head and the opening of the gods
holy eyes (cuvmi tirukkan tiruttal). Once life and sight are restored
and a few finishing touches applied, the head can be brought out for
daran (ritual viewing) and procession.
Outside, excitement mounts. Sleepers wake and assemble, jammed
together in front of the shed to await the gods emergence for his
public eye-opening and placement on the portable ktyam. This
palanquin is set about eight meters from the shed entrance, and faces
north, toward which the procession will begin. The Vanniyars kvil
mlam (temple band) circumambulates the temple and goes off to
return with the tiruvcci, the flower decoration, an arched scaffold of
orange and red flowers. It must be carefully erected at the back of the
ktyam, beneath a royal umbrella, to await the placement of the head
in its arc. As the parties position themselves for the climactic moment,
the Vanniyar temple band aligns itself in the eastward passageway
between the shed and the ktyam, while the Paraiyar kings band
stands ahead of the ktyam to precede it north once its procession
begins.10 The Alis are nowhere in evidence at this moment.
Signalled by three volleys of firecrackers, the head is finally brought
out from the shed at about 3:40 a.m. The crowd shouts Govind!
(a cry common in Draupad cult possession scenes; Hiltebeitel 1988a:
326; cf. 1988b: part 1) as it is raised onto the ktyam, on which five
headmen are perched to receive it. The procession then begins. The
Harijan petromax lantern-carriers and rja mlam lead; the Vanniyar
kvil mlam with a relay of karakam carriers in its midst comes next;
and the Cmi on the ktyam comes last in the position of honor
(Hiltebeitel 1991: 44950; Herrenschmidt 1982: 4549). Once the
Cmis head has set forth, the ktyam is flanked by additional plat-
forms, also called tr (chariot), one of which bears Kttnt avars

Of the two bands, one could generalize that, whereas the Vanni yars temple band
serves and subordinates itself only to the deity, the king of his temple-palace (kvil),
the kings band serves and subordinates itself to the king: that is, the Vanniyars
collectively, and in the persons of their headmen, as regional landed dominants.
226 chapter eleven

garland-draped chestplate (mr patakkam) and epaulets (pujam),

which have been brought from the nearby village of Kirimtu, and the
other his flower karakam, which has been danced through the eve-
ning. The mr patakkam is so important that it is said there can be
no festival without it. (cf. Hiltebeitel 1988a: 82; 1991a: 109). This train
then proceeds around the four streets of the village, from about 3:45
till daybreak. As the Cmi sets off, it is said that one can see in his face
that he is happy (cantsam) and smiling.
The climax of the festival begins around 5:30 a.m. Awaiting the
return of the ktyam procession, a vast crowd fills the open area
among the village temples. On top of the Kttnt avar temple, cocks
are offered: since the recent prohibition on animal sacrifice, no longer
cut but thrown up to the temple roof. Finally, the post or kampam
on Kttnta vars tr stands ready for the building of the gods body.
Already in place are the horses, forearms, feet, bells, and two flags
(koticclai) from Nttan hamlet, and straw to fill out the frame. The
tiruvcci and umbrella (kutai), removed from the ktyam, are danced
toward the tr and raised to a support atop the kampam, followed in
similar fashion by the umbrella, from Cevaliyankulam village, and the
winged half-tubular epaulets, raised to shoulder position. While the
head is danced around the tr, the gods shoulders, beneath the epau-
lets, are filled with flowers: strings of jasmine thrown up to the devo-
tees assembling the deity on the body-scaffold. Penultimately, the mr
patakkam is raised to the chest. Last, the head is raised and set on the
poles top. A great boom of firecrackers goes off; 200 kilograms of smoky
camphor is lit in front (east) of the deity; the entire concourse flings
strings of jasmine, spectacularly filling the air before the flowers land on
the gods body as two lakhs worth of garlands (200,000 rupees). The
Alis fling garlands removed from their hair (Nrull 1990: 40). Devotees
on the gods frame gather the flowers onto and into his body.
Aravn thus goes forth in heroic pose, ready to fight, his frame-and-
straw body complete, decorated, draped and filled with flowers. The
giant image is drawn forth as a vran to repeat the march around the
village streets. Now, however, he is said to begin looking sad: happy
as a head, he is sad as embodied. Exhibiting tears and sweat, he is
like Nala, as one informant put it.11 As he prepares for the rites that

The association is that Nala showed sweat and fading garlands, enabling the clever
Damayant to distinguish him from the four deities who showed no such human traces,
dying before the mahbhrata war 227

recall his kalappali and eighth-day fight, his embodied state registers
his readiness for sacrificial suffering and death. While his tr goes
around the village, he takes on an increasingly dead look, and when
he reaches the southern street where the Alis rent, the Alis have a mar-
ried man with them who provides them with a white saree for their
collective widowhood. Here they begin to ritualize Aravns death,
undeterred by the additional deaths to follow.
Around noon, having circuited the village, the chariot turns on
the unpaved lane to the north, taking the route that was opened up
momentarily when the head was danced through the streets on the
opening night. According to some, this point of turning north is what
marks the occasion of Aravns kalappali. This is the moment when
he faces north toward Kuruksetra (Hiltebeitel 1991a: 28994, 30117),
as he does at the Draupad cult patukalam: the ritual battlefield (lit.
place of lying down, of death) equivalent to the Kttn ta var cult
alukalam or weeping ground (20813, 299302). His march from
that point would then represent his entry of Kuruksetra on the wars
eighth day to fight Alambusa. Alternately, others insist that there is
no precise ritual enactment of kalappali, but it is recalled homonymi-
cally as akin to Aravns final sacrifice (katai pali): the dismantling
of his body that will take place at the Kl temple on this evening after
night-fall. The rituals are thus relatively unfixed in relation to the epic
myths, just as the myths are relatively unfixed in relation to the rituals.
Aravns deaths and rebirths multiply, but his kalappali remains the
most universally known mytheme, since it is recalled minimally as the
reason that everyone is wearing tlis to marry him.
The tr now moves in stately fashion through dried paddy fields
toward the weeping ground, in N. Devanayakar Chettiyars handsome
grove. In 1982, before HRCE intervention, cocks were still sacrificed
as votive offerings before the chariots wheels. The widowed Alis, now
(mostly) with hair dishevelled (Nrull 1990: 40), carry on ritualized
womens laments (opprai) for the goddead for them, doomed for
othersand continue their dirges in approaching the alukalam. On
the way, attendants hanging from the giant images arms toss his fad-
ing and browning garlands to those below, indicating that as Aravn

and to choose him as her husband. One wonders whether the comparison extends
to themes of marriage, suffering, and mortality. On sweat and other down-to-earth
bodily substances in the localizations of South Asian folklore, see Ramanujan 1986:
6467; 1993: 103, 114.
228 chapter eleven

Figure 1. Kttnta var on his chariot shorn of jasmine strings, approaching

death and the weeping ground.

approaches the weeping ground he is still losing vigor and vitality (see
figure 1). Alis weep for Aravn because he has performed kalappali. As
Elumalai says, Alis come to Kvkkam mainly for the kalappali, and
the weeping ground is a terrific place for the Ali contingent (1989:
47) (see figure 2). Non-Ali informants, however, indicate no such uni-
formity on the cause for weeping. Some say it is because Aravn cuts
his flesh performing kalappali, while according to others it is because
he is dying (or dead) from his fight with Alambusa. I will return to
such varied motivations in closing.
Most Alis who lament Kttnta var along his northward route go to
the alukalam long in advance of the slow chariot procession. There,
completing their weeping, now widowed, they cut their tli strings and
leave them on top of a bamboo post, the velli kkl (silver leg). They
also break their bangles, leaving the broken bits of colored glass strewn
at the posts base. Most of them then go to a bathing tank, don white
widows saris, and begin to head home. They should wear white for
dying before the mahbhrata war 229

Figure 2. Alis lamenting Kttnt avars battlefield sacrifice as his chariot passes
through the fields toward the weeping ground.
230 chapter eleven

thirty days before they can again don colored sarees, bangles, tilakam
or forehead dot, etc. Most of them are thus gone when the chariot
arrives at the alukalam grove in mid-afternoon.
The rest of the day involves the huge regionally gathered crowd
(around 20,000 people), and then narrows to the entirely local cer-
emonies that concern villages around Kvkkam. The velli kkl rites
continue through mid-afternoon. Rather than the Alis, those who now
cut their tlis and break bangles there are adult males from Kvkkam
and surrounding villages who have vowed to marry Kttn t avar.
Aravn watches these tli-cuttings, something he cannot do for the
Alis, who have left the velli kkl before his tr arrived. Presumably his
last vigil coincides with the gods third death and his second boon
of watching the war to its end. It would thus appear that whereas the
Alis lament Kttnt avar only through the half-death of his first death,
and while the Pratiyr maintains attention on his second, the villagers
undergo their widowing rites at the close of his third and last.
We shall pursue the significance of these distinctions in chapter 13,
but for the moment it might suggest another mythic resonance behind
the villagers widowing rites. On the one hand, the Alis lamenta-
tion and widowing rites focus mythically on Aravns first death, his
kalappali. The coincidence of the ordinary male devotees widowing
with Aravns war-ending vigil could, on the other hand, suggest that
while the villagers marry Aravn, they wait for his third deathat the
end of his eighteen-day vigilto cut their tlis. In waiting till the end
of the war, their widowhood could then be assimilated to the grieving
by the widows (and mothers) of all the heroes slain at Kuruksetra. That
is, for the villagers, the velli kkl rites could be equivalent to the epics
Str Parvan, the postwar Book of the Women and their lamentations.
This is what happens in the Draupad cult, where Aravnsurviving
as a head at the end of his eighteen-day vigilis among the heroes
lamented by Draupad and other women at the patukalam (Hiltebeitel
1991a: 32931). The Draupad cults main core area fire-walk myth
supports this theme, averring that Draupad, at Krsn as behest, led the
Kaurava widows into the fire to end their inauspicious weeping and
turn them into involuntary suttees (Hiltebeitel 1988a: 44042; 1991a:
406; and Shulman 1993 on the theme of filicide). Kvkkam villagers
do not make this latter analogy. Just like the Alis, they don tlis to
marry Aravn, and could thus not be anyone elses widows. But if their
mythic prototype is still Mohin, I have not heard any of them identify
her as Krsn as Ali avatra. Local men at Kvkkam and other villages
dying before the mahbhrata war 231

with Kttnta var temples speak of Krsn a as assuming only a womans

guise (pen vtam), not an Ali avatr. Such accounts have come only
from Alisa matter we shall explore further in chapter 12.
After these rites, a cloth is placed over Kttn t avars head to cover
his dead eyes, and the chariot, now a death house, is drawn from the
alukalam grove into sugarcane fields behind it toward an open-air Kl
temple by a lakeside. There, in ceremonies we shall describe in chapter
13, Kttn ta vars body is disaggregated, his head is revived, and the
head and body parts are then danced back to the their repective vil-
lages of origins as the main event of the seventeenth daythe head
going to the Kttnta var temple in Kvkkam. Finally, on the eigh-
teenth-day morning, the Cmis head is decorated and taken around
Kvkkam only: this time, without dancing. In the evening, it is deco-
rated as Dharmarja (Yudhisth ira) and coronated in the sanctum.

Heads, Bodies, and Ancestors

We can now turn to our second version of South Arcot mythol-

ogy about Aravn becoming reborn as Kttn t avar. This version is
provided by the cirpi (icon sculptor and repairer, working in both
wood and masonry), N. Dandapani, whose native village neighbors
Tailpuram (Tindivanam Taluk, South Arcot), home of a thriving
Kttnt avar temple and festival. A great raconteur, Dandapani weaves
together both the local knowledge of Tailpuram festivals and a wider
regional knowledge that comes from working at both village and Brah-
manical temples all over northern Tamilnadu. His account, treated
more extensively in chapter 13, informs us further on the dynamics of
Kttn ta vars reembodiment from a head.
When the demon Tantavakkiran (Dantavaktra) died, his vi (spirit,
ghost) split into two, going to two different worlds. The two vis prayed
to Nrada to create a child (pilla i) by bringing them back together.
Nrada made the two vis into a small cicu (child) who would be called
Kttcran.12 Nrada asked the child to perform tapas. This attracted iva,
who offered the child a boon. The child requested that it should be born
in a way that no one could defeat him who had a complete body, and
further, that his conqueror could only be another child who should be a
head from a body with thirty-two marks, and, like Kttcran himself,

See chapter 13 for Dandapanis explanation of this name.
232 chapter eleven

be born from water and not from a mothers womb, and have the grace
of god.
Kttcran ruled, torturing the sages. The gods appealed to Mahvisnu
(= Krsn a), who foretold that the child could only be defeated by the
Pn davas child Aravn. With the Mahbhrata war imminent, Aravn
would be sacrificed, and on the eighth day beheaded by Alampcran;
his head would fall into the poykai nti (a river pond), where he would
grow into a man who could defeat the Asura. Aravn then went to war,
did kalappali, and was beheaded on the eighth day. His head fell into the
Carapanka River, where it changed form into a child. He got a new life
(uyir) and the name Kttan, and when he fought and killed Kttcran,
he got the name Kttn t avar.
There is nothing about Candragiri or the founding of the Kvkkam
temple. What takes prominence is the rich compacting of themes
relating birth and death. Both the demon and the deity regenerate
new bodies, probably recapitulating, at least in the case of Aravn, a
transmutation of the funerary ritual of sapin dkaran a, in which rice
balls (pin das) representing the head are projected into full bodies for
the ancestors to occupy in the afterlife.13 In fact, several of Aravns
myths and rituals evoke rites for the dead: his new moon-night bat-
tlefield sacrifice coincides with a time for ancestral offerings, and his
last-minute marriage makes him an eligible, if extraordinary, Pitr, after
whom village children may be named. Aravns head grows a new
body, however, not to become a Pitr, but to reenter the world so that
he can defeat a demon and protect his devotees. However, we know
from Kttcrans boon that Aravns regenerated head cannot turn
into a full body. As a reborn childs head (ciracu kulantai), at the
most he grows only into a child (cicu, kulantai; see n. 8 above). It
is in this happier, floating and dancing form that he kills Kttcran
and receives devotion.

Gender, Caste, Region, and Devotion

One could, of course, go back and forth over many other resonances
of the mutilated body and the severed head. For instance, in discuss-
ing Aravn, Shulman (1980: 12729, 347) and I (1988a: 332) have
invoked the originally Vedic theme of the head of the sacrifice as the

I owe this insight to a suggestion by David Gordon White (Paris, May 1992). On
sapin dkaran a, see Knipe 1977.
dying before the mahbhrata war 233

remainderseed of sacrificial regeneration that can represent the

vied-for goods of life (Heesterman 1993: 4344). Aravn is clearly
offered as an alternative, or surrogate, for Krsn a and for his father
Arjuna, who are usually the only two others who are mentioned as
having the thirty-two bodily marks that would make them accept-
able battlefield sacrifices to Kl.14 His serpent affinities define his low
status in relation to these two alternative but inconceivable victims,
while they also evoke his Nga connections with underworld resources
and powers (cf. Hiltebeitel 1988a: 202; Przyluski 1925). As substitute
for these two, he is the double sacrificial surrogate for the deity and
the ideal king. As the former, he is akin to the sacrificial victim of
the site (vstupurusa) of a temple, with the provisosignificant for
Draupad festivalsthat the site is not an ordinary temple, but the
ritual battlefield as a Kl temple (Hiltebeitel 1991a: 31019). As the
substitute for the king, he reminds one of the condemned regicide in
whose flesh-removal and dismemberment Foucault sees the symmet-
rical, inverted figure of the king: that is, a victim whose disarticulation
defines by its ritual negation the surplus articulation of the kings
two bodies, as analyzed by Ernst Kantorowicz.15 The Kttnt avar cult,
however, and the Kvkkam festival in particular, situates the rampant
symbol of the severed head in a political anatomy that draws short
of requiring any full corporation. It would seem that Kttn t avars
head has the advantage of being freer when noncorporate to traverse
opposed orders of meaning.
First, there is a gendered opposition between Vanniyars and Alis.
On the one hand, the severed head relates to the super-male figure
of Aravn, the great warrior of thirty-two bodily perfections, a vir-
tual Mahpurusa, who offers his head to the goddess as the culmina-
tion of his thirty-two self-mutilations. He is an embodiment of the
Ksatriya ideal that is part of the heritage and identity of his Vanniyar

In Villiputtr lvrs Makpratam (5.7.3), Krsn a tells the Pn davas that he is
the only other choice. Cf. Elumalai 1989: 11, who, possibly following a folk tradition
about Villi, says that in Villi, Krsn a says this directly to Aravn. Some informants add
alya as another possibility, but he is by now on the Kaurava side.
See Foucault 1979: 34 (flesh removal and dismemberment of the attempted
regicide Damiens in 1757), 2829; cf. Kantorowiczs (partially) transposable language
(1957: 36): Bit by bit he [Shakespeares Richard II] deprives his body politic of the
symbols of its dignity; (26165): unilateral sacrifice for the head by the members of
the body politic as though the life itself or the continuity of life rested in the head
alone, and not in head and members together; (363): the corporate character of the
crown; (41936): effigies, juxtaposition of the lugubrious and the triumphal (431).
234 chapter eleven

devotees. This goes back to the head offerings to the goddess that, like
the Aravn myth, represent a warrior ideal of the Pallava period. On
the other hand, the symbol also relates to the involvement of the un-
manned Alis, who marry a male who shares something of their ordeal,
but retains his masculinity. For them, lamenting only Aravns first
death, it is the half-lives and half-deaths of his dancing head, from
which a whole heroic body is reconstituted in their presence and which
they abandon, as widows, while it is still fully embodied, that would
seem to evoke the powerfully remembered half-lives of what they have
Second, similar to the opposition that one finds in the exchange
of heads between princess and Untouchable in myths of the village
goddess Ren uk-Mriyamman (Brubaker 1977; Assayag 1992: 4246,
1023, 14041), is a reflection of the body politic of village castes in
the interplay between Vanniyars and Paraiyars. Whereas Renuk-
Mriyamman has her high-caste head and low-caste body, the situa-
tion with Kttnt avar is just the reverse.16 At Kvkkam, it is said that
his head was found in a field of irki (a cereal) by Paraiyars under a
thorny bush. In a certain sense, the Paraiyars provide (and later care
for) the head while the Vanniyars provide the body: both as the bod-
ies that dance it and as the patrons of the rituals of his bodys yearly
dis- and re-aggregation. Here again full corporation is undesired.
As a symbol of the goods of life, the head is a ripe symbol for every
type of contestation, while the bodies put together with it can only be
provisional. Nonetheless, they replicate the social inequalities of village
life and the temporary heightened cooperation and (always danger-
ous) mixing of castes that occurs at festivals. At other villages, the
tensions differ. At Pill aiyrkuppam, they are between rich Cettiyrs,
whose ancestors found the head, and Vanniyars, who again (figura-
tively) supply the body (Elumalai 1989: 3637). At Kalankiyam, eigh-
teen villages form two groups to pull Aravns body apart in a tug of
war. Alis have no part in these village-level oppositions.
On the regional level, if we return to the myth of the origins of
Kttnta vars Kvkkam temple, we should note its latent opposition
between our two regional Mahbhrata cults and the varying ways

The question of Kttn ta vars relation with Mriyamman requires fuller treat-
ment than is possible in this chapter, to discuss Kttnt avar temples and festivals at
other villages. At Kvkkam, the Kttn ta var festival opens with ceremonies honoring
Mriyamman and the Harijans. See now chapter 13.
dying before the mahbhrata war 235

they activate the epic. Let us recall that it is Draupad Parakti who
grants Kttnta var the residence for his Kvkkam temple. South
Arcot, the only place where Krsn a takes his Mohin avatra to marry
Kttn t avar, is also the region where Draupad takes on her Gingee
avatra to rescue one of the ancient kings of Gingee (Hiltebeitel 1988a:
65; 1991a: 48289). South Arcot Vanniyars, a landed dominant caste
through our main area, would seem to have linked their martial ideol-
ogy in both stories, through the Mahbhrata and the lunar (Candra
or Soma) dynasty (note the name Candragiri), with the histories and
mythologies of important regional fortified kingdoms (Narayana Rao
1986: 14243). Candragiri, in Chittoor District of Southern Andhra
Pradesh, about twenty-five miles southwest of Tirupati, is, like Gin-
gee, not only in the larger core area of the Draupad cult, but tied in
with that areas regional history during what may be the formative
period of both cults in the Vijayanagar period. There are differences,
however, in the ways these two fortified capitals seem to be used as
symbols. Draupads Gingee avatra is linked with the virtual found-
ing and ongoing protection of the Gingee kingdom, however mythical
it has become. Kttn ta vars link with Candragiri is, at least in these
myths, only episodic, and ultimately no more than a prelude to his
obtaining Draupads blessing for the foundingin the Gingee area
she protectsof his Kvkkam temple. If we can risk imputing his-
torical reminiscences to these symbols, Gingee, even though it was
ruled by Telugu Nayaks, would seem to denote the Tamil heartland
of the regions power as one that developed from local regional tradi-
tions and specifically the warrior peasant traditions of Vanniyars. On
the contrary, Candragiri, at the northern periphery of this area, would
recall only the residue of Vijayanagar imperial power in its period of
retreat to that fort after the Muslim takeover of the Vijayanagar capital
at Hampi. Gingee represents the beginnings of the regions medieval
tradition of regional Hindu little kingship and resistance to northern
Islamic sway (Hiltebeitel 1988a: 1723), whereas Candragiri marks its
end. Gingee, through Draupad, links Vanniyars (and related castes)
with the full martial ideology of the Mahbhrata and with rituals and
dramas that recall all its heroes and reenact the entire war. Candra-
giri, through Kttnta var, links Vanniyars (and related castes) with a
depleted and failed martial tradition that singles out one hero only and
his multiple miracles of revival. Although both cults are Mahbhrata
cults, with eighteen-day festivals typical of each, the former is thus
more focused on martial themes than the latter. At the present stage
236 chapter eleven

of this research, it may thus be hypothesized that the Draupad cult

represents the initial core area activation of the Mahbhrata during
the heightened Hindu revivalism of the early Nayak period; that the
Kttnta var cult registers the depletion of the martial ideology that
supported that revivalism but retains in concentrated form the regen-
erative theme within it; and that the Alis who have infused themselves
into the Kttn ta var cult have appropriated this latter theme in a dis-
tinctively modern and emerging postmodern context.
Let us then reflect further on the increasing attraction of trans-
sexuals to Kttnta var festivals, and particularly to Kvkkam. Here
we must extend our first opposition between Alis and Vanniyars to a
more general one between Alis who come from afar and villagers who
are local. For both, the impersonation of Krsn a-Mohin is important
and compelling because it is some kind of bhakti. But for neither is it
the type that Shulman attributes to Kttn t avars myths and festivals,
in which, he says, the soul pictures itself as a woman longing for her
lover, or identifies himself with the earthly bride who seeks union
with the divine (1980: 306, 307). As Shulman says, The major dis-
tinction in the cult is that the worshipers here direct their rites to the
victim rather than the Bride who claims his life. How then can the
transvestite worshiper . . . be assimilated to the goddess who is mar-
ried, or yearns to be married, to this deity? (307). One hears nothing
of Krsn as yearning. The identification with Krsn a-Mohin is not one
of love in separation, or of identification with the feminine in order
to worship a male divinity. To hazard a formulation, the bhakti must
go two ways: a love not only of Aravn, through Krsn a as bride (and
also Kl as mother), but of Krsn a (and Kl), through Aravn. In this
respect, it is a bhakti in which both the villagers and the Alis iden-
tify themselves, however differently, with the god [Krsn a] as well as
with the devotee [Aravn], so that they can bring forth the devotees
unique, but also multiple, experience of the god.
For local area villagers, whose vow to marry Kttn t avar for a
day encompasses all three of his deaths, their temporary transvestit-
ism retains the minimum of its martial epic setting, and the Krsn a
whom they personify retains something of his full and devious epic
range of character. They identify with a god who takes a womans
guise to lure Kttn ta var into marriage for a day so as to guarantee
his self-mutilating death; who makes possible the serpentine means
of his continued life, but also dictates its winged withdrawal so as to
guarantee his heroic death in battle; and who grants vision to his sur-
dying before the mahbhrata war 237

viving head so that he can see the war to its end. What Kttn t avar
sees while villagers cut their tlis and break their bangles before him
is the completion of their one-day marriage with rites of widowhood
on the ritual battlefield. Their daran of Kttn t avars dying head is
thus a seeing through his eyes of his impeccable folk bhakti vision of
Kuruksetraa kind of vran bhakti bestowed by Krsn a himselfthat
Aravn describes when he is asked what he saw: Ayan, your conch
shell held the heads of those sens, and your cakra cut their necks.
This is what I saw with my eyes.17 What Aravn sees is in fact these
villager-devotees, who in completing their vows to marry him in their
womens guise, have, in their identification with Krsn a, enabled the
war that he sees, and the vision it supplies, to take place. He is their
loving and devoted sacrificial victim. From him, and their vows to
him, they obtain whatever goods of the world they have sought: health,
good crops, good business, good work, educational enhancement, sex-
ual fertility, and other desiderata. In the death rites that follow, their
husband vran becomes an ancestor; and through the revival rites with
Kl, he returns to protect their villages.
For the Alis, things are different (see figure 3). If one asks them why
they come to Kvkkam, and especially why they have come recently
in such increasing numbers, one most often hears three things: that
the increase began with the first visit of Tvi; that, since then, it has
become a great convergence-point for Alis to renew ties, have fun, and
air grievances; and that the deity Kttnta var there, and the place itself,
have great power (akti) for them. These matters are complex. We get
some sense of Kvkkams attraction to the Alis from Nrulls inter-
view with Tvi. Although one cannot tell how much of this account
is from his interview and how much is from Kvkkam informants
recollections about her visit, it shows signs, in either case, of becoming
a piece of Ali mythology about the place.
Tvi heard of Kvkkam and its rites by word of mouth. She arrived
on the full moon of Cittirai at Villupuram, went from there by train to
Parikkal, and walked the last seven miles on foot, with the sun beating
down in the middle of the road, to Kvkkam. There she collapsed on
the pyol in front of the house of a barber named Rju, who offered her
hospitality with great affection. That house, which was then a shack, is

In one form or another, this saying circulates widely in both Draupad and
Kttnt avar cult contexts. Sometimes Aravn reports that he has also seen Kls kapla,
skull-bowl, collecting all the blood (Shideler 1987: 186; Hiltebeitel 1988a: 445).
238 chapter eleven

Figure 3. An Ali at the weeping ground, hair dishevelled, impersonating Krs nas
Ali avatra as Mohin with a flute.

now made into a building. And if Tvi comes to Kvkkam today, it

is in this house that she stays. When she came for the very first time
to Kvkkam, Tvi prayed to Cmi with fear and devotion (payapatti;
Winslow: reverential awe). She took part in the rites. She did all the
necessary things. Within the next year, she had become a millionaire
(lakstipati). She believed that it was by the grace of Kttnt avar that she
obtained this wealth. Therefore, the second year, she came to the festival
with fifteen people. The fourth year, four hundred Alis came. Tvi is
the mother of Shanta, who is from Bombay. Through this, Alis from
Bombay also began to come to Kvkkam. . . . [Now] thousands of Alis
tie their tlis at Kvkkam village (Nrull 1990: 4142).
One finds here the initiatory journey, the hospitality from the villag-
ers that becomes proverbial, the fear and devotion that comes with
participating in the rites, the very worldly goods of life that flow both
to Tvi and the barber, and a suggestion that Alis nourish family rela-
tions of their own at the Kvkkam festival. Clearly, Ali families are
different from those of villagers. The mother-daughter relation
dying before the mahbhrata war 239

referred to in Tvis story is one of adoption, and entails a guru-chela

(disciple) relationship as well (Nrull 1990: 1523). Narull men-
tions that Tvis first impression was that she felt that the festival
was like [made for?] her own in am (kindred, kind, species, gender)
(tan atu in attaippn ra vilvka iruppatka avar unarntr; 41). This
statement can be read as implying that she felt that her own kind,
the Alis, were like the Kvkkam festival. Perpetuating their in am
is a basic Ali concern (15, 34, 4748, 55). Elumalai, using the same
term, reports that Alis who come from afar (Kerala, Mysore, Delhi,
Gujarat, Bombay, Calcutta, etc.) may be seen at the Kvkkam festival
in their home-based groups (kulu), moving under their own big Ali
(periya ali) leaders with a group mentality that lends them protec-
tion. Having no other opportunity to enlarge their in am, they say that
they protect newcomers this way, and nurture their in am by changing
them into Alis (alikal ka mrruvatu) and raising them (1989: 57). The
festival thus provides a context in which Alis who have left Tamilnadu
converge to replenish their in am and nurture its continuity. Aravn
cannot be an ancestor, but his festival, and specifically his sacrifice, like
that of the Alis, helps the Alis sustain their species.
Here also, receiving local attention and even state and national press
coverage (Nrull 1990: 9; Bhaskar 1989; Shetty and Singh 1990), Alis
find a rallying point to speak out on modern political issues at vil-
lage, state, and national levels.18 Many being from cities, they com-
plain of the facilities for bathing, changing clothes, and sleeping at
Kvkkam: We have petitioned the authorities . . . and the people at
the temple, but there is no evidence that anybody has ever listened. So
we are not being respected by the society we were born into, by our
family, or by the government (Elumalai 1989: 58). As we have seen,
the local authorities now include the state government HRCE Board-
appointed temple trustees. The second biggest Kttnt avar festival at
Pill aiyrkuppam, which attracts mainly Madras Alis (63), serves as an
Ali convention center on Tamilnadu state issues. And with evident
whimsy, Ali informants at Kvkkam address national concerns: The
government should help us according to the law. They must give us a
substantial portion of the funds that are earmarked for family planning.

It is fitting to add, as a new note here, a recommendation that readers simply
Google the names Kuvakkam and Kuttantavar to see how Alis, who since this essay
was written have come to prefer the name Aravnis, gain yearly attention through
multimedia coverage of the Kvkkam festival.
240 chapter eleven

We cant give birth to children like other women, [but] having [femi-
nine] physical parts, we provide a way of outlet for the [sexual] feel-
ings of men. In this way we are covertly helping the family planning
program of the government (5758; cf. Narull 1990: 69).
It is an obligatory custom (katt ya marapkum), Elumalai learned,
that on the night before kalappali, Alis have physical relations with
somebody (1989: 65). The revelries of this climactic night, when they
have fun with men in the fields and join in dancing Kttnt avars akti
karakam, also coincide with what many Alis take to be the consumma-
tion of Mohin and Kttnt avars marriage. (All agree, including local
people, that this night celebrates the marriage, but some say it is not
consummated.) One of them, interviewed amid this fiesta, supplied a
myth and an invitation to my assistant: What happens after the mar-
riage? Because Vramankli [Kl] was demanding Aravns sacrifice,
Krsn a Paramtm [as Mohin] said, If you dont allow me to consum-
mate my marriage, youll have the same situation and wont be able to
have fun with your man. So Kl thinks, Oh, Im also a woman. She
should have fun. Thats why every year they allow Krsn a and Aravn
to have sex. Did they have a child? No. Well, maybe the sperm didnt
go to the right place. There could be a child. Lets go do that . . . .
All in all, then, Kvkkam remains central to the cult for its akti and
bhakti, and for the hospitality its villagers extend to the Alis. Rejecting
the Aliphobic attitudes of locals from other villages (Elumalai 1989:
5859), the Panchayat President N. Periyacmi, an astute and presti-
gious patron of the Kvkkam festival, puts it as follows in an inter-
view with Elumalai:
Poor Alis (pvam alikkal). Only here in Tamilnadu are they not respected.
The troubles that they have here are completely absent in the north (cf.
n. 9). On the contrary, there they have a good income. Its a mistake to
think that theyre walking on a wrong path. The mistakes that a few make
are attributed to all of them. In truth, the Alis are people who are full of
bhakti (pakti niraintavarkal). We conduct family and agricultural lives.
Even among us you cannot say we are bigger bhaktas. And even then
the bhakti they have for Aravn is indescribable (aravnmitu pakti . . . col-
limutiytu) . . . They dont come to Kvkkam at their own expense to do
unsavory things that they could just as well do in Bombay and Calcutta.
Are they crazy to come from places so far away to light camphor in the
festival? (1989: 59).
Let us then close on what is distinctive about this fearfully rever-
ent and indescribable bhakti of the Alis. Strictly speaking, unlike
dying before the mahbhrata war 241

local villagers, Alis do not make temporary vows to marry Aravn.

Their marriage to him is a permanent condition. As Elumalai was told,
Every Ali thinks of Aravn as husband (aravn kan avan) (1989: 54).
They identify with Krsn a not in a temporary womans guise (vtam),
but, through a condition that they have made permanent themselves,
with his Ali avatra; thus their extended widowhood of thirty days.
In focusing only on Aravns first death, his kalappali, they affirm
their transformation both by identifying with this form of Krsn a, who
appears as such only in this episode, and, as I have tried to suggest,
by worshiping a hero whose self-sacrifice is in so many ways similar
to and, to them, as heroic and sad as, their own. Their worship of
Kttnt avar allows them to ritually reenact a kind of marriage between
their present and former selves, and, in relation to Kttnt avars story,
to participate in an initiatory ritual that affirms such a marriage col-
lectively so that their inam may flourish.



Two Tamil folk cults rooted in the Indian epic mythology of the
Mahbhrata may be contrasted in their treatments of hair.1 One is the
cult of Draupad, the epics chief heroine and wife of the five Pn dava
brothers, the epics main heroes. The Kttnta var cultthis chapters
main focusis that of a minor epic figure named Aravn (in Tamil)
or Irvat (in Sanskrit), and called Kttn t avar by his worshipers: the
son of the middle Pndava Arjuna and a snake princess named Ulp
or Ngakkan n (Serpent maiden), with whom Arjuna has a brief affair
sealed by an estranged marriage. Each is worshiped as an adjunct fig-
ure in the others cult. Draupad temples are found throughout Tamil-
nadu and up into southern Andhra Pradesh and western Karnataka;
Kttn t avar temples only in a belt across north-central Tamilnadu
from South Arcot to Coimbatore.2 But both cults are concentrated in
the districts of South and North Arcot.
The Draupad cult builds upon a pan-Indian popular tradition
that is not explicit in the Sanskrit Mahbhrata, but already alluded to
there (Hiltebeitel 1981: 186201): the menstruating Draupad vowed
after she was abused at the pivotal epic dice match by the Pndavas
rivals, the Kauravas, that she would keep her hair disheveled until
she could dress it with blood from the Kaurava king Duryodhanas
thigh.3 These images are all traceable to a classical-medieval-folk
continuum within the public culture of India, and decipherable in the
context of a code of womens hair styles. Loose hair or dishevelment
is for widows and menstruating women; the triple braid (triven ) and

This chapter appeared in a book on hair in Asian cultures (Hiltebeitel and Miller
1998), which followed up a 1993 Association of Asian Studies panel I and Barbara
D. Miller co-chaired on Hair in Asia.
See map in chapter 11 and maps cited in chapter 11 n. 3. The Kttnt avar cult
belt is wider than this indicates here, extending to Dharmapuri District. For help with
various Tamil sources used in this chapter, I would like to thank Abby Ziffren, Pon
Kothandaraman, J. Rajasekaran, and S. Ravindran.
For variants with illustrative photographs, see Hiltebeitel 1988a and 1991a. See
also Hiltebeitel 1981.
244 chapter twelve

chignon (Tamil kntal or kon ta i) are for auspicious married women;

the single braid (ekaven ), a kind of pony tail, is for women sep-
arated from their husbands (virahins). These conventions can be
reconstructed from classical Sanskrit and Tamil texts (Hiltebeitel 1981:
18485; 198081; 1985a). Even Draupads gruesome demands for a rib
comb and a garland or hair ribbon of guts have counterparts in public
rituals like Dasar, the fall festival for Durg, or the iconography of
Kl (Hiltebeitel 1991a: 388, 436). The deep motivations that shape
Draupads hair ritualsthe thirteen-year extension of her menstrual
purification terminated by her use of womblike blood from a male
thigh (1991a: 39698)do not seem to require elaborate interpretative
routes through psychoanalytic theory. When I observed that no one in
the night-long dramas performed at Draupad festivals ever mentions
that Draupad is menstruating when she makes her vow, and asked
my chief actor informants why not, suspecting Id found a suppressed
or unconscious theme, I was told that there was no need to mention
it: Everybody knows that (1988a: 23435). When women and men
identify with Draupad in rituals of possession and vow taking, even
where it includes loosening, letting fly, and binding of hair, there is
thus no reason to think that the hair symbol is personally motivated.
It is, at least predominantly, in Obeyesekeres terms, a psychoge-
netic public symbol rather than a deeply transformative personal one
(1981: 1318).
The Kttn t avar cult is different. It too has pan-Indian epic and
ritual associations, but they are much harder to find, and have left no
early echoes in the Sanskrit Mahbhrata. Rather, our oldest sources
are from a distinctive Tamil Mahbhrata tradition. Kttn t avar tem-
ples are also more isolated from each other. Cult ritual and mythology
vary so much that there is probably not much of which anyone can
say, Everybody knows that. Certain hair-related themes have wide
currencyfor instance snakelike hair. But on the level of field observa-
tion, the details are little if ever explicitly in evidence, and for the most
part must be discerned in a network of associations. There is no one
main metaphoric hair theme, but bits and pieces, deployed metonymi-
cally from one locale to another.
In South Arcot, however, there is a distinctive participation in the
cult by Alis, Tamil-speaking transsexuals or so-called eunuchs, the
wider Indian (Hindi) name for whom is hijr. Here one may suspect
deep motivations in appropriations of the private or personal charac-
hair like snakes and mustached brides 245

ter of the cults mythic and ritual symbolism, which begins with the
item that leads every Freudian mythologists seemingly endless list of
castration displacements: a severed head. I have not done in-depth
interviews of Alis that would allow me to probe the personal character
of this symbolism as far as I would like. But I have interviewed enough
Alis and villagers to probe two questions, each connected with hair:
What happens on the level of personal participation in a cult where
a presumed conventional symbolism of displaced castration is shared
with Alis who have really been castrated? And, to the extent that
we can trace Alis contribution to the cults symbolism, what is that

Hair that Stands on End

Aravn is specially suited for self-sacrifice before battle because he

has thirty-two beautiful body marks. But one trait stands out among
them. One first meets it in late-nineteenth-century folkloric texts: in
one, Duryodhana chooses Aravn because he needs a person for bat-
tlefield sacrifice (kalappali) who has hair that stands on end (etir
urmakkrarkavum); in the other, Krsn a says that Aravn is the
ideal victim for sacrifice to Kl because he is handsome, truthful,
observant of customs, and has hair that stands on end (etirrman ).4
What is etirrman ? Here we face our hermeneutical divide. Do we limit
ourselves to emic interpretations specific to what people say about fes-
tivals they participate in themselves? Do we admit information from
wider Kttn t avar cult circles? Do we consider Draupad cult associa-
tions, since the two cults overlap? Do we look at other cults, some also
with regional overlap, others more distant, with similar usages? Do
we allow interpretative routes through other regional Mahbhrata
folklores that appear to be variations on this one? This essay makes a
case for all these lines of inquiry. Panregionally, much of the mythol-
ogy that Tamils associate with Aravn is associated with a virtual dou-
ble of Aravn in other areas of India. And regionally, participants at a

Respectively, Can mukakkaviryar 1969: 2, Uttiyka Paruvam, 199 (a prose ver-
sion of the eighteenth-century Tamil Mahbhrata of Nallppilla i), and the folk drama
Aravn Katapali Ntakam (1977: 40; Hiltebeitel 1988a: 321). Can mukakkaviryar,
active around 1890, was probably familiar with both cults, or at least their folklore
(Hiltebeitel 1988a: 172). The drama is traceable to 1897 (ibid., 160).
246 chapter twelve

particular festival include not just local villagers but performers who
have intercultic and wider regional knowledgeability, and in our pres-
ent case Alis who come from far and wide. But if one looks beyond
participants explanations, one must have reasons for positing some
kind of selective memory or interpretative principle, which in some
cases will amount to a misrecognition or mconnaissance. In the
case of Kttnta var cult hair symbolism, participants explanations
seem to be most unknowing precisely on matters where informa-
tion about similar usages from the other contexts open possibilities
of psychoanalytic interpretation. If we straddle this divide, we must
distinguish between what people say about festivals they participate in
themselves, and what we draw from wider associations.
According to informants at his main cult center at Kvkkam,
South Arcot, Aravns etirrman is a mole with hair growth. The hair
surrounds the mole on the right shoulder and stands on end. It is an
auspicious beauty mark, as are all of Aravns thirty-two marks. Only
a few other warriors have these features: Arjuna, Krsn a, and a Kaurava
ally named alya. But these three are ruled out as victim for one reason
or another.5 Granted the spotty nature of our information, it is striking
that Draupad festival informants give different explanations. Locals
at a Tindivanam Draupad festival in South Arcot said that etirrman
denotes that Aravn is too hairy, which makes him an ideal victim
because his abundant body hair will stand on end at the thrill of
self-sacrifice. This trait would seem to evoke the requirement in south
Indian animal sacrifices that the victim must shiver before it can be
beheadedas provoked by pouring water on its hair. On the other
hand, in the folklore of Draupad cult dance dramas or terukkttus,
Arjuna and the god iva each have a fish-shaped mark (maccarkai)
on their right sides, and ivas wife Prvat and a Brahman woman
named Krp have them on their right thighs, which fills iva with lust
for both of them. These moles are erotic beauty marks (Hiltebeitel
1988a: 19193, 2089, 308).
I suggest that Kvkkam informants reject the animalian and
demonic implications of excessive hair and self-sacrificial thrill, and
draw upon the beauty-mark tradition to interpret Kttnta vars
etirrman as a sign of their deitys handsomeness, beauty, and even

Information from Kvkkam July 1994, especially from Ganesan, age fifty-two.
hair like snakes and mustached brides 247

erotic appeal. If Kttnta var is often portrayed with large canine teeth
(he does not have them at Kvkkam), he is still not the grotesque
with demonic facial features that he often is as Aravn in Draupad
cult iconography (see plates in Hiltebeitel 1988a: 249; 1991a: 25660).
I propose that Kvkkam informants restrict or condense hair that
stands on end to a beauty mark, and sublimate rather than eliminate
the animalian and demonic associations of excessive and bristling hair,
which reappears in other forms.
I will thus at times see hair where Kvkkam informants see other
things. The hypertrichosis of the Mahbhratas prewar victim, it
turns out, is a secure folkloric motif that cannot be left as a mole, or
left to Aravn . In Telugu, Hindi, and Rajasthani traditions, we meet
his double: Barbark/Barbareeka, a son of either Ghatotkaca (a son of
Bhma, the strongest and second oldest Pn dava, and the wild demo-
ness [Rksas] Hidimb), or of Bhma himself and another Serpent-
maiden (Ngakany). In either case, his hairiness marks him as a
typically othered missing link. In Sanskrit, barbarkafrom barbara,
barbarianmeans curly hair or a particular mode of wearing the
hair (Monier-Williams 1899: 722).
In Andhra folklore, where Barbareeka is Ghatotkacas son, in return
for his head offering to Kl, Krsn a grants him the boon of being able
to see the entire eighteen-day war from a mountain top with the sev-
ered head (Subba Rao 1976: 27273; Hiltebeitel 1988a: 317). One learns
nothing about his hair from Subba Raos brief account, but S. Nagaraj of
the Centre for Folk Culture Studies, University of Hyderabad, tells me
that it is mainly a question of head hair: to all those below the moun-
tain top, its spread looks as vast as a cloud (personal communication,
January 1995). In Rajasthan, his prewar sacrifice is known in popular
Mahbhrata traditions (Dominique-Sila Khan, personal communi-
cation, 1993). According to a Hindi book on Kt Shym, the main
village site near Jaipur of a regional Mahbhrata folk cult, Barbark,
as Ghatotkacas son, had at his birth the hair naturally disheveled and
standing on end, that is why he was given the name barbareeka. In
certain areas of Rajasthan, his head is worshiped during the fall festival
to the goddess (Navartra or Durg Pj) and immersed in a tank or
well on its tenth day of victory (Vijaydaam) that celebrates the
triumph of the goddess.
It is probably not incidental that in one version of the drama on
Aravn s Battlefield Sacrifice, Yudhist hira (the eldest Pndava and
248 chapter twelve

king among them), links Aravns battlefield sacrifice with the hon-
oring of weapons (yudha pj), a subrite of the same fall festival
of Dasar (which means tenth day). There would seem to be a link
between Aravn s horripilation and the raising up of weapons for war
by those who sacrifice him. Dasar is also a festival at which buffaloes
can be sacrificed to the goddess. Accordingly, Yudhisth ira asks Krsn a
whether it might not suffice to offer a wild buffalo, elephant, boar,
horse, cock, sheep, or deer instead of Aravn (Hiltebeitel 1988a: 321;
Aravn Katapali Ntakam 1977: 40). Here the low-status victim-hero
is clearly assimilated to animal victims: especially maned ones that are
supposed to shiver to signify their sacrificial consent, and ones that are
hairy and tusked (Hiltebeitel 1988a: 327).
Aravns double, Barbark, is also found in parts of Uttar Pradesh,6
as well as in Madhya Pradesh at Mahoba, where I found that he is the
son of Bhma and the Serpent-maiden and has the distinctive trait
that the hair of Babrk is like a snake.7 I have not heard this said of
Aravn, but it must be considered admissible evidence in interpret-
ing his hair, even if it is gathered far from Tamilnadu. Although they
have different fathers from the Pndava faction, both Barbark and
Aravn, at least in this version, have a Serpent Maiden (Nakakkan ni,
Ngakany) for their mother.8 Yet here we have an obvious problem.
In all cases, one can see how paternal genes could contribute to the
heros hairiness: from Ghatotkaca, low Rksasa genes; from Bhma
the same, with his Rksasic disposition; and from Arjuna, the famous
trait that his body trembles and hair stands on end when he overlooks
the two armies drawn up for the battle of Kuruksetra (Bhagavad Gt
1.29).9 Like father, like son, mutatis mutandis. Indeed, both Arjuna
and Aravn s hair stands on end at the beginning of the epics sacri-
fice of battle: a trope for the Mahbhrata itself, since the war and the
battlefield are repeatedly described as lomaharsan a, making the hair
bristle or thrill, and one of the epics bards is named Lomaharsan a,
from the hair-raising nature of the tale. It is more difficult to see

Khan, personal communication, 1993; William Sax, personal communication,
Babrk ke kesh ng-saman th: as described in December 1995 by Baccha Singh
of Mahoba, a celebrated singer of the Hindi folk epic lh that is linked with the
The sources identifying Ghatotkaca as the father do not mention the mother.
I have not, however, found that Arjuna has etirrman in this context.
hair like snakes and mustached brides 249

how Aravn or Barbark would derive his hairiness from an ances-

try of hairless snakes. Yet clearly they have. Perhaps we may say that
whereas paternal genes are dominant, maternal genes are recessive.
We will thus be justified in looking for additional folkloric connec-
tions between snakes and hair.10
Further exploration of Aravns hair must take us into the unfold-
ing of his boons. First and oldest is the boon that enables him to die a
hero on the wars eighth day in his fight with the Rksasa Alambusa.
Though there is nothing explicitly about hair in any account of this
story, there is in all accounts a major role for snakes, who seek to pro-
tect their descendant, the Serpent Maidens son. This is a theme that
receives rich elaboration in folkloric accounts, and deserves consider-
ation as a bodily, and specifically hair-related, symbol. In the account
of the temple icon sculptor N. Dandapani, a well-traveled expert on
Kttn ta var and Draupad cult folklores, Aravn has been reduced to
a skeleton in the eight days since his sacrifice. How can he best use
the fifteen minutes Krsn a gives him to fight? He calls on his serpent
mother, who implores her father Nkarjan , King of Snakes, to give
her son strength:
So you know all those cobras, snakes, different sizes, suiting different
parts of the bodyfingers, arms, thighsthey came and wound around
the different parts of Aravns body. His skeleton became stiff (viraippu)
now with all the snakes wound around him. He finished off one army
division completely in an hour and a half. Krsn a thought, What is this?
We wont have anything left for others! . . . So he did another trick, and
called Garuda to appear in the sky. When even the shadow of an eagle
falls, the cobras will disappear. . . . While Aravn was losing all his strength
with the snakes falling off, Alampucan [Alambusa] cut off his head.
Here the hair-snakes combination relates specifically to an alternating
stiffness and droopiness in Aravns combat persona.

Mustached Brides and Masquerades

Aravns second and third boons also involve his hair. The second,
his war vigil, is ritualized most prominently in the Draupad cult, in
which his clay or wooden head overlooks the cults ritual battlefield

The Mahbhrata opens with the story of Kadr, mother of snakes, who orders
her brood to hide in the tail of the horse Uccaihravas (1.18).
250 chapter twelve

(patukal am) (Hiltebeitel 1991a: 287302). In the Kttn t avar cult, it is

treated more variously, and usually less conspicuously, but always in
conjunction with his third boon of marriage to avoid the postmortem
penalties of bachelorhood. As we saw in chapter 11 while describing
festivals in South Arcot, whereas local villagers highlight the comple-
tion of the war vigil, Alis focus entirely on the third boon, Aravn s
marriage, and leave the rituals recalling the wars endwith the final
death and revival of Aravns headentirely to villagers.
In North Arcot, where Alis, at least as a group, do not participate in
the Kttn ta var cult, the combination of the third and second boons
presents us with a mustached bride. At Putr village (Vellore Taluk),
before dawn on the main night of his three-day festival, Kttn ta vars
head is placed on a swing (ucil) along with the head of Krsn a the
eunuch (Kan n an pti). This night has four main phases. First, around
9 P.M., there is a goat sacrifice (by some, identified as Aravn s battle-
field sacrifice) and Kttnta var possesses his oracles. Second, when
the two heads exit the temple, they are placed on a double-bullock
cart chariot and taken on procession late into the night. Household-
ers make offeringsmostly vegetarian, some additional goatsas the
chariot stops before their homes, and worship Krsn a the eunuch
and Kttn ta var because, like partners, they created the whole
Mahbhrata war. For Krsn a, the war is like play (oru vilaiytu
mtiri). Third, before sunrise, the two heads swing from a tall stone
cross-beam, and are worshiped for their favor of having conducted
the whole war like a dance (or game: kttu, as in Kttnt avar, Lord
of the Dance) and brought about justice in the end. If the sun should
touch Kttn ta var, it would bring diarrhea, cholera, and other dis-
eases. Swinging makes the pair happy, as at a marriage and relaxes
them, since they must be tired after conducting the war. But the
village has performed no actual marriage for them for a hundred
years, since it would cost ten thousand rupees. Male villagers, young
and old, are supposed to come (none did in 1990) before the swing-
ing heads to press swords against their chests (see Hiltebeitel 1991a:
829, 351523). Before daybreak, the two heads must be returned to
the temple.
Clearly, the swinging ceremony suggests a condensation of
Kttnta var and Krsn a the eunuchs wedding. But the procession
is no ordinary wedding march. The goat sacrifices and war allusions
evoke a sacrifice of battle, and the newlyweds reach the swing as dis-
embodied heads. Rather than enjoying a marriage, they seem to dance
hair like snakes and mustached brides 251

their way through the war (in the procession) and then watch it (the
sword feats) from the swing. Other North Arcot Kttnt avar festivals
have similar swinging rites for Kttnta vars head alone. At two such
temples, the head swings facing the Mahbhrata reciter (Pratiyr)
for the eighteen days (equivalent to the war) that he sings the epic.11
Where Kttn t avars head is alone, the swinging thus refers to the
war vigil only; where he is with Krsn a the Eunuch (or Krsn a the
Woman), it seems to represent the vigil and marriage together.
What is important here is that both of the swinging heads at Putr
have mustaches (see figure 4). Despite the apparently female form he
assumes to marry Kttn t avar, Krsn a the Eunuch has a mustache.
One finds a similar pair of heads at another North Arcot Kttn t avar
temple, but there the mustached female is called (at least by the
priest) Krsn a the Woman (Kan n an pen ).12 This distinction may be

Figure 4. Krsn a the eunuch with mustache on left, and with Kttnt avar on
right, at Putr. Aiyanr, son of iva and Visn u transformed into Mohin, is
represented on the poster in the middle. Photo by Alf Hiltebeitel.

At Crpappattu (Chengam Taluk) and Tvan ur (Tiruvannamalai Taluk).
At Vrn antal, Chengam Taluk.
252 chapter twelve

significant. The eunuch identification could reflect awareness of the

relatively recent publicity about Alis at Kttnt avar festivals in South
Arcot, particularly Kvkkam. But Thurston and Rangachari indicate
a comparable role eunuchs might have played at swing ceremonies in
small-scale south Indian village festivals: At the annual festival of the
Gadabas of Vizagapatam, thorns are set on a swing outside the shrine
of the goddess. On these the priest or priestess sits without harm. If the
priest is masculine, he has been made neuter. But if the village is not
fortunate enough to possess a eunuch, a woman performs the cere-
mony ([1904] 1965, 3: 29192; 2: 25051). Krsn as eunuch identifica-
tion at Putr as swing partner of the soon-to-be-mutilated Kttn t avar
could thus be an iconic counterpart (with the Mahbhrata war vigil
as painful) to the swinging of eunuchs, or alternately women, on
thorny swings.
Krsn as iconography as a divine eunuch or woman with a mus-
tache has its counterpart in cases where the divinity is not just a
mustached bride but a mustached goddess.13 The goddess Yellamma-
Ren uk at her mountain temple at Saundatti (Parasgad Taluk, Bel-
gaum District) in northern Karnataka presents the most comparable
profile. For her daily decoration (alam kra), the icon in her sanctum
is fitted every morning with multiple sarees, tiara, marriage neck-
lace (tli), neck capsule for her little phallus or ista linga (worn by
Lingyat sectarians), and a silver mustache in the form of double lunar
croissant (Assayag 1992b: 394, 398 [plate 32]). Like Kttn t avar
at Kvkkam, this Yellamma attracts transsexuals: especially males
(Jgappas), but also, among their female counterparts (Jgammas),
some who masculinize their dress and manner (231, 234; Bradford
1983: 318; Tarachand 1991: 98100). The cults myth is a folk version
of a Mahbhrata story about the Brahman Jamadagni, his Ksatriya
(warrior class) wife Renuk (who is the same as Yellamma), and their
five sons, youngest of whom is Visn us avatra Paraurma, Rma
with the axe. When the four brothers refuse Jamadagnis command
to slay Renuk for a slight lapse of marital chastity, Paraurma obeys
the paternal command to behead both his mother and brothers, and
then uses his fathers boon to revive them (Assayag 1992b: 4244;

Beyond the cases discussed below, cf. Brckner 1987: 2526; Claus 1979: 1056
in Tulunad, the typical family deity of royal Bants, the androgynous nature-
controlling goddess Jumadi, is mustached, as is her male attendant Ban t e.
hair like snakes and mustached brides 253

Tarachand 1991: 9899, 103). Jgammas identify with Ren uk (Yel-

lamma) in the status of secondary divine brides of Jamadagni (iden-
tified with iva). Jgappas see themselves as descendants of the couples
four oldest slain and revived sons.
Jgappas are transvestites, taking on womens dress and man-
nerisms, sexually impotent and indifferent to women, theoretically
asexual and celibate, who display a stereotype of aggressively erotic
femininity (Bradford 1983: 316; Assayag 1992b: 22934). Only those
few who have atrophied penises become eunuchs, taking it as a sign
to undergo the operation from the goddess Bahucr, a multiform of
Kl whose Ahmedabad temple in Gujarat has pan-Indian recognition
among transsexuals (Assayag 1992b: 26, 22936; cf 1992a; Bradford
1983; Nanda 1990: 145). Jgappas dance for Yellamma more often
than Jgammas, many of whom are prostitutes. Some Jgammas
receive their sign of divine calling from the goddess at a very young
age by the miraculous appearance of locks of matted hair (jte or
jedi; Assayag 1992b: 21721, plates 17 and 18). Their youth aside, this
spontaneous appearance is like that of the female ascetics Obeyesekere
describes in Medusas Hair. The matted locks are variously explained.
Assayag sees unsanitary rural and low-caste conditions as a factor
(21718). For many non-Jgammas, it is a sign of their impurity.
Jgammas, who affect renunciation despite being prostitutes, explain
that Yellamma takes hold of their sexuality to yield back to them this
extraordinary capillary vitality, a sign of their exceptional religious
fecundity, and a particle of the power of the Mother.14 When their
matted locks inevitably break, they offer them to Yellamma in a pj
that metonymically combines them with weapons, garlands, snakes,
wild vegetation, and hanging tree roots (21821). Some Jgappas also
have matted hair, and appear together with the Jgammas in proces-
sions (218 and plate 20).
As to this goddesss mustache, pjrs call it ms, mustache.
Assayag found no explanation for it (personal communication, May
1992), but contextualizes it in relation to Yellammas royal and cos-
mic power (1992b: 39495) and Supreme Energy (akti), which lies
at the base of all gender categories and their transformation (235).
Tarachand indicates that the mustache is of a type that is offered to

Tamil Alis also consider themselves to be particles (am as) of the same goddess
Bahucr (Tamil Pakucramt; Nrull 1990: 52).
254 chapter twelve

the goddess votively by her devotees: Those who want children vow
to offer [a] miniature form of a cradle made in silver. Those who want
a son offer a silver model of mustaches to the deity (1991: 90). Thus
while Yellamma is the source of all children, the mustache offering
generates boys from her male aspect.
Back at North Arcot Kttnta var festivals, one finds a similar inter-
changeableness in the multiple identities of the second mustached
head. Only at three villages were they identified as female: those already
mentioned and a third.15 But even at one of these, Virnantal, the priest
was contradicted by other informants, who identified the second head
as Municuvaran, a male guardian deity. At still other North Arcot
Kttn t avar temples, there were further explanations. Both heads are
Kttn t avar: the smaller goes on procession, the larger stays in the
temple and rides the swing to listen to Mahbhrata recitation.16 Or,
a red head is Aravn , a gold one Kttn t avar; once someone tried to
steal the original red Aravn, went blind, had family problems, and left
it on a hill to be recovered. To regain his eyesight, he commissioned
the new gold Kttn t avar for the temple.17 Or, a small head is Aravn ,
a larger one Kttnta var, while two other heads belong to Mohin,
the Enchantress. All four fit bodies that are taken on procession on
the festivals last day, celebrating Aravns marriage, battlefield sacri-
fice, and the lamentation for his ordeal. Earlier in this festival, on the
third day, men dress as women and put on a tli to marry Aravn
like Mohin, who is understood to be Krsn a in womans guise; later
they take off the tli and bangles and throw them in a tank. As in
South Arcot, Mohin is the form Krsn a assumes to marry Aravn as
his last-minute bride.18 But this is the only North Arcot temple where
a voluntary connection was made between Krsn as female (or eunuch)
guise and the Mohin myth. North Arcot informants who knew the
story connected it with the South Arcot Kvkkam festival. But most
did not know the story. It is thus possible in this fragmented situation
that North Arcot identifications of Krsn a the woman or eunuch
might have originally been independent of the South Arcot myth.

Pulimtu (Vellore Taluk, very near Putr), where the second head belonged to
Kannan pen, Krsn a the Woman.
At Kilvan ampt i, Chengam Taluk.
At Kumraman kalam, Vaniyampati Taluk.
At Clavaram, Vellore Taluk.
hair like snakes and mustached brides 255

Virtually all of these North Arcot Kttnt avar festivals com-

bine their worship of Kttnta var with that of Kenkaiyamman , the
river goddess Gang (Ganges). It is again Kenkaiyamman, known as
Gangamma just across the North Arcot border in Andhra Pradesh,
who has a big festival in Tirupati (Chittoor District) in which she is
worshiped as Veshalamma, the Mother of Costumes or Goddess of
Disguises. Her guises reenact myths of her male transformations to
seek out a vile lusting landlord who despoils budding maidens on the
night before their weddings, and turns his lust on Gangamma when
he sees her drying her hair after a bath. To expose and kill him, she
adopts various disguises, several male, in the last of which she becomes
a foreign prince (Dora) to behead him, and then momentarily revives
him to show him her true supreme female form. Her transformations
have their counterpart in ritual masquerades by men in womens guise
during her festival, which replicate her own disguises. The landlord,
in processions, has a mustached red-painted wooden head, while the
goddesss terrifying image has only protruding fangs (Handelman
1995: 22933). But her princely disguise includes a mustache (Anand
Akundi, personal communication, January 1995). A eunuch also fig-
ures in the story: her true form is also a head, lost when thrown into
a deep well by a crazed eunuch. Handleman takes this to indicate a
male feminized in his loss but not made female, and therefore in dis-
junction, rather than in harmony, with the female (1995: 293). This is
slippery territory. I would suggest that the crazed eunuch is an allo-
morph of the goddess: a joint image of what lies at the source of gen-
der transformations, like the Supreme akti at Saundatti, as described
by Assayag.
On the same arc but now back in western South Arcot, just south-
east of the range of the North Arcot temples and west of Kvkkam,
one finds Draupad festivals where men dress up as women. In the
village where this ceremony is said to be finest,19 two groups of about
fifty men each don both male and female guises. One group represents
the Pn davas and Draupad in the guise of low caste foxy gypsies
(narikuruvarvsam) going to Duryodhanas court during their year
incognito (cf. Hiltebeitel 1988a: 3019). The other portrays Krsn as
sports with his mistress Rdh and the cowherd women (Gops)
in the guise of cowherds (itayarvsam). The sequencing of these

Mutiyanr, Kallakuruchi Taluk.
256 chapter twelve

masquerades is structurally identical to those at Kvkkam: both occur

on the night before Aravns sacrifice. In one case men don womens
guise to sport with Krsn a; in the other men don a guise of Krsn a to
marry Kttnta var.
North Arcot thus lies at a crossroads of masquerading and gender-
mixing rituals focused on different deities (Draupad, Gang, Krsn a,
Aravn ) whose worship coincides in North Arcot Kttn t avar tem-
ples.20 If such triangulation allows us to read behind the few accounts
that identify the double-mustached heads as Kttnt avar and Krsn a as
his divine bride, it is clear that our explanation is beyond our infor-
mants horizons. Not surprisingly, the question why a mustache?
provoked only puzzlement and amusement when asked at Putr:
Twenty years ago, Krsn a had no mustache. The man who painted it was
a potter from Arani, an icon painter by profession.
About twenty years back someone asked, why this mustache? Someone
from Vaitiyantankuppam said, No no, he must have a mustache.
How do we know about such things? So its painted once in fifteen or
twenty years. We pay about one hundred and thirty rupees.
How old are the images?
I wouldnt know. Even my grandfather wouldnt know. Will they put
another mustache on when it is repainted? It should not have a mus-
tache. If we say no, perhaps they will leave it off.
One informant thus attributes a new mustache to the professional
expertise of an icon painter (like Dandapani), while the other recalls a
debate concluding that an old mustache should not be removed. Both,
however, agree: it is twenty years since anyone had any idea as to why
the mustache is there, that those who once might have known were
from outside the village, and that no one living there today has a clue
about it.
Perhaps the mustached companion is without a stable mythical
identity because the identity derives from masquerading rituals in
which identities are already multiple. Perhaps it is there because all
kinds of people take the vow to marry Aravn , and because Aravn ,
given his well-known, last-minute predicament, is willing to marry all
kinds of people in multiple masquerades. These are vague possibili-
ties, but they raise precise questions. Even though the identifications

North Arcot Kttnta var temples may also jointly be Draupad temples, be linked
with Draupad temples in one village (as at Vrnantal), or perform Draupad-type
festivals (as at Pulimtu).
hair like snakes and mustached brides 257

of Krsn a the woman or Krsn a the eunuch may evoke the Mohin
myth, they might reflect an older stratum of the cults ritual in which
a mustached divine bride, as a symbol of this fluid state of affairs, is
an intelligible incongruity.

Alis at Kvkkam

It is in eastern South Arcot, and especially at Kvkkam, that large

numbers of visiting transsexual Alis join temporary village transves-
tites in donning the tli to marry Kttn t avar. It is not clear how
old this practice is. But it is intriguing that Francis, writing in 1906,
mentions no Alis in attendance. As we saw in chapter 11, he mainly
describes a ceremony on the eighteenth night: An enormous num-
ber of fowls are sacrificed to the god while men who have vowed to
marry him dress like women, pay the priest to tie tlis around their
necks, and then join in laments for their husbands sacrifice (Francis
1906: 376).
One cannot tell from this nonmention whether Alis participated
ninety years ago. But some things have changed: prominence of the
eighteenth day rather than the sixteenth; the massive sacrifice of cocks;
the type of sacrificial enactment. Kvkkam informants today differ
as to whether the festival was ever without Alis, but they agree Alis
came in only small numbers until the visit of an Ali named Tvi about
twenty or thirty years ago. As cited in chapter 11, Nrull provides
an account of Tvis visit (Alis refer to themselves with feminine pro-
nouns or gender-neutral honorifics). Arriving during festival time on
the full moon of Cittirai, she
walked the last seven miles on foot, with the sun beating down in the
middle of the road, to Kvkkam. There she collapsed on the pyol in
front of the house of a barber named Rju, who offered her hospitality
with great affection. That house, which was then a shack, is now a build-
ing.21 And if Tvi comes to Kvkkam today, she stays there. When she
first came, Tvi prayed to Cmi with fear and devotion. She took part in
the rites. She did all the necessary things. Within the next year, she had
become a millionaire. . . . (Nrull 1990: 4142).

Kvkkam villagers I interviewed did not know that the barber Rju figured in
Tvis story. When he was alive his family was well off, but is so no longer. His son is
a Kttn t avar temple drummer.
258 chapter twelve

It is intriguing that Tvi collapsed on the pyol of the barbers shack.

The story suggests an unconscious affinity between hair cutting and
castration, drawing together Tvis initiatory journey and the barbers
traditional role of performing tonsure ceremonies (and, according to
Thurston, the castration operation itself for some eunuchs).22
Nrull prefaces this account by saying that Tvi felt an affinity
between the festival and her own in am (kind, species, gender). Alis
gather at Kvkkam and other South Arcot Kttn t avar festivals in
part to strengthen their in am (Elumalai 1989:57). By marrying Aravn ,
whose battlefield sacrifice mirrors their own castration, they partake in
a ritual that nourishes an inam that cannot reproduce itself biologi-
cally. As noted in chapter 11, there are transparent affinities between
Alis and this self-mutilating god, but there is never a hint that Aravn
cuts his penis among his thirty-two cuts.23 As the beautiful and heroic
male whom the Alis impersonate Mohin to marry, Aravn evidently
retains his full masculinity, but the Alis seem to have picked up
on his ordeal as a theme that draws their sympathies, sadness, and
No doubt Alis who might have preceded Tvi to Kvkkam would
have felt the same affinity with the god as her. In Franciss time, small
numbers could thus have been drawn to a masquerading type of fes-
tival more like the type encountered presently in North Arcot, where
Krsn as eunuch guise, with mustache, alternates with his more straight-
forward guise as a mustached woman. Today at Kvkkam, prior to
the large-scale rituals of transvestitism that reenact Aravn s marriage
on the fourteenth through sixteenth days, there is a second-day double
ritual wedding in which two men impersonate the bride (one each
from the Van n iyars and Paraiyar-Untouchable communities) and
marry two men (from the same groups) who impersonate Aravn.
Since we will look into this ceremony more closely in chapter 13, it
will suffice here to note only that Kvkkam residents consider this
double wedding ancient. But at a Salem District Kttnt avar festival
where the bride is not identified as Mohin, an Untouchable likewise

Thurston and Rangachari [1904] 1965, 3: 289 mention that barbers perform the
castration for south Indian Muslim eunuchs called Khjas (apparently Nizrs).
A Draupad temple pcri says these are his head 1, point between brows 1, 2,
temple 1, nose 1, earlobes 2, lips 2, chest 1, knuckles 2, wrists 2, elbows 2, shoulders
2, stomach 1, knees 2, insteps 2, and toes 10 (Hiltebeitel 1988a: 327).
hair like snakes and mustached brides 259

dresses up as a woman to marry Aravn at the festivals beginning.24 In

this temporary womans guise (vsam), he marries the deitys icon,
wears a tli for eighteen days, and then removes it for the rites of
battlefield sacrifice, which include public lamentation. In both situa-
tions, the bride is an Untouchable: an indication that the guise may
not only mix genders but castes, since, on his paternal side, Aravn
is a Ksatriya. Salem District informants say that Alis do not come to
their festivals at all.25 Sexual masquerading would thus seem to be an
older tacit, or at least nonmythicized, ritual expression of marriage to
Aravn, in which the brides status can be low: a category open to both
Alis and Untouchables.
Whether their part in the Kttn ta var cult is foundational but origi-
nally small and only recently spectacular, or recent only, one may in
any case hypothesize that since Tvis visit to Kvkkam, Alis have
given strong impetus to the reshaping there of earlier mythic and
ritual traditions. Nowhere can their contribution be suspected to be
more crucial than in their distinctive formulation of Aravns third
boon: his marriage to Mohin. The oldest known literary version of
this myth refers only to Krsn as transformation into Mohin as a
woman (Mkin ippen ). This is found in the chapbook publication of
the drama Aravn s Battlefield Sacrifice, whose earliest known print-
ing is 1897 (Aravn Katapali Ntakam 1967:7071; see n. 4 above).
Thus nine years before Franciss description of the Kvkkam festi-
val, which refers only to local villagers taking on the appearance of
women, the drama tells of Krsn as transformation into Mohin in the
form of a woman, but gives no hint of Alis. The dramas in this collec-
tion seem to reflect Draupad cult performative traditions of the mid-
to late-nineteenth centuries (Hiltebeitel 1988a: 16062, 324).
According to Ali informants today, however, Mohinat least in this
appearanceis not Krsn a the Woman or even Krsn a the Eunuch,
but Krsn a in his Ali avatra (Nrull 1990: 4344, 61; see chap-
ter 11). Rather than a temporary guise, avatar status defines a divine
mode of being that the god makes permanently worshipful, and is in
some sense permanent in a way that a guise is not. This designation

At Panaimatal village, Attur Taluk.
The same is said of Coimbatore District festivals, but one Ali conspicuously
attended the 1995 Kttnt avar festival at Cin kanallr near Coimbatore, far from South
Arcot. Festival functionaries told me she had no connection with the ceremonies she
attended, but that might be a matter of interpretation. See now chapter 14.
260 chapter twelve

reflects the Alis own marriage to Kttnta var, which extends beyond
the festival in rites of widowhood that are unique to them, and can be
renewed at the festival each year. Most Kvkkam villagers have a dif-
ferent view. Like North Arcot villagers who knew some version of the
story, they regard Krsn a as Krsn a the woman. In temporarily becom-
ing Mohin to marry Aravn, he becomes not a eunuch, but simply a
woman. Both theological positions are thus found at Kvkkam: not
surprisingly, the Alis voicing the eunuch avatra, local villagers gen-
erally upholding the woman identification.
The Ali avatra identification may, however, be older than our
sources reflect, and only have surfaced with the Alis heightened activ-
ity in the cult since the visit of Tvi and, of course, recent fieldwork.
Their increased participation in South Arcot could have encouraged
them to popularize their older understanding of Mohin. In any case,
their contribution to the Mohin-Aravn marriage myth has by today
taken the latent tension in the two images of Mohin toward a fusion
of horizons between the Alis perceptions and those of local and
regional villagers.26 This tension and fusion provide us with a pos-
sible glimpse into the cults deep motivational power for Alis at the
personal level.

The Ali Ambience at the One-Day Marriage

To portray the Alis festival participation as it relates to hair, we must

consider Kttn ta vars embodiment and disembodiment in three pro-
cessions. First, he is carried on the heads of dancers; second, his head
is carried on a portable platform (ktyam); and third, he is embodied
on his full chariot (tr).
The first occurs after nightfall of the opening day, and is repeated for
six evenings. Kttn t avars head is brought outside the temple to be
danced through the village. Nine Van niyars dance in relay, their bold
strides said to recall medieval Van niyar warriors. But it is Kttnt avar,
a great dancer, who is said to really be dancing. Dancers say they are
not possessed, but they are clearly in the role of possessed god-dancers

Although it was defended by the Kvkkam pcris older brother in a fraternal
disagreement, the Ali avatra explanation was volunteered only by Alis.
hair like snakes and mustached brides 261

called cmiytis. Cut pieces of plantain stalk fixed to Kttn ta vars

crown are thickly covered with jasmine flowers, while strings of jas-
mine hang down from the crown and fly about the dancing head.
Villagers consider the ensemble a special flower-anointing (cttuppiti
p alankram) that has nothing to do with hair. But when the long
strings fly about, they look like snakes or dreadlocks (see Figure 5).
The god thus first appears in a mode that looks as if it could reflect his
proverbial hairiness and his maternal snake ancestry. Alis have not yet
arrived at the festival.
From the thirteenth day onward, preparations are made for the two
other processions. After a secret ceremony in the sanctum, in which the
gods life (uyir) is temporarily removed from his head into a vessel,
the head, now mere wood, is wrapped in red silk and taken to a shed
in the back of the temple. Every year a carpenter and painter remove
Kttnt avars jewels and mustache, peel off all the paint-on-cloth that
covers the head, and attach a new white cloth as the base for repaint-
ing. Aravns mustache is again important. It is one of the last signs
of life removed and one of the last features restored before his lifes
renewal (see figure 6). As elsewhere across many cultures, hair, which
continues to grow (like nails) after death, carries life of its own even
beyond death. According to the sculptor Dandapani, the mustache has
characteristics that make it one of Aravn s thirty-two marks: broad
chest, tallness, long hands and legs; bow-shaped eyebrows; fish-like
eyes; mustache twined very tightly: there is a principle that a mustache
should be slightly twined and long so that you could stick a lime on
the pointsthese principles have been handed down by our elders;
wide open eyes; eyelids that should never flicker. Other than the mus-
tache, this partial list could recall the thirty-two marks that identify the
Buddha as a Mahpurusa, among which is the twentieth: that every
hair on his body [is] detached, ascending upwards, and turned to the
right (Foucher 1963: 258).27 Dandapani gives fullest attention to the
bristly mustache, mentioning it along with two other facial hair traits:
the bow-shaped eyebrows and eyelids that never flicker. The spiking of
limes on the end of the mustache evokes the conventional impalement
symbol of limes on upraised weapons, and its bristling condenses, for

The Buddha has other hair-related marks (Foucher 1963: 25859), and in the
Mahyna light rays (rather than hairs) spread from the pores of his skin.
262 chapter twelve

Figure 5. Kttn t avars head being carried on the heads of dancing Van n iyar
devotees at Kvkkam on the first night of the festival. The jasmine strings of
his flowered crown fly about, looking like dreadlocks. Photo by Alf Hiltebeitel.
hair like snakes and mustached brides 263

Figure 6. Kttn t avars head after it has been repainted at Kvkkam, with
the mustache, still askew, as one of the last features restored, to signify the
heads revival. Photo by Alf Hiltebeitel.
264 chapter twelve

the head, the trait of hair that stands on end. The mustache is thus
another exemplar of Kttnt avars hairiness, and, through its evocation
of impalement and sacrifice, a condensation of his suitability for the
battlefield sacrifice that, again, evokes the worship of upraised weap-
ons at Dasar.
On the fourteenth through sixteenth days, those who vow to marry
Kttnta var can begin to buy tlis. On the fifteenth, a festive atmo-
sphere builds from evening on, centering on the arriving Alis. They
have year-by-year arrangements to rent houses on the southern street
of Kvkkam for the nights of their stay. An overflow camps out in
surrounding fields; others fill hotels in nearby towns. At the Rolex
Hotel in Villupuram, before busing to Kvkkam, Alis dress each oth-
ers hair into flowered chignons (see figure 7). At Kvkkam, on the
pyols in front of their rentals, they put make-up on each other like
cinema actresses (Nrull 1990: 39).
After midnight, last touches are applied to the painting of the deitys
head. By now it is dark on the sixteenth morning, and the sale of tlis
and strings of jasmine (malar carankal) for the hair is brisk. After

Figure 7. Alis at the Rolex Hotel in Villupuramprior to the bus trip to

Kvkkam during the 1982 Kttn t avar festivaldress each others hair in
flowered chignons. Photo by Alf Hiltebeitel.
hair like snakes and mustached brides 265

an eye-opening ceremony restores life to the gods head, the last

touches are another flower decoration and a refitting of the mustache:
hair and vitality are thus restored to the icon. Meanwhile, excitement
mounts outside as the faithful gather in front of the shed to await
the gods emergence for his public eye opening. After a big display of
fireworks in the fields, the head is brought out at about 3:40 a.m. and
set onto the portable platform. It is then carried forth flanked by two
additional platforms: one bearing Kttn t avars garland-draped chest-
plate and epaulets, the other his flower karakam, a flower-decorated
pot containing his power or akti, which has been dancedby both
villagers and Alisthrough the night. This procession tours the village
till daybreak. As the god sets off, it is said, much as it is of the swinging
heads at Putr, that Kttnta vars face shows him happy (cantsam)
and smiling.
At about 5:30 a.m. the festival then reaches its climax. Awaiting the
arrival of the portable platform, a vast crowd assembles in the open
area among the village temples. On top of the Kttnt avar temple,
cocks are offered: no longer cut, but thrown up onto the temple roof.
Officers of the Tamilnadu State-governed Hindu Religious and Chari-
table Endowments Board have instituted this change since the early
1980s by posting signs, even within the temple, discouraging animal
sacrifice. But villagers continue to offer cocks away from the temple.28
One will recall the enormous number of fowls sacrificed on the
culminating day in 1906. Today, only cocks are offered: If it doesnt
crow, God (Cmi) wont take it. As we shall see, for the Alis, at least,
cock sacrifice resonates with castration.
Finally, the post on Kttnta vars chariot stands ready for the
assemblage of the gods body. The epaulets and chestplate are danced
forward, and while the head is danced around the chariot, the epaulets
are filled with flowers and the chestplate positioned. Finally the head is
lifted into place. There is a grand display of firecrackers. Two hundred
kilograms of camphor are lit in front of the deity, and the entire con-
course flings strings of jasmine, spectacularly filling the air with two
hundred thousand rupees (two lakhs) worth of flower strings. These
are gathered onto and into the deitys body by men positioned on

They keep saying dont cut cocks, we keep doing it. The government just takes
our money. They dont do much. All the hard work is ours. Whatever they want to do
they have to get permission (June 1994 [informants name withheld]).
266 chapter twelve

the chariot and its scaffolding. The head is supplied by the Kvkkam
temple; the other wooden bodily parts and insignia (chestplate, flags,
umbrella) come from nearby villages. But the strings of jasmine that
fly snakelike through the air come from those who have vowed to
marry him: in particular the Alis, who fling the flower strings that
they have just removed from their hair (Nrull 1990: 40).29 The gods
temporary ritual body, made for this one-day manifestation as an epic
hero, is thus the joint production of villagers, who supply solid bodily
parts, and Alis, who offer snaky garlands from their hair.
Aravn is thus depicted in heroic pose, prepared for battlefield sac-
rifice and armed to fight.30 But he is already looking sad, sweating,
and shedding tears. No longer happy as a head, his embodiment is
painful. On his body, the jasmine strings offered from the Alis hair
now take on the position, and to me the look, of excessive body hair
(see figure 8). If so, they also remain analogous to snakes, since, as
one recalls from Dandapanis account, as the god sets forth for battle,
his skeleton is rigidified by snakes: snakes who, like the flowers, fill
his every limb and digit. Snakes like flowers, flowers like hair, hair
like snakes. As Kttn t avar marches off toward the weeping ground,
the Alis now let their ungarlanded hair hang loose to begin the phase
of the ritual that identifies them as his widows, and as he marches on
and takes on an increasingly dead look, his jasmine garlands fade to
brown. This is a telling continuum. While the widows hair of the Alis
falls loose, the flower-strings from their hair represent Kttn ta vars
loss of vigor and vitalitya theme associated, as we have seen, with
hair itself.31 Moreover, the snaky jasmine strings that the Alis have
offered from their hair, which now fill Kttnt avars body and hang
from his limbs, are removed from his body, during his march, as sym-
bols of his self-mutilation.32 Unlike the snakes that rigidify his body
and then fall off, the jasmine strings, before they are removed, hang
limp and wilt. Meanwhile, villagers still sacrifice cocks before the char-
iots wheels along the march.

When I saw this amazing flinging scene from the Kttn t avar temple roof, I was
too stunned to take a picture.
As noted in chapter 11, the march represents both Aravns prewar kalappali and
his eighth-day fight with Alambusa, according to different informants.
Kvkkam informants mention this association in connection with hair and nails
used in black magic (pillicn iyam), against which Kttnt avar protects people.
They are tossed as prasda (leftovers from an offering, denoting the deitys grace)
to devotees who follow the chariot.
hair like snakes and mustached brides 267

Figure 8. As Kttnt avars embodied effigy marches toward his battlefield

sacrifice with jasmine strings draped from his limbs, he has an exceedingly
hairy look. Photo by Alf Hiltebeitel.
268 chapter twelve

Most of the Alis who lament Kttnt avar go to the weeping ground
long in advance of the slow procession. There they complete their weep-
ing, and, like all the rest who have married Aravn, leave their severed
tlis and broken bangles at a bamboo post.33 When the effigy arrives,
entirely divested of garlands, the hanging roots of a great peepul tree
behind him, the huge image looks skeletal and is said to be all but
deadhis only life hovering in his eyes, which oversee the tli and
bangle offerings that continue to be made, but now only by villagers.
The Alis have gone to a pond, donned white widows sarees, and, their
hair left loose, begun to head home. They should wear widows white
for thirty days before they can once again don colored sarees, bangles,
forehead dot, and dress their hair with flowers.
Eventually, the huge regionally gathered crowd thins out, and by
evening, the ceremonies concern villagers only from in and around
Kvkkam. Ultimately, the head is revived at a Kl temple, redeco-
rated, and danced back to the Kttn t avar temple in the fashion of
the first six nights with what, again, looks like snaky dreadlocks of

Mohin, the Ali Avatra

Let us now consider the convergence of perceptions between Alis and

local villagers that has, in the last thirty or so years, probably reani-
mated the symbol of Aravn s head at Kvkkam and the festivals
distinctive deployments of hair. First, there is an opposition between
local villagers (particularly Van n iyars) and Alis. The severed head
relates to the super-male figure of Aravn, the great warrior and man
of thirty-two bodily perfections, a virtual equivalent of the Great
Man (Mahpurusa) of Sanskrit lore (a title given, among others, to
the Buddha with his thirty-two perfections), who offers his head to
the goddess as the culmination of his thirty-two self-mutilations.34 He
is an embodiment of a Ksatriya ideal that is part of the heritage and

Cf. Nrull 1990: 41: A pcri cuts their tlis with a scythe. He snatches off the
garlands of blossoms on their heads and throws them. He completely erases the red
mark on their foreheads. Then they stand there and finish crying.
In what appears to be a northern adaptation of these themes, the legendary
emperor Vikramdityas thirty-two marks make him the ideal offering to the goddess
Kmks or Kmkhy, who, appreciating his willing supplication, relents and fulfills
his wish without the offering (Edgerton 1926, xxiii: 18286).
hair like snakes and mustached brides 269

identity of his Van niyar devotees, who consider themselves Ksatriyas.

But the symbol also relates to the involvement of the Alis, who imper-
sonate Krsn as Ali avatra to marry this deity. They marry a male who
shares something of their own ordeal, but evokes what they have irre-
vocably lost by retaining his masculinity. One cannot miss the impli-
cations of Freudian theory: the presumed displacement of castration
anxiety from the penis to the severed head, which would account for
the general Ksatriya perception of Kttn ta var, would be reoriented
for the Alis, who, having been castrated, would displace (or double)
their formerly male identification with Kttnt avar by a feminized Ali
identification with Krsn a-Mohin.
Here one might reflect on Mohins wider mythology. One must
first think of the classical Mohin, who embodies the bewitching power
of illusion (my) that inhabits Visnu when he sleeps during the night
of cosmic dissolution, and whom he projects as an Enchantress
(Mohin) to seduce demons, or iva, in various mythsin popular
South Indian accounts, to bear their child Aiyanar or Hariharaputra
(Son of Visnu and iva).35 Yet I submit that the image of Mohin at
the genesis of the Kttnta var cults configuration is that of Mohin
as succubus, a seductress who appears in dreams and fantasies, and
lures men, especially young onesnot only Hindu and south Indian
Christian (Caplan 1989: 57) but Sri Lankan Buddhist ones (Obeyesek-
ere 1981: 13940)into debilitating activities and fantasies. Although
Kvkkam villagers do not identify Mohin as an Ali, they speak of
Krsn as Mohin avatra as including the Mohin who comes when
a young man is sleeping to arouse sexual desires in dreams. This they
distinguish from Piccu (Ghost or Demonic) Mohin, whom
they consider a mad (pitta) spirit (Ganesan and other informants,
Kvkkam, July 1994). The distinction is theological rather than psy-
chological. According to Caplan, Mohin mainly appears to young
bachelors, most typically on their way home from a late night film,
wearing a white saree and flowers in her hair. She seeks to lure such
a man in order to live with him as his wife, take his strength, and
finally kill him. In Madras, among both Christians and Hindus, she is

On these aspects of Mohin, see Bhgavata Purn a 10.88:1446; discussion and
sources in OFlaherty 1973: 2829, 32, 5051, 22829; 1975: 27480; Shulman 1980:
30714; Biardeau 1991: 83.
270 chapter twelve

generally referred to as Mohin-peey (1989: 57).36 In Madurai, she is

known as Mohin-piccuyou smell jasmine, hear anklets; a piece of
charcoal or a rusty nail can keep her away (J. Rajasekaran, personal
communication, July 1994).
It is not just that Mohin marries Aravn. By agreeing to marry
him, she seduces him into his self-mutilation, his wasting away, his
death. Similarly, it is frequently said that one mode of recruitment of
Alis is for other Alis to seduce them; and several Alis interviewed at
Kvkkam said that they underwent the sexual operation and trans-
formation because they felt they were women trapped in mens bodies.
The Alis thus identify with the seducer and the seduced, with both
Mohin and Aravn. They are seduced by both, and identify with each
as seducer.37 It may well be that a transsexual homo-eroticism is latent
in classical Mohin myths as well,38 but nowhere else is it combined
with an allurement to self-mutilation. It is probably not possible to
account for the genesis of such a symbol, and its operative unfolding
in South Arcot Kttnta var festivals (cf. Obeyesekere 1981: 13739),
without this subjective imagery of Alis as a contributing factor. And
it is probably the unrecognized transsexual, if not precisely the un
acknowledged Ali, in every man, that explains the general enthu-
siasm among South Arcot villagers for marrying the self-mutilating
god as Mohin, even if they generally favor an identification with
her as Krsn a the [temporary] Woman rather than the Ali avatra.
At Salem and Coimbatore Kttn t avar temples, it is still Krsn a who
secures Aravn a last-minute bride; but he always finds him a woman
plain and simple, and never a male turned something else. We would
seem to need the Alis in combination with the regions festival mas-
querades to understand the unique forms that Kttn ta vars marriage
takes in South and North Arcot.

Py, like its more Sanskritic equivalent piccu, mentioned below, denotes ghost
and/or demon.
Writing of the inner life of Alis (alikalin antarankam), Nrull observes that
many have strong desire for men (n mtu cai), and may even shed seed (vntu)
after some time in a loving relationship (1990: 2425). Approaching the festival, some
Alis even seem to be casting a net for good-looking youths (alakn a vliparkalai;
46; cf. 6061, 66).
Folklorists Palani and S. Rajarathinam, responding to an earlier presentation
of this chapter, insisted that homosexuality is also implied in Visnus appearance as
Mohin to seduce iva (personal communications, 1994).
hair like snakes and mustached brides 271

One should also note that in other Mohin myths, Mohin is a form
of Visnu rather than of Krsn a. The Arcot mythologies thus provide the
only instance I know of where the transformation into Mohin becomes
one of the tricks (tantiram) of Krsn a in the Mahbhrata. Here we
must recognize that the story can be interpreted as a multiform of the
sort of Indian Oedipal pattern discussed by Goldman and others in
which a real or surrogate son is punished, typically by castration or
impotence, for intruding on the sexual life of his fatheror, in varied
transpositions, by the coercive and potentially castrating power of
dominant males such as fathers, older brothers, gods, gurus, and
sages (1993: 391, 395). Drawn into the Mahbhratas ubiquitous
concern with the central but often disguised triangle of father, mother,
and son (392), Aravn certainly does not intrude intentionally into
his fathers sexual life, but he does provide a means to revive it. The
Pndavas, who must authorize Aravns sacrifice, cannot restore mari-
tal relations with Draupad until they have won the war and she has
put up her hair. Aravns self-offering to mother Kl is thus also one
that satisfies mother Draupad and his five Pndava fathers. But it is
really Krsn a whose trick is most troubling: perhaps even more trou-
bling than usual. He is the supreme male (purusottama), the great man
(peruml), who truly instigates and requires this sacrifice, and who
could be the victim himself, yet becomes delusively female to guar-
antee it. To Aravn he is both the ultimate castrating father and the
surrogateas momentary wife and loverfor the seducing, castrat-
ing mother. We should remember, however, in defense of a devi-
ous divinity, that Krsn a is not omnipotent, that in the Mahbhrata
the will of God may intervene in whatever way it can (Matilal 1991:
40815), and that here, as always, Krsn a works through the desires and
fantasies of those who hear and tell his stories.

Hair, Flowers, Snakes, Cockscombs

It is, of course, problematic to ascribe the genesis of particular symbols

in the Kvkkam festival, and in South Arcot, to Alis. But the case is
made more tempting by numerous symbols that reinforce this Mohin
complex. It is striking that so many of the cults major symbols, which
can be found at festivals where Alis do not regularly participate, and
even, as we have seen, in rituals at Kvkkam that anteceded their
increased participation, take on deepened resonances for them.
272 chapter twelve

Let us start with the roosters. The cock is the vehicle of Bahucr
Mt, the goddess of the main temple for Alis and Hijras in Ahmeda-
bad, Gujarat. The Tamil term kucu, penis, also means little chick.
According to Nrull, a preliminary to the Alis castration operation
is the severing of a piece from the cockscomb of a rooster to give it to
the candidate on the night of his castration (1990: 61). This red piece
of cockscomb is called kon ta i, hair bun or chignon: the term used
to describe a womans hair (for example, Draupads) when she puts it
up, or before she loosens or dishevels it. Ali informants thus make con-
nections between hair put up, the piece of the cockscomb, castration,
and Aravn s battlefield sacrifice, at which their own hair comes down.
The operation, which is called nirvnam, connoting nakedness,39 is
by and large conducted on the night of the full moon of the month
of Cittirai, the night on which Krsn a took the Ali avatra to marry
Aravn the hero (61). When the penis is cut off, the cut piece (tun tu )
and blood are placed in a bucket of water, or a new pot. According to
one account, no blood should touch the ground. The tun tu swells up
and bobs around or quivers (tuti) with a half-life (pti uyir); it throbs
or dances like Kttn ta vars head. Later it is taken and buried in a
small pit. According to another, The Alis . . . hold that new pot which
contains the blood with the candidates member, place it on their
heads and go dancing and singing continuously in a procession. Then
they go and release it in the sea (Nrull 1990: 63). These accounts
derive from Alis who worship Kttn ta var. However momentarily,
like Kttnt avars head, the severed penis takes on a new life. At
Kvkkam and several other Kttnta var temples, Kttnt avars head
was first discovered floating down a river, or on the sea. Happy, as a
head, he floats, dances, sees, and hears. Sad, he suffers in the embodied
state, however fleeting it is for him (one days daylight in an eighteen-
day festival), so that he can cut himself in thirty-two places, marry,
make love, die in combat, and feel with his toes and fingers.
In Kttn ta vars worship at Kvkkam, symbols of head and body
hair alternate between hardness and softness. He is said to be the
ideal person for self-sacrifice because he has etirrman : a hair-circled
mole on his back that concentrates into one beauty mark the theme

Nrull 1990: 6162; cf. Nanda 1990: 2632, 6669: they must be as naked as
the day they were born (27). Nothing indicates whether Alis are ritually shaved for
the operation, but Olivelles mention (1998) of the renouncer as bearer of the form
he had at birth suggests a parallel. The new Alis head hair is regularly attended for
a forty-day period after the operation (Nanda 1990: 2829).
hair like snakes and mustached brides 273

of excessive hair that stands on end at the thrill of self-sacrifice to

Kl. His mustache bristles so sharply that one can spike a lime on
each end. Alis at the festival depilate or pluck out whatever trace they
may have of a mustache, and wear their hair like women. But seeing
them at the festival, these are matters of degree. Some have beautiful
long hair and soft faces. Facial and underarm hair will cease to grow
for some after their operation (Nrull 1990: 10). Others are bald (see
figure 9), or their hair hangs only from below their bald spots. Some
have pancake faces and shave regularly. Nanda mentions Hijras who
nostalgically recall their former mustaches (1990: 16, 60). Villagers,
meanwhile, must shave their mustaches before they don the tli to
marry Aravn.
If hair standing on end is an indication of Aravns predisposition
toward the thrill of self-sacrifice to Kl, it is hard not to see both the
stiff hair and stiff snakes in Dandapanis account as heightened phallic
symbols of his masculinity. Alis, on the other hand, are known teas-
ingly as Omptu (more correctly, Onpatu), by which is meant the
number 9, on which the lower portion of the number always dangles
and never gets erect (Nrull 1990: 9); they are regarded popularly as
having been males who could never get erections (cf. Nanda 1990: 67).
Their contribution to Kttnta vars body and hair is not stiff snakes
but dangling flower strings from their own hair.
Hair, snakes, cockscombs, and strings of jasmine thus take on a new
look at Kvkkam. Again, one can follow the chain of associations
back to the Alis castration ritual. A bobbing penis survives with a half-
life that is assimilated to a decapitated dancing head: a happy head,
forever revivable, that, even before the Alis arrive and after they have
left, is danced through the streets with snaky dreadlocks. But while the
Alis are at the festival, they are there to marry and lament a god whose
embodiment is tragic and sad. The Ali life is full of regrets (Nrull
1990: 72; Nanda 1990: 6869), and their laments reflect this, refer-
ring to their situation as one of cut hands (kaiyaru nilai) and arms
(Elumalai 1989: 64). Kttnta vars hair thrills for battlefield sacrifice.
When the Alis let their hair down and fling their strings of jasmine
onto his body, he sets forth to cut that body in thirty-two places as
an offering to Kl. When Alis are castrated, their hair must be long
enough for it to muffle their cries as they hold it between their teeth
(interviews, 1982, 1990; Nrull 1990: 62).
Flowers like snakes, hair like flowers, snakes like hair: all like penises,
they are what Alis have had to offer.
274 chapter twelve

Figure 9. Artists depiction of an Ali on the cover of r. Nrulls Alikal

Vlkkai, Long Live Alis, a pamphlet-sized book sold at the 1990 Kvkkam
Kttnt avar festival. Photo by Alf Hiltebeitel.


Kttnta var or Aravn, a Tamil village deity usually represented

by a fierce red severed head, has associations with death that extend
from the village to the Mahbhrata.1 In the overlapping pantheons
of Tamil village gods and epic heroes, he is in each case a relatively
inconspicuous figure.
It is sometimes said that Kttnta var is a village deity (kirmatvatai),
whose worship is general to the whole village, rather than a clan deity
(kulateyvam, kulatvatai), though the latter terms are also used, par-
ticularly in South and North Arcot by some of his main worship-
pers from the Vann iyar community,2 and curiously, by Alis, so-called
eunuchs, who worship him at certain South Arcot festivals as a deity
who helps them increase their in am (kindred, kind, species).3 His festi-
val worship is often conjoined with that of village goddesses (especially
Mriyamman , Ken kaiyamman, and Kli yamman), and includes goat
and rooster sacrificesboth as general (potu) offerings sponsored by
the village, and individual offeringsat key points. His fearfulness is
often highlighted. At Panaimatal (Attur Taluk, Salem District), at the
ritual of his eyeopening, no one will stand in front of him for fear of
suffering from his glance.4

This chapter benefits from presentations at the Association of Asian Studies
Annual Meeting, Washington, D.C.; the cole Pratique des Hautes tudes, 5th Sec-
tion, Sciences Religieuses, Paris; the Center for Asian Studies, University of Virginia
(all 1992); and the South Asia Seminar, University of Texas, Austin (1993). Fieldwork
in 1990 was sponsored by the American Institute of Indian Studies. Writing was facili-
tated by support from the George Washington University Committee on Research. I
thank E. Sundaramurthy, N. Daivasundaram, Pon Kothandaraman, Lee Weissman,
J. Rajasekaran, and Abbie Ziffren for their help with Tamil. I thank my son Simon
Hiltebeitel for his preparation of the map that originally appeared with the version in
Hiltebeitel 1999b: 277.
See Hiltebeitel 1992b for an analysis of this ethnography as it relates to the same
issues in the Draupad cult. Vanniyars are the most prominent community in the
highest density area of both cults: that is, in South Arcot and adjacent areas.
See chapters 11 and 12.
According to Grard Colas, Vaikhnasa priests say that the eye-opening cere-
mony is always dangerous; immediately on opening, the god has a wide gaze, and
should see auspicious things (personal communication, May 1992). At Draupad
276 chapter thirteen

At Kuricci (near Coimbatore) the goat sacrifice outside his tem-

ple prevents death in the village. At Tailpuram (Tindivanam Taluk,
South Arcot), the four offering stones at the corners of his temple
receive offerings to prevent death pollution in the village if a death
occurs at his festival. At Vellalr (near Coimbatore) pieces of his
body brought back from the festival protect families from diseases.
At Palaiyacramankalam (near Salem city), he protects children from
coughs that are caused by Whooping Cough (Kakkuvn) Mriyamman .
At Vrnantal (Chengam Taluk, North Arcot) and Putr (Vellore
Taluk, North Arcot), he is connected with chicken pox, diarrhoea, and
cholera. At Putr, when his head is brought back from his nightlong
procession and swung from a swing, he must be returned to his temple
before sunrise or these diseases will afflict the village. At Vrnantal,
the guardian deity Pttu Rja is sent out from Kttnt avars temple to
combat cholera if it breaks out. At Tvanr (Tiruvannamalai Taluk,
North Ascot) and Knalr (Gingee Taluk, South Ascot), he has not
only a head that resides in his village temple but an even more fero-
cious and dangerous one at a hill temple that must not be disturbed
except at the opening of his festival. The Mahbhrata would appear
to be far from view here, but the rich unfolding of Tamil traditions
about his death, or better deaths, as a hero, clearly ties in with his
mastery over death and diseases in villages.
As noted in chapter 11, Tamil accounts provide Aravn with a new
story of his noble self-sacrifice at the beginning of the war. In its ear-
liest telling by the ninth-century Pallava poet Peruntvanr, Aravn
agrees to sacrifice himself on the new moon night to Kl, goddess of
the battlef